Report No. 16253-ES El Salvador Rural Development Study Technical Annexes Volunre It August 7, 1997 Central Amicri(a [Deptt i)ament Latin America and the Caribbean Region Document of the World Bank CURRENCY EQUIVALENTS Currency Unit = Col6n (0) $1.0 = 08.75 FISCAL YEAR January 1 to December 31 WEIGHTS MEASURES Metric System 1 hectare (ha) = 1.4 manzana 1 ton = 22.22 quintales (qq) MAIN ABBREVIATIONS AND ACRONYMS AGDP Agriculture Gross Domestic Product BCR Central Bank (Banco Central de la Reserva de El Salvador) BFA Agricultural Development Bank (Banco de Fomento Agropecuario) CACM Central American Common Market CCs Credit Cooperatives (Cajas de Credito) CEM Country Economnic Memorandum CENTA National Center for Agricultural and Forestry Technology (Centro Nacional de Tecnologia Agropecuariay Forestal) CGLAR Consultative Group on International Agricultural Research CNR National Registry Center (Centro Nacional de Registros) DGRNR General Directorate of Renewable Natural Resources (Direcci6n General de Recursos Naturales Renovables) D.L. Legislative Decree (Decreto Legislativo) EDUCO Community Education Program (Educaci6n con Participaci6n de la Comunidad) FEDECACES Federation of Salvadoran Credit Unions (Federacion de Asociaciones Cooperativas de Ahorro y Credilo de El Salvador) FEDECREDITO Federation of Credit Banks (Federaci6n de Cajas de Credito) FOB Free on Board FOSAFFI Financial Sector Recapitalization Fund (Fondo de Saneaniento y Fortalecimiento Financiero) FINATA National Financing Institution of Agricultural Land (Financiera Nacional de Tierras Agricolas) FUSADES Salvadoran Foundation for Economic and Social Development (Fundaci6n Salvadoreha para el Desarrollo Economico y Social) GOPA Agrarian Policy Options Group (Grupo de Opciones de Politica Agraria) IRA Supply Regulating Institute (Insuituto Regulador de Abastecimientos) ISTA Salvadoran Institute of Agricultural Transformation (Institulo Salvadoreino de Transformacion Agraria) MAG Ministry of Agriculture and Livestock (Ministerio de Agriculturay Ganaderia) MOPW Ministry of Public Works (Ministerlo de Obras Ptiblicas) OCTA Coordinating Office for Agrarian Themes (Oficina Coordinadora del Tema Agrario) PROCAFE Coffee Producers' Association PUT Land Transfer Program (Programa de Transferencia de Tierras) RER Real Exchange Rate SSF Superintendency of the Financial System (Superintendencia del Sistema Financiero) VAT Value Added Tax Resnonsible Manafers and Staff Vice President Mr. Shahid Javed Burki Director Ms. Donna Dowsett-Coirolo Staff Member Ms. Cora Shaw EL SALVADOR RURAL DEVELOPMENT STUDY Technical Annexes ANNEX 1: Diagn6stico Agropecuario: Factores Determinantes de su Desempefno ANNEX 2: The Agriculture Sector: Pricing Policies and Competitiveness ANNEX 3: Rural Poverty: A Quantitative Analysis ANNEX 4: The Rural Non-Agricultural Sector and Poverty ANNEX 5: Who uses Basic Services in Rural El Salvador? ANNEX 6: Rural Land Markets ANNEX 7: Land Degradation Problems in El Salvador DIAGNOSTICO AGROPECUARIO: FACTORES DETERMINANTES DE SU DESEMPENO 1. Desempeflo Econ6mico del Sector Agropecuario 1.1. Comportamiento del PIB agropecuario El sector agropecuario constituye el uinico sector de la economia que reporta una tasa de expansi6n econ6mica negativa, como promedio en los uiltimos veinte afios (1975-95), las actividades agropecuarias se caracterizan por su desempeiio econ6mico exiguo y comportamiento erratico, el cual en promedio registra una tasa de variaci6n negativa de -0.2%, en el periodo. Por el contrario, la economia en su conjunto reporta un crecimiento positivo promedio de 1.3% para el periodo. El comportamiento sectorial refleja que si bien el sector agropecuario no experiment6 las mayores tasas de contracci6n econ6mica promedio, como fue el caso de los sectores industria, comercio y construcci6n, si es la actividad que presenta el crecimiento promedio mas bajo en el periodo de la posguerra (1990-95). Es preciso destacar que comparativamente, el sector agropecuario no registr6 las mayores tasas de contracci6n econ6mica entre 1975-85, pero que ha operado bajo las condiciones mas adversas en terminos de inseguridad, limitaciones de infraestructura, productividad de la mano de obra, catastrofes naturales, precios internacionales desfavorables y otros factores, que dan cuenta del decrecimiento del sector. En efecto, el sector agropecuario no ha acompafiado el crecimiento econ6mico extraordinario que ha experimentado la economia salvadorefia de 6.1% promedio para el periodo 1990-1995. La tasa de recuperaci6n econ6mica en las actividades agropecuarias es relativamente lenta, comparada con el dinamismo de las actividades de comercio y construcci6n, las cuales crecen a tasas promedio superiores al 7% para el periodo de posguerra. Asimismo, el sector agropecuario no s6lo registra la tasa de expansi6n promedio mas baja, sino que ademas es el unico sector con tasas de variaci6n negativas entre 1991-95 (Cuadro No. 1). Cuadro o1: PIB Sectorial (tasas de crecimiento real base 19901) AAos Agropecuari Industria Construccio Comercio Servicioe PIB 1970-75 4.4 2.8 14.9 4.6 5.3 4.6 1975-80 0.5 -2.3 -2.7 -2.5 3.0 0.0 i980-85 -3.2 -4.7 -4.0 -4.8 -0.6 -2.0 1985-90 0.3 2.7 2.2 1.7 1.7 2.1 1990-95 1.5 6.0 7.5 8.4 4.5 6.1 1975-85 -1.3 -3.5 -3.4 -3.6 1.2 -1.4 1975-95 -0.2 0.3 0.6 1.7 2.1 1.3 1990 6.5 3.2 -12.8 3.1 2.8 4.8 1991 -0.3 5.9 10.3 7.0 1.9 3.6 1992 8.0 9.9 6.4 11.5 2.6 7.5 1993 -1.4 8.3 8.3 6.5 3.3 7.4 1994 -27 7.9 10.1 8.6 6.7 6.0 1995 5.1 7.0 6.0 8.0 5.0 6.1 Variaci6n promedio obtenkba aplicando la tasa cornpuesta para el periodo. 2Comprende Servicios del Gobemo. comunales y ponales, prestados a las empresas, bancos y seguros, transpolte, almac. y comunic., explotaci6n de minmas y cantems, eecticdad, gas y agua. FUENTE: Eaborado con base en cffras del BCR. El sector agropecuario reporta un crecimiento exiguo en los uiltimos quinquenios principalmente por el comportamiento negativo de las actividades de cafe, algod6n y ganaderia, asi como por no haber superado su limitada estructura productiva. La tasa de expansi6n de la actividad agricola, excluyendo cafe, es de 4.1% para el periodo 1990-95, la cual es substancialmente superior al 1.7% del total de productos agricolas (Cuadro No. 2). El cafe, la principal actividad productiva del pais y rubro de exportaci6n, registra tasas promedio de variaci6n negativas en los ultimos tres quinquenios, coyuntura desfavorable que s6lo es compartida por la actividad de algod6n, la cual desapareci6 como actividad productiva en este periodo, rubro que represent6 la segunda actividad productiva del pais. La actividad del algod6n se afect6 negativarnente por diversos factores, entre ellos: 2 Annex 1: Diagnosis conflicto militar en la zona oriental en los anos ochenta, asistencia crediticia, precios bajos de la fibra, aumento en el precio de los insumos agricolas, mal control de la calidad y uso inadecuado de los pesticidas para el control de plagas.1 De las actividades agropecuarias, la avicultura es la u4nica que reporta un crecimiento real sostenido, aiun en el periodo de crisis econ6mica nacional. Entre las actividades agricolas con un desempefio econ6mico positivo en los uiltimos dos quinquenios se tienen: granos basicos, cafna de azucar y los otros productos agricolas. Las actividades agropecuarias han perdido imnportancia relativa en la estructura productiva del pais. En efecto, el peso relativo del sector agropecuario sobre el PIB declin6 en el periodo 1975-95, pasando de 13.8% a solamente 9.0%. Las iunicas actividades que lograron aumentar su importancia relativa en el PIB son granos basicos y avicultura, expansi6n que no logra compensar la declinaci6n del resto de actividades agricolas y pecuarias. Las actividades de granos basicos y avicola tienen la caracteristica de que ambas son destinadas principalmente al mercado nacional, y corresponden a bienes esenciales en la dieta de consumo popular en El Salvador, las cuales incluso en el periodo de guerra no experimentaron deterioro, excepto los granos basicos en el periodo de 1980-85, con una leve contracci6n promedio en su actividad de -0.9%, mientras que los productos agricolas declinaron a una tasa anual promedio de -4.30/o, para este periodo. Cuadro No. 2: PIB de Actividades Agropecuarias (crecimiento real base 1990 y relaci6n del PIB1) ACTIVIDADES 1970-75 19754 198048 1985-9 1990-9 % del PIB 1975 1995 AGRICOLA 4.3 0.0 -4.3 -0.7 1.7 13.79 9.01 CaN 3.8 2.7 -4.6 -0.2 -2.3 4.89 3.01 Algod6n 6.0 -3.6 -15.8 -27.6 -100.0 3.11 0.00 Granos Basicos 4.2 1.6 -0.9 3.9 5.1 2.55 3.19 CaAa de azOcar 14.8 -4.1 6.2 -1.6 5.3 0.58 0.59 Otros prod. agrlc. 2.0 -2.0 -1.0 0.3 4.3 2.66 2.22 GANADERIA 4.3 2.9 -3.2 4.4 -2.0 2.31 1.96 AVICULTURA 9.0 5.2 2.2 3.4 5.4 0.86 1.47 SILVICULTURA 3.5 0.7 1.6 0.5 0.6 0.86 0.79 PROD. CAZA Y PESCA 1.2 -5.6 4.8 -5.0 5.2 0.64 0.47 PI8 AGROPECUARIO 4.4 0.5 -3.2 0.3 1.5 18.46 13.70 AGROP. EXCL CAFE 4.5 -0.3 -2.6 0.6 2.7 13.57 10.69 AGRICOLA EXC. CAFE 4.6 -1.5 -4.1 -1.0 4.1 8.90 6.00 PIB TOTAL 4.6 0.0 -2.0 2.1 6.1 100.00 100.00 ' Vaiaci6n promedio estimada do ls tasa compuesta del perlodo. Fuente: Elabormdo con base on cifas del 8CR. El desempefio econ6mico del sector agropecuario se puede apreciar tambien desde el punto de vista del comportamiento de las exportaciones originadas en el sector. La economia salvadorefia, previo a los afnos noventa, se ha caracterizado por su alta dependencia de las exportaciones de bienes agricolas, en especial del cafe, de ahi que se le identific6 dentro del grupo de economias monoexportadoras. El caf6 continua siendo el principal producto de exportaci6n del pais, pero su importancia relativa dentro de la canasta de bienes de exportaci6n se ha reducido. En efecto, las exportaciones de cafe constituyeron entre un 55% y 60% del total exportado por el pais para los afios setenta y ochenta, pero su importancia relativa ha declinado a solamente un 22%, pam el primer quinquenio de los ainos noventa (Cuadro No. 3). La perdida de importancia del caf6 es producto de la combinaci6n de la reducci6n del valor exportado por este producto, asi como por el incremento del valor total exportado por el pais. En los ultimos quinquenios se destaca el crecimiento de las exportaciones al Mercado Comun Centroamericano (MC CA) y de maquila, y en menor magnitud, las no tradicionales a otros mercados. Vease CENTA (1 995a), CENTA (1994), Basagoitia (1989), para mayor explicaci6n de los problemas del cultivo del algod6n. Annex 1: Diagnosis 3 Cuadro No.3: Exportaciones de Bienes (Millones de US$) No Tradicionale Aflos Tradicionales Cafe MCCA Otros Mercado MAQUILA1 TOTAL 1978_ 679.5 558.7 265.3 58.3 n.d. 1,003.1 1981-85 519.0 441.4 159.7 56.6 n.d. 735.3 1986-90 384.5 349.0 137.2 85.0 7952 638.5 1991-95 303.9 245.6 303.6 138.7 342.4 1 088.6 ~~~~~~~............................................................: : .....................................................I ......... % del totl 1978-80 67.7 55.7 26.5 5.8 n.d. 100.0 1981-85 70.6 60.0 21.7 7.7 n.d. 100.0 1986-90 60.2 54.7 21.5 13.3 n.d. 100.0 1991-95 27.9 22.6 27.9 12.7 8.9 100.0 1995 25.6 21.8 25.7 9.2 31.5 100.0 ....... i ......... ..........i ............................................................................... .... . _.... 1980 726.0 615.2 295.8 53.5 n.d. 1,075.3 1985 525.7 463.7 95.7 73.7 n.d. 695.1 1986 593.6 546.8 91.0 70.3 n.d. 754.9 1987 386.4 351.5 119.6 84.9 n.d. 590.9 1988 393.5 358.0 139.8 75.5 n.d. 608.8 1989 252.8 228.6 160.6 84.1 78.0 575.5 1990 296.2 260.2 175.0 110.3 81.0 662.5 1991 272.0 219.5 193.7 122.3 132.0 720.1 1992 214.2 149.5 248.0 125.1 198.0 785.3 1993 283.4 226.3 309.2 139.1 295.0 1,026.7 1994 324.1 270.9 340.4 154.5 430.4 1,249.4 1995 425.6 361.8 426.6 152.4 656.7 1,661.4 Tradicoale oorresponde a caf6, algodn, cahia de azacr y camar6n. I Solcluye maqula para preso4var consistencia con infonmaci6n del BCR. 2El promo es considemndo la inforrnaci6n disponible de 1989.1990. Fuente: Ebborado con base en cifras del BCR. La tendencia de pdrdida de importancia de las exportaciones de cafe respecto del total se mantiene excluyendo el valor de maquila; en este caso, su peso relativo se reduce de un promedio de 60% en el periodo de 1981-85 a solamente un promedio de 32.9% para el periodo de 1991-95. Asimismo, la economia salvadorenia registra como caracteristica una canasta de productos de origen agropecuario (cafe, algod6n, cafna de azucar y camar6n), como bienes de exportaci6n tradicional. Este subconjunto de productos agropecuarios experimenta una tendencia similar a la del cafe, que consiste en una perdida de peso relativo en las exportaciones totales y una disminuci6n en el valor exportado. Las exportaciones no tradicionales se disminuyen en importancia relativa de un promedio de 70% en el periodo de 1981-85 a solamente 27.9% para el periodo de 1991-95; la caida, considerando como total el valor exportado excluyendo maquila alcanza un promedio de 32.9%, para este ultimo periodo. En terminos de valor exportado entre el periodo de comparacion considerado, las tradicionales se han reducido en mas del 100%, pasando de un promedio de $679 millones a solo $303 millones, respectivamente (Cuadro No. 3). En la declinaci6n de las exportaciones tradicionales ha contribuido la tendencia desfavorable de los precio de cafe en los mercados internacionales, asi como la desaparici6n de la producci6n y exportaci6n de algod6n, las cuales en 1988 alcanzaron un valor exportado de $98 millones y para 1984 unicamente se exportaron $9 millones. Las exportaciones de azucar constituyen el rubro excepcional de los productos tradicionales, que registra una tendencia creciente en su exportaci6n en los ultimos quinquenios. La canasta de productos de exportaci6n se ha amnpliado en los ultimos anios, en especial por el incremento notorio registrado por las exportaciones no tradicionales destinadas al MCCA, las cuales de niveles de $95 millones en 1985 han pasado a un total de $426 millones para 1995, y para el primer quinquenio de los anos noventa report6 como promedio $303 millones. Tambien se destacan las exportaciones de maquila que reportan una tendencia creciente, en particular en los anios noventa, pasando de $78 millones en 1989 a un total de $656 millones en 1995, y con promedio de $342 millones para el periodo de 1991-95. Es preciso indicar que las exportaciones de maquila en t6rminos de servicios (exportaciones-importaciones) s6lo alcanzan un valor de $250 millones para 1995. 4 Annex 1: Diagnosis Por otra parte, el pais import6 productos agricolas por el orden de $84 millones en 1994, de los cuales corresponde un 77% a granos basicos. La economia salvadorefia registra una tendencia de declinaci6n de la importancia relativa de bienes importados de origen agricola, la cual se reduce de 6.5% respecto del total, en promedio para 1981-85, a solamente un peso relativo promedio de 3.3%, para el periodo de 1991-94. El valor importado de legumbres y hortalizas presenta una tendencia declinante, excepto para 1994 que exhibe un repunte inportante duplicando el monto internado al pais (Cuadro No. 4). Cuadro No. 4: Importaci6n de Productos Agricolas (Millones de US$) Granos Legumbres Frutas Totai Total % Bfsicos Hortalizas Agricola Importaciones' Agricola/ Total 1981 -85 27.6 22.9 9.9 60.4 934.6 6.5 1986-90 18.4 10.4 4.8 33.6 1072.0 3.2 1991 - 94 36.2 8.4 6.4 64.5 1816.3 3.3 1990 17.9 6.8 5.4 30.1 1262.5 2.4 1991 59.8 8.3 5.9 73.9 1405.9 5.3 1992 38.0 6.6 5.6 50.2 1698.5 3.0 1993 35.4 5.6 7.9 48.9 1912.2 2.6 1994 65.4 13.3 6.2 84.8 2248.7 3.8 'Total excluyendo maquila Fuente: Elaborado con base en infonnaci6n del BCR, Revista Trimestral, varios n6meros y DGEA, Anuario de Estadisticas Agropecuarias, varios nuimeros. 1.2. Area cultivada El desempeflo del sector agricola se afecta negativamente con la decisi6n empresarial de disminuir la extensi6n dedicada a los cultivos; este tipo de decisiones puede estar motivada por opciones mas atractivas de inversi6n, o incluso cambio de cultivo hacia otros mas rentables, o factores negativos como problemas fiancieros y otros. La extensi6n agricola utilizada por los principales productos agricolas tradicionales y granos bisicos reporta una disminuci6n neta de aproximadamente 20 mil manzanas, en el periodo comprendido entre 1980-95. Los productos tradicionales de exportaci6n experimentan una disminuci6n en la extensi6n agricola en forma neta de 82.8 miles de manzanas, lo cual es determinado principalmente por la desaparici6n de la actividad del algod6n. No obstante, la cafia de az2ucar y los granos basicos (maiz, fiijol y sorgo), logran en el periodo un incremento en la area sembrada, la cual no compensa la menor extensi6n utilizada por los cultivos tradicionales (Cuadro No.5). El caso del algod6n, como se sefial6, corresponde a diferentes factores que propiciaron una situaci6n de perdida de rentabilidad y condiciones de inseguridad para el desarrollo de la actividad. Por su parte, en el caf6, la menor utilizaci6n de area sembrada se relaciona con factores de abandono de las explotaciones agricolas por las condiciones adversas del conflicto, precios desfavorables del producto y un menor porcentaje dedicado a urbanizaci6n de terrenos. La cania de azuicar, registra un incremento significativo en la extensi6n sembrada del cultivo en los uiltimos quince anos, es decir, es el unico cultivo en expansi6n del grupo de los tradicionales. Es preciso hacer notar que la expansi6n del area cultivada en este producto se registra en el periodo 1991-95, lo cual ocurre en un periodo en el cual se han mantenido reglas e incentivos permanentes para este cultivo. La reducci6n del area cultivada dedicada a productos tradicionales (cafe, algod6n y caiia de azuicar) constituye la principal perdida de extensi6n agricola, siendo esta decisi6n empresarial que ha afectado negativamente el desempefno del sector agricola, asi como la disminuci6n del valor exportado. Si bien se ha incrementado la extensi6n agricola dedicada a granos basicos, en especial en maiz y frijol, el aumento en el valor agregado o la contribuci6n de estos productos no logra compensar la perdida que representa para el pais, en terminos de valor, la producci6n de cafe y algod6n, y por su generaci6n de divisas. Annex J: Diagnosis 5 Cuadro No. 5: Area Sembrada de Productos Agricolas (en miles de manzanas) PRODUCT | 1981-85 1986-90 1991-95 1980 1995 Var. absolut CAFE 253.0 244.1 234.3 265.8 234.2 -31.6 AZUCAR 48.0 50.9 70.2 38.0 70.0 32.0 ALGODON 57.9 16.0 5.3 83.2 0.0 -83.2 MAIZ 358.2 393.3 447.5 417.0 451.8 34.8 FRIJOL 79.2 90.7 105.1 75.0 87.8 12.8 ARROZ 20.1 19.2 21.6 24.0 17.4 -6.6 SORGO 164.5 176.0 188.8 170.7 191.6 20.9 TOTAL 1,073.7 1,052.8 -20.9 Fuente: Elaborado con base en cifras de la DGEA-MAG. La contribucion de productos agricolas en el PIB nacional se ha visto afectada negativamnente por la menor area cultivada de productos tradicionales de exportaci6n, la cual ha sido compensada parcialmente por la ampliaci6n de producci6n de granos basicos y cania de azucar, los cuales son de menor valor agregado. En virtud de que la extensi6n reducida en forma conjunta dedicada a productos tradicionales y granos baisicos no es significativa (0.002% del total) se considera que no es el factor mas importante para explicar la situaci6n negativa que ha experimentado el sector agricola. 1.3. Rendimiento de productos agricolas La productividad por manzana de los principales productos agricolas no reporta un cambio significativo para el conjunto de bienes tradicionales y granos bAsicos, a excepci6n del arroz y frijol. En los iiltimos quince anios el incremento promedio en la productividad del arroz es de 8.1% anual y en frijol, el rendimiento por manzana se ha incrementado a un ritmo de 7.1% anual. La productividad del arroz se ha generado principalmente en el periodo de 1991-95, mientras que, el frijol obtiene las ganancias en rendimiento en el quinquenio 1986-90. Este ultimo cultivo reporta en el periodo previo al aumento de rentabilidad un deterioro ligero que es ampliamente compensado (Cuadro No. 6) Los cambios en los rendimientos de productos agricolas pueden ser influenciados por diferentes factores, entre ellos decisiones empresariales de caracter tecnol6gico, es decir, aplicaci6n de semillas mejoradas, de fertilizantes, calidad del suelo, asi como por factores climatol6gicos (sequias, prolongaci6n de inviemos, huracanes). Estos ultimos factores negativos se han registrado en el pais, y en los anios noventa ha sido frecuente el Fen6meno del Ninlo (bajo nivel de Iluvia). Cuadro No. 6: Rendimiento de Productos Agricolas ( lmanzana) Var % Var. % Var % Var. % PRODUCTO 1981-85 1986-90 1991-95 1981-85 1986-90 1991-95 1981-95 CAFE 14.0 12.5 13.6 -2.4 8.3 -0.4 1.9 AZTCAR' 60.5 57.1 53.9 3.7 -0.5 -1.9 -1.0 ALGODON 12.0 11.2 13.0 -1.2 2.8 4.8 1.8 MAIZ 28.9 30.9 29.1 2.0 2.4 3.8 1.9 FRIJOL 11.1 11.0 12.7 -3.5 21.3 0.3 7.1 ARROZ 36.6 41.2 59.8 3.3 1.2 16.8 8.1 SORGO 16.8 15.8 22.2 0.7 -15.2 3.9 2.2 Rendimiento medido en forma de tonelada corta por manzana Fuente: Elaborado con base en cifras de la DGEA-MAG. Los principales cultivos cafe, cania de azuicar, algod6n y maiz, en los ultimos quince afios no experimentan cambios de productividad significativos, siendo el incremento promedio anual inferior al 2%, y para la cania de azuicar, que en el periodo 1980-95 prActicamente ha duplicado la extensi6n de area sembrada, experimenta rendimientos promedios negativos. Los unicos cultivos que reportan incrementos importantes en el rendimiento o productividad por manzana son frijol y arroz. 6 Annex 1: Diagnosis En general, el sector agropecuario registra tendencias negativas en su desempefio econ6mico, el cual se ha mostrado a traves de su crecimiento en t6rmninos reales, importancia respecto del PIB y dinamica de sus exportaciones e importaciones, las areas cultivadas y productividad por producto. El siguiente apartado esta dedicado a encontrar los factores explicativos del comportamiento del sector, distinguiendo entre factores de caracter end6geno, ex6geno, infraestructura y de politica econ6mica. 2. Determinantes end6genos de la actividad agropecuaria Los determinantes end6genos del crecimiento del sector agropecuario se vinculan con las condiciones o caracteristicas principales asociadas con las decisiones empresariales para el desempefio de la actividad productiva, es decir, las caracteristicas de gerencia de las explotaciones agropecuarias en cuanto el enfoque de su mercado (subsistencia o transable en segmentos de mercado intemo y extemo), utilizaci6n de insumos quimicos, tenencia de la tierra e inversiones en el recurso escaso, y otros elernentos inciden sobre la productividad y rentabiidad de las unidades de producci6n agropecuarias. 2.1. Producci6n de subsistencia no empresarial Las explotaciones agricolas, en su mayoria, se manejan con practicas tradicionales y sin enfoque de negocio o bajo las condiciones minimas de manejo gerencial apropiado de un negocio, es decir, las explotaciones agricolas (en especial de granos basicos) tienen un enfoque de producci6n de subsistencia, bajo tdcnicas de producci6n y gerencia de microempresas familiares agricolas. Ello constituye un factor fundamental que se convierte en una limitante significativa para la determinaci6n de la producci6n, selecci6n de cultivos de producci6n, y en especial sobre el rendimiento por unidad del recurso suelo, el cual es un factor escaso en el pais. La mayoria de las explotaciones agricolas tienen un enfoque de producci6n orientado a la producci6n de bienes de sustentaci6n alimenticia, es decir, para subsistir econ6micamente y no para generar ingresos en una actividad productiva empresarial. Esto significa que los pequefios productores agricolas operan al margen de un enfoque empresarial, y su principal interes es generar condiciones minimas de subsistencia alimenticia por medio de su autoproducci6n y no buscando obtener el maximo rendimiento del recurso productivo, el recurso suelo. Las explotaciones agropecuarias que tienen un enfoque empresarial corresponden a las producciones de bienes que participan del comercio internacional, siendo ellas las identificadas como bienes de exportaci6n tradicional: caf6, canta de az2ucar, arroz, camar6n y producciones no tradicionales como follajes, pifia, melones y otros productos. Los productos no tradicionales de exportaci6n tienen la caracteristica de que son proyectos empresariales con un enfoque serio de estrategia de negocio, el cual enfrenta las lirnitaciones de infraestructura, politica macroecon6mica, economia de escala y otras vicisitudes e incluso riesgos personales. Tarnbidn se destaca la labor reciente (en los anos noventa) que desarrollan las explotaciones de ganaderia dedicadas a la generaci6n de leche y came. En tal sentido, en general las explotaciones agricolas de micro y pequefios empresarios operan no como ernpresas sino como unidades de producci6n para autoconsumo, esto significa, que el enfoque de producci6n no emplea los criterios minimos de una unidad empresarial y no se emplea el recurso con orientaci6n de eficiencia. Para 1995 el 70% del area de uso agricola se destin6 a la producci6n de granos basicos, las cuales corresponden a las tipicas explotaciones de micros y pequenios empresarios agricolas y que operan como producci6n familiar de autoconsumo, sin un enfoque de negocio. La economia salvadorenia depende de este tipo de producci6n para generar el 3% del PIB, con la utilizaci6n del 70% del area dedica a fines agricolas, el tipo de suelo tambien corresponde a las calidades de menor potencial agricola empleadas en los granos basicos. Los productores agricolas dedicados a la generaci6n de granos basicos corresponden a un 68% del total en labores agricolas, de ellos, una proporci6n del 63.7% opera con una extensi6n agricola inferior a 2 hectareas, y practicamente el 82% de los productores de granos basicos poseen explotaciones con terrenos de hasta 5 Annex 1: Diagnosis 7 hectareas. En un estudio de CADESCA se determin6 que el minimo requerido en una explotaci6n de granos basicos para contratar empleo diferente al familiar es de 13 manzanas (Cuadro No. 7) 2 La extensi6n agricola total dedicada a granos basicos en un 34.5% corresponde a explotaciones inferiores a hectareas y el subconjunto de explotaciones con terrenos inferiores a 5 hectareas es equivalente a 54.5% del total empleaciones en la generaci6n de granos basicos. Ello significa que por el tamafio de la rnayoria de las explotaciones el enfoque principal de la actividad agricola es de naturaleza de susbsistencia, y es relativamente bajo la proporci6n dedicada como unidad empresarial. Los productores de granos basicos que participan en operaciones de venta de su producci6n tienen una proporci6n diferenciada en funci6n del tipo de grano basico, las relaciones mas bajas de productores que venden parte de su producci6n se registra en sorgo, en donde el rango de frecuencias se ubica entre 32 y 45% para los diferentes tarainos de explotaci6n. Para el maiz el promedio de productores que vende parte de su producci6n es de ahrededor del 55%, y no existe diferencia significativa entre los diferentes tananios de explotaci6n, excepto para las productores que cuentan con extensiones supenores a 50 hectireas por unidad, en donde los productores que venden producci6n es de un 70%. La extensi6n agricola destinada a granos basicos y que se desarrolla en exploataciones superiores a las 50 hectareas es de solamente un 11.3% del total cultivado en granos basicos (Cuadro No. 7) Cuadro No. 7: Caracteristicas de Producci6nd de Granos Basicos' Tamafio de tenencia (hectireas) 0- 2 2-5 S-20 20-50 mas de5 Total No. de productores 104,221 29,707 23.153 4,785 1,770 163,636 % respecto a productores de Granos Basicos 63.7 18.2 14.1 2.9 1.1 100.0 esect aod 6resrIcas9.7 67.6 72.4 57.1 40.6 68.7 Areadeprod. Granos 100,610 58,231 70,362 29,553 33,002 291,759 %Areadegranos bAsicos 34.5 20.0 24.1 10.1 11.3 100.0 % irea agricototaI 76.3 70.8 67.9 36.0 16.9 49.1 S de productores que venden produc. Malz 54.8 55.2 56.4 55.9 70.5 FnJol 84.4 68.7 55.3 63.7 81.1 Sorgo 39.0 45.2 36.1 32.1 35.8 Las coopeativas de la refonna agaraa son tratadas como preductores individuales En todos los casos, la tiear uilizada con cultivos asociados no se le ha cmtado -padmnmte Fuente: AID, (1989), El SalvadorAgrinculwal Land Use andLandTenure Swv. 2.2 Calidad del suelo utilizado en actividades agricolas La calidad del suelo constituye una de las limitaciones mnas importantes para el desempeiio de las labores agricolas en El Salvador. En efecto, el pais cuenta con limitaciones significativas en la calidad del suelo, por el deterioro que ha experimentado el recurso, principalmente porque no se ha desarrollado una practica apropiada de conservaci6n del recurso, y por el contrario, la utilizaci6n del suelo ha profindizado las condiciones de deterioro de este recurso. El Salvador registra una tasa del 58% del total de la extensi6n en condiciones de superficie sujeta a erosi6n, las cuales representan aproximadamente un total de 1,629 miles de hectareas bajo deterioro. Adicionalmente, los suelos con mayor presi6n de deterioro por erosi6n corresponde a las extensiones donde se desarrollan labores agricolas de granos basicos, como se muestra en el Cuadro No. 8. En tal sentido, el recurso suelo en El Salvador se encuentra amenazado por la presi6n de generar producci6n de subsistencia, ain en los terrenos que no reuinen las condiciones mimnias de protecci6n ecol6gica. 2De Calder6n V., y San Sebastian C. CADESCA/CEE (1991). "Caracteristicas de los productores de granos basicos en El Salvador." 8 Annex 1: Diagnosis Cuadro No. 8: Calidad del Suelo Utilizado en Granos Basicos y Sujeto a Erosi6n (miles de hectareas) Tirpo de suelo Total Nacional Utilizado-Granos Basicos Superficie suieta a erosi6 i 13.7 5.6 If 105.9 19.9 III 237.5 38.1 166.2 IV 339.6 88.7 266.3 V 35.9 7.1 VI 201.0 61.8 110.5 Vml 858.6 175.1 644.0 VIil 253.7 16.9 TOTAL 2,045.9 413.2 1,187.0 Fuenre Eldaxnci6n prpia con base en inoamwci6n divem del OSPA/MAG. 2.3. Generaci6n y transferencia de tecnologia agropecuaria El sector agropecuario presenta dos sectores marcadamente diferentes en varios aspectos, uno de ello es la tecnologia incorporada. Mas de 80% de los agricultores explotan fincas de menos de 3 hectireas, que comprenden cerca del 25% de las tierras cultivadas. Mientras que un 3% de los agricultores, trabajan el 44% de las tierras con vocaci6n agricola, en explotaciones de mas de 30 hectAreas. Asimismo, se encuentran grandes extensiones en propiedad de cooperativas, que surgieron como producto de la Reforma Agraria realizada en los anios ochenta. Los principales productos del pais, de acuerdo con la superficie utilizada son: maiz, cafe, sorgo, frijol, azicar y arroz. Los minifundios o propiedades pequenas por lo general se utilizan en el cultivo de granos basicos, y presentan niveles de productividad bastante bajos, que reflejan la condiciones rudimentarias con que son explotadas y la escasa tecnologia que utilizan. Las explotaciones mAs grandes son destinadas principalmente al cultivo de los exportables no tradicionales: frutas y hortalizas, o son ocupadas para la explotacion del hato ganadero. En estas propiedades se observa una mayor tecnologia y, por lo mismo, niveles de productividad mucho mnas altos. En cuanto al sector reformado, varias de las haciendas que fueron expropiadas se han dedicado al cultivo de granos bisicos, manteniento una buena proporci6n de sus tierras sin utilizar. Los ingenios que en la pasada decada y en los primeros anios de los noventa fueron explotados en forrna cooperativa, estan siendo actualmente privatizados, presentando un rezago tecnol6gico importante, y bajos niveles de productividad. El sector agropecuario ha sido uno de los sectores mnas afectados; primero, por la Reforma Agraria, y luego por el conflicto armado que duro na-s de doce anios, sucesos que incidieron grandemente en el deterioro de las explotaciones, y en su rezago tecnol6gico. Los incrementos ganados en productividad se han dado en forma demasiado lenta, y continian siendo bajos (Cuadro No. 6), y si en algunos anios se observ6 un crecimiento de la producci6n de algunos bienes agropecuarios, este hecho estA ligado mAs que todo a un incremento en la superficie cultivable, que a niveles de productividad naAs altos. La investigaci6n y transferencia de tecnologia en el sector agropecuario se remonta a hace varias decadas, pero los resultados alcanzados han sido bastante limitados, en parte condicionados a los sucesos mencionados, pero tambien por deficiencias en las instituciones estatales originados de la falta de coordinaci6n, poco personal calificado, escasez de fondos, falta de un plan de trabajo claro, excesiva burocratizaci6n, y poca participaci6n de los beneficiarios en el disenio de los programas. En la presente decada, el gobiemo ha tratado de dar impulso a las actividades de investigaci6n y transferencia de tecnologia en el sector agropecuario, para lo cual ha realizado una serie de transformaciones en su marco legal e institucional. Como primer paso, se ha concentrado en la asistencia tecnica basicamente del pequeflo y mediano productor, que generalmente tiene mas dificultades en el acceso a la nueva tecnologia, y ha dejado las actividades de invesfigaci6n del cafe y azucar al sector privado, a traves de la Asociaci6n de Productores de Cafe (PROCAFE), y el Centro de Investigacion de la Cafia (CENCICA), cultivos correspondientes principalmente a unidades de extensiones mayores. Al mismo tiempo, la Fundaci6n Salvadorefia pam el Desarrollo Econ6mico y Annex 1: Diagnosis 9 Social (FUSADES) se ha dedicado a la experimentaci6n de cultivos agricolas no tradicionales de exportaci6n, que comprenden una variedad de frutas y legumbres. Por el lado gubernamental, las principales instituciones involucradas con actividades de desarrollo de nuevas tecnologias para el sector agropecuario, son el Ministerio de Agricultura y Ganaderia (MAG), el Centro Nacional de Tecnologia Agropecuaria y Forestal (CENTA), y el Consejo Nacional de Ciencia y Tecnologia (CONACYT). Tambien puede mencionarse a la Escuela Nacional de Agricultura (ENA), dedicada a la forrnaci6n de profesionales con conocimientos te6ricos y practicos sobre agricultura, y el Banco de Fomento Agropecuario (BFA), instituci6n que en algunos momentos ha condicionado el otorgamiento de creditos al uso de semilla mejorada. Ademas, el CENTA ha recibido colaboraci6n de parte de algunos de los centros internacionales de investigaci6n del sistena CGIAR, como el CIMMYT para el maiz, y el CIAT para el fiijol. El Ministerio de Agricultura y Ganaderia realiza su labor investigativa en favor de la actividad agropecuaria y pesquera a traves de sus dependencias: Direcci6n General de Recursos Naturales Renovables (DGRNR), el Centro Nacional de Desarrollo Pesquero (CENDEPESCA), la Direcci6n General de Riego y Drenaje, la Direcci6n General de Ganaderia, y la Direcci6n de Sanidad Vegetal y Animal. El CONACYT tiene como objetivo general formular y dirigir la politica nacional en materia de desarrollo cientifico y tecnol6gico, y tiene dentro de sus funciones asesorar al gobiemo en la programaci6n de la inversi6n y en la asignacion presupuestaria de actividades de investigaci6n. Algunos de sus programas relacionados con el sector agropecuario son: Red de Lacteos para la Regi6n Centroamericana, Plan de Desarrollo Tecnol6gico del Sector Alimentos, y Gestion Total de la Calidad en la Industria de Alimnentos, este iiltimo realizado en los sectores lacteos y canico. Sin embargo, la instituci6n que ha estado mucho ma.s involucrada en las actividades de investigaci6n y extensi6n de nueva tecnologia para el agro es el CENTA. Esta instituci6n se cre6 en 1969 como una dependencia del MAG. En 1993, la instituci6n se transform6 en un organismo aut6nomo, buscando superar muchas de las limitaciones, y mejorar el alcance de sus investigaciones. En su ley de creaci6n se establece su responsabiidad sobre la investigaci6n de cosechas y ganado, y que sus beneficiarios deberan ser pequefios y medianos agricultores. El CENTA todavia enfrenta los problemas de destinar parte de sus presupuestos para sufragar gastos administrativos, lo que limita la asignaci6n de recursos a actividades propiamente de investigaci6n. Asimismo, s6lo un niumero muy reducido de su personal posee titulo de maestria o de doctorado, lo que evidencia la falta de personal adecuadamente capacitado para realizar tareas de investigaci6n. Pero se ha tratado de mejorar muchos aspectos, y sobre todo de vincular la actividad investigativa con la actividad de extensi6n, cuyo debil nexo ha limitado grandemente en el pasado la aplicaci6n de sus investigaciones. La dinamica de trabajo que esta observando el CENTA se apoya en la organizaci6n de grupos vecinales, y en la utilizaci6n de extensionistas que logren dar una mayor cobertura a sus programas. En sus inicios, el incremento en los rendirnientos logrados por el CENTA estuvieron bastante relacionados con el uso de fertilizantes y otros agroquimicos, nuevas semillas y nuevas practicas de manejo. Aunque ahora continua trabajando con la experimentaci6n de semilla mejorada, ha dejado la parte correspondiente a su comercializacion a la empresa privada. La instituci6n hace investigaci6n sobre granos basicos, ganado y forraje, frutas y vegetales, fibras y semillas oleaginosas, con especial enfasis en los granos basicos. En cuanto a la participaci6n de las universidades, estas han reducido en fornna notoria el papel en la generaci6n de nueva tecnologia. Su actividad se ha centrado en la ensefnanza, en menoscabo de la experimentaci6n y la investigaci6n. Por otra parte, las bajas remuneraciones han provocado la fuga del personal mas capacitado y apto para estas actividades. Asimismo, las carreras como ingenieria agron6mica o economia agricola, han visto seriamente memiado el nimero de estudiantes, ante las pocas oportunidades y los bajos 10 Annex 1: Diagnosis iicentivos monetarios en el mercado de trabajo, junto con el excesivo niumero de materias de prograrna de estudios en algunos casos. En uno de los centros universitarios de mayor prestigio, este tipo de carreras han sido incluso clausuadas. Cuadro No. 9: Asistencia Tecnica Recibida y Extensi6n con Vocaci6n Agricola % Asistencia Areas de uso agricola Departamentos T6enical (Miles de ha) Ahuachapan 22.7 69.0 Santa Ana 12.0 79.4 Sonsonate 20.0 82.6 Chalatenango 11.8 27.9 La Libertad 20.0 77.7 San Salvador 28.6 31.2 Cuscatlan 60.0 31.7 La Paz 18.6 70.1 Cabafas 35.3 24.5 San Vicente 7.7 47.9 Usulutan 27.6 123.2 San Miguel 11.1 81.4 Morazian 13.0 24.7 La Union 7.5 34.6 Total general 19.2 805.8 ' Respuesta afunativa de asistencia t6cnics en la encuesta de FUSADES. Fulnte: FUSADES- Banco Mundial. Encuesta de Desarnollo Rural 1996 y Hidraulica Intemacional, S.A. de C.V. Mejoramiento de vias nrsales y su impacto en l desarrollo agropecuario de El Salvador 1996. En geral, aunque se han dado algunos pasos inportantes en el impulso a la investigaci6n, los frutos de la misna s6lo ban beneficiado a un escaso niumero de agricultores, tal como lo evidencia la encuesta de desarrollo uWal en la parte de tecnologia, donde en 1995 menos del 20% de los agricultores encuestados reportar6n haber recibido asistencia tecnica, siendo los propietarios de parcelas de menos de dos hectareas los que tuvieron menos acceso a esta asesoria (Cuadro No. 9). Ademnas, los resultados positivos de los programas de introducci6n de nuevas variedades de semillas y nuevos metodos de cultivo, en un gran nuimero de casos han sido temporales, ya que por la filta de concientizacion sobre las ventajas y las mayores ganancias que representa para el agricultor la incorporaci6n de nueva tecnologia, al dejar de recibir la asistencia y los incentivos los beneficiarios del programa, se vuelve a las viejas t6cnicas de cultivo. Por otra parte, los fondos que el gobiemo ha destinado a la investigaci6n tecnol6gica, son claramente insuficientes, considerando que en teminos reales se encuentran por debajo de los montos invertidos por las instituciones a finales de los setenta (Cuadro No. 10). Cuadro No. 10: Gasto Publico en Instituciones de lnvestigaci6n y Desarrollo (Millones de colones constantes base 1978) DGRNR' CENTA2 DIGESTYC' Periodos Gasto Real Gasto Real Gasto Real 1978 - 1985 13.7 13.6 3.4 1981 -1985 8.7 10.0 1.4 1986- 1990 1.8 5.5 0.7 1991 - 1994 2.4 3.4 3.8 1994 4.0 3.4 1.2 Direccion General de Recursos Naturales Renovables; 2 Centro Nacional de Tecnologia Agropecuaria 3 Direcci6n General de Estadistica y Censos Fuente: Diario Oficial, ailos 1978-94 S6lo mediante la investigaci6n cientifica, puede garantizarse el crecimiento del sector agropecuario, que siendo clave para el desarrollo econ6mico del pais, se encuentra peligrosamente limitado por la estrechez territorial del nuismo. Un objetivo central de toda estrategia agropecuaria debera ser incrementar la productividad, la cual, por las dimensiones del pais, marcara el potencial de expansi6n de esta actividad. 2.4. Utilizacion de insumos quimicos en la agricultura La utilizaci6n apropiada de insunmos quimicos constituye uno de los factores end6genos relevantes para explicar el desemperlo de las actividades agricolas, asi como para preservar las condiciones ecol6gicas del recurso suelo. La agricultura en El Salvador muestra una fuerte dependencia de insumos quimicos. Tradicionalmente, el Annex 1: Diagnosis 11 uso de agroquiricos en el pais ha sido mas elevado que en el resto de los paises del istmo centroamericano. Coma se puede apreciar en el Cuadro No. 11, el consumo de fertilizantes en superficie agricola por hectarea, en los anIos 70 y 80, fue significativamente nas alto en El Salvador que en los paises vecinos. Cuadro No. 11: Consumo de Fertilizantes en Paises Seleccionados Consumo en superficie Consumo en tierras de agricola, por hectirea (kg) labranza, por hectirea (kg) Palses 1973 1988 1973 1988 1 El Salvador 87.1 60.8 168.7 11.3 Costa Rica 31.3 35.6 130.6 191.1 Guatemala 17.7 39.4 31.3 68.6 Honduras 6.2 8.7 15.1 21.1 Nicaragua 9.5 11.0 44.7 56.8 Panama 12.9 20.2 41.1 67.0 Estados Unidos 40.6 41.2 92.8 95.6 Fuente: Arrnuolles Boutet, 1992 El principal factor que explica el patr6n de uso de agroquimicos en el pais es el uso intensivo de la tierra. La presi6n demogrifica sobre el recurso tierra ha significado un uso mas intensivo de los recursos en el campo que en los demrns paises de la regi6n. Por otro lado, el trabajo del Centro Nacional de Tecnologia Agropecuaria y Forestal (CENTA) se concentr6, desde los afios 60, en la introducci6n de nuevas variedades de granos basicos de mayor rendimiento que son mas intensivas en el uso de insumos quimnicos que las variedades "criollas". Los paquetes tecnol6gicos conocidos colectivarnente como la "revoluci6n verde" han perrmitido aumentar los rendimientos de los granos, basicos pero a la vez han incrementado la dependencia sobre estos insumos. El uso intensivo del recurso suelo en El Salvador tambien ha resultado en la degradaci6n del recurso. A medida que la erosi6n se ha intensificado se ha recumdo a insumos quimicos como los fertilizantes para co*msar la perdida de productividad de los suelos. Lo mismo se puede mencionar sobre el uso de insecticidas. El uso intensivo de los recursos naturales en las areas rurales del pais tambien han modificado los equilibrios biol6gicos que mantienen especies que constituyen plagas para el sector agropecuario bajo control; por lo cual se observa una crociente dependencia de insumos quimicos para mantener las plagas bajo control. Informaci6n recopilada recientemente a traves de encuestas, muestra que a nivel nacional la percepcion de los agricultores es que la productividad de los suelos en los ultimos afios ha disminuido. Tamnbi6n se considera que el problema de las plagas es mayor actualmente que hace 10 anios. El Cuadro No. 12, contiene las respuestas de agricultores encuestados sobre estos temas. Estas percepciones explican por que puede haber una creciente dependencia sobre insumos quiniicos en la agricultura salvadorefia. Cuadro No. 12: Tendencia de Rendimiento y Plagas en Explotaciones Agricolas, 1996 Tendencia de rendimient Tendencia de plagas Mis alto de hace diez aflos 8.9% 77.5% Mis bajo de hace diez aflos 56.6% 6.0% Igual que hace diez aflos 24.8% 8.3% Igual, pero se necesita utilizar mas fertilizantes (pdrdidas) 9.6% 8.3% Fuente: FUSADES, Encuesta de Desarrollo Rural, 1996 2.4.1 Uso actual de insumos quimicos Infonnaci6n reciente sobre el uso de insumos quimicos en el campo muestra una fuerte dependencia sobre estos productos. En terrninos generales, la mayor parte de los agricultores encuestados (mas del 78%) indicaron que ulizaron fertilizantes en 1995. El Cuadro No. 12, resume estos resultados para los distintos cultivos. Como se puede apreciar en este cuadro, los cultivos tradicionales, como el maiz, el frijol, el arroz, el cafe y la caila de azucar mantienen una alta dependencia sobre este tipo de insuno. Para el caso del maiz, se observa que el 97.9% de los encuestados utilizaron fertilizantes en 1995. Dentro de los granos basicos, el cultivo del sorgo parece ser el menos intensivo en el uso de estos insumos. Sin ernbargo, es posible que esto se explique porque a menudo el sorgo se cultiva en asocio con el maiz o despues de la cosecha del maiz y, por lo tanto, las nocesidades de fertilizaci6n pueden ser menores. 12 Annex 1: Diagnosis Pam varios cultivos no tradicionales tambien se report6 una alta dependencia de fertilizantes. Todos los agricultores encuestados que se dedican al cultivo de mel6n, sandia, pepino, chile verde y ejote utilizaron estos insumos. Por otro lado, los agricultores que se dedican a cultivos como el guineo de seda, mango, maraft6n, pltao y henequen, reportaron que no utilizan estos insumos. Otro punto de interes sobre el uso de fertilizantes los fertilizantes organicos. Las respuestas de los agricultores encuestados indican que la gran mayoria de ellos utilizan fertilizantes inorgarnicos. Como se puede apreciar en el Cuadro No. 13, solamente el 2.5% de los agricultores que utilizan fertilizantes para sus cultivos usan ferilizantes organicos. Cuadro No. 13: Utilizacion de Fertilizantes Producto % Utiliz6 Tipo de Fertilizante Fertilizantes % Organico % lnorganico Caf 73.1 10.5 94.7 AzOcar 75.0 0.0 100.0 Malz 97.9 1.4 98.9 Frijol 76.6 0.0 100.0 Arroz 90.0 0.0 100.0 Sorgo 46.0 0.0 100.0 Total | 78.6 2.5 99.0 Fuente: FUSADES- Banco Mundial. Encuesta de Desarrollo Rural 1996 Aunque las muestras de agricultores son pequeflas, para cultivos como el coco y los citricos, las respuestas de estos indican que el uso de fertilizantes organicos es comun en estos cultivos. De los cultivos tradicionales, s6lamente el cultivo del cafe parece hacer un uso significativo de fertilizantes organicos, factor que puede ser explicado por las exigencias de algunos mercados internacionales de este producto. Entre 1992 y 1996 se ha registrado un aumento muy significativo en el cultivo de cafi organico, el area ha pasado de 0.3 miles de hectireas a 7.4 miles, de acuerdo con infonnaci6n de UCAPROBEX. Existen esfuerzos por trasladar experiencias similares a cultivos como el ajonjoli y el marani6n. El Cuadro No. 14, muestra el gasto en fertilizantes por manzana para los distintos cultivos. De los cultivos tradicionales, se observa que el maiz, el arroz y el cafe son los mats intensivos en el uso de fertilizantes por manzana. Sin embargo, estos parecen ser superados por algunos cultivos no tradicionales como la sandia, tomate de mesa, chile verde y maiz amarillo. Cuadro No. 14: Costos de Insumos Qurmicos Promedio (¢ por manzana) Producto Fertilizantes Insecticidas Herbicidas Fungicidas Cafe 847.12 196.13 68.75 131.40 AzOcar 208.83 215.50 178.42 - Malz 642.97 124.06 180.52 135.35 Frijol 459.02 106.44 191.78 73.44 Arroz 983.01 126.30 388.61 164.00 Sorgo 356.16 96.64 125.80 8.50 Total | 619.78 154.56 180.87 218.90 Fuente: FUSADES-Banco Mundial. Encuesta de Desarollo Rural 1996 Pama el caso del maiz se puede deducir que el uso de fertilizantes representa aproximadanente un tercio de los ingresos brutos del cultivo. Si consideramos una producci6n promedio de entre 25 y 30 qq por manzana y un precio al productor de 075/qq3, podemos concluir que el gasto en fertilizantes es equivalente a entre 28 y 34% de los ingresos brutos por nanzana del agricultor. Esta relaci6n disminuye para el frijol; si suponemos una producci6n de 12 qq/mz y un precio al productos de 0250/qq, vemos que el gasto en fertilizantes representa aproximadamente el 15% de los ingresos brutos por manzana del cultivo. 3 Precios del primer timestre de 1996. Annex 1: Diagnosis 13 El uso de insecticidas, herbicidas y fungicidas para el control de plagas es menor que el uso de fertilizantes, al comparar el numero de agricultores que reportaron utilizar este tipo de insumos. Para e caso de los insecticidas, el 46.4% de los agricultores los utilizaron en 1995, siendo su uso mis coman pam cultivos tradicionales como el maiz, frijol y arroz. Al igual que para el caso de los fertilizantes, el cultivo de productos no uadicionales como sandia, mel6n, pepino, ejote y maiz amarillo, tambi6n depende de insumos quiniicos pam su producci6n. Sin embargo, es importante mencionar que las muestras de agricultores que se dedican a estos cultivos son mucho menores que para los cultivos tradicionales. Cuadro No. 15: Utilizaci6n de Insecticidas, Herbicidas y Fungicidas Producto Insecticidas Herbicidas Fungicidas Cafe 17.3 3.8 9.6 AzOcar 50.0 50.0 0.0 Matz 62.5 77.4 2.1 Frijo 48.1 40.3 9.1 Arroz 50.0 90.0 20.0 Sorgo 21.0 37.0 1.0 Total 46.4 51.5 5.2 Fuente: FUSADES- Banco Mundial. Encuesta de De5arrollo Rural 1996 El patr6n de uso de herbicidas es similar al reportado para los insecticidas, observandose que los agricultores que se dedican a cultivos como el maiz y el arroz tienen una alta dependencia de estos productos. Sin embargo, el uso de herbicidas en el cultivo de productos no tradicionales parece ser menos comun que el uso de insecticidas. Por otro lado, el uso de fingicidas es mucho menos comun que el uso de insecticidas o herbicidas, siendo imporante para cultivos como el arroz, el mel6n, la sandia, la papaya, el tomate de mesa, el pipian y el chile verde (Cuadro No. 15) Tnformaci6n reciente sobre el gasto en estos insumos muestra que en este rubro tambien es elevado. Para el caso del maiz, el gasto promedio en iisecticidas y herbicidas por manzana, pam quienes utilizan estos productos en sus cosechas es de aproximadamente 0300. Esta cifra representa un poco menos de la mitad de lo invertido en ltiFlizantcs. Utlizando los mismos supuestos que para el caso de los fertilizantes, esta cifra representa entre el 13 y el 16% de los ingresos brutos por manzana dedicada a este cultivo. Al sumar los gastos en fertilizantes, insecticidas y herbicidas, se observa que el desembolso en insumos quinicos puede se equivalente a mas del 45% de las ganancias brutas de los agricultores que se dedican al cultivo del maiz. 2.4.2 Importaciones y exportaciones de insumos quimicos El Salvador importa la mayor parte de los insumos quimicos que consume para su producci6n agricola. Sin embargo, El Salvador exporta varios plaguicidas. Como se muestra en el Cuadro No. 16, durante los ultimos diez affos, la cantidad de fertilizantes e insecticidas importados al pais ha aumentado significativamente. Cuadro No. 16: Consumo Interno y Flujo Comercial de Agroquimicos (en millones de kilogramos, miles de hectareas y kilogramos por hectarea) CONSUMO DE AGROQUIMICOS Importaci6n fertilkzantes' Exportaci6n Plaguicida Area Ag. FertJAg PestJAg (millones de kg) (millones de kg) AMo (miles ha) (Kolha) (KglAg) 1986 700 305.53 3.80 214.1 1.6 1987 753 289.59 2.90 218.1 2.3 1988 758 243.99 3.37 184.9 0.9 1989 756 312.60 3.69 236.4 1.2 1990 779 44.60 2.81 34.8 0.9 1991 797 270.03 4.08 215.4 1.6 1992 847 257.50 6.56 218.2 1.5 1993 819 342.90 4.08 281.0 1.4 1994 796 346.87 5.08 276.3 1.1 Conespond. a abonos naturales y qulmicos Funte: Elbborado con base on informaci6n del 5CR y DGEA 14 Annex 1: Diagnosis Al dividir las importaciones de fertilizantes e insecticidas por la extensi6n de tierra utilizada para la agriculura se nota que la relaci6n cantidad de insumos por hectarea estA aumentando, lo cual muestra que se esta haciendo un uso mas intensivo de los recursos disponibles o que se estA cambiando a cultivos que utilizan mas insumos quimicos por hectarea. 2A.3. Impacto de los agroquimicos sobre la producci6n agricola La informaci6n disponible sobre la rentabilidad de varios cultivos muestra que el uso de agroquiniicos es un factor decisivo pam las ganancias del agricultor. Pam el caso del maiz, por ejemplo, el uso de agroquimicos puede incrementar significativanente las ganancias por hectarea, pasando de perdidas netas a una ganancia de 32 centavos por cada d6lar invertido. Resultados similares se obtienen para cultivos como el frijol y el cafe. El Cuadro No. 17 muestra el impacto que el uso de agroquimicos puede tener sobre las ganancias de los agricultores. Cuadro No. 17: Impacto del Uso de Insumos Qu[micos Sobre la Rentabilidad de Cultivos Seleccionados Cultivo Rentabilidad por US S invertido No tecnificado Tecnificado' Malz -0.24 0.32 Frijol 0.03 0.08 Cafe 0.41 0.72 CaAa de azbcar 0.00 1.75 Tomate 0.45 1.07 Esto incluye el uso de insumos qulmicos y semilla mejorada Fuente: Elaborado por FUSADES, con base en datos de Bayer Sin embar8o, existe poca informaci6n sobre la posibilidad de sustituir estos insumos quiuicos por insumos organicos y sobre el impacto de insumos organicos sobre las ganancias de los agricultores. Algunas estimaciones prelininares sugieren que el uso de insumos organicos si puede resultar en mejoras sustanciales para el agricultor, especialmente a largo plazo. Sin embargo, a pesar de los esfuerzos de varias ONGs, existe poca actividad en matexia de investigaci6n y extensi6n, especialmente por parte del Centro Nacional de Tecnologia Agropecuaria y Forestal (CENTA), la principal instituci6n de extensi6n en el pais. En vista de esta falta de informaci6n es posible concluir que el uso de insumos quimicos en el agro salvadorefio continuara siendo muy difundido en el corto y mediano plazo. Anteriormente se mencion6 que el uso de insumos quimicos es muy comun en el sector agricola y que su uso es un factor significativo pam la rentabilidad del agro. El hecho que existe poca informaci6n sobre el potencial de uso de insumos organicos como compleomento o sustituto de insumos quimicos, implica que este patr6n de uso no cambiara significativamente en el corto plazo. Ante esta situaci6n y bajo un contexto de apertura comercial, la disponibilidad de insumos quiimicos y su precio se vuelven factores clave para la competitividad del sector. Cuadro No. 18: Precios de Fertilizantes por Quintal (promedio anual de 1995 en $) Urea Sulfato Pais Perlada de Amonio 15-15-16 Costa Rica 16.24 11.49 12.82 El Salvador 16.32 8.05 13.79 Guatemala 14.00 9.96 14.37 Honduras 12.83 10.03 12.86 Mexico 6.73 4.19 6.37 Nicaragua 17.38 10.93 13.59 Panama 14.71 10.60 - Rep. Dominicana 13.66 8.48 10.88 Fuente: Elaborado por FUSADES, con base en informaci6n del CORECA Como se muestra en el Cuadro No. 18, los precios de los insumos seleccionados no difieren sustancialmente de los precios observados en los demas paises de la regi6n. Para el caso de la urea perlada, el precio en El Annex 1: Diagnosis 15 Salvador es similar al observado en Nicaragua y Costa Rica, pero superior a los precios de venta en Guatemala y Honduras. Sin embargo, para el sulfato de arnonio, el precio en El Salvador es el mas bajo de centroamerica. 2.4.4 Impacto del uso de agroquimicos sobre los recursos naturales A excepcion del impacto de los agroquimicos utilizados para el cultivo del algod6n en los anios 70 sobre el medio ambiente y la salud humana, existe poca informaci6n sobre la relaci6n entre el uso de estos productos y la calidad ambiental. Se cree que uno de los principales problemas ambientales es el lavado de fertilizantes y otros insumos por escorrentia, lo cual contamina los recursos hidricos y trastorna los ecosistemas acuAticos. Por otro lado, se considera que hay poca educacion sobre el manejo apropiado de plaguicidas en el campo, lo cual resulta en problemas de salud humana como envenenanientos y enfermedades causadas por la toxicidad de estos productos. La acumulaci6n de plaguicidas t6xicos en las cadenas alirnenticias tambien ha sido sefialado como un problema ambiental serio. Sin embargo, existe poca informaci6n precisa sobre la magnitud de este problema. Encuestas realizadas en el primer trimestre de 1996 a nivel nacional, muestran que el uso de agroquirnicos en el pais es elevado, mostrandose una alta dependencia en cultivos tradicionales como el maiz, el frijol y el cafe. Aunque las muestras de los agricultores que se dedican a cultivos no tradicionales encuestados son menores, la informaci6n disponible indica que para estos cultivos tambien existe una alta dependencia de insumos quimicos. Esto refleja, en parte, los esfuerzos tradicionales de diseminar paquetes tecnol6gicos con veriedades de mayor rendimiento que dependen de una mayor cantidad de insumos quimicos. Sobre los fertilizantes, es posible concluir que el uso de insumos organicos es poco comuin, mientras que el uso de insumos quiniicos es muy difundido. El poco conocimiento sobre el uso de insumos organicos podria ser un factor que explica su bajo nivel de uso. Asimismo, la inexistencia de analisis sobre el impacto del uso de insumos orgAnicos sobre la rentabilidad de distintos cultivos podria explicar el hecho que estos no se utilicen de una manera mrs amplia. En vista de la fuerte presi6n que existe sobre los recursos naturales en El Salvador y el posible impacto negativo de estos productos sobre la calidad ambiental, este deberia ser un punto prioritario de investigaci6n en el Area de productividad y rentabilidad agricola. El gremio de agricultores que parece estar aprovechando las oportunidades presentadas por el uso de insumos organicos es el de la caficultura, observandose un aumento significativo en el area dedicada al cultivo del cafe organico. Para orientar la politica agraria en materia de insumos quimicos y velar por los intereses de largo plazo del sector, tambien debe ser un punto prioritario de investigaci6n el impacto del uso de insumos quimiicos sobre la salud humana y ecosistemas naturales. 2.5. Distribuci6n y limitaciones del mercado de tierras La economia salvadorenia ha experimentado grandes transformaciones en el regimen de tenencia de la tierra, derivado de un proceso de Reforma Agraria y programas especiales de distribuci6n de tierras como parte de los Acuerdos de Paz entre las partes involucradas en el conflicto. En tal sentido, dos programas asociados con el proceso de guerra han originado cambios dramAticos en la tenencia de la tierra, al punto que en el pais no existen explotaciones privadas con extensiones superiores a las 500 hectareas; ello constituye uno de los problemas relevantes no s6lo por las limitaciones que significa el no contar con explotaciones que utilicen las ventajas de las economias de escala, sino ademAs, porque la Reforma Agraria afect6 una proporci6n del 20% de la tierra con mejor vocaci6n agropecuaria del pais. En virtud de ello, el factor de expropiaciones y distribuci6n de la tierra en el pais incorpor6 el elemento de inseguridad juridica, bajo nivel de inversi6n privada en el sector agrario, el surgimiento de procesos de cartera morosa y de condonaci6n de la misma, asi como, debilidad en la capacidad ernpresarial para la gerencia de las explotaciones cedidas en formia de cooperativas, que incidi6 negativamente en el desernpenlo del sector agropecuario desde el inicio de este proceso en marzo de 1980. 16 Annex 1: Diagnosis La Refonna Agraria comprendi6 la fase I, en la cual se expropiaron las explotaciones con extensiones superiores a las 500 hectareas, que afect6 aproximadamente 215 miles de hectAreas, lo cual represent6 2/3 del total comprendido en el proceso de Refornma Agraria (Decreto Legislativo No. 154, marzo de 1980)4, las cuales se distribuyeron cooperativas estatales por un total de 346 cooperativas. El segundo componente de importancia en el proceso de Reforma Agraria con la fase III, que se desarrolla a traves de la adjudicaci6n (traspaso) de las parcelas a sus cultivadores directos, en diciembre de 1980 (Decreto Legislativo No. 207), identificados como Finateros, por su la instituci6n creada para administrar, esta fase Financiera Nacional de Tierras Agricolas (FINATA). La operaci6n de la fase III comprendi6 un total de 47 miles de hectAreas y como beneficiarios directos participaron 80 miles de personas. El tercer componente del proceso de distribuci6n de tierras se gener6 a partir de 1992 con la implementaci6n de los Acuerdos de Paz, en el cual se consigna entregar tierras a los excombatientes de ambos bandos, asi como a los pobladores seguidores del moviniento de izquierda, y que ocupan propiedades ilegalmente. Bajo esta operaci6n de transferencia de tierra se asigna un total aproximado de 65.6 miles de hectareas y el nurmero de beneficiarios directos es de cerca de 30 mil personas. Las explotaciones agricolas comprendidas de la fase I, en la cual se organizaron en 346 cooperativas han sufrido un proceso de descomposici6n, habiendose desintegrado hasta fines de 1992 12 cooperativas y transformado en propiedades de tenencia individual; asimismo, se encuentran 10 bajo el sistema mixto, es decir, individual y cooperados, las cuales representan en su conjunto un 7% del total de las asociaciones colectivas de producci6n. Adicionalmente a los problemas de fraccionamiento y desmembramiento de las cooperativas, se ha experimentado el problema de extensiones agricolas sin cultivar, este tipo de casos para la cosecha 1993/94, era de aproximadamente 10.3 miles de hectAreas, que representa el 6% de lo asignado en la fase I. Adicionalmente, bajo la categoria de tierras en abandono o sin cultivar por parte de los beneficiarios de transferencia de tierra de los Acuerdos de Paz se tiene un total de 5.2 miles de hectareas, que alcanza la proporci6n del 27.7% de las tierras entregadas a los excombatientes y pobladores de zonas exconflictivas. Es significativa la proporci6n de tierra que en El Salvador se ha encontrado en calidad de abandono o bien sin cultivar y en poder de los beneficiarios de los programas de redistribuci6n de la tierra, lo cual ha perjudicado el desempeflo de la actividad agropecuazia (Cuadro No. 19) Los niveles de utilizaci6n de la capacidad productiva del recurso suelo en poder de las cooperativas es mermado por diferentes factores, entre ellos se tienen como los principales: la falta de credito, areas en descanso, fen6menos naturales, la incapacidad empresarial, y la perdida general de competitividad del sector agropecuario. Las limitaciones de credito se derivan de una trayectoria constante de los afnos ochenta de cartera morosa, la cual ha tenido que refinanciar el Estado e incluso condonar (Cuadro No. 20). 4Anexo No. 3,comprende una resefia de principales acuerdos relacionados con el proceso de Reforma Agraria. Annex 1: Diagnosis 17 1993194 USO DEL SUELO 1988/89 1989190 1990/91 1991192 Colec. Tra FAES FMLN Area.agrf9ola sin cultivar 19.7 14.5 14.0 16.5 10.3 0.150 5.1 TOTAL 199.5 191.7 191.1 189.3 171.2 0.373 18.4 ...............i i ......I......... .............. F .................... i i ..................... ... .......... ......... .... Cooperativas Censadas 328 325 327 319 294 4 56 Fuente: OSPA. XII Evaluaci6n del proceso de la Reforma Agrada 1993; IX Censo de las cooperativas de la primera etapa de la Reforma Agraria, 1994. Las inversiones en compra de tierras por el sector privado en los iiltimos anos han sido relativamente pequeinas, y el principal operador del mercado ha sido el Estado, desarrollando diferentes fases de compras para proceder a las transferencias de tierras. No obstante, el sector privado ha desarrollado diferentes inversiones en extensiones rurales para la implementaci6n de proyectos de construcci6n de parques o zonas francas dedicados a maquilas, proyectos de urbanizaci6n y actividades turisticas. Cuadro No.20: NOmero de Casos de Cooperativas y Areas que no se Trabajaron, por Tipo de Raz6n para no Cultivar Cosecha 1993/94 (Superficie en miles de hectareas) TOTAL RAZON CASOS AREAS Falta de cr6dlto 33 4.4 Area en descanso 22 2.1 Fen6menos naturales 10 0.3 Incapacidad empresarial 7 0.4 Falta de recursos 5 0.5 Recientemente activada 5 0.1 Otros 26 2.4 TOTAL 108 10.3 Fuente: OSPA-DISE. Noveno conso de las cooperativas de la la etapa de la Reforma Agraria, 1994. El Gobiemo, en las operaciones de compra de tierras, imput6 valores a las propiedades conforme el valor declarado para fines fiscales por las personas expropiadas. En las primeras transacciones el Estado compro 10 haciendas, con un precio promedio por manzana de 10.7 miles de colones ($4.2 miles por manzana). Las haciendas administradas en formna de cooperativas experimentaron cartera morosa que se adiciona al valor de la propiedad, para derivar el valor de recuperaci6n monetario de equilibrio en una operaci6n de compra-venta, se llega a un promedio de 14 miles de colones que se deben de recuperar en promedio por manzana, en la cual invirti6 el Estado. La valoraci6n comercial de las haciendas es funci6n de diversos factores, entre ellos: accesibilidad, zona exconflictiva, distancia a puntos de interes comercial o costa, y otros. El rango de precios para 1995 en la muestra de las 10 haciendas se ubica entre 6 mil y 600 mil colones, las de menor valor comercial corresponden a las zonas de mayor actividad conflictiva en el pasado, es decir, de mayor inseguridad real de la tenencia, en las cuales respecto al valor inicial de compra en la Reforma Agraria se ha reducido en alrededor del 50% su valor comercial (Cuadro No. 21) La liberalizaci6n de las transacciones de compra venta de tierras en poder de los beneficiarios de la Reforma Agraria se ha desarrollado mediante una reforma juridica reciente, expresada en la Ley del Regimen Especial de la Tierra en Propiedad de las Asociaciones Cooperativas Comunales y Comunitarias Campesinas (Decreto Legislativo 719, de mayo de 1996), en la cual se faculta al sector reformado a la venta de tierras en forma libre, ademas se otorga la oportunidad a los productores para que del producto de la venta se permita reducir sus respectivos saldos en mora con el sistema financiero u otra instituci6n del Estado. El valor de la deuda se reduce en un 70% por pronto pago, para incentivar las transacciones de tierra. Este paso fundamental en la libre opci6n y capacidad de enajenaci6n de las tierras en manos de los beneficiarios de los procesos de distribuci6n de la tierra, pennitira contar con mercado mas activo de inversiones en operaciones de compra venta y de arrendaniento productivo del recurso. 18 Annex 1: Diagnosis Cuadro No. 21: Valor de la Tierra de Cooperativas Seleccionadas de la Reforma Aararia (miles de ¢/Mz) Cooperativas 1980 1996 1995 Venta Departament Agau Fria 12.5 15.7 290.0 La Libertad Florenciade R.L. 13.3 15.0 340.0 La Libertad Pasatiempo 18.6 35.0 600.0 La Libertad Corral de Mulas 4.8 5.9 55.0 Usulutan La Carrera 11.9 15.5 30.0 Usulutan San Andres 16.4 22.4 400.0 La Libertad Astoria 5.9 6.2 90.4 La Paz Chanmico 3.1 3.6 50.0 La Libertad Valle San Juan 13.1 15.9 6.0 Usulutin Santo Tomas 7.7 7.7 20.0 La Paz Incluye intereses mora cr6dito de avio cartera FRAP 2_ Precio indicativo Fuente: IICA, Potencial Social Econ6mico y Ambiental de la Desregulaci6n a las Asociaciones de Cooperativas de la Reforma Agrara 3. Determinantes ex6genos del desempefno agropecuario Para discriminar entre los diferentes factores, se entienden como ex6genos aquellos sobre los cuales no tienen la capacidad de incidir los empresarios, es decir, los factores que se determinan fuera de la influencia de los empresarios agropecuarios. Por su importancia sobre el desemnpefio econ6mico del sector se consideran como ex6genos, la guerra civil de El Salvador desarollada principalmente en las areas rurales, el precio intemacional del cafe y las remesas familiares. 3.1 La confrontaci6n social que afect6 el desempeiio agropecuario La situaci6n de guerra que experiment6 el pais no s6lo modific6 las condiciones de seguridad personal, sino que afect6 directamente las explotaciones agricolas en forma directa. La guerra propici6 el abandono de explotaciones agropecuarias, perdida y deterioro del patrimonio invertido en las explotaciones y la solvencia financiera de los empresarios agropecuarios, a tal grado que se disent6 un programa de refinanciainiento para los empresarios afectados por el conflicto (Decreto 242). Sin duda, esta variable ex6gena a las decisiones de los empresarios agropecuarios determin6 el desempefio del sector agropecuario. Ademas, en el marco de este proceso de guerra del pais se irnplement6 una Reforma Agraria, y como parte de los acuerdos de paz, se procedi6 a otro proceso de transferencia de tierra a los excombatientes, lo cual modific6 las condiciones de tenencia de la tierra. El deterioro del sector agropecuario de los afnos ochenta se evidenci6 en la primera parte de este estudio. 3.2. Precio internacional del cafe Por la importancia de la actividad cafetalera en el pais, una de las principales variables ex6genas de caracter internacional que inciden en la economia salvadorefia la constituye los precios intemacionales del cafe. Entre los mecanismos mais importantes a traves de los cuales el comportamiento de los precios del cafe inciden en la economia se tienen: el impacto de los flujos de divisas sobre el mercado camnbiario, el poder de compra de los agentes que participan en la actividad, la situaci6n de rentabilidad del cultivo y la mora que incide sobre la politica crediticia para este sector. En los ultimos quince anlos se experimenta una tendencia marcada a la baja en los precios internacionales del cafe, la cual alcanza su punto minimo con un precio internacional de alrededor de $50 por quintal en 1992; en el periodo de 1993 a 1995 se registra una recuperaci6n en los precios, habiendo alcanzado su punto maximo en 1995 con un promedio de $150 por quintal. En la decada de los ochenta los precios internacionales del cafe mostraron una tendencia a la baja, a excepci6n de 1986, cuando se report6 un repunte con un precio que super6 ligeramente los $175 por quintal; sin embargo, a este precio excepcional le sigui6 una tendencia a la baja en el periodo de 1987-93 (Grafica No. 1). Annex 1: Diagnosis 19 En el periodo de liberalizaci6n de la economia, los precios intemacionales han presentado su escenario mas negativo en los uiltimos quince anios, a excepci6n del repunte sefialado para 1995, es decir, durante el periodo de 1990-94 los precios internacionales del cafe se cotizaron en promedio por debajo de los $100 por quintal, e incluso para el bienio 1992-93 por debajo de los $75 por quintal. El comportamiento a la baja en los precios internacionales tambien ha estado acompanado por una tendencia similar en el volumen de producci6n del cafe, para el bienio de 1979-80 se exportaron en promedio 4,282 miles de quintales y para el periodo de 1994-95 el volumen promedio exportado es de solamente 2,505 miles de quintales, ello significa que el volumen exportado se ha contraido en 41.5%. Esta magnitud de contracci6n en el volumen de exportaci6n se explica en buena medida por la tendencia desfavorable en los precios del caf6, asi como por otros factores que se exponen en este documento, entre ellos: reforma agraria, tipo de cambio real, eliminaci6n de subsidios en la tasa de interes y otros. 3.3. Remesas familiares Las remesas familiares se han constituido para la economia salvadorefia en una variable ex6gena relevante para explicar la estabilidad de la economia en los uiltimos quinquenios y en especial para el periodo 1991-95. La emigraci6n de la poblaci6n salvadorefia en los anios setenta y ochenta fue significativa, como una opci6n de progreso para la mayoria de grupos familiares rurales y urbanos de bajos ingresos, a tal punto que el pais en los afnos ochenta se caracteriz6 en variables demograficas elevadas de emigraci6n y defunci6n, ambas vinculadas con el conflicto que experiment6 el pais. La tasa de emigraci6n paso de un promedio de 5.2 a 14.2% entre 1975 y 1985. A nivel de grupo familiar en el pais, pricticamente la mayoria de hogares salvadorenios cuentan con al menos un miembro (de primero o segundo grado de consanguinidad) emigrado hacia los Estados Unidos, afortunadamente los lazos familiares de los salvadorenios son fuertes, al punto que las remesas familiares o divisas que ingresan en forma de ayuda familiar han liegado a superar el valor de las exportaciones totales, excluyendo maquila. La tendencia de las remesas familiares es creciente en forma continua, particularmente desde 1986 que comienza a mostrar tasas de crecimiento suaves, hasta 1989. En el periodo de 1990 a 1995 las remesas familiares experimentan una tendencia de crecimiento acelerado que no guardan proporci6n con la experimentada en los anios ochenta. Una de las hip6tesis que podrian explicar el incremento en el flujo de las remesas familiares podria ser la confianza en la estabilidad y pacificaci6n del pais que motiv6 a los salvadorefnos emigrados a ayudar a sus familiares para progresar en su pais, con la expectativa de retomar en el futuro a El Salvador y ademas por las condiciones negativas que se han generado en terminos de asistencia social y empleo en Estados Unidos para los grupos de emigrantes indocumentados. Las remesas familiares, que desde 1986 exhiben un comportamiento creciente, han influenciado la estabilidad del mercado cambiario, es decir, han aumentado la oferta de divisas en el pais, la cual ha propiciado una presi6n constante para mantener el tipo de cambio apreciado en terminos reales, en la cual se relaciona el indice de tipo de cambio real y las remesas familiares para el periodo 1980-95. Es notorio el grado de asociaci6n entre las variables, particularmente desde 1991, en donde a medida que se aceleran los flujos de remesas familiares al pais, se acelera la apreciaci6n real del tipo de cambio. En tal sentido, las remesas familiares han incidido modificando la relaci6n de rentabilidad entre bienes transables y no transables (tipo de cambio real), en detrimento de las actividades de bienes transables y por constituir las actividades agropecuarias como los bienes tipicos transables, el comportamiento de la variable ex6gena remesas familiares ha incidido negativamente sobre las actividades agropecuarias. Adicionalmente, la ayuda familiar por medio de las remesas, ha generado una cultura de hAbitos de consumo y de aspiraciones de trabajo en la poblaci6n rural joven, que no corresponden con las oportunidades que brinda el sector agropecuario, principalmente en labores agricolas. Esto significa que la 20 Annex 1: Diagnosis poblaci6n rural joven presiona por oportunidades en las principales zonas urbanas del pais, de esta manera la poblaci6n con mayor potencial de productividad para labores agricolas no estA acompafiando el esfuerzo de este sector. Asimismo, existe un desincentivo econ6mico a participar en las labores agricolas, en la medida que la asistencia por remesas familiares permite cubrir las necesidades basicas para la poblaci6n rural, la cual tiene muy pocas expectativas de superaci6n por sus condiciones de vida. En tal sentido, este es otro elemento que ha afectado negativamente la actividad agropecuaria por el comportamiento de las remesas familiares. Los flujos de divisas destinados al sector rural tambien han generado beneficios en termino que aumentan el poder de compra de las familias rurales y de esta forma se han incrementado principalmente las actividades de comercio y servicio en otras ciudades, por ejemplo San Miguel, Santa Rosa de Lima y otras ciudades, principalmente de la zona oriental, que corresponden a las zonas donde mayor emigraci6n se report6, por la misma situaci6n que se experiment6 de gravedad del conflicto. 4. Infraestructura de apoyo a la producci6n agropecuaria El sector agropecuario ha desarrollado su actividad productiva bajo condiciones de infraestructura sustancialmente diferentes a la red vial, de energia y telefonia, con que operan las empresas de industria, comercio y servicios, principalmente las ubicadas en la cuidad capital, las cuales en el periodo del conflicto se registr6 la tendencia de concentraci6n geografica, en este centro productivo del pais. Desafortunadamente, las actividades agropecuarias disponian de una debil infraestructura de apoyo a la producci6n , la cual con el conflicto se deterioro, convirtiendose en una de las principales barreras para el desempenio empresarial en El Salvador en la actualidad. 4.1. Infraestructura vial La infraestructura vial en el sector rural es un factor determinante de acceso a los mercados de bienes, capital, recurso humano y financiero, asi como de acceso a la tecnologia e informaci6n. En El Salvador las carreteras rurales superan el 60% del total de la red, seguidas de las vias terciarias, con un peso relativo del 17%. El diagn6stico de la red vial actualizado para 1993 por la Direcci6n General de Caminos del MOP, determin6 que las principales categorias de carreteras: especial, primaria y secundaria, constituyen alrededor del 20% de la red del pais. En el inventario se estableci6 la condici6n superficial de las carreteras clasificadas en categorias de: Muy Bueno (MB), Buena (B), Regular (R), Mala (M) y Muy Mala (MM). Derivandose el resultado de que aproximadamente el 26% de la red vial se encontraba en estado satisfactorio de conservaci6n, un 30% en estado regular y 38% en mal o muy mal estado. Los mayores indices de deterioro en la condici6n superficial de las carreteras corresponden a las vias secundarias y terciarias. En 1990 se desarroll6 un ejercicio similar de evaluaci6n del estado superficial de la red vial, con un resultado no estadisticamnente diferente de 23.7% en buen estado de las carreteras. Ello significa que los programas de inversi6n no propiciaron un cambio notorio en el estado de las carreteras de la red. La red vial en los anios ochenta, experiment6 un descuido prolongado de las carreteras, que condujo al deterioro progresivo y acelerado de los pavimentos. Ello se combin6 con el periodo de vencimiento de la vida util de las carreteras, al grado que las actividades de conservaci6n de la red, han sido calificadas de excesivas y costosas debido a la ejecuci6n de tareas de mantenimiento de emergencia para mantener la circulaci6n. 5GOES-USAID. (1995)."Estudio de Optimizaci6n del Sistema de Transporte' Annex 1: Diagnosis 21 Entre las principales causas de la deficiente conservaci6n de la red se sefiala: falta de financiamiento para la conservaci6n prevista en las leyes del presupuesto, ineficiente programaci6n de la inversi6n, ineficiente la institucion piiblica responsable, escasa atenci6n a los efectos del mal estado de las vias sobre los costos de operacion vehicular, las cuales se agravaron en el periodo del conflicto. Los costos de mantenimiento tradicionalmente en El Salvador han sido considerados como de funcionamiento de la entidad encargada. Durante el periodo de 1993-95, alrededor del 90% del presupuesto ordinario de la Direcci6n General de Caminos se asign6 a pago de salarios. Las asignaciones para materiales son insignificantes representaron en 1994, unicamnente 0 110 ($12.57) por kil6metro por afio. En general, la politica se manejo absorbiendo las reducciones del presupuesto, a traves de ajustes en la compra de materiales y mantenimiento de la nomina de empleados fijos. En la encuesta FUSADES-Banco Mundial de 1996, se determin6 que la distancia promedio de las explotaciones agricolas es de 6 kms., lo cual es un indicativo de las limitaciones de acceso a mercados para las explotaciones agropecuarias, que deben operar con significativas limitaciones de apoyo a la producci6n. 4.2. Infraestructura de riego El Salvador cuenta con aproximadamente un total de 805.8 miles de hectareas con vocacion agricola, de las cuales se utiliza para el cultivo del cafe 195.7 miles (24.2%) y el remanente para otros fines agricolas. La tierra con mejor aptitud agricola (Clase I a V) se encuentra localizada en la zona costera. Estas condiciones geogrificas han demandado diferentes estudios para ofrecer obras de riego, drenaje y control de inundaciones, con el proposito de intensificar su productividad. El Plan Maestro de Desarrollo de Recursos Hidraulicos determin6 en los ochenta que aproximadamente 273 miles de hectareas tenian potencial de ser objeto de areas irrigables, lo cual se detalla en el Cuadro No. 22. Cuadro No. 22: El Salvador: Superficie Agrfcola Irrigable (en miles de hectareas) Superficie agricola irrigable (miles de Ha.) Zona Clase 1-4l Clase IV Clase V Total Proyecto Rlo Paz 4.48 1.13 0.40 6.01 Proyecto Sonsonate-Bandera 5.31 1.47 0.30 7.08 Proyecto Comalapa 11.57 1.38 0.50 13.45 Proyecto Bajo Lempa 28.96 6.03 15.0 50.0 Provecto San Miu .l 25.03 3.88 6.00 34.91 Subtotal 75.35 13.91 22.20 111.46 Otras zonas del pals 92.89 50.00 19.17 162.07 ~~~~~~~~~~~~~~~~~~~~~~...............................................................................................................................................................................: TOTAL 273.535 Fuente: Hidraulica Internacional (1996). Actualmente, el total de area que dispone de infraestructura de riego es de 34.4 miles de hectareas, en las cuales se incluyen los tres grandes distritos irrigables del pais y ejecutados por el Estado: Zapotitan, Atiocoyo y Lempa-Acahuapa. En tal sentido, la infraestructura de riego disponible es solamente el 12.6% del potencial de extensi6n irrigable. El apoyo a la produccion agricola con infraestructura es sumamnente debil y limitado en los diferentes campos de servicios, como el caso de riego, red vial, energia y telefonia. 5. Factores de Politica Macroecon6rnica 5.1. Apertura Comercial La economia salvadorefia ha transitado aceleradamente a un esquema de liberalizacion comercial como parte de la estrategia de establecer una economia de mercado, constituyendo un elemento muy importante 22 Annex 1: Diagnosis de esta reforma la relativa a la apertura comercial. El Salvador se ha adheri6 al GATT hasta mediados de 1990; por otra parte, ha desarrollado un proceso de reducci6n de aranceles en forma acelerada. La apertura comercial se desarroll6 teniendo como objetivos aumentar el grado de eficiencia del aparato productivo intemo, es decir, de las industrias que compiten con las importaciones, asi como reducir los sesgos de antiexportaci6n que prevalecieron sobre el sector agropecuario, en especial sobre el caf6. Adicionalmente por las tasas de aranceles relativamente altas en bienes de consumo la apertura comercial se consider6 como un instrumento apropiado para incidir sobre la tasa de inflaci6n y mejorar de esta manera las condiciones de vida de la poblaci6n. Cuadro No. 23: El Salvador: Programa de Reducci6n de Aranceles, 1989-99 Ahios No. de Tasas Rango Asta Septiembre 1989 25 0-290 Septiembre 19891 9 1, 5, 20, 25, 30, 35, 40, and 50 Abril 1901 6 5, 10, 20, 25, 30, and 35 Junio 19912 5 5,10, 20, 25, and 30 Diciembre 19912 4 5,10, 20, and 25 Marzo 19921 5 5,10, 20, 25, and 30 Diciembre 19941 3 5,10, and 20 Programa Oficial enero julio 1995 1 - 20' 1996 0 - 20 1997 0- 19 1998 0- 18 0- 17 1999 0-15 0- 15 lncluyendo Parte lli; a Bienes de capital 1 %; 2 Excluyendo Parte III; Actualizado con base en informaci6n del Ministerio de Economia. Fuente: World Bank (1995), El Salvador. Meeting the Challenge of Globalization. Table No. 3.12 El proceso de reforma de aranceles se inci6 en septiernbre de 1989, partiendo de una estructura que contaba con un rango de aranceles de entre 0 y hasta 290%, con 25 diferentes tasas, de formia que la estructura era sumamente compleja y comprendia niveles de dispersi6n arancelaria superior al 100%. Por ello, la reforma se busc6 bajar las tasas de protecci6n por la via de los aranceles y disminuir la distorsi6n que comprendia la dispersi6n arancelaria. A finales de septiembre de 1989, en la primera etapa de la reforma, se logr6 un avance importante que consisti6 en reducir a 9 tasas de aranceles ubicadas en un rango entre 1 y 50%. Se puede sefialar que en esta primera fase, pricticamente se elimin6 el agua en el arancel que prevalecia en la protecci6n nominal (Cuadro No. 23). Entre el periodo de abril de 1990 y diciembre de 1994, se aplicaron cinco fases de reducci6n de aranceles que permitieron cuLminar con un rango de aranceles de entre 5 y 20% y con uinicamente tres diferentes tarifas (5, 10 y 20%). Como se aprecia en el Cuadro No. 24, la reforma de reducci6n de aranceles se desarroll6 sistematicamente y en forma continua en un periodo aproximado de cinco anios. A pesar de la reducci6n drastica en la tasa de aranceles, la actual administraci6n puiblica tiene definido un programa de reformas buscando alcanzar un rango de aranceles entre 1 y 15%, contando con un horizonte final en 1999. Las tasas nominales de aranceles de las actividades agricolas, previo a la reforma, contaban con un promedio de 39% como tasa nominal, las cuales en 1990 habian descendido a 17.2% y a junio de 1991 eran uznicamente de una protecci6n nominal de 12.1% para el conjunto de productos agricolas. Las tasas nominales promedio de aranceles de las actividades agricolas no son significativamente diferentes para las reportadas para la economia en su conjunto como se muestra en el Cuadro No. 25. Es notorio en la informaci6n de este cuadro que las tasas nominales de aranceles relativamente altas previas a la reforma se encontraban en los bienes clasificados de consumo y originados en el sector de manufacturas, el cual registra una tasa promedio de 59% como tasa nominal en 1988. Annex 1: Diagnosis 23 Cuadro No. 24: El Salvador: Tasa y Rango de Aranceles, 1989-99 (promedios simples) 1998 1989 1990 1991 1992 Jun-94 Ener95 Economia Total 36.9 20.4 17.8 17.7 12.5 11.6 10.7 Agricultura 39.0 20.4 17.2 13.9 11.5 12.1 12.1 Mineria 10.4 6.9 6.9 7.5 7.3 6.3 6.3 Industria 37.2 20.6 18.0 14.9 12.6 11.6 10.7 Bienes de Consumo 59.2 29.5 24.2 20.0 16.2 18.8 16.7 Bienes Intermedios 21.8 14.4 13.8 11.6 10.3 7.8 7.5 Bienes de Capital 21.5 14.0 13.3 10.8 9.8 6.3 6.3 Fuente: World Bank (1995) Table No. 3.12 A nivel de las tasas de protecci6n efectiva se tiene que las actividades agricolas se beneficiaron sustancialmente con la reducci6n de la tasa de aranceles. El cafe corresponde a las actividades mas beneficiadas de la reforma, tomando en cuenta que contaba con una tasa de desprotecci6n del -28%, en terminos de tasa de protecci6n efectiva y pas6 en 1993 a una tasa de -1.3%, este cambio se experiment6 por la eliminaci6n del impuesto a las exportaciones del cafe que era del 25%. Los otros productos agricolas pasaron de una tasa de protecci6n efectiva de -2.6% a una protecci6n del 7%. Las actividades productivas que perdieron niveles de protecci6n efectiva son las industrias de alimentos procesados, bebidas y tabacos, textiles, vestuario, articulos de pieles y calzado, muebles y maderas, papel e impresi6n, las cuales gozaban de tasas de protecci6n efectiva superior al 100%, las cuales fueron reducidas a niveles inferiores de protecci6n efectiva del 30% (Cuadro No. 25). Cuadro No. 25: El Salvador: Tasas de Protecci6n Nominal y Efectiva (porcentajes) Actividades Tasas Nominales Tasas Efectivas 1987 1993 1987 1993 Cafe -25.0 0.0 -28.1 -1.3 Algod6n 0.0 5.0 -11.1 4.0 Otros agricolas y minerfa 4.8 7.3 -2.6 7.0 Alimentos procesados 50.4 13.4 199.1 30.4 Bebidas y tabaco 225.7 16.8 1409.0 21.2 Textiles 54.2 13.0 121.7 20.4 Vestidos, artic. de piel y calzado 86.6 26.7 225.6 18.2 Maderasymuebles 124.8 13.5 371.1 18.6 Papel e impresi6n 52.8 4.9 105.1 2.4 Qulmicos 14.3 5.7 7.1 4.6 Caucho, plasticos y minerales 37.6 9.3 69.7 6.1 Industria de hierro y del acero 17.6 8.6 17.3 9.3 Maguinaria y Equipo 29.0 9.3 56.6 10.3 Fuente: Banco Mundial (1989) para 1987, y Abrego (1994) para 1993. La reforma comercial, sin duda, ha propiciado beneficios a la economia en terminos de reducir la distorsi6n de precios entre los internos y los internacionales, asi como ha inducido a mejoras en la eficiencia en la producci6n, asi como beneficios para los consumidores por elevar la calidad de la canasta de bienes que se transa en el mercado intemo. No obstante, el crecimiento de las importaciones ha sido notorio al grado que la economia salvadorefna presenta la caracteristica de mantener un estilo de crecimiento econ6mico en consumo (en especial importado) alimentado con la demanda de los influjos de divisas, el cual tambien se ve apoyado por la apreciaci6n del tipo de cambio real y la expansi6n del credito. 5.2. Liberalizaci6n de precios Como parte de las reformas iniciadas en 1989, se liberalizaron los precios de los bienes y servicios, en vista que la economia salvadorefia reportaba una canasta de mas de 200 precios que eran objeto de regulaciones y establecimiento de precios por politica puiblica. En los anios ochenta, se buscaba proteger el poder de compra de la poblacion a travds de mantener congelado el precio de los bienes y servicios de una canasta basica, la cual correspondia principalmente a bienes alimenticios y servicios basicos. Los resultados de este tipo de politicas no lograron su prop6sito en el sentido de proteger el poder de compra de los grupos mas vulnerables de bajos ingresos, y mas bien su poder de compra se vio deteriorado por el surgimiento de mercados negros o paralelos a las politicas oficiales. 24 Annex 1: Diagnosis La reforma de liberalizaci6n de precios en la parte de bienes agricolas, comprendi6 la liberalizaci6n de precios de granos bAsicos y otros agropecuarios, asi como liberalizar la comercializaci6n intemacional de los bienes de exportaci6n. Es decir, la reforma elimin6 la nacionalizaci6n del comercio exterior de los principales productos de exportaci6n, asimismo, elimin6 el Instituto Regulador de Abastecimientos (IRA), el cual operaba como un comprador mayorista y distribuidor a nivel del consumidor con puntos de venta propios del IRA, y generalmente en lugares diferentes a los principales centros de comercio del pais. El indice de precios al por mayor para la economia salvadorefia, excluyendo el cafe, y el indice de precios de granos basicos al productor, construido tomando 1990 como afio base y compuesto por la canasta de maiz, frijol, arroz y sorgo, ilustran la tendencia creciente de ambos indices en los ultimos anios; sin embargo, en el periodo que opero el IRA, el indice de precios de granos basicos tenia un menor nivel respecto a los precios al por mayor de la economia salvadorefia. Ello significa que la politica del IRA mantenia los precios al productor artificialmente por debajo de los niveles inflacionarios de la economia en su conjunto y al margen de las condiciones de oferta y demanda de estos productos. En el periodo de liberalizaci6n de los precios de los granos basicos se muestra un comportamiento de tendencia creciente en especies; sin embargo, tambien se reporta mayor comportamiento erratico en los precios de los granos bAsicos, los cuales se asocian a las condiciones de oferta y demanda de productos en cl mercado interno. De tal forma que la liberalizaci6n ha permitido que los precios de los granos basicos absorban las condiciones de oferta y demanda a nivel del productor. La estrategia de liberalizaci6n tambien se diseni6 con instrumentos de cobertura y estabilidad para los productores de granos basicos, por medio de una banda de precios que busca neutralizar por medio de modificaciones en la tasa de aranceles los cambios en los precios intemacionales de los granos basicos, la cual se desarrolla en el siguiente apartado. 5.3. Banda de precios de granos bisicos La banda de precios como instrumento de suavizamiento de los cambios en los mercados intemacionales de bienes agricolas se implement6 a partir de 1990 y oper6 hasta fines de 1994. La aplicaci6n de la banda de precios iinicamente se puso en vigencia para maiz, arroz y sorgo. El mecanismo consistia en tasas de aranceles variables en funci6n de los precios intemacionales de los granos basicos; para el maiz, por ejemplo, se oper6 con un promedio de referencia de $145/TM al cual corresponde el arancel base de 20%, para precios por arriba de este el arancel se ajusta hacia la baja, por el contrario, para promedios inferiores al de referencia el arancel se corrigia con incrementos proporcionales a la baja. En el perido que oper6 la banda de precios el rango de variaci6n de los aranceles se mantuvo entre 18% y 30%, para estabilizar el mercado intemo, y de esta fonna mantener estabilidad para el productor y consumidor de granos basicos. Desafortundamente, la medida se modifico para el maiz y arroz, por irregularidades administrativas en la aplicaci6n de la disposici6n, pasando a un esquema de arancel fijo de 20%, desde fines de 1994. La banda de precios ha sido un instrumento importante para neutralizar los comportamientos errAticos de los precios intemacionales de granos bAsicos. A pesar de ello, los efectos de estabilidad son mas notorios en los precios en plaza para los consumidores, en relaci6n con la estabilidad esperada para los productores. Es notorio que en los anios de vigencia de la banda de precios se ha logrado estabilidad relativa en los precios de los granos basicos, y tambien en el periodo se ha experimentado un crecimiento real del PIB de estos bienes. En el caso del maiz y sorgo cuando se liberaliza la banda de precios se reporta un repunte en los precios de estos bienes, por la interacci6n mAs libre a las condiciones del mercado intemacional. Annex 1: Diagnosis 25 Asimismo, el indice de precios al productor de los granos basicos presenta mayor inestabilidad en los afios de eliminaci6n de la banda de precios, es decir entre 1994 y 1995, los cambios erraticos de los precios de granos basicos constituyen la raz6n fundamental para la vigencia de la banda de precios, como mecanismo de estabilizar los precios para interes de los consumidores y productores. 5.4. Tipo de Cambio Como parte del proceso de liberalizaci6n de la economia, se estableci6 un regimen de tipo de cambio libre, es decir, determinado por las condiciones de oferta y demanda en el mercado, el cual se autorregul6 bajo este sistema bajo el periodo comprendido entre 1989 y 1992. A partir de 1993 se presenta una tendencia de mayor influjo de divisas. Lo cual propicia que las autoridades monetarias establezcan como prop6sito de politica mantener un tipo de cambio nominal estable. La presencia de flujos de divisas sustanciales en las economia gener6 una situaci6n en que el Banco Central operaba en el mercado de divisas bajo la modalidad de comprador neto de divisas para sostener la cotizaci6n a un nivel de 8.75, es decir, para evitar una apreciaci6n nominal del tipo de cambio, lo cual efectivamente ocurri6 entre 1993 y 1994. La politica de sostener un tipo de caznbio estable se ha mantenido en 1995, afio en el cual las condiciones de operaciones de compra y venta de divisas del Banco Central registran un cambio, en el sentido que el BCR oper6 en forma neta como vendedor de divisas. Esto significa que para este afno mas bien se evit6 una depreciaci6n nominal del tipo de cambio. Como indicador del tipo de cambio real se ha construido un Indice del Tipo de Cambio Efectivo Real (ITCER), tomando 1980 como afio base y ponderado por la importancia relativa del comercio de los principales socios del pais: Estados Unidos, Alemania, Jap6n, Guatemala y Costa Rica. 6 El desempeono de las exportaciones y las actividades que compiten con las importaciones se ha visto afectado en forma negativa por el comportaniiento del tipo de cambio en terminos reales. El comportamiento del Indice del Tipo de Camnbio Efectivo Real (ITCER) refleja una tendencia continuada de apreciaci6n real a partir de 1991, la cual registra un promedio de apreciaci6n real de 6.4% anual para el periodo 1991-95. El nivel acumulado de apreciaci6n a fines de 1995 respecto al anio base de 1980 es equivalente a un 81%, es decir, la economia ha experimentado un deterioro de perdida de competitividad medido por este indicador de tipo de cambio real. En el primer quinquenio de los afios ochenta, la economia salvadoreiia experiment6 un proceso de apreciaci6n moderada, la cual se corrigi6 practicamente con las depreciaciones reales experimentadas entre 1984 y 1985, en donde se tuvo una politica cambiaria nominal muy activa, tratando de recuperar los niveles de competitividad de 1980, lo cual se logro tomando en cuenta que a fines de 1985 la apreciaci6n real acumulada fue de -0.5%. Se destaca los afios iniciales de la reforma de liberalizaci6n econ6mica (1989-1990) en donde se generaron cambios importantes en terminos de depreciaci6n real del tipo de cambio, en especial 1990, que se registr6 22% de depreciaci6n real. No obstante, desde 1991 se experimenta una situaci6n de marcada tendencia de apreciaci6n real, a tal punto que la importante depreciaci6n registrada en 1990 se ha anulado con la apreciaci6n acumulada del periodo 1991-95 de 31.2% (Cuadro No. 27). En el quinquenio 1991-95, que se experiment6 un deterioro en la competitividad de las exportaciones, por la apreciaci6n real acumulada en el periodo, tiene la particularidad de corresponder a tasas de inflaci6n relativamente controladas y de alrededor del 10% a excepci6n de 1993, que la tasa de inflaci6n experiment6 un repunte de 18.8% anual, lo cual se gener6 por un incremento de precios desmedido por la introducci6n del impuesto del IVA. La tasa de inflacion de El Salvador es muy similar al promedio que registran sus 6 Vase pam mayor ampliaci6n de este tema la aplicaci6n desarrollada por: Saca, Nolvia N. (1995) Black Market Exchange Rate, Unification of the Foreign Exchange Markets and Monetary Policy: The case of El Salvador, University of Kiel, Kiel. 26 Annex 1: Diagnosis principales socios de comercio, e incluso en algunos meses entre 1994 y 1995 el promedio de la inflaci6n de los socios de comercio supera ligeramente a la de nuestro pais. Cuadro No. 26: Variaciones del Tipo de Cambio Efectivo Real' (base 1980=100) Tipo de cambio Variaci6n % Depreciaci6n (.) Depreciac. (. lnflaci6n nominal Tipo de cambio apreciaci6n (-) apreciac. () ES acumulada 1981 2.50 0.0 -4.23 -4.23 14.9 1982 2.61 4.4 -8.24 -8.24 11.7 1983 2.81 7.7 -0.12 -8.37 13.1 1984 3.16 12.4 -8.31 -15.68 11.8 1985 3.84 21.5 -6.79 -8.32 22.2 1986 4.96 29.2 -7.77 -0.52 32.1 1987 5.00 0.8 -20.72 -26.78 25.0 1988 5.00 0.0 -9.41 -39.96 19.8 1989 5.53 10.6 -13.64 -62.06 17.6 1990 7.54 36.3 2.05 -58.80 24.2 1991 8.04 6.6 22.01 -30.15 14.5 1992 8.41 4.6 -5.91 -38.33 11.1 1993 8.74 3.9 -6.78 -48.40 18.8 1994 8.73 -0.1 -10.03 -64.93 10.6 1995 8.75 0.2 -2.72 -69.54 8.9 Estimado utilizando la metodologia en Rhomberg, Rudolf R. (1976). 'Indices of Effective Exchange Rates". IMF Staff Papers 30. 1 (marzo),88-112. La tendencia del indice del Tipo de Cambio Efectivo Real y Nominal, asi como la relaci6n de precios de El Salvador y sus principales socios de comercio, en forma mensual desde 1980 hasta 1995, es de marcada apreciaci6n real, lo cual incide negativamente en los sectores exportadores y actividades que compiten con las importaciones. El sector agropecuario se caracteriza por ser una de las actividades tipicas de bienes transables, es decir, tiene un fuerte componente de bienes exportables y/o importables, la economia salvadorefia se caracteriza por tener una estructura de oferta exportable concentrada en productos de origen agricola. En tal sentido, el deterioro del tipo de cambio real ha afectado negativamente el desempefno del sector agropecuario. 5.5. Politica de salario minimo En El Salvador se tiene como politica publica establecer salarios minimos por actividades econ6micas, los ajustes de dichos salarios se fijan en funci6n de los niveles de inflaci6n del periodo de ajuste, es decir, no responde a incrementos de productividad o decisiones de rentabilidad de las actividades productivas. El salario minimo en las diferentes actividades productivas es aplicado en forma general, sin embargo, existen excepciones donde no se atienden las disposiciones de salario minimo, principalmente son actividades en zonas geogrificas distantes donde existen liuitaciones de informaci6n, oferta amplia de trabajo, asi como en actividades informales, el salario se establece al margen de la politica sobre esta materia. En general, los ajustes nominales del salario minimo se fijan por debajo del promedio de inflaci6n del periodo y ademAs los ajustes se otorgan con un rezago del termino del periodo de correcci6n. En virtud de ello, los salarios reales normalmente presentan una situaci6n de incrementos reales negativos, es decir, el poder de compra de los asalariados se disminuye por las correcciones aplicadas en el salario minimto. En tal sentido, la politica de salarios minimos seguida en El Salvador, si bien es una distorsi6n para los niveles de empleo y salarios establecidos en condiciones de oferta y demanda, no constituyen un instrumento de politica econ6mica que perjudique sensiblemente la rentabilidad de las actividades econ6micas a nivel de la economia. No obstante, las actividades agropecuarias por experimentar una situaci6n de deterioro en terminos reales en el periodo de 1991-95, la politica de salarios minimos podria estar perjudicando la relaci6n de rentabilidad en las actividades agropecuarias, tomando en cuenta que la productividad por manzana en la mayoria de productos agricolas no experimenta un cambio importante en el periodo. Annex 1: Diagnosis 27 En el Cuadro No. 28, se presentan los incrementos promedio por quinquenio de los salarios minimos para las diferentes actividades econ6micas, en el cual se refleja que los ajustes en los salarios nominales en las actividades agropecuarias, generalmente se encuentran por debajo de los ajustes aplicados a las actividades de industria, comercio y servicio. El promedio de ajuste nominal en el salario minimo de las actividades de cafe y cafia de azucar es de alrededor de 4% para el periodo 1991-95, mientras que para el sector de industria, comercio y servicio, tiene un promedio cercano al 13%, ello genera una situaci6n en la que los asalariados del sector agropecuario experimentan mayor perdida en el poder de compra. Cuadro No. 27: Salarios Minimos de Principales Actividades Econ6micas (tasas de crecimiento promedio) SALARIOS 1971-76 1976-80 1981-86 1986-90 1991-95 NOMINALES Industria, comercio y servicios 13.8 11.6 3.6 10.4 12.9 Agropecuaro (mayores de 16 anos 7.0 11.3 0.0 18.8 11.7 Cafe (recolecci6n) 9.9 22.9 0.0 6.8 4.1 AzOcar (recolecci6n) ........... 9.6 18.3 0.0 7.5 4.7 ............. . R EA LES.......... ............................................................................................... REALES lndustra, comercio y servicios 5.1 -1.2 -11.1 -13.3 -0.1 Agropecuario (mayores de 16 aflos -1.7 -1.5 -14.7 -4.9 -1.3 Cafe(recolecci6n) 1.2 10.1 -14.7 -16.9 -8.9 Az(icar (recolecci6n) 0.9 5.5 -14.7 -16.2 -8.3 INFLACION -IPC- (promedio) 8.7 12.8 14.7 23.7 13.0 Fuente: Elaborado con base en informaci6n del Consejo Nacional del Salado Minimo. En el periodo 1981-85 no se registran incrementos en los salarios minimos para las actividades agropecuarias y la inflaci6n promedio del periodo es de 14.7%. Este es un tipico ejemplo de que los ajustes salariales registran un rezago respecto al periodo de correcci6n por inflaci6n. En los uiltimos anios se ha tratado de mantener al menos las condiciones de terminos reales constantes, por lo que los ajustes nominales del salario siguen pricticamente en una relaci6n uno a uno con la inflaci6n. Este tipo de ajuste perjudica las actividades agropecuarias en la medida que el precio de sus productos presentan un menor ajuste o peor aun en el caso del cafe en donde se han registrado deterioros en su precio internacional de exportaci6n. 5.6. Tasa de interes Como parte de la liberalizaci6n financiera, se desarrollaron diferentes reformas en el sector financiero, para tratar de revertir la situaci6n calificada de represi6n financiera, la cual se caracterizaba por tasas de inter6s fijadas por la autoridad monetaria, siendo estas inferiores a la inflaci6n del periodo, asi como por definir cupos crediticios por actividad econ6mica, es decir, se asignaban los recursos financieros a traves de cupos crediticios conforne los objetivos politicos. Con la liberalizaci6n financiera, se pas6 a un sistema donde la tasa de interes se fija libremente por los intermediarios del mercado de dinero y el credito se asigna en funci6n de la administraci6n de riesgo de las instituciones de credito privadas. El sistema financiero opera bajo las condiciones descritas de liberalizaci6n financiera, a partir de mediados de 1992, donde se deja el establecimieento de las tasas de interes plenamente a discreci6n de los intermediarios financieros. Las tasas de interes de los aflos ochenta se caracterizan por ser reales negativas que a niveles de alrededor entre -5 y -10% en t6rminos reales, los mayores subsidios a traves de esta politica se destinaban para nuevos productores de granos bAsicos, a quienes correspondia tasas reales negativas de alrededor del 10%, como se muestra en el Cuadro No. 28. La reforma financiera de los alios noventa pretendi6 elininar los subsidios a traves de las tasas de interes, en virtud de que los beneficiarios no correspondian a los sectores objetivos de la politica, como se demostr6 en el estudio The Interamerican Management Consulting Corporation (1989), en el cual se concluye: "que el credito dirigido es un mecanismo ineficiente para canalizar recursos a sectores 28 Annex 1: Diagnosis prioritarios, asi como que la distribuci6n del cr6dito estaba a favor de las empresas mras grandes". Por otra parte, una politica de tasas de interes reales negativas desestimula el ahorro financiero intemo y reduce el potencial de intermediaci6n financiera. Cuadro No. 28: Tasas de Interds Nominales y Reales Aplicadas a Usuarios Finales (tasas anuales simples) TASAS DE INTERES 1980 1985 1988 1990 1992 1995 NOMINALES Granos Basicos 8.0 13.0 13.0 17.0 19.0 18.0 Cafla de AzOicar 13.0 14.0 15.0 20.0 19.0 18.0 Cafe avlo) . .... ...........?13,0 13.0 15.0 20.0 19.0 18.0 NOMINALES Granos Basicos -9.5 -9.2 -8.9 -7.1 7.8 7.9 Cafna de Az(icar -4.5 -8.2 -4.9 -4.1 7.8 7.9 Cafe (avio) -4.5 -9.2 -4.9 -4.1 7.8 7.9 INFLACION -IPC- (promedio) 14.9 22.2 19.9 24.1 11.2 10.1 Fuente: Elaborado con base en informacion del BCR, Revista Trimestral, varios nimeros. Las tasas de interes nominales en el periodo de liberalizaci6n financiera, generalmente se han establecido por arriba de la tasa de inflaci6n, y en terminos reales corresponden a un nivel de alrededor del 8%, para las tasas de credito aplicado para las operaciones del sector agricola, las cuales no muestran diferencias importantes respecto a las tasas aplicadas al resto de actividades productivas del pais. El cambio en la politica de tasas de interes ha afectado la rentabilidad de las actividades productivas y en especial de aquellas que contaban con las mayores tasas de subsidio por medio de este instrumento, las cuales corresponden a las actividades agricolas. Las tasas nominales, bajo el periodo de liberalizaci6n financiera, comparativamente con las tasas nominales de 1980, se han aumentado entre 5 y 10 puntos nominales, lo cual se ha experimentado en un periodo de tasas de inflaci6n relativamente mAs bajas y controladas respecto a los afios de represi6n financiera en los ochenta. La critica sobre las tasas de interes del sistema financiero salvadorefio en el sentido de considerarlas relativamente altas, no es una apreciaci6n exclusiva de los sectores agropecuarios, sino de las actividades productivas en general, lo cual se explica por esa combinaci6n de mayores tasas nominales en un periodo de menores tasas de inflaci6n, bajo un esquema de eliminaci6n de susbsidios por esta via. 5.7. Politica de credito La liberalizaci6n financiera permiti6 eliminar la politica de cupos crediticios por actividad economica, pasando a un sistema de asignaci6n de fondos en funci6n de los riesgos que administra la instituci6n financiera, esta reforma se implement6 entre 1989 y 1992, periodo en el cual se trat6 de homogeneizar las condiciones de credito por actividad econ6mica y eliminar las practicas de cupos crediticios para posteriormente dejar la administraci6n de credito a los intermediarios, en la medida en que las instituciones financieras se privatizaron (principalnente en 1991-93). El destino del credito otorgado a las diferentes actividades productivas se modific6 sustantivamente, teniendo una menor participaci6n el sector agropecuario respecto a la registrada en los afnos de la represi6n financiera. La importancia relativa del sector agropecuario en el PIB en forma similar tambien ha declinado en los iiltimos afios. El credito destinado al sector agropecuario pas6 de tasas de importancia relativa superiores al 40% para los primeros anios de la decada de los ochenta, llegando a tener una importancia ligeramente superior al 12% para el periodo de 1994-95, ello significa que la participaci6n del sector agropecuario en el credito ha declinado en alrededor de un 28%. Los sectores que han ganado participaci6n en la asignaci6n de los recursos crediticios son: comnercio, construcci6n e industria. La administraci6n de los riesgos crediticios en forma privada por las instituciones del sistema financiero, ha llevado a una practica del credito en donde el sector agropecuario es considerado como uno de los destinos con mayor riesgo crediticio. Esta categoria del sector agropecuario se le ha imputado por las diversas experiencias negativas de moda en el sector en los anios ochenta y noventa, asi como por la Annex 1: Diagnosis 29 perdida de competitividad percibida por el comportamiento del tipo de cambio real y el precio de los principales productos agropecuarios. Las instituciones financieras han desarrollado diferentes practicas para defenderse del mayor riesgo que constituye el sector agropecuario, aplicandose politicas como las siguientes: creditos en funcion de la rentabilidad de la actividad, pero con una garantia hipotecaria que permita una sobrecobertura respecto al riesgo total del credito; normalmente se exige una hipoteca abierta y la exposici6n de creditos se procura que no exceda del 70% del valor de los bienes hipotecados. Este tipo de practicas financieras ha llevado a una situaci6n en la cual los deudores agropecuarios perciben que el credito no es por la actividad o riesgo financiado, sino mas bien por el valor que representa la hipoteca, es decir, cr6ditos en funci6n de la hipoteca y no por su destino. 5.8. Politica de Impuesto al Valor Agregado (IVA) Como parte de la reforma tributaria en el pais, se ha simplificado la estructura de impuestos, eliminando los impuestos ineficientes en terminos de su recaudaci6n y por su concepto inapropiado. La reforma a mediados de 1993 con la introducci6n del Impuesto al Valor Agregado (IVA), se alcanza una estructura simple que cuenta con tributos principales de Renta, Aranceles e IVA, en el pasado se tenian impuestos que grabavan el patrimonio, las exportaciones y tributaci6n en cascada el consumo como el impuesto de timbres y papel sellado. Sin duda, la reforma de introducci6n del IVA no s6lo es apropiada sino beneficiosa para lograr mayor eficiencia en la recaudaci6n tributaria, y con impuestos mAs neutrales. No obstante, el IVA en el pais se aprob6 con excepciones para un grupo de bienes entre ellos las medicinas y los bienes agropecuarios. Los uinicos productos agropecuarios que pagan IVA son los bienes de exportaci6n, asi como los agropecuarios procesados como queso, lacteos y otros de origen agropecuario procesados. La tasa del IVA se estableci6 en 10% en septiembre de 1993, pero en julio de 1995 se aumento a 13%, con ello se eleva la desventaja por IVA para los productores agropecuarios. Es decir, las actividades agropecuarios y medicinas, tienen que pagar IVA en sus compra de bienes y servicios para generar su valor agregado, pero no pueden cargar IVA en la venta de sus productos, esto significa que estas actividades operan con insumos con IVA, pero no pueden recuperar este sobrecosto para vender su producto sin WVA, o tambien pueden vender su producto a un precio que les permita recuperar su margen de rentabilidad deseado, asi como el IVA que conlleva en sus costos de generar su valor agregado, en este uiltimo caso opera a costos superiores del 13% que sus competidores intemacionales. 5.9. Politica de Draw Back La politica de de "draw back" implementada en marzo de 1990, a traves de la vigencia de una Ley para la Reactivaci6n de las Exportaciones (Decreto Legislativo No. 460) busca compensar el sesgo antiexportador, mediante el reintegro del 6% del valor FOB de las exportaciones no tradicionales destinadas a mercados fuera de la regi6n centroamericana. Las actividades que quedan excluidas de los beneficios corresponden a la industria de maquila, las tradicionales agricolas: cafe, algod6n y azucar, y las exportaciones no tradicionales al Mercado Comin Centroamericano. No obstante, se aplica el beneficio del reintegro del 6% para las exportaciones tradicionales que logran incorporar un valor agregado nacional en la fase de transformaci6n superior al 30%/o, casos tipicos el az2ucar refinado o el cafe soluble.7 Las cooperativas de la reforma agraria tambien son excluidas de los beneficios del "draw back" en vista de que gozan de exenci6n de aranceles (arancel tasa 0%) para las importaciones de bienes intermedios y de capital destinados para el desarrollo de su actividad productiva, segun consignado en la Ley de Asociaciones de Cooperativas (Decreto Legislativo No. 481, julio de 1992). 7vease para mayor ampliaci6n de este tema la investigacion de Angel, A., Los Sistemas de Reintegro Arancelarioy Devoluci6n del IVA. Chemonics International, Inc.-AID. Agosto de 1996 30 Annex 1: Diagnosis La estrategia con el "draw back" es desarrollar una politica neutral para remover el sesgo antiexportador, que permita compensar los impuestos pagados por la internaci6n de las materias primas imnportadas (aranceles). No obstante, se discrimina arbitrariamente entre las categorias de bienes tradicionales y no tradicionales, asi como entre el destino reginal o extrarregional de las exportaciones. Adicionalmente, existen criterios diferentes y contradictorios entre la Ley de Reactivacion de las Exportaciones y la politica de condiciones fmancieras mas blandas que administra el Banco Multisectorial de Inversiones (BMI), en tanto este Banco califica como creditos elegibles bajo la categoria de no tradicionales al cafe organico, para obtener los beneficios del Fondo de Credito para el Medio Ambiente, FOCAM, con tasas de interes en moneda nacional al 6% y a 20 afnos plazo, con 6 afios de gracia contemplados. En la medida que se avance en la reforma de reducci6n arancelaria prevista, de llevar los insumos al 5% y 10% para los bienes intermedios que compiten con producci6n local, a mediados de 1999, el esquema de reintegro del 6% para compensar el costo de los aranceles por importaciones incorporadas en el valor exportado, pierde su sustentaci6n y se convierte en un instrumento de subsidio con un esquema que discrimina arbitrariamente a las exportaciones de origen agricola, calificadas de tradicionales. En las condiciones vigentes posterior a agosto de 1996, con tasas de aranceles vigentes de 3% materia prima y de 10% para bienes intermedios, y 15% bienes intermedios que compiten con la producci6n local, la tasa de reintegro del 6%, de hecho se esta convirtiendo en un subsidio para la mayoria de exportadores beneficiarios, excepto para los casos con una estructura de costos del 40% o mas con una tasa de aranceles promedio del 15%. Por ello, en las medidas de reactivacion de la economia anunciadas en junio de 1996 (12 puntos por cumplir) comprende la medida de mantener el "draw back" como un incentivo para la producci6n mas que como una politica de compensaci6n. . . . . . . . . . . . . . . . . . . ... ~ . . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . THE AGRICULTURE SECTOR: PRICIN{G POLICIES AND COMPETITIVENESS I. INTRODUCTION Partly as a result of the structural reforms that began to take place in 1989, and partly as a result of the end of the civil war, El Salvador's economy has been showing impressive signs of improvement in most of its economic indicators. GDP growth averaged 6% per year during 1990-95, which compares very favorably with the poor performance of the 1985-89 period (1% growth per year). Growth has been led by the services, industry, and mining sectors. In sharp contrast with this dynamic economic environment; the agricultural sector is lagging behind the rest of the economy, with little signs of reversion from this negative trend. As a result, the share of the agricultural sector in total GDP has systematically declined, from a high 17% in 1990 to less than 14% by 1994. Although this negative trend started much earlier-by 1975 the agricultural sector accounted for more than 30% of GDP- the question remains whether there is room for a more dynamic agricultural sector within El Salvadors current economic framework. The poor performance of agriculture raises questions about the sustainability of growth in the future, the long run capability of the economy to generate employment and foreign exchange eaTnings, and ultimately, about future political stability, as a large proportion of the poor line in rural areas. Fiaure 1: Growth and Aaricultural Share of GDP {1985-94) 95 19 Growt .. .......... ......... sectoris.th.combned.rsult.f.phai ............ l c h ng. ............................... i c r o a e .. ...............o.........n............................. ........... 3...........................ty o h e t rv sav sot e e t r ft ee.. *. General.. nacroeconomic..onditions:....ch...........he..ggregate..evel..f..inancial resources(e.16, ...... savings), theaggregat...............umancaptalintheeonomy,et Growllasthe i aonyd iveons soectnorogisa th oge s cobned resultio of pyiaandw tehumno capitaescmltin.ogte with... te. ooia change................trn.te.ac.t.hih.umn.ndpysca.cpial.a.el.a.tcholgyi incorporated..............depends on:........... *.... Relaiveprics:.hic.affct.he.elatve.roftabiityof.te.scto.visa.vs.oher.ect...f.th.ecnom ....... ..ene...l..ma..roeconom.. IcI onitos:wh.h.nsranth.agrgteleelo..nacalreoucs.... asawe ly s asslto the odtosfr teuchnoogial preomies tand beaopin tof naew tlaeholog adpatles.rsut f h 2 Annex 2: Competitiveness If one looks at all three determinants, it becomes fairly evident that the agricultural sector in El Salvador is facing significant constraints to growth. To begin with, agricultural prices relative to other goods have been steadily declining since 1990. As Table 1 shows, most importable agricultural goods of El Salvador have seen their real price decreased by percentages in the range of 20% to 34% between 1990 and 1995, with the most important part of the decline being concentrated during 1993-95. There is little doubt that this negative evolution of real agricultural prices must have influenced the intersectoral allocation of physical and human capital, therefore implying a negative constraint to future growth. Table 1: Agricultural Prices, Selected Importables Goods (1990=100) Year Maize Rice Milk Live Cattle 1990 100 100 100 100 1991 100 158 102 112 1992 87 106 87 101 1993 70 80 81 81 1994 97 87 94 71 1995 66 80 67 67 Source: Ministerio de Agrcultura y Ganaderia, El Salvador. One of the objectives of this paper is to analyze the various causes underlying the decrease in relative agricultural prices, assessing the relative importance of the real exchange rate vis a vis other factors. In effwet one of the most important factors underlying the decline in real agricultural prices has been the decrease in the real exchange rate, which declined at 7.5% per year during 1990-95 (Figure 2). In turn, the decline of the real exchange rate has several causes. Among them, the most obvious seem to be the return of confidence-with the associated consequences in terms of demand for real estate and other non tradables-as well as the increasing flow of remittances, and common shocks affecting the Latin American region. However, and as will be argued later, the lack of infrastructure, storage capacity and adequate credit intermediation have also had an important impact on low agricultural prices, particularly at harvest times. In any case, the flow of physical and human capital into the sector has been discouraged by the negative evolution of real agricultural prices. Figure 2: Real Exchange Rate and Remittances 100 - - 95 0- 9 C so - ~ \ \ 1 n ^ 11 1>- - 85 80 - ~~~~~~~~~~~~80 70 - ~~~~~~~~~~~~70 60 - 6 ~ 65 50 60 _i _ -n Nn 'qt ^^att r -Re m ittan c e s -Real Exchange Rate Looking now at macroeconomic constraints, the striking feature that comes out from the data are the very low levels of national savings. In effect, national savings stand at around 5% of GDP with foreign remittances forming the bulk of the financing of national investment, which nevertheless reaches only 12% of GDP. Therefore, not only low relative profitability must have negatively influenced the flow of capital into the agriculture sector, but the aggregate level of investment of the economy is constrained by low national savings. In Annex 2: Competitiveness 3 other words, even if relative agricultural pffces had been higher during 1990-95, aggregate investment levels would have imposed an important constraint on capital accumulation, at any rate. Finally, if one focuses the attention on the institutional and market organization constraints, there are various barriers and constraints which fulrther discouraged capital accumulation in agriculture. To begin with an evident barrier is the scale of exploitation. Agricultural land in El Salvador is highly divided into very small plots, which acts as an important constraint on capital investments. In effect, since capital investments tends to be indivisible, there is a positive relationship between capital investment and optinal scale of production. Hence, the lower the scale of exploitation, the higher the "shadow cost" of capital investments. Although rural cooperatives could be seen as an alternative to the scale problem, in practice, the experience with most cooperatives has not been successful and an important proportion of them have undergone a bankruptcy process or are facing serious financial problems. But scale is not the only non-price barrier. Ideological constraints have restricted the incorporation of private capital into agro-industry. For example, there are limits to the maximum share that private capital may get in privatized sugar mills. Also, and according to most interviews, the rural areas are still a "dangerous place" to live in, since delinquency, cnme and robbery have substituted guerrilla and terrorism in recent year. Given this environment, the poor performance of agriculture in El Salvador should not be surprising. The sector is affected by lower historical human and physical capital, due to migration and physical deterioration during the civil war. The post-war period has not yet provided conditions adequate to a quick recovery of agricultural activities. Relative prices have reduced agricultural profitability which, coupled with unfavorable non market constraints (scale of exploitation, political restrictions, etc.), have implied that little of the scarce capital accumulation of the economy is poured into the sector, thereby severely constraining future growth. The present report looks with greater detail at one of the elements singled out above: relative prices. The broad issues posed here concern the determinants of relative agricultural prices and the extent to which these may be distorted. It begins in the next section with a brief review of the major direct price intervention policies that have taken place in El Salvador since the early 1990s. It continues in section m with a "price decomposition" exercise, analyzing the determinants of the decline in relative agricultural prices, including an the relative importance of the real exchange rate, international prices, protection levels (commercial policy) and domestic marketing factors. Section IV looks in greater detail at the real exchange rate and analyzes the extent to which the currency may be overvalued. Section V focuses on the relative protection levels of various agricultural goods, looking at potential distortions within the price structure of agricultural goods. Finally, Section VI summarizes the main results and policy recommendations. II. AGRICULTURAL PRICING POLICIES: In what follows, and to the extent that available information allows, we will concentrate the analysis on three kinds of importable products: grains, milk and bovine meat; two exportables: sugar and coffee; and one non traded product: chicken meat. Table 2 gives the relative importance of each of these agricultual goods measured by their participation in total value added of production and in foreign trade. The selected products account for about $951 million of value added of production, representing 85% of total agricultural value added and over 35% of total exports of the country. General Trade Policies During 1990-95 El Salvador reduced its imnport tariffs, non-tariff barriers, and tariff dispersion. Thus, while in 1989 the tariff range was 0-290% and there were 25 tariff rates, toward 1995 the range had been reduced to 5-20% and the number of tariff rates to 3 (World Bank, 1995). The average import tariff of the 4 Annex 2: Competitiveness economy went down from 17.8% in 1990 to 10.7% as of January 1995. These efforts however, although important have not yet had a sizable effect on the degree of openness of the economy. Measuring the degree of openness of the economy by the ratio of volume of trade over GDP, it has remained roughly constant at about 38% over the last five years. It is expected that for a small countiy like El Salvador, long-term volume of trade levels should be significantly above 50%. Table 2: Value Added and Consumption-Production Ratio, Selected Products, 1994 Product Share Value Added Consumption-Production Ratio hipoftable Goods Grains 18% Maize n.a. 1.13 Rice n.a. 1.14 Sorghum n.a. 1.03 Livestock 15% Bovine Meat n.a. 1.14 Milk n.a. 1.51 Exportable Goods Coffee 27% 027 Sugar Cane 4% 0.68 Non Tradable Goods Chicken 21% 1.00 TOTAL SECTOR 100 0.82 Notes: 1. (agricultural GDP+agricultural impots-agricultural exports)/agricultural GDP source: Ministerio de Agricultura y Ganaderla de El Salvador, 1996 Part of the explanation for the slow response of the volume of trade to the opening of the economy, has to do with several constraints, some of which are related to cumbersome and non-transparent import procedures. To some extent then, the reduction in import tariffs hidden non tariff import costs which are of an administative nature. Also important has been the fact that some exports which used to be significant in the early eighties, like cotton, have virtually disappeared'. Table 3: Tariff of the Economy and Degree of Openness Year Tariffs Openness 1990 17.8 38% 1991 14.7 37% 1992 12.5 38% 19941 11.6 38% 19952 10.7 n.a. Notes: Openrness is defined as (exports+importsYGDP, agricultural exports considered sugar, coffee and cotton 1. June-1994: 2. January-199 Sources: World Bank( 1995) and Fusades(1994a) As will be discussed later, the free trade regime wiin the Central American zone, combined with a rather high protection in some agricultural goods, gives rise to a distorted pattern of relative prices. The strategy toward a more open economy has been combined by a policy encouraging non traditional exportables. Thus, in 1990 the Ley de Reactivaci6n de las Exportaciones was enacted which affects all finms exporting goods and services, with the exclusion of coffee, cotton and sugar, which account for 35% of total exports of the country. The export promotion law, distnguishes three categories of finns: * Firms which have some export activity: they are granted a 6% drawback and are exempted from the rubber and stamp tax on exports, and other indirect taxes. * Firms which export 100% of their production: they are granted a 6% drawback are exempted from the ' According to recent studies, cotton production is no longer competitive. Additionally, it has a negative environmental impact because an intensive use of dangerous agrochemicals (Mnisterio de Agricultura, 1996). Annex 2: Competitiveness 5 rubber and stamp tax on exports as well as any tax on patrimony. * Firms which export traditional products (coffee, cotton, sugar) which have been subject to a transformation process incorporating as a minimum a 30% of additional value added: they are granted a drawback of 6%. If additionally, the company works on "maquila" it can apply to the regime of temporal admission of foreign goods (free of import taxes). It is interesting to note that this export promotion regime discriminates against traditional exportables, despite the fact that these products have not had a particularly good performance in the recent past and are in need of new investments to keep them competitive. There seems to be little ground for favoring non traditional over traditional exports. Table 4 shows the evolution of coffee, cotton, sugar, and non traditional exports. Table 4: Export Performance of Cotton, Coffee, Sugar and Non-Traditional Exportables (US$m). Non- Year Coffee Coffon Sugar Traditional' 1984 450 9 26 221 1985 464 29 23 169 1986 545 5 25 161 1987 352 2 12 205 1988 358 0 19 215 1989 229 1 14 245 1990 260 1 20 285 1991 220 1 32 316 1992 151 2 45 380 1993 235 0 35 446 1994 271 0 28 497 Notes: 1. Non-Traditonal exports=total exports-(sugar exports + coffee exports + cotton exports + shnmp eports) Source: Fusades(1994a) Also important is the fact that non efficient bureaucratic procedures imply that the drawback is typically paid to exporters with a significant delay, usually exceeding the stated 45 days period. Although Table 4 shows a dynamic pattem of non traditional exports, most of the growth is accounted by non agricultural exports. Non traditional agricultural exports, mostly tropical fruits, face various problems which have precluded further development. Among them, the most important are high air transport costs-around 40% of those exports are sent by air; low critical mass of producers-e.g., there are only four main water melon producers; and fierce competition from other tropical countries with an entrenched competitive position in the major international markets. If anything, the policy encouraging non traditional exports has benefited exports of a non agricultuiral origin. Value Added Tax During 1992 a VAT of 10% was applied to most goods. However, beans, white maize, rice, fresh fruits and vegetables, and liquid and powder milk were exempted from the VAT. As a consequence, a negative effective protection effect was generated, since these goods have to pay the VAT on their inputs (e.g., ferdlizers and pesticides) without the possibility of recovering it in the final price. The VAT was increased to 13% in 1995. Agricultural Pricing Policies During 1990, El Salvador, like several other countries of Latin America, implemented a system of price bands, which involved variable specific import duties on a number of importable agricultural goods. As usual in this scheme, the variable duties were imposed depending on the level of intemational spot prices as compared with historic prices. The price band system was applied to rice, maize, sorghum and operated jointly with a flat import tax of 20% which was subject to contingent reduction in case the import cost of the good was above the ceiling 6 Annex 2: Competitiveness price. However, after three years the system was dismantled in 1994 and only the flat ad-valorem import tax of 20% prevailed. Later in 1995 import taxes on yellow and white maize were reduced to 5%. Although a complete analysis of the costs and benefits of the price band regime is not available, the intemational experience suggests that price bands in agricultural products trend to generate only modest welfare gains, and that happens only when they are applied to products of extreme intemational volatility like sugar (Quiroz and Valdes, 1993). As will be argued in the next section, most of the price fluctuations in the products that used to be subject to price bands in El Salvador, come either from intemal market conditions or from real exchange rate movements. Hence, it is likely that the price band regime in El Salvador was of little use to ameliorate real producer price risks. On the other hand, the elimination of the price band avoids complicated relative price distortions that arise with derivative products, e.g., a price band in maize raises the need to adopt some compensatory scheme for chicken, pork and other derivatives. Marketing Policies In 1991 the Regulating Institute of Rice Supply (IRA) was dismantled and privatized, eliminating thus the govemment intervention in marketing. In maize, the Strategic Grain Reserve was eliminated in 1995. Currently, there is almost no government intervention in domestic marketing. However, in the case of milk, there are stricter-than-usual quality standards to the final product, like the prohibition imposed during 1995 to sell liquid milk reconstituted from powder milk, and the prohibition to market milk, cheese and cream elaborated with adulterants. It is uncertain whether these restrictions have gone beyond strict health considerations. III. UNDERSTANDING THE EVOLUTION OF RELATIVE AGRICULTURAL PRICES The Determinants of Relative Agricultural Prices For agents producing a specific commodity i, what matters is the relative price of the good (P,) relative to a given numeraire (P). If we denote by gP* the international price of the good i, E the nominal exchange rate (colones per dollar), P the general price level abroad, ti the tariff of the product and 0b as "others factors" intervening in the arbitrage between border and producer prices, we can decompose the domestic real price of good i into four components: Pf (;) . (EP ) . (I + i). (1) where the first component is the real international price of the good; the second is the real exchange rate; the third is the equivalent tariff of the product; and the last one summarizes the "others factors".2 Taking logs to (1) and applying first differences we obtain: pi = pi + e + Ti + T (2) where: P. -P. ~ EP pi = Aln-I p,P = Aln i*,e = Aln ,T=Aln(l +ti), = Aln06 The above decomposition, when applied to real agricultural prices, allows disentangling the various 2 See Quiroz and Valdes (1993) for further details. Annex 2: Competitiveness 7 determiinants of them. Therefore, using tbis identity we can asses the relative importance of the real exchange rate, international prices, domestic tariff protection, and "other factors" in the evolution of real agricultural prices. "Other factors" include primarily all the miscellaneous determinants that affect domestic marketing-that is, factors that intervene in the arbitrage between import costs and domestic prices in the case of an importable good, or the arbitrage between fob price and the producer price in the case of an exportable. The results of this decomposition exercise are presented in Table 5 .for rice, coffee, and sorghum. White maize was not included because there does not exist an international price for this product. The same observation applies to milk products and live cattle. The following important facts emerge from the table: Table 5: Decomposition of Rate of Growth of Domestic Prices of Selected Products, 1991-95 item 91-92 92-93 93-94 94-95 91-94 91-95 Change of the Domestic Real Price Rice -40% -28% 8% -9% -60% -69% Coffee 0% 72% 27% -60% 99% 39% Sorghum -15% -8% 15% -33% -8% -41% Change of the International Real Price Rice -6% -4% 31% -26% 21% -5% Coffee -44% 14% 54% 22% 25% 47% Sorghum -2% -5% 4% 8% -3% 5% Change of the Real Exchange Rate -6% -11% -9% -6% -26% -32% Change of the Tariffs Rice 18% 3% -22% 5% -1% 4% Coffee 0% 0% 0% 0% 0% 0% Sorghum -1% 1% -2% 1% -1% 0% Change of "Others" Rice -45% -16% 7% 19% -55% -36% Coffee 50% 69% -18% -76% 100% 24% Sorghum -7% 6% 22% -36% 21% -15% Notes: Elaborated based on equation (2) and information of the Ministerio de Agrcultura y Ganaderia, FUSADES o The real prices of the two importable products decreased significantly over 1991-95. The decline was about 69% in rice and 41% in sorghum3. Although in the case of rice the comparison is somewhat misleading because of the high real price in 1991, the base year of Table 5, the fact that real prices declined by 20% or more is relatively robust to the choice of the base period (Table 1). * Coffee faced a better situation: its real price increased by about 39%4 during 1991-95. * The sharp decrease in the real prices of importable goods cannot be attributed to the development of international markets. The international prices of the goods under study evolved in a positive way: they decreased by only 5% in rice (1991-95), and they increased modestly in sorghum (5%) and significantly in coffee (22%). This suggests that price bands would have had little impact had they remained in place. * The major factors underlying the decrease in the real prices of importable goods were the real exchange rate and "other factors" (domestic marketing conditions). In the case of rice, an increase in the gap between import cost and producer prices was responsible for about half the total decline in the real price observed during 1991-95. Therefore, although the real exchange negatively affected the evolution of real agricultural prices, at least for the two goods studied, domestic marketing conditions played an equally important role. 3 Formula (2) states the price decomposition in tenns of log changes. For changes like those during the period of analysis, the change in the log differs from the real percentage change (the log change magnifies the percentage change). That explains the difference between the "percentage" changes of Table 5 and the true percentage changes that can be inferred from the real price indexes of Table 1. 4 The same comment of the previous footnote applies. 8 Annex 2: Competitiveness * In the case of coffee, marketing factors ("other factors') played a positive role during 1991-95, being responsible for a significant part of the improvement of producer prices. Not surprisingly however, "other factors" exhibited a positive correlation with international prices-that is, when international price increased, the markup between fob prices and producer prices diminished. This is most probably a reflection of the prevalence of fixed cost in the processing and export process. Thus, when the international price rises, the gap between the producer and the fob price diminishes proportionally. The findings that emerge from the price decomposition exercise suggest the importance of examining more carefully both the marketing process, which stands as a crucial deterninant of producer prices, and the real exchange rate. The rest of this section analyzes the marketing processes, while the next section focuses on the real exchange rate. Explaining the Behavior of Producer Prices Specific econometric models were estimates to gain more insights on the effect of "other factors" on the price decomposition. We begin with the simple hypothesis that the domestic price of any agricultural importable is influenced by both, the international price and the domestic quantity of the good produced. This must be so because of storage costs: the higher the quantity produced relative to total demand, the lower the price received by the producer. The crucial point however is to asses how important is the amount harvested in the determination of the (spot) producer price. The greater the lack of storage facilities, financing conditions and imperfections in the marketing process, the higher the importance of the quantity harvested in determining the producer price. Figures 2 to 4 provide an interesting first approach to the problem of price determination and the importance of the quantity harvested. The figures show the normalized quantity harvested of rice, sorghum, and maize together with their producer price. Producer prices decline strongly at the peak of the harvest exhibiting an impressive volatility. For example in rice, during 1993, at the peak of the harvest period, producer pricesfell to two standard deviations from the sample mean price, only to increase a few months later to two standard deviations above the sample mean.' This behavior points to a major distortion-with appropriate storage facilities and with adequate access to the capital market, such behavior would be impossible, for speculators would arbitrage prices intertemporally. The absence of intertemporal price arbitrage strongly suggests that great distortions and imperfections prevail in these markets. Figure 2: Quantity Harvested and Domestic Price of Rice 2 -2- 92:01 92:07 93:01 93:07 94:01 94:07 95:01 95:07 i~:. Quantity Harvested - Domestic Price 5The figure shows normalized values, e.g., for any variable x, it shows x minus the sample mean divided by the standard deviatiorL Annex 2: Competitiveness 9 Figure 3: Quantity Harvested and Domestic Price of Sorghum 3 2 -2 . . , . . .. i ... 92:0'1 92:07 93:01 93:07 94:01 94:07 95:01 95:07 E}] Quantity Harvested Domestic Price Figure 4: Quantity Harvested and Domestic Price of Maize 3 2 92:01 92:07 93:01 93:07 94:01 94:07 95:01 95:07 Quantity Harvested Domestic Price Table 6, shows an econometric estimation of a model determining the producer price of rice. The dependent variable is the producer price measured in nominal dollars and the explanatory variables are the lagged producer prices, current and lagged international prices, current and lagged domestic output, and seasonal dummies. The most striking result is the great importance of the quantity harvested in the price determination. One cannot reject the null hypothesis that the coefficients associated to the lagged quantities sum to zero, and therefore, one can conclude that the long-term elasticity of the producer price with respect to output is extremely high, since the lagged producer price coefficients are almost equal to one, reflecting high persistence to shocks. On the other hand, the international price level although important, carries less weight than the quantity harvested. The striking conclusion then is that, at the end, the behavior of dollar-measured producer prices had little to do with the international prices (which did not decline strongly in the period) nor with the import tariffs- the crucial determinants were related to marketing conditions, with the price of rice behaving almost like a non tradable (highly influenced by the quantity harvested). 10 Annex 2: Competitiveness Table 6: Explaining the Domestic Price of Rice Dependent Variable: Pdt(1992:01-1995:11) Independent Variables Coefficient t-stafstio Constant 1.80 1.64 Pdt 0.69 6.85 PIt4 0.30 3.58 Qa_- -0.25 -2.68 AQat.2 -0.34 -2.24 Qa,5 4-0.42 -4.32 Q't4 0.74 4.88 AP t-1 0.47 4.01 p't3 0.32 2.83 Pt-4 -0.21 -1.87 Adi-R-0.85 D.W.=2.20 F-Statistic=1 3.58 Notes: Po- logarithm of the domestic price of rce measured in dollars in the month tj Qa t logarithm (1 +quantity harvested of rice in the month t-j); Pt= logarithm of the international price of rice - CIF Acajutla -measured in dollars in the month t-j; aX= )4-).1. In addition, monthy dummies were used. Table 7, presents a similar analysis for sorghum. The econometric analysis confinns, again, the importance of the quantity harvested, at least in the short run, but in tius case, international pnce variation was found to play a more important role.6 Table 7: Explaining the Domestic Price of Sorghum Dependent Variable: Pdt (1992:02-1995:12) Independent Variables Coefficient t-statistic Constant -0.50 -0.75 Pd,1 0.94 14.99 Pt,2 0.18 1.88 Q, -0.11 -6.31 AdjlR =0.89 D.W.=1.74 F-Stalislic=39.24 Notes: P j= logarithm of the domestic price of sorghum measured in dollars in the month t-j; Q -= logarithm (1 +quantity harvested of sorghum in the month t-j); P t= logarithm of the intemational price of sorghum - CIF Acajutla -measured in dollars in the month t-j; aX= Xt-Xt; In addition, monthly dummies were used. In sum, the econometric analysis reveals that the quantity harvested plays an important role in the determination of producer prices, sometimes even more important than the international price. By all measures, the major influence of the quantity harvested on the determination of producer prices suggests great deficiencies in the marketing process. If the credit market were deep enough, the price reduction at harvested would be much lower han observed. The behavior pattern displayed by the data correspond, in the case of rice for example, to an almost complete "intertemporal autarchy". Interviews with small rice and maize producers suggest that this is indeed the case, and that the few available credit lines seldom reach the producer at the right time. To this situation one must add the need of investment in the mnarketing process of rice, which is characterized by obsolete equipment (CIRAD/iCA, 1993). The situation that emerges from this analysis suggests some straightforward policy options which could be both non-distortionary and at the same time improve producer prices. An obvious agenda along these lines seems to suggest the possibility of introducing tradable warehouse deposit certificates which could be used as collateral warrants for credit lines at harvest times. Even if those credit lines had a significant spread over the overall deposit rate, the price depression observed at harvest times could be significantly reduced. In this regard, efforts should be made to remove whatever institutional restriction precludes financial flows into grain storage. IV. The Real Exchange Rate As explained in section Im, the decline in the real exchange rate was, together with the internal marketing conditions, a very important factor underlying the evolution of real domestic agricultural prices. The real exchange rate, measured as a relative price between the U.S. CPI and El Salvador's CPI, declined by 36% 6 As the table indicates, in the case of sorghum the level of the quantity harvested is not an explanatory variable but only the change in harvest On the other hand, the level of international prices does play a role in producer price determination, particularly m the long nm. Annex 2: Competitiveness 11 between 1991-95. Undoubtedly, such a decline had an indirect negative effect on the prices of all tradable goods, particularly on commodities, which comprise the great majority of the agricultural sector. Although the real exchange rate is always an important determinant of real agricultural prices, the policy implications that can be derived from this fact are far from obvious. To begin with, one cannot overemphasize the fact that the real exchange rate is a relative price, and as such, its variations cannot be directly controlled by the authorities. The degree to which active fiscal and monetary policies can affect the real exchange rate in a sustainable way, depend to a large extent on whether the real exchange rate is or not outside its long term equilibrium path. In turn, the real exchange rate can be considered outside its long term path if there are other real fundamental macro variables that show evidence of being at unsustainable levels. The important corollary of this line of reasoning is that the decline in the real exchange rate does not imply per se that the real exchange rate is below its long run equilibrium path. istad, rather than the evolution of the real exchange rate, the crucial variables to look at are the external accounts and the way they are being financed, global export trends, unemployment levels, and the evolution of the real exchange rate in other countries in the region. For example, a high level of unemployment coupled with an unsustainable current account deficit and stagnant exports, would be evidence of real exchange rate misalignment, even if the real exchange rate were "high" by historical standards. For the case at hand, the data do not show evident misalignment in the real exchange rate. To begin with, the rate of growth of non traditional exports has been 15% during 1990-95, a relatively dynamic trend. On the other hand, the large trade deficit is financed by remittances which are expected to remain in place over the medium term. In other words, the problems that affect the agricultural sector, where non traditional exportables have not been able to gain momentum, can be considered sector specific instead of symptoms of macro disequilibria. Finally, as Table 8 and Figure 5 show, the real exchange rate decline observed in El Salvador during the 1990s, can be considered part of a regional trend that has characterized most Latin American countries. This global phenomenon is linked with the renewed confidence that the international capital markets had on the region ailer the "lost decade" of the 1980s. Within this contex, the decline of the real exchange rate in El Salvador appears as moderate (in the threshold of the lower third of the sample), and perfectly understandable for a country that has just come out of a long civil war. Table 8: Real Exchange Rate, Selected Countries (1 990=1 00) Year Costa Rica El Salvador Guatemala Haiti Honduras Nicaragua Rep. Dornin. 1990 100 100 100 100 100 100 100 1991 108 98 88 93 108 105 101 1992 103 98 87 96 102 105 101 1993 101 87 88 110 113 108 97 1994 101 84 85 98 125 113 95 1995 98 82 83 86 113 118 94 Source: CEPAL In sumri, the empimcal evidence does not allow to conclude that the real exchange rate is grossly misaligned, and therefore, active policies aimed at increasing its level may have little real long-term effect. Hence, although the evolution of the real exchange rate did have an important effect on the behavior of real agricultual prices, there does not seem to be significant scope for modifying this situation through active economic policies. In terms of policy prescriptions then, it seems more advisable to focus on policy areas where real changes can be achieved and where distortions are far more evident, like the internal marketing conditions referred to in the previous section. 12 Annex 2: Competitiveness Figure 5: Growth of the Real Exchange Rate, Selected Countries, 1990-95 15 0%_ 10.0%_ 1:X ! .0 % I 0.0% 0 .0% 5 .0 %~~~~~~~~~~~~~~~ 1I 0.0% V. THE STRUCTURE OF PROTECTION WITHIN THE AGRICULTURAL SECTOR The discussion of section m suggests that looking at noninal protection rates as indicators of protection in some importable goods may be misleading. As explained, domestic marketing conditions make the same products behave almost like non tradables, with the quantity harvested exerting a large influence on domestic prices. Yet, it is always useful to carry out this analysis because, if the distortions that imply almost a non tradable price behavior of importable goods like rice are removed, the structure of protection will become the relevant variable. Table 9 shows the nominal protection of various agricultural goods for the period 1990-95. Most importable agricultural goods, with the exception of maize, face an ad-valorem tariff of 20% (10% for maize). This apparently uniform protection structure is not such when one looks more carefully at each good. Table 9: Tariffs for Selected Products Maize Rice Sorghwn Ad- Price Ad- Price Ad- Price Live Year Total Valorem Band Total Valorem Band Total ValoreFn Band Milk' Cattle2 1991 17% 20% -3% 16% 20% -4% 20% 20% 0% 5%-35% 20% 1992 20% 20% 0% 35% 20% 15% 19% 20% -1% 20% 20% 1993 20% 20% 0% 39% 20% 19% 21% 20% 1% 20% 20% 1994 19% 20% -1% 16% 20% -5% 19% 20% -1% 15% 20% 1995 10% 10% 0% 20% 20% 0% 20% 20% 0% 15%-20% 20% Notes: 1. Range corespond to different types of milk; 2. The trade with Carbbean counties is excepted of tariftf Source: MAnitero de Agricultura y Ganaderia, El Salvador In principle, if one considers that the 20% rate is the highest tariff rate in El Salvador, one might conclude that the effective protection rate of agricultural goods which have a 20% nominal protection should be 20% or higher. That is because, in the worst case, all tradable inputs used in the production of these goods could have also the highest tariff rate and therefore, the effective protection, at worst would also be 20%. Actually, all agricultural inputs and agricultural equipment are in the 5% category, and therefore, in the absence of any other qualification, effective protection rates ought to be higher than 20%. This conclusion however must be qualified by the fact that the value added tax applies to agricultural inputs but is exempted in several agricultural outputs. More specifically, it can be shown that the effective protection rate of an agricultural good i having a Cobb-Douglas production function (which implies constant shares of inputs in total output value in the long run) and having all inputs with a tariff rate b can be approximated by: Annex 2: Competitiveness 13 (ti- sh tj - shij VAT) F-Pi =(I -shyj (3) where ti is the nominal protection rate of good i; shij is the share of inputs j in the output value of good i; 4 is the nominal protection rate of inputs and VAT denotes the value added tax which we assume that does not apply to good i. If good i is also subject to VAT, the VAT would have to be added to the numerator. Using the cost figures reported in OAPA/MAG (1995), Table 10 shows estimates of effective protection rates for: (i) agricultural importables, rice, milk, white maize and sorghum; (ii) traditional exportables, sugar; and (iii) non traded, chicken, which is treated here as potentially importable product for the purpose of effective protection calculation. Although in several cases the agricultural good is exempted from VAT while the inputs are not, the effective protection rates are still quite high, with the notable exception of maize. In almost all the products the effective protection rate is higher than the 20% threshold value. Table 10: Effective Protection Rate, Selected Products, 1995 Product t sh1 VAT Effecfive Protection Rate Rice 20% 5% 56% 10% 26% Sugar 20% 5% 36% 10% 23% Milk 15% 5% 77% 10% 15% WhifeMaze 10% 5% 58% 10% 3% Chicken 20% 5% 76% 10% 35% Sorghum t 20% 5% 68% 10% 31% Notes: "erZOAPAIMAG, 1995 Therefore, on a first approxirnation, one should conclude that effective protection of importable agricultuiral goods in El Salvador is quite high. This result does not imply that agricultural importables are in a good competitive position however. As the analysis in section m mnade clear, the various constraints and distortions of the marketing process affect negatively, and in a major way, various producer prices. This qualification explains why the effective protection estimates of OAPA/MAG (1995) are lower than ours: OAPA/MAG did not use the actual nominal protection rate to the output in their calculations but the implicit nominal protection that results from comparing the intenational price with the domestic one. We rather prefer not mxing up what belongs to the trade and tax policy areas with other considerations which affect the marketing process and happen to be the result of various other factors (infrastructure for storage facilities, financial conditions, oligopolistic practices, etc.). In any case when the effective protection calculations presented here are combined with the evidence of section m1, these results and those of OAPA/MAG (1995) become broadly consistent: the negative effect induced by the internal marketing conditions dominate over the apparent high effective protection rates that result from the sector's tariff structure. The above results are subject to some important qualifications: * Livestock: since there is free trade with Central America, and Honduras is a very competitive exporter, the producer price of live cattle in El Salvador is more or less equivalent to a free trade situation, being even lower than $1 per kg. of live animal. In other words, the 20% flat import tariff rate is non binding. These lower producer prices, however, are not transferred to the consumer. Again, marketing and processing explain e difference. As Table 11 shows, the carcass/livestock price ratio of El Salvador is the highest of all the sample of Latin American countries considered. The implications are straightforward: while in Chile the price of livestock paid to the producers is about 25% higher than the price received by Salvadoran farmers, the carcass price is about 7% lower in Chile than in El Salvador. And this happens despite the fact that Chile is by no means an example of efficient industrial processing in the sector (GERENS, 1994). 14 Annex 2: Competitiveness Table 11: Intemational Evidence for Livestock Sector, 1994/95 (US$/Kg). Uve Ankal Carcass Ratio Country (1) (2) (2)1/(1) Costa Rica 0.88 1.68 1.91 El Salkdfo 0.96 2.50 2.62 Guatefmla 1.25 2.41 1.92 Honduras 0.91 1t69 1.85 M64dco 1.00 1.59 1.60 Nkaragua 0.82 1.55 1.89 Panami 1.02 1.95 1.91 Rep. Dorn. 1.34 2.14 1.59 Chie 1.21 2.38 1.97 Argefnta 0.68 n.a. n.a. Sources: Ministerio de Agricuftura y Ganaderia, El Salvador ODEPA and Agroeconornica Consultora * Milk: with the price of milk, the opposite occurs. The producer price of milk in El Salvador is one of the highest in Latin America (Table 12), which is partly explained by the fact that a substantial part of milk production is processed into fresh cheese, which is basically a non tradable. In addition, the quality regulations inposed in 1995 may contribute to further isolate the price of milk in El Salvador from its international opportunity cost. The dichotomy between milk and livestock prices should be a matter of concerm-the prices of livestock are consistent with prices faced by large extensive producers, while the price of rnilk is more in line with that of a highly-protected sector. This suggests that in relative terms, the country may be better prepared to compete in livestock for meat consumption instead of milk. Table 12: Intemational Evidence for Dairy Sector, 1994/95 (US$/ft) Producer Price Consuner Price Ratio Country (1) (2) (2)1(11 Costa Rica 0.30 0.42 1.41 El Savador 0.42 0.71 1.68 Guatemala 0.44 0.63 1.43 Honduras 0.30 0.47 1.54 Mixico 0.24 0.40 1.71 Nicaragua 0.32 0.51 1.59 Panami 0.24 0.74 3.05 Rep. CDon. 0.33 0.96 2.94 Peru 0.23 n.a. n.a. Brazi 0.21 n.a. n.a. Argentna 0.20 n.a. n.a. Uruguay 0.17 n.a. n.a. Chile 0.23 n.a. n.a. Sources: Ministero de Agricutura y Ganadera, El Salvador. FEPALE varous Lssues and ODEPA * Maize, Chicken Meat, and Chicken Parts: chicken meat basically behaves like a non tradable in El Salvador. The argument for protection of the chicken sector in El Salvador is similar to those advanced in other Latin American countries-that is, the sector should be protected against the excess supply of chicken legs that would come from the U.S. if a freer trade regime in chicken parts prevailed. However, protectionism to chicken producers has gone too far, regardless of the validity of the U.S. chicken-legs argument. The producer price of chicken in El Salvador is $1.87 per kg. compared with $1.23 in Chile, which also protects the sector from U.S. chicken leg exports and, like El Salvador, has to import maize. The explanation: while the chicken-maize price ratio in El Salvador is 9.55, in Chile it is only 7.17 (Table 13). Again, the cause has to be sought in the processing and agro-industrial chain rather than in output or producer prices. Furthermore, in El Salvador the protection to yellow maize is low because, it has been argued mostly by chicken producers, since the country does not produce yellow maize no protection should apply, regardless of the fact that the country does produce white maize. Empirical research tells a different story: the evolution of yellow maize pnices does influence the evolution of white maize prices (see econometric estimation presented in Table 14). Therefore, the picture iat begins to emnerge is the following-with the argument that yellow maize prices do not matter for the determination of white maize prices, they are given a low protection rate. Since yellow maize prices do affect white maize prices, in practice protection to white maize producers is negatively affected. Then, with the argument of the Annex 2: Competitiveness 15 excess supply of chicken legs in the U.S. the sector is virtually isolated from international competition, with the result that chicken meat prices are the highest in Latn America. Table 13: Intemational Evidence for Poultry Sector, 1994/95 (US$/Kg) Price of Chicken Price of Maize Ratio Country (1) (2) (2y(1) Guatemala 1.97 0.19 10.28 El Salvador 1.87 0.20 9.55 Chile 1.23 0.17 7.17 U.SA 1.07 0.08 13.93 Sources: Ministerio de Agricultura y Ganaderia, El Sahrador; and Asociaci6n de Productores Avicolas de Chile Table 14 Explaining the Domestic Price of White Maize Dependent Variable: Pdt (1992:05-1995:12) Independent Variables Coefficient t-statistic Constant -1.30 -1.33 Pd F1I 0.87 9.33 Pdt.3 0.33 1.93 F)dp4 -0.17 -2.55 p^, 0.21 2.38 AQs1 -0.05 -3.54 AQS 2 -0.01 -2.51 AdjeRz0.89 D.W.=2.03 F-Statistic=45.50 Notes: P',F logarithm of th domestic price of white maize measured in $ in month t-j; P1'tF Pt + P't+i; QS-F bgarithm (1+maize quantity harvested in the month tj); P te logarithm of the intemational price of yellow maize-CIF Acajuta-measured in dollars in the month tj + intemational price in the month tj-2; AX= X-X-1; In addition, monthly dummies were used. VI. CONCLUSIONS AND POLICY RECOMMENDATIONS The economic performance of the agncultural sector in El Salvador has been poor, both in absolute tems and in comparison with the rest of the economy. Unfortunately, there are significant constraints to future growth, among which the most important are: (i) the evolution of relatve agricultural prices, which have been declning systematically during the 1990s; (ii) the low aggregate level of savings and investments in the economy; and (iii) institutional and market organization constraints, the most obvious being the scale of exploitation. The above consbaints restrict in an important way the flow of physical and human capital into the agricultural sector and also the process of technology adoption, all of which are crucial determinants of future growth. The most important conclusions that come out of this study are: * Real prices of importable agricultural goods have diminished significandy during the 1990s (20% or more). Due to the particular evolution of international prices, the prospects of coffee producer prices have been better. * One of the most important determinants of the decline of real prices has been the evolution of the real exchange rate. Although the real exchange rate was not the main focus of this report, it suffices to say that there are good reasons to believe that no quick reversion from this decline can be expected in the near future, and if anything, the negative trend might continue for a while. To a large extent, the decline in the real exchange rate is the result of increased confidence in the economy and remittances. Available evidence do not suggest a gross real exchange rate misalignment. * This study revealed also that there are major imperfections in the marketing chain connecting international prices with producer prices. In cases like rice the absence of intertemporal arbitrage is evident and is reflected as an important influence of the quantity harvested on the spot producer price. A similar phenomenon happens in sorghum, although there the imperfections are smaller. * Also important is the fact that there still prevail important distortions in the structure of relative prices within the agricultural goods basket. Producer prices of livestock are close to the free trade prices of an exporter 16 Annex 2: Competitiveness country while producer prices of milk are among the highest m Latin America. Protection to white maize is lower than it seems because the low protection to yellow maize does influence white maize prices. On the other hand, the effective protection to chicken producers is extremely high, and again, chicken meat prices in El Salvador are among the highest in Latin America. Finally, non traditional exports are partially compensated through a drawback from protection to importables but there is little ground for not protecting traditional exports on the same grounds. The above findings provide a basis for policy recommendations of two types: 1. Marketing Chains: It is urgent to improve the agricultural marketing chains in El Salvador, particularly in agricultural importables like rice. Basically, policy reforms would have to improve the conditions in order to bring capita into the sector. For example, in the case of rice marketing, it is necessary to attract "speculators" who can undertake intertemporal pricing arbitrage which is virtually nonexistent today. In order to bring this kind of agents, several steps must be carried out: (i) introduce quality standards compatible with broader trading (heterogeneous products are not compatible with outside trading); (ii) introduce appropriate waffant mechanisms to allow credit to flow smoothly into this activity; and (in) improve information and trading infrastructure. 2. Rationalization of Protection: * Make the VAT applicable to all goods: there is discrimination against certain products which have to pay the VAT at the input level but are exempted at the output level. * Make the drawback applicable to all exportables: non traditional exports are not eligible for the drawback but there is litte ground for this kind of discrimination. * Reduce protection to chicken producers, and/or, increase protection to white maize to align it with the tariff on yellow maize: effective protection to chicken producers has risen far above other agricultural goods, certainly beyond the level necessary to protect the industry from excess supply of U.S. chicken legs. In part this has been done at the expense of white maize producers, since there are clear links between yellow and white maize prices. Finally, it must be stressed that these recommendations will not by themselves bring back agriculual growth, but should be part of a comprehensive and longer-term sector strategy. Annex 2: Competitiveness 17 REFERENCES Avelaneda, R- , Ramirez, J. and E. Echeverry (1995). "Informe Sobre la Caficultura de Costa Rica, Guatemala, El Salvadory Honduras: 1995". Federaci6n Nacional de Cafeteros de Colombia, Santafe de BogotA. CIRAD/l1CA (1993). "Situaci6n Actual, Oportunidades y Desafwos de la ActividadArrocera en El Salvador". FUSADES (1994a). i C6mo esta Nuestra Economia?. FUSADES. FUSADES (1994b). "Una Estrategia de Desarrollo Agricola para El Salvador: 1994-2000". Documento de Trabajo N° 37. FUSADES. Ministerio de Agricultura y Ganaderia (1995). "Agenda Nacional de Concertacion para la Reactivaci6n del Sector Agropecuadrio ". Ministerio de Agricultura y Ganaderia (1996). "La Competitividad del Cultivo del Algod6n en El Salvador". Internal report. OAPA/MAG (1995). "Plan Econ6mico de El Salvador 1995-1999, Anilisis Efectos en el Sector Agropeacuario". Politica Agricola N° l. Ministerio de Agricultura y Ganaderia, El Salvador. Ramos, H. (1995). "Evaluaci6n del Sistema de Bandas de Precios de Inportaci6n en El Salvador". Informe de Coyuntura Dic. 1995, OAPA/MAG. Quiroz, J and A. Valdes (1993). "Price Bands for Agricultural Price Stabilization: The Chilean Experience". Documento de Investigacion 1-64. Programa de Postgrado en Economia, ILADES/Georgetown University. The World Bank (1995). "El Salvador: Meeting the Challenge of Globalization, Country Economic Memorandum/Private Sector Assessment". Report N° 14109-ES. I RURAL POVERTY: A QUANTlTATTVE ANALYSIS' Introduction Almost 50% of the population of El Salvador is considered rural and the extent and intsity of poverty has been shown to be much worse in rural areas than in urban ones according to the World Bank, 1994. Although the above study has provided a thorough characterization of both rural and urban poverty in El Salvador, there is very litfle quantitative understanding of the deminants of rural poverty. What is the role of education, access to land and capital as well as of other demographic characteristics in determining rural household income? How does rural infrastructure affect the potental income of rural households? How responsive is household income to greater participation in the labor force by women and children? These are important questions that may have significant policy implications. Apart from the emphasis on trying to determine the quantitative imnportance of certain key variables on household income, this study will also provide a detailed analysis of the relative importance of agricultural vis-a- vis non-agriculural sources of income for the various income groups. Additionally, efforts will be made to relate poverty to the functional sources of income. That is, to analyze how the propensity of being poor changes across various rural classes, namely, farmers, landless agricultural workers and landless non-agricultural workers. The Data The data used is based on a survey of rural households implemented in early 1996 as part of this study. The survey covered a sample of about 630 rural households randomly obtained from all regions of El Salvador. The sample was designed to be representative of the rural population at 10% level of significance. The idea was that the sample would reflect the rural households' characteristics according to their main economic activities, i.e., self-employed farmers, landless agricultural workers and landless non-agricultural workers. The sample was based on the labor force survey implemenied in 1992 to reflect the current rural population. According to ths survey, about 32% of the rural labor force are farmers that live in rural areas, about 43% are agricultural landless workers and 25% are landless workers in non-agricultunal activities.2 The random sample reflects this same distribution. The sample was straified by department (El Salvador has 14 departments) to reflect the distribution of each one of the three groups of households by department. Thus, each department is represented in the sample according to the proportion of the labor force in each group that reside in each department, according to the 1992 labor force census. ' This background paper was written by Ram6n L6pez (Department of Agricultural and Resource Eonomics, University of Maryland. Excellent research assistance was provided by Claudia Romano. Comments by Ana-Maria Arriagada and Cora Shaw and Alberto Valdes on an earlier version were very helpful. The field work was developed by FUSADES under the direction of Margarita de San Feliu. Financial support for this research was provided by ICEG, USAID, El Salvador and the World Bank. 2 This classification reflect the primary activity. Many people classified as farmers also do off-farm work in agriculture or in other sectors and many landless farm workers also do non-agricultural activities and may have a small plot of land. 2 Annex 3: Rural Poverty The survey obtained information of a wide range of demographic characteristics, location and income variables. Additionally, the survey for self-employed households included a detailed description of the revenues and costs of the activity and it also accounts for the value of the goods produced by the enterprise and consumed directly by the household. For all households the income questions were highly disaggregated by specific categories and, whenever possible, by each household member. This allows us to calculate the household income from a variety of specific categories and individuals within the household instead of relying on aggregate income estimates provided by the household head as in many surveys. Furthermore, apart from the income data, the survey provides a detailed account of the assets, both physical and financial, owned by the household as well as of the production characteristics of the enterprises owned by households. Main Characteristics of the Rural Households Table 1 provides a summary characterization of the rural households surveyed, classified by per capita income quintiles. The average annual per capita income for all the rural households was 4040 colones or approximately US $460 per capita. This is well below the national per capita income which according to recent estimates is about US $1200 per capita (World Bank, 1995). In most developing countries, rural per capita incomes are well below the national income. Thus, the large difference between our estimates and the national estimates may not necessarily reflect a serious problem of income underreporting. If we assume that the potential underreporting affects agriculture and non-agriculture incomes similarly, it is possible to probe the plausibility of our estimates vis-a-vis the national per capita income. According to our sample, about 50% of the rural GDP is originated in agriculture and the other 50% comes from non-agricultural activities. Thus, given that the share of agriculture in El Salvador's GDP is approximately 11% according to National Accounts, the share of the rural GDP in total GDP would be about 22%. The total GDP of El Salvador in 1995 was about US $7.1 billion and, therefore, the rural GDP would be about US $1.56 billion (0.22 x 7.1). Since the total rural population is estimated at about 2.8 million, this would imply that the per capita income of the rural population is about US $560, that is US $100 above the per capita income provided in our survey. Hence, the income underreporting was about US $100 per capita and, thus, it would be necessary to correct the estimates by about 21.7%. Interestingly, the correction used by the World Bank, 1994 was quite similar to this, amounting to 17.6% for the rural incomes. Thus, using the usual assumption that the rate of income underreporting is similar across income groups we can still compare the various groups. The correction will only affect the estimations of poverty. One important finding shown in Table 1 is that the poorest rural households are much more dependent on agricultural sources than the wealthier households. While agriculture wage income, for example, accounts for almost 50% of the total income of the poorest two quintiles, it constitutes less than a quarter of the total income of the richest two quintiles. The combined share of agriculture wages and agriculture self-employment income is about 70% among the poorest 40% of the rural households. Thus, for the poorest rural households, the performance of agriculture is crucial for their well being. A possible explanation for this could be related to differences in education between poor and non-poor households. As shown below, education appears to be of some importance in raising income in non-agricultural activities while it has much less of an effect on increasing income in agricultural activities. This implies that people with low education, the poor, would obtain higher returns in agriculture than in non-agricultural activities and, therefore, the poor are much more engaged in agricultural activities than the non-poor. Another aspect illustrated in Table I is the fact that remittances from migrants from abroad appear to play a much smaller role in boosting the income of the poorer households than that of the richest ones. Annex 3: Rural Poverty 3 This is just a corroboration of the fact that migration is mainly a middle or lower middle class phenomenon.3 The evolution of the demographic characteristics of the households across the income groups confirms the findings of previous studies in other countries. Family size and the number of children rapidly decline as households become richer. Schooling is highly correlated with income. The average years of education of the household heads for the whole sample is extremely low, less than 3 years and almost 37% of the household heads have no formal education at all. The education differences for adults among the five income groups are very large, with the top quintile having roughly twice as much education as the bottom one. An important result is that the comparisons of the educational levels of adult members and of children show a dramatic intergenerational increase in schooling. The average schooling years completed by the 13-18 year old cohort (many of which are still increasing their schooling) are more than 5 years compared to about 3 years for the adult population. Moreover, it appears that the educational differences across income groups for the children are much smaller than the differences for the adults. In fact, while the adults in the richest quintile have almost twice as much education than those in the lowest quintile, the 13-18 year old children in the highest quintile have less than 50% more education than those in the poorest quintile. Access to services is in general quite low in rural El Salvador, with only 55% of the households having electricity and about 20% having access to potable water. Further, only approximately 20% of all rural households are registered in either a public or private health system.4 As expected, the poorest quintiles fare considerably worse in all these aspects than the richest quintiles. Similarly, despite the reduced territorial size and the relatively high rural population density of El Salvador, on average rural households seem to be quite isolated with reference to distances to schools (3.3 Kin), doctors (5.2 Kin) and to a paved road (5.8 Km). The degree of isolation is much worse the poorer a household is. The Farmers Table 2 provides a similar characterization of the farmers' sub-sample, except that we have grouped them in 3 income levels instead of 5 given that the sample size is smaller. On average farmer seem to be better-off than the average rural household, with a per capita income of 4900 colones or US $560, which applying the correction factor for income underreporting would be about US $680. The poorest third of the farmers are, however, very poor with a per capita income of only 836 colones (less than US $100) which is only slightly higher than the per capita income of the poorest quintile in the complete rural sample. Poor farmers are much more dependent on off-farm employment in agricultural activities than better-off farmers. For the poorest third, off-farm agriculture wage income constitutes more han 20% of their total income while for the richest third this is less than 10%. The opposite occurs for off-farm wage income in non-agricultural activities, which is more important for richer farmers. The average farm size is quite small reaching only 4.4 manzanas. Somehow surprisingly, land is not very unequally distributed. In fact, the average farm size for the richest third is only 6.6 manzanas compared to 2.6 manzanas for the lowest third. This is clearly the effect of the land reform and, perhaps 3 See, for example, the evidence provided by Reed (1994) for Brazil. 4In principle, the public health system covers all the uninsured. The issue of registration in a health program is, however, important because people registered are likely to obtain better service. 4 Annex 3: Rural Poverty more importantly, of the fact that the vast majority of the farmers that own large land areas do not live in rural areas. Since the survey considers only rural inhabitants, the fanilies of the large farmers are not represented in the survey. Thus, this part of the survey should be considered representative of the farmers that live in rural areas and not of all farmers. The land titling situation seems extremely good by comparison to other Latin American countries (see, for example, L6pez 1995a and 1995b for Chile and Honduras, respectively). More than 80% of the farmers reported having legal titles on at least part of their land. As expected, land titling and income are correlated, with the poorest third of the farners having only 73.4% of title rate compared to 89% for the richest third. Technical assistance reaches less than 20% of the farmers, with only 11% of the poorest farmers receiving any technical assistance at all. The credit situation appears quite negative, with only 31% of the households receiving any credit between 1991 and 1995. Table 2 also provide some insights on the structure of agricultural production of farmers. The largest share in the total value of output is covered by subsistence crops (maize, corns and others) which is 50% for the whole sample, followed by cash crops (mostly coffee, fruits, vegetables and sugar cane) which accounts for 20% of the value of output and next by livestock products, accounting for 15% of the output value. The poorest farmers are much more dependent on subsistence crops. The share of subsistence crops declines from 71% among the bottom third of the farmers to 25% among the richest third. Similarly, cash or export crops are much less important for the poorest group (4%/O) than for the highest income group (43 %). Of the 300 farmers surveyed, only 12 of them were part of the land reform cooperatives. This is consistent with the small importance of the members of these cooperatives in the total rural labor force. Their per capita income is about 5% lower than the rest of the sample and they have less income per capita and land and received less credit than the rest of the sample. Soil erosion seems to be a widespread problem among farmers, with almost 44% reporting that at least some of their land suffering from erosion problems. Soil erosion affects poor farmers more frequently then better-off farmers. On the positive side, almost 50% of all farmers surveyed used some form of soil conservation practices. Interestingly, the greatest incidence of soil conservation occurs among the middle income farmers (57%/o) and the lowest among the richest third (42%). Comparing Farmers and Landless Table 3 compares the average per capita incomes and various other characteristics among farmers, landless employed primarily in agriculture, landless mixed (i.e. employed in agriculture and non-agriculture activities) and landless employed outside agriculture. The Table also provides the statistical significance of the mean differences among the four groups. From the view point of average per capita income the landless employed in agriculture are the worst-off among the four groups. In fact, this group's income is statistically significantly lower than the average income of each of the other groups. The wealthiest groups on average are the farmers and the landless employed in non-agricultural activities which have similar per capita incomes. Farmers are generally older than all the others and the head of households among farmers and landless employed in agriculture have significantly less schooling than the head of households in the other groups. Also farmers and landless in agriculture have a mush greater incidence of illiteracy than the other groups. Thus, in terms of education of adults, it seems that farmers and landless in agriculture are generally worse-off than the rest. Interestingly, farmers' children aged 13-17 are not less educated than the children in the other groups. That is, farm families may be closing the education gap with other groups, but this is not true for the landless employed in agriculture which still show a low rate of schooling for their Annex 3: Rural Poverty 5 children. In terms of access to services, farmers are better-off than the landless employed in agriculture particularly in terms of access to electricity and to the public health system. Although on average farmers are better-off than the landless employed in agriculture, the lowest third in both groups are not much different. That is, the poorest third of the farmers are as poor as the poorest third of the landless. It is among the higher income subgroups that farmers are better-off. Table 4 compares the statistical significance of the differences between farmers and landless in agriculture. As can be seen, there are practically no statistically significant differences in any of the characteristics considered, including per capita income, among the poorest third in each group. All the differences occur in the higher income group where the farmers are clearly better-off than the landless. Comparing Poverty Among the Farmers and the Landless Table 5 provides an overview of the rural poverty situation in El Salvador and the comparisons across the three classes of rural inhabitants. The government poverty lines for total poor and extreme poor in rural areas were used. The estimates in Table 5 use per capita income corrected for underreporting using a 1.216 conversion coefficient. The head count method indicates that approximately 46% of the rural population falls below the poverty line (about US $360 per capita per annum) and that almost 20% is below the extreme poverty line (defined at about US $180 per capita per annum). These estimates are not directly comparable with the World Bank (1994) estimates of rural poverty in El Salvador because the adjustment for income underreporting used in that study is 17.6% while our correction coefficient is slightly higher, 21.6%. The figures in brackets in the first two columns of Table 5 are calculated using the World Bank correction. The World Bank estimate of rural poverty for 1992 based on the Head Count method was 55.7% for total poverty and 14.3% for extreme poverty. Thus, it appears that between 1992 and 1995 rural poverty in El Salvador has declined by almost 20%. However, extreme poverty appears to have become worse, increasing from 14.3% to 20.2%.5 The largest incidence of poverty occurs among the landless employed in agriculture where according to the unadjusted head count method is almost 80%, followed by farmers with 51% total poverty and the landless not in agriculture with 38%. The poverty picture changes dramatically when we use a consumption adjustment based on the Rothbarth adult equivalency scale. The idea is that children do not have the same needs in terms of food and clothing as the adults, and so the household poverty line should be adjusted according to the demographic composition of the family.6 With this adjustment total poverty falls to 35% and extreme poverty to less than 15% as shown in the 2nd row of Table 5. Still, in terms of total poverty, the landless in agriculture are worse-off than farmers but this is not the case in terms of extreme poverty. The incidence of extreme poverty is almost 50% higher among farmers than among landless in agriculture! On the other hand, extreme poverty among the landless employed outside agriculture is practically negligible. Table 5 also provides estimates of poverty under the assumption that there are economies of scale in consumption as family size increases. The consumption basket within the household includes public goods (shelter, etc.) and increasing household size (and heterogeneity) implies less waste of food, clothes and other goods. In this case the true per capita income of the household in terms of actual consumption is S It is important, however, to indicate that the World Bank (1994) study calculation of self-employment income, particularly concerning products not marketed, is not as exhaustive as ours and the sample used at the time excluded the regions severely affected by the war, which supposedly are the poorest. Thus, one should not attach too much accuracy to the poverty comparisons between the two studies. 6 The Rothbarth equivalency scale assumes that children aged 0-4 have consumption needs equal to 15% of an adult, for children 5-10 the equivalency is 20% and for children 11-15 is 43%. 6 Annex 3: Rural Poverty y/N9, where y is total household income, N is household size in adult equivalent and 0 is a scale coefficient, with 0 < 0 < 1. If 0 = 1 then there are no scale economies and as 0 declines, economies of scale become more important. The last three rows in Table 5 simulate the poverty measures under 3 values of 0, 0.9, 0.8 and 0.7. The important thing is that the relative ranking of the three groups do not change with 0. In all cases, landless in agriculture exhibit the greatest incidence of total poverty followed by farmers while landless outside agriculture have the lowest incidence of total poverty. In terms of extreme poverty incidence, farmers are worse-off than the landless employed in agriculture regardless of the assumed value of 0. The stability of the rankings of poverty and extreme poverty among the rural three social classes considered gives us confidence in the conclusion that the incidence of extreme poverty is worst among the farmers, followed by the landless employed in agriculture. The incidence of poverty among the landless employed outside agriculture is the lowest of all 3 groups. The Econometric Results: The Household Income Table 6 presents the results of the estimation of per capita household income functions using a GLS estimation (all continuous variable are in log form). In the specification, income corresponds to the total returns to the factors of production owned by the household, namely, land, physical capital, human capital and labor. That is, income is analogous to the household value added concept. Naturally, the household value added is a function of the factors of production owned by the household as well as of factors that may affect the productivity of such factors including geographical location, access to infrastructure and others. Human capital is likely to play a dual role, it is a factor of production and also is a productivity factor that may affect, for example, the management ability of the household and, hence, its productivity. The specification of the income fimction is the following, (1) Yi =AiKxLOTYeSi where Yi is total value-added of household i, Ki is capital owned, Ti is land owned, Li is the total hours worked by household members, Ai is an index of total factor productivity and ci is a random disturbance. Total factor productivity Ai is, in turn, a function of education of the household members, age, location variables, social class, etc. These are variables that may affect technical productivity or the ability of the household to negotiate prices, wages, etc.7 Thus, (2) Ai = F(Ei,Zi)I where Ei is education and Zi is a vector of location, age and other characteristics. To estimate equation (1) we normalize all variables by family size, Ni, thus allowing us to express (1) in per capita terms, (3) yi = AikiPtYNO+P+Yvle;i The household wealth, say Ki, Ti, can also imnprove the ability of the household to obtain credit and to negotiate better terms of trade. Annex 3: Rural Poverty 7 where yi is per capita household income, k, is the per capita stock of capital, £i is hours of work per capita and t, is land owned per capita. Note that the coefficient of Ni will measure economies of scale. If a + f3 + y < 0 then the household value added is subject to decreasing returns to scale. The main difference between the estimates in column (1) and (2) in Table 6 is that the average level of education for all household members is used as an explanatory variable in column (2) while only the education of working members is used in column (1). Also, the estimates in column (2) do not control for gender of the household head. The results are similar with the only exception of the coefficient of education, which is significantly smaller when only average education of working members is used as an explanatory variable. The effect of education of the working members on per capita income is highly significant but quantitatively not very large (column 1). The coefficient of education is 3 times larger if the average level of education of all household members, whether working or not, is used instead of only the education of working members (column 2). This is likely to be due to the demand-for-education effect of income. That is, the coefficient of education is much more contaminated by "reverse causality" when all household members' education (including children) is used. This confirms the findings in other countries (i.e., L6pez and Vald6s, 1996) where the education demand effect appears much larger than the effect of education as a source of income. The income elasticity of education of working household members is quite robust, with a value of approximately 0.04. This means that an increase in one year of schooling by working household members, i.e. an increase of about 30% given that the average schooling of working household members is about 3 years, would cause an increase in per capita income of 0.9%. That is, the average annual per capita household income would increase by about 40 colones or about US $4.50. Even doubling the schooling level would only increase per capita income by less than US $15 per annum. This is indeed a very small return to education. It is important, however, to emphasize that this is the return to education in the rural sector. It is possible that people that migrate to the urban sector may obtain a much greater return to schooling. In fact, it is likely that most of the benefits of education consist in allowing rural people to obtain skills that will allow them to obtain high paying jobs in the urban sector. Thus, an important implication of the above result is that the rural economy has not been able to generate activities of sufficient human-capital intensity that could reward schooling more highly. It appears that only by migrating into urban areas have educated people a chance to exploit their human capital more fully. Another important result is the lack of (joint) statistical significance of the 14 regional dummy variables in explaining per capita income. Although the raw data indicate that poverty tends to be regionally concentrated, with certain departments exhibiting lower incomes than others, once we control for household specific characteristics the regional effects disappear. The only variable related to location that shows a robust effect is the distance of households to a paved road. This variable could reflect quality of infrastructure and proximity to markets. According to the estimates, increasing the proximity to paved roads by one kilometer, ceteris paribus, is likely to increase per capita income by about 1%. The fact that the dummy variables for landless are all positive and significant, implies that farmers are able to extract a lower income out of their resources than non-farmers. The source of this phenomenon is not immediately obvious, but it may be related to the same problem observed for long periods of time in developed countries where farmers tended to obtain below normal returns for their assets. Access to land appears to be one of the most important determinant of household income. Increasing land ownership by 10% for the average rural household can boost its per capita income 8 Annex 3: Rural Poverty according to the estimates by almost 4%. The large effect of land on income is in part associated with the fact that El Salvador is land scarce. The land/labor ratio in particular is very small and consequently land is relatively much more scarce than labor. This is also reflected on the generally low responsiveness of income to labor supply. In fact, the income elasticity with respect to hours worked is only 0.06 for male and 0.02 for female workers. That is, given the limited resources of the average rural household, even a very large effort to reduce leisure and, hence to increase labor supply, would have a very modest effect on per capita income. The income elasticity with respect to male work is about 3 times larger than with respect to female work. The reason for this is not necessarily that male workers are more productive than female workers but rather that women are much more engaged in house work such as child bearing, cooking, and house caring, outputs that are not accounted for in the income variable. As shown below, this is also due to lower wages for salary work that women receive. Rural Wage Employment and Labor Market Participation As shown earlier, wage employment is a vital source of income for a large proportion of the rural population, particularly for the poorest. In this section we provide some econometric evidence on the determinants of participation in salary employment and of the hourly wage. The econometric method used is the Heckman two-stage procedure as presented in Heckman, 1979 and Greene, 1981. This procedure allows one to obtain consistent estimators of the coefficients of the wage equation and, at the same time, to correct for the biases of the standard errors in the wage equation. Table 7 presents the estimates of the labor market participation equation and the wage equation. The employment or participation equation is estimated using a Binomial Probit model. This equation is important in its own terms and also provides the Mills' ratios to correct for sample selection bias of the wage equation. The Probit employment equation (column 1 of Table 7), explains the probability of wage employment as a function of characteristics of the individual, characteristics of the household to which the individual belongs and locational variables. All right-hand-side continuous variables are in log form. Age has, as expected, a non-linear effect on the probability of wage employment. The probability of wage employment increases until age 31 approximately and afterwards decline. The level of schooling has no statistically significant effect on the probability of employment. Having access to land and capital is likely to decrease the probability of participating in the labor market mainly because living in a household with more land and capital increases the opportunity cost of work outside the household self-employment activities, such as farming. Thus, the reservation wage for participation in the labor market is likely to increase with the amount of land and capital owned by the household. Similarly, if a family receives remittances, the reservation wage is likely to increase and the probability of participating in the labor market diminish. As can be seen in Table 7, per capita family land has a negative and statistically significant effect on the probability of employment outside the household activities. The effect of remittances is also negative, significant and quantitatively very strong. The effect of family capital is also negative as expected, but is not statistically significant. The estimates also suggest that females and children are much less likely to participate in salary employment than men even after controlling for individual and household characteristics. The lower rate of female participation is doubtless related to the demands for house chores and children care that women are subject to. Geographical location is also a very important determinant of participation in the labor market. This is likely to be due to the diverse degree of development of the labor market in the various regions. Annex 3: Rural Poverty 9 The 2nd column of Table 7 shows the determinant of the hourly wage using a Mincer-type specification, using age as a proxy for work experience. Education has a highly significant effect on the hourly wage with an elasticity of 0.024. This implies that one additional year of schooling is likely to increase the hourly wage by less than 1%. As in the case of total income, the returns to schooling do not appear to be quantitatively very strong in determining the hourly wage. Thus, this result is highly consistent with the finding that the returns to education in the rural sector is very low compare to those encountered in other sectors. The positive and significant effect of household capital on the hourly wage is consistent with the hypothesis that the reservation wage is likely to increase with the household wealth. The negative and significant effect of land, however, is surprising. The land variable appears to pick-up a "farmer" effect.! It seems that famers tend to obtain lower retums for all their factors of production including labor. The reason for this is not clear. Another important result is the very large and significant effect of the gender dummy variable in explaining labor market participation. Not surprisingly, the rate of participation of women in wage employment is much lower than that of men even after controlling for several individual and household characteristics. The estimates suggest that females are not paid less than males after controlling for schooling, location and other individual and household characteristics. This result is very important, particularly in view of the fact that the average hourly wage of women is about 10% below the corresponding rate for men. Thus, the estimates in Table 7 suggest that women demographic characteristics are able to explain the observed wage gap. Finally, the geographical and location factors do not play any role in determining wages. This is shown by the fact that neither the regional dummies nor the distance to paved roads dummy are statistically significant (they are not jointly significant either). Thus, it appears that there is little regional segmentation in the rural job market in El Salvador. The high significance of the dummy vanable for women suggests that the women's labor market participation may be structurally different. For this reason we estimated a separate participation equation for females (Table Al., Appendix). The most important finding of the female labor participation equation is that for women education does have a positive and significant effect on female participation. The Farmers' Production Function Farmers are an important segment of the rural population in El Salvador and, as we indicated earlier, the rural poor are highly dependent on the agricultural sector for their subsistence. For this reason it is important to consider certain aspects of the determinants of farm output. We do this by estimating a Cobb-Douglas production function. Table 8 provides the instrumental variable estimates of the farm production function. In general the coefficients have plausible signs and are quite robust to changes in the specification. The factors of production are disaggregated into 3 types of family labor (adult men, adult women and children less then 15 years old), hired labor, variable inputs (fertilizers, pesticides, seeds, fuels, etc.), farm capital and land area. Total factors productivity is assumed to vary with the level of education of the household members that work, age of the head of the household, technical assistance, distance to paved road, and the geographic location of the farm. 8In fact, when we use dummy variables to control for economic classes, the land variable ceases to be significant. 10 Annex 3: Rural Poverty The largest contributors to farm production appear to be land and variable inputs. Together their share in total output is more than 70%. The high share of land is consistent with the overall scarcity of land in El Salvador while the high contribution of variable inputs suggests the importance of intensification as a source of greater production. Consistent with the relative labor abundance of the country, the estimates indicate that the contribution of labor to farm production is quite modest. The combined contribution of all labor including family and hired labor is less than 10% of the value of output. A result that is counterintuitive is the low contribution of capital to the agricultural output. In fact, the coefficient of capital though positive is not significant. The farm production technology exhibits decreasing returns to scale in the factors of production. This is shown by the fact that the sum of the coefficients is equal to 0.8 and, in fact, significantly less than 1. This result is consistent with the findings in several other Latin American countries, where recent studies have found retums to scale in the range of 0.7 to 0.9.9 A surprising result is the lack of significance of education as a factor determining total factor productivity. This is, however, consistent with the idea discussed above concerning the low value of education in the rural sector and, in particular, in agricultural activities. Farmers that receive technical assistance, on the other hand seem to be significantly more productive than those that do not. This, however, does not necessarily mean a direct causality. It could be that farmers that seek technical assistance are the most productive in the first place. A deeper understanding of the causality issue may have important policy implications. Coffee producers seem to be dramatically more productive that those that do not produce coffee. The coffee producer dummy could indeed capture some of the value of farm capital given that coffee producers tend to be more capital intensive than those that do not. Also, coffee producers are more likely to be integrated in the input and product markets, what can be positive for productivity. Farm location does not seem to play a very important role in determining farm productivity. Among the department dunmmy variables only a few had some significance but their quantitative values were rather small. Additionally, the variable distance to paved road, which may also capture aspects related to the degree of isolation or regional clustering of farmers is not statistically significant either. Conclusions The major findings that emerge from this study are: * About 45% of the rural population in El Salvador can be considered poor and almost 20% is extremely poor. * Poverty is highly concentrated among the landless agricultural workers and small farmers. Approximately 30% of the landless and 27% of the farmers are extremely poor. Poverty is much less frequent among the landless occupied in non-agncultural activities, of which only 7% is extremely poor. * Although on average the farmers per capita income is much higher than the landless farm workers, the 30% poorest farmers are as poor as the bottom 30% landless. 9For Honduras, for example, Lopez and Romano (1995); for Chile, Lopez 1996, for Paraguay, see L6pez and Thomas, 1995 and for Mexico, see Lopez, Nash and Stanton, 1995. Annex 3: Rural Poverty 11 * The poorest rural households are much more dependent on income originated in agriculture than the non-poor. Almost 75% of the income of the poorest 40% of the rural households is originated in agriculture compared to about 36% for the richest 40%. * Remittances from migrants are important for the middle and higher income groups but are much less important for the poorest 40%. * The average level of schooling of the adult population is very low but there are clear signs of rapid improvement as reflected by the education of children. * Access to land is a very important factor that affects per capita income. * Although education significantly affects per capita income, its quantitative value is quite low. The value of education in the rural sector is in general low. This could be due to the fact that to get high returns to schooling there is a need to migrate to urban areas. It appears that the rural economy has not created activities of sufficient human capital intensity that could reward schooling more highly. * Regional location of the households does not play a very important role in affecting their income once proper control of household characteristics is implemented. Thus, lack of rural infrastructure appears to have played only a modest role in affecting the income potential of rural households. * Remittances appear to play an important role in reducing labor market participation and in increasing the reservation wage of the household members that receive them. * Females are less likely to have wage employment then males even after controlling for individual and household characteristics. * There is no evidence of wages discrimination against women in El Salvador. 12 Annex 3: Rural Poverty REFERENCES Heckman, J. "Sample Selection Bias as a Specification Error," Econometrica 47(1979):153-161. Greene, W. "Sample Selection Bias as a Specification Error: Comment," Econometrica 49(1981): 795- 798. L6pez, R. and A. Valdes, Rural Poverty in Latin America, 1996, forthcoming. L6pez, R. "Determinants of Rural Poverty: A Quantitative Analysis for Chile," Cuademos de Economia, 1996. L6pez, R. and C. Romano, "Rural Poverty in Honduras: Asset Distribution and Liquidity Constraints," Technical Department, Latin America and the Caribbean Region, the World Bank, 1995. L6pez, R. and T. Thomas, "Rural Poverty in Paraguay: The Determinants of Farm-Household Income," unpublished, University of Maryland, College Park, 1995. L6pez, R. "Land Titling and Agriculture Productivity in Honduras," unpublished, Departnent of Agricultural and Resource Economics, University of Maryland, College Park, 1996. Reed, D. "Migration in Brazil: Evidence of Credit Constraints," Departnent of Economics, Yale University, November 1994. The World Bank. El Salvador: The Challenze of Poverty Alleviation, Report No. 12315-ES, Latin America and the Caribbean Region, June 1994. Annex 3: Rural Poverty 13 Table 1. El Salvador (1995): Distribution of Means of Rural Population Characteristics Across Income Groups' I Wnome Levels An I Low Mid-low UMid Mid--h;g High Income Per capita income 4,040 784 1,836 2,846 4,310 10,447 Originated in agriculture 1,881 580 1,156 1,593 1,680 4,411 Originated in non-agriculture 1,774 147 556 1,018 2,186 4,979 Others 385 57 124 235 444 1,057 Sources of Income (as % of total income) Agriculture wages 34.2 47.9 50.1 36.3 24.1 12.4 Farn self-employment 19.4 25.3 14.1 19.6 15.3 22.6 Non-agriculture (wages and self-employment) 37.5 19.3 29.3 35.8 50.1 53.3 Remittances from abroad (including all households in 5.3 3.4 1.5 5.6 8.2 7.6 this group) Remittances from El Salvador (including all 1.2 0.4 3.0 1.3 0.6 0.7 households in this group) Subsidies, pension, rental 2.3 3.4 2.0 1.4 1.5 3.3 Percent of households that receive remittances from El 5.6 4.8 8.7 4.0 5.6 4.8 Salvador Percent of households that receive remittances from 14.6 7.2 6.3 13.5 19.8 26.4 abroad Demographic Characteristics Total number of people in household 5.9 6.7 6.5 5.8 5.5 4.7 Women adults (16 and above) 1.7 1.7 1.6 1.6 1.8 1.6 Men adults (16 and above) 1.8 1.7 1.6 1.7 1.8 1.9 Children (under 16) 2.4 3.3 3.3 2.4 1.8 1.2 Numberofchildrenofheadofhousehold 3.1 3.9 3.6 3.0 2.5 2.2 Average age of head of household 46.1 46.6 42.3 44.1 47.5 49.9 Average age of fanily 26.3 23.7 21.2 26.3 28.6 32.1 Dependency Ratio (No. of non-workers divided by No. 0.73 1.0 0.98 0.68 0.61 0.35 of workers) Education Percent of illiterates for people aged above 12 29 38 32 30 27 20 Average years of education of head of household 2.9 2.2 2.4 2.5 3.3 4.3 Average years of education of household males older 3.4 2.4 2.8 3.1 4.0 5.0 than 18 Average years of education of household females older 2.8 1.8 2.2 2.5 3.2 4.2 than 18 Percentage of household members between 6 and 17 30 40 32 26 27 20 not attending school 14 Annex 3: Rural Poverty Table 1: El Salvador (1995): Distribution of Means of Rural Population Characteristics Across Income Groups' (continued Income Levels All | Low Mid-low Mid Mid-high |High Access to Services Percent of households with access to electricity 54.5 43.2 45.2 52.4 63.5 68.0 Percent of households with access to water 19.9 14.4 14.3 15.9 26.2 28.8 Percent of households with at least one member 15.3 4.8 4.0 14.3 25.4 28.0 covered by public health insurance Percent of households with at least one member 5.1 4.0 4.8 3.2 4.8 8.8 covered by private health insurance Average distance to nearest school (Km) 3.3 3.6 3.6 3.4 2.8 2.9 Distance to the nearest doctor (Km) 5.2 5.5 6.2 5.2 4.1 4.8 Average distance to the nearest paved road 5.8 7.5 6.3 6.2 4.8 4.1 Percent of households that received credit since 1991 26.1 36.8 23.8 19.1 25.4 25.6 Total credit received since 1991 per household (only 5,756 4,048 3,395 4,867 4,052 12,575 for households which received credit) Employment Characteristics Percent of people working in salary activities 33 27 26 34 34 44 Percent of people working in self-employment 2.1 1.5 1.3 0.3 4 7.8 activities (excluding household activities) Average wage rate of salary workers 5.9 4.2 4.9 5.5 6.2 8.8 Gender Aspects Percent of households headed by a woman 8.1 4.8 9.5 7.9 10.3 8.0 Average hours worked per year by women 16 and 2,859 2,860 2,916 2,887 2,773 2,857 older (house, farm and off-farm work) Average hours worked per year by men 16 and older 1,763 1,476 1,767 1,765 1,864 1,946 (house, farm and off-farm work) Average hourly wage for women 16 and older 5.6 3.9 4.7 5.5 4.7 8.3 (colones 1995) Average hourly wage for men 16 and older (colones 5.9 4.3 4.9 5.4 6.6 9.3 1995) -Number of households 628 125 126 126 126 125 1 Sample includes 192 farmers (randomly selected from 302 households surveyed), and 436 landless: 166 with > 66% income from agric., 55 b/w 33% and 66% income from agric., and 215 with <33% income from agric. All money values in colones 1995. Annex 3: Rural Poverty 15 Table 2: El Salvador (1995): Distribution of Means for Farmers' Characteristics Across Income Groups Income Levels All Low Mid High Household income 24,443 6,005 17,958 49,367 Per capita income 4,916 836 2,957 10,954 Percent of income from self-employment Percent of income from off-farm work in agriculture 14.6 20.7 14.1 9.1 Percent of income from off-farm work in non- 20.3 15.5 22.1 23.3 agriculture Percent of income from remittances from El Salvador 2.2 3.7 2.1 0.8 (including all households in this group) Percent of income from renittances from abroad 10.9 6.0 15.3 11.4 (including all households in this group) Percent of income from subsidies and pensions 2.7 5.8 0.8 1.5 Percent of households that receive remittances from El 6.3 7.8 6.3 4.7 Salvador Percent of households that receive remittances from 25.0 7.8 35. 31.3 abroad Percent of households received credit since 1991 30.7 35.9 23.4 32.8 Credit received since 1991 per household 7,970 3,305 4,166 15,827 Capital attached to land per household 4,470 2,140 2,533 8,737 Capital non-attached to land per household 36,307 9,266 16,455 73,917 Number of people in household 6.0 6.7 6.0 5.3 Percent of females head of household 7.3 4.7 9.4 7.8 Number of children of head of household 3.1 3.6 3.2 2.4 Percent of household members self-employed (in house 66.2 60.2 68.9 69.4 or own land) Percent of household members working off-farm 19.6 14.4 19.4 24.9 Average age of family 28.6 26.7 27.5 31.6 Age of head of household 52.9 52.4 51.1 55.3 Years of education of head of household 2.4 1.7 2.0 3.4 Percent of heads of households without formal 44.3 56.3 42.2 34.4 education Average education of household members between 6 2.0 1.5 2.4 2.3 and 12 years old Average education of household members between 13 5.3 4.3 5.6 6.2 and 18 years old Average education gap of children between 6 and 18 2.8 3.2 2.7 2.3 years old (age - 6 - years of school completed) Percentage of children (between 6 and 18) in ideal 17.7 20.6 15.4 16.9 grade or better Average education of household males older than 18 3.5 2.1 3.1 5.2 Average education of household females older than 18 2.8 1.7 2.6 4.3 Land size (in manzanas) 4.4 2.6 3.9 6.6 Percent of households holding title to land 82.8 73.4 85.9 89.1 (continued) 16 Annex 3: Rural Poverty Table 2: El Salvador (1995): Distribution of Means for Farmers' Characteristics Across Income Groups (continued) Income Levels All | Low | Mid | High Percent of total farm value from cash crops3 21 4 16 43 Percent of total farm value from subsist crops4 50 71 53 25 Percent of total farm value from other crops5 1 1 1 1 Percent of total farm value from livestock production 15 12 15 18 Percent of households receiving technical assistance 17.2 10.9 20.3 20.3 Percent of households with access to electricity 60.4 45.3 64.1 71.9 Percent of households with indoor running water 17.7 12.5 14.1 26.6 Percent of households reporting safety problems 29.7 28.1 28.1 32.8 (e.g.,burglary of crop and animals, vandalism, personal threats, payment demands, land invasion) Percent of households with at least one lot of land of 64.1 48.4 64.1 79.7 good quality soil Percent of households with at least one lot of land 4.2 3.1 4.7 4.7 irrigated Percent of households with at least one lot of land of 15.1 14.1 17.2 14.1 steep slope Percent of households with at least one lot of land with 43.8 50.0 45.3 35.9 erosion problems Percent of households which apply soil conservation 48.4 45.3 57.8 42.2 practices Distance to nearest school (Km) 3.5 4.0 3.6 3.0 Distance to nearest doctor (Km) 5.1 6.0 4.2 5.0 Distance to the nearest paved road 6.0 7.3 6.5 4.3 Percent of households with at least one member 10.9 4.7 10.9 17.2 registered in the public health insurance Percent of households with at least one member 3.1 3.1 1.6 4.7 registered in the private health insurance Number of households 192 64 64 64 (1) Sample includes 192 farmers randomly selected from the total number of farmers surveyed (302). (2) Money values in colones 1995. (3) Cash crops include coffee, fruits and vegetable, sugar cane. (4) Subsistence crops include maize, beans, soybean and rice. 17 Annex 3: Rural Poverty Table 3. El Salvador (1995): Comparison of Means of Rural Population Groups' Groups T-Tests: Ho=Means are equal2 Landless Landless in Landless non- 1x2 1x3 1x4 2x3 2x4 3x4 Farmers agriculture mixed agriculture ________________________________________________________ . (1) (2) (3) (4)_ Per capita income 4,916 2,255 2,756 4,963 * *** ** *** *** Number of people in household 6.0 5.7 6.2 5.7 Percent of households headed by a woman 7.3 5.4 5.5 11.6 * Number of children of head of household 3.1 3.1 3.0 3.0 Average age of family 28.6 25.9 26.5 24.6 ** Years of education of head of household 2.4 1.9 2.9 4.2 * *** *** *** *** Percent of head of households without formal education 44.3 51.8 25.5 21.9 *** *** *** *** Average years of education of household members 5.3 4.2 5.3 5.5 *** * between 13 and 18 years old Average years of education of household males older 3.5 2.2 3.3 4.5 *** *** *** *** *** than 18 Average years of education of household females older 2.8 1.7 2.5 3.6 *** ** ** *** ** than 18 Average education gap of children between 6 and 18 2.8 3.2 3.1 2.7 * years old Percentage of children (between 6 and 18) in ideal 17.7 15.1 12.1 21.3 ** ** grade Percent of households with access to electricity 60.4 39.2 60.0 59.5 *** *** *** Percent of households with access to water 17.7 16.3 27.3 22.8 * * Average distance to nearest school (Km) 3.5 3.4 2.6 3.0 Average distance to the nearest doctor (Km) 5.1 6.1 4.2 4.8 * *** ** Average distance to the nearest paved road 6.0 6.1 6.2 5.2 Percent of households with at least one member covered 10.9 1.2 12.7 30.7 *** *** *** *** *** by public health insurance Percent of households with at least one member covered 3.1 5.4 1.8 7.4 * ** by private health insurance Number of households in sample 192 166 55 215 1 Sample includes 192 farmers (random selected from the total of 302 farmers surveyed), and 436 landless: 166 with > 66% income from agric., 55 between 33% and 66% income from agriculture, and 215 with less than 33% income from agricultural activities. Money values in colones 1995. 2 Means are different at *p < 10%, ** p < 5%, *** p < 1%. 18 Annex 3: Rural Poverty Table 4. El Salvador (1995): Comparison of Means Between Farmers and Landless in Agricultural Sector, According to Income Level Groups T-Tests: Ho = Means are equal' Between Low Between Mid Between High Income Groups Income Income Groups Per capita income Number of people in household Percent of households headed by a woman Number of children of head of household Average age of family Years of education of head of household Percent of head of households without formal education Average years of education of household ** members between 13 and 18 years old Average years of education of household males older than 18 Average years of education of household females ** older than 18 Average education gap of children between 6 and 18 years old Percentage of children (between 6 and 18) in ** ideal grade Percent of households with access to electricity Percent of households with access to water Average distance to nearest school (Km) Average distance to the nearest doctor (Km) ** Average distance to the nearest paved road Percent of households with at least one member * ** covered by public health insurance Percent of households with at least one member covered by private health insurance 1/ Tests are done between the two rural groups for low, mid and high income levels. * p<10%, **p<5%,***p<1% 19 Annex 3: Rural Poverty Table 5. El Salvador (1995): Distribution of Poverty and Extreme Poverty'" 2 All Rural Farmers Landless in Landless NOT Agriculture in agriculture Total Extreme Total Extreme Total Extrem Total Extreme Poverty Poverty Poverty Poverty Poverty e Poverty Poverty ________________________________________ ~~~~ ~~~~~Poverty _ _ _ _ _ _ _ _ _ Head count 45.4 20.2 46.7 27.2 71.1 29.5 27.9 7T0 (46.5) (14.3) Consumption-adjusted (theta= 1)3 25.3 10.7 37.4 18.9 38.0 12.7 10.2 4.7 (26.9) (10.8) Consumption-adjusted (theta = 0.9) 21.5 8.0 32.5 15.6 31.9 9.0 8.4 3.7 (22.1) (9.1) Consumption-adjusted (theta = 0.8) 18.5 5.3 28.5 13.9 25.9 3.6 7.0 1.9 (19.3) (5.4) Consumption-adjusted (theta = 0.7) 15.0 4.5 24.5 12.3 19.3 3.0 5.6 0.9 ___________ __________________ .______________ (16.1) (4.5) . I Percentage of population in poverty; calculated using the value of a rural Basic Food Basket (BFB) for extreme poverty (annual income per capita < US $180) and 2 BFBs for total poverty (< US $360). 2 Estimates using per capita income corrected for underreporting using a correction factor of 21.6%. Figures in brackets correspond to the underreporting correction factor of 17.6% as used by the World Bank for the poverty estimates for 1992. 3 The consumption adjusted estimates use the Rothbarth equivalence scale 20 Annex 3: Rural Poverty Table 6. El Salvador (1995): GLS Estimates of Total Income per Capital (Dependent variable: log of total income per capita) Parameter estimates Independent Variables' (standard errors in parentheses) (1) (2) Intercept 12.510*** (2.400) 13.049*** (2.412) Total annual hours of work by adult men 0.065*** (0.012) 0.054*** (0.010) Total annual hours of work by adult 0.016*** (0.004) 0.016*** (0.004) women Total annual hours of work by children 0.004 (0.005) 0.004 (0.005) Capital attached to land 0.017* (0.010) 0.018* (0.010) Other capital 0.016*** (0.006) 0.015*** (0.006) Size of land owned and operated 0.390*** (0.076) 0.369*** (0.075) Dummy for household headed by a 0.142 (0.116) NA Average education of working household 0.037*** (0.011) NA Average education of household members NA 0. 136*** (0.031) Average education (above 12) squared NA 0.017*** (0.006) Age of head of household -2.604** (1.323) -2.944** (1.330) Age square 0.396** (0.180) 0.442** (0.181) Distance to nearest paved road -0.008** (0.004) -0.009** (0.004) Dummy for landless in agriculture 3.071*** (0.576) 2.964*** (0.571) Dummy for landless not in agriculture 3.777*** (0.577) 3.590*** (0.573) Dummy for landless of mixed income 3.327*** (0.095) -0.287*** (0.094) Sarnple size.. 623 625 1. The Breusch-Pagan test rejects homoskedasticity at the 1% level of significance. Correction for heteroskedasticity was performed using family size, land size, dummies for belonging to differenct rural groups, and regional dummies. 2. All continuous variables are in per capita and log format. 3. Twelve regional dummies were included in regressions but are not reported. Of those only one is significant at the 5% level; a joint test of significance for all regional dummies does not reject Ho = all regional dummies = 0. *p< 10%, **p<5%, ***p< 1%. Annex 3: Rural Poverty 21 Table 7. El Salvador (1995): Two-stage Estimates of the Labor Market Participation and Wage Functions Independent Variables' Parameter estimates (standard errors in parentheses) ........................................................................................................................................................................................................................................ Participation2 Wage equation' Intercept -22.06*** (2.05) -5.70** (2.91) Years of education -0.004 (0.009) 0.02*** (0.005) Age 13.18*** (1.21) 4.05** (1.64) Age squared -1.92*** (0.18) -0.57** (0.24) Capital (total for household) -0.007 (0.007) 0.01*** (0.004) Size of land owned and operated (total -0.13*** (0.05) 0.03*** (0.03) household) Remmittances (total in household) -0.02*** (0.005) -0.003 (0.004) Distance to nearest paved road -0.0003 (0.0006) -0.0004 (0.0004) Dummy for women adults (16 and above) -1.63*** (0.07) -0.29 (0.19) Dummy for children (15 and under) -0.63*** (0.14) -0.02 (0.12) Dummy for landless in agriculture -0.37 (0.39) 0.24 (0.23) Dummy for landless not in agriculture -0.57 (0.39) 0.58** (0.24) Dummy for landless of mixed income -0.44 (0.40) 0.33 (0.24) Log-likelihood -1190.4 -730.1 Sample size 2634 1038 Two-stage Heckman sample selection model (Heckman, 1979 and Greene, 1981). 1 Explanatory continuous variables are in log format, and are per capita (unless related to individual education and age). 2 Dependent variable is dummy = 1 if participates in off-farm job market, 0 otherwise. Twelve regional dummies were included but are not reported. Of those only one is significant at the 10% level. 3 Dependent variable is log of hourly wage from all off-farm work, in both agriculture and non-agriculture. Twelve regional dummies were included but are not reported. Of those 7 are significant at the 5% level. *p< 10%, **p<5%, ***p< 1%. 22 Annex 3: Rural Poverty Table 8. El Salvador (1995): Instrumental Variable Estimates of the Production Function' (Dependent variable: log of total value of farm output per capita) Parameter Estimates Independent Variables2 (Standard errors in parentheses) (1) (2) (3) Intercept 4.29 (6.06) 4.24*** (1.53) 5.88** (1.38) Dummy for poor (adjusted)3 NA NA - (0.10) Annual hours of work on farm by 0.03 (0.02) 0.03 (0.02) 0.03* (0.016) adult * Annual hours of work on farm by 0.02* (0.01) 0.02** (0.01) 0.02* (0.008) adult * Annual hours of work on farm by 0.02** (0.01) 0.02** (0.01) 0.01* (0.007) Hired labor 0.04*** (0.01) 0.03*** (0.01) 0.03* (0.007) Variable inputs4 0.33*** (0.11) 0.35*** (0.12) 0.21* (0.10) Farm capital 0.01 (0.02) 0.01 (0.02) 0.02 (0.02) Land size per capita 0.36*** (0.10) 0.36*** (0.11) 0.41* (0.09) Average education of members 0.03 (0.02) 0.03 (0.02) 0.01 (0.02) Age of head of household 0.39 (3.08) 0.34 (0.25) 0.13 (0.23) Age of head of household squared -0.0002 (0.40) NA NA Dumnmyfortechnicalassistance 0.32** (0.14) 0.30** (0.15) 0.41* (0.14) Dummy for coffee producer 0.83*** (0.16) 0.85*** (0.15) 0.71* (0.15) Distance to nearest paved road -0.003 (0.006) NA -0.002 (0.006) Title to land 0.07 (0.11) NA NA Family size -0.25*** (0.06) -0.22*** (0.06) - (0.05) Sample size 300 300 299 1. Restriction applied: coefficient to log family size = sum of coefficients of log of family labor, hired labor, log land size, log of attached and non-attached capital and log of variable inputs, minus one. The Breusch-Pagan test rejects homoskedasticity at the 5% level of significance. Heteroskedasticity is corrected using land size, dummy for coffee producers, and all regional dummies. 2. All continuous variables are per capita and in log format. Twelve regional dummies were included in regression but are not reported here; of those 5 were significant at least at the 10% level in regression (1) and (2) and 11 were significant in regression (3). 3. Dummy = 1 if per capita income of household (consumption adjusted according to Rothbarth equivalence scale) is below the total poverty line (= $359). 4. Variable inputs are instrumentalized using family size, age of head, average education of members above 12, capital, technical assistance, electricity and regional dummies. THE RURAL NON-AGRICULTURAL SECTOR AND POVERTY1 I. Introduction Poverty is the subject of much discussion in El Salvador. There is a sense that without concerted attention to poverty issues, a full and sustained transition from decades of violence will remain elusive. It is also feared that equity-motivated programs, such as the agrarian reforms, have not succeeded in eliminating poverty altogether; either because implementation has not been as effective as hoped, or because at least part of the poverty problem is linked to issues beyond access to land and tenure security alone. There is a feeling that more must be done. While few dispute the importance of addressing poverty in El Salvador, there is considerable debate surrounding certain key aspects of the poverty problem. There is, for example, no unfiversal consensus on how widespread poverty is in El Salvador, where poverty is concentrated, and which household characteristics are most closely linked to poverty. Much of this debate is prompted by differences in methodological approaches, and by shortcomings in available data sources. Nonetheless, there does appear to be broad agreement that poverty in rural areas deserves particularly close attention. Most approaches to the measurement of poverty tend to indicate that the rural poverty problem is particularly pressing. In order to understand the causes of rural poverty and to design policies which address these, one must ezxamine in some detail the operation of the rural economy. It becomes quickly apparent that the rural economy e:xtends well beyond agriculture. The non-agricultural sector in rural areas is highly heterogeneous (to such an *extent that it is basically defined in terms of what it isn 't: agriculture), but can represent a very important part of thie rural economy in terms of incomes and employment generated. While, in the past, this sector has not received the same level of attention as the agriculture sector, there is a growing appreciation of its potential in terms of both poverty alleviation and growth more generally. Recent analyses of the sources of growth in East Asia have stressed the cenreal role played by the non-agricultural sector in rural areas2. A question of considerable interest is whether the non-agricultural sector can play a similar role in stimulating rural growth in El Salvador, and Latin America more broadly. -This paper examines data from two recent household surveys in El Salvador to assess to what extent the non-agricultural sector might be able to contribute to rural poverty alleviation. The paper indicates that in El Salvador, as in other Latin American countries, this sector provides a significant share of total employment and income. Employment in this sector is strongly associated with lower poverty, suggesting that it might offer a route out of poverty for those who gain access to non-agricultural employment. Of particular interest are the possible policy interventions which might be able to stimulate such beneficial outcomes. The paper suggests that for policy purposes, attention should be paid to geography, infrastructure and education. The plan of the paper is as follows. The next se,ction discusses poverty in El Salvador, and presents some tentative estimates of poverty based on consumption expendlitures. Section m introduces some quantitative evidence on the size of the non-agricultural sector in rural El Salvador, and the range of activities which comprise this sector. This section also considers what relationship exists between poverty and non-agricultural This paper was written by Peter Lanjouw, World Bank 2 Aoki, Murdock and Okuno-Fujiwara (1995)jargue that the East Asian success in utilizng cheap labor in rural areas, in sectors outside oftraditional fianing, was one of the most important elements of East Asian development" (page 40; see also Hayamii, 1995). 2 Annex 4: Rural Non-Agricultural Sector employment. Section IV turns to an examination of the factors which appear to influence the involvement of rural households in the non-agricultural sector and also the earnings associated with those activities. In Section V we draw on the preceding sections to discuss possible policy implications of the findings. Before proceeding, we make a brief note about the data sources underlying the analysis. Two sources of data are being used in this paper. The Encuesta de Hogares de Propositos Multiples, 1994-IE, (EHPM) is a nationally representative household survey fielded by the Ministerio de Planificaci6n y Coordinaci6n del Desarrollo Econ6mico y Social (MIPLAN) in El Salvador. The EHPM is an annual survey, fielded throughout the year in four "waves". In total, roughly 20,000 households are covered. The analysis in this paper is based only on the third wave of the 1994 survey (the waves are designed to be amenable to self-standing analysis) covering 4229 households in total (1743 in rural areas). This wave is somewhat special in that it contained a detailed consumption module (for a sub-sample of households) which permits an analysis of poverty based on consumption expenditure rather than income (see below)3. The second source of data comes from a rural household survey fielded by Fundacion Salvadoreiia para el Desarrollo Econ6mico y Social (FUSADES). The survey covered a sample of about 630 rural households from all regions of El Salvador, stratified on households' characteristics according to their main economic activities, i.e. the self-employed, agricultural workers, and non-agricultural workers. The survey was designed to be representative of the rural population at the 10% level of significance (Lopez, 1996). The FUSADES survey obtained information on a wide range of demographic characteristics, location and income vaniables. The level of detail in the infornation collected has permitted the calculation of a comprehensive measure of income which is less likely to suffer from important omissions than is conventionally the case with income surveys. II. Rural Poverty in El Salvador Poverty has been the focus of attention in many studies in El Salvador (recent examples include FUSADES, 1993, World Bank 1994 and MIPLAN, 1995a). However, to date, there is no clear consensus as to the magnitude and dimensions of the poverty problem in El Salvador. There are numerous methodological and data-related issues which stand in the way of precise, quantified poverty rate calculations in El Salvador; for the country as a whole, and amongst various population sub-groups. These issues are briefly noted below, but space prevents discussing them in detail4. In El Salvador, as in many other Latin American countries, poverty analysis has generally been carried out on the basis of income as the household-level indicator of well-being. This is in contrast with conventional practice in other parts of the world, where consumption expenditures are commonly taken as the welfare indicator. The principal reason for choosing consumption is that experience has shown that these are measured with greater accuracy than incomes - particularly for the poor, who are most likely to consume a relatively narrow range of goods and services.5 Many of the household surveys fielded in Latin American countries are modelled on labor force surveys designed to measure household eamings. Such surveys are generally weak in capturing incomes from non-employment sources (such as self-employment) and from agricultural activities. These omissions may be particularly pertinent to the measurement of poverty. 3. The EHPM contains household weights wiich allow one to aggregate up to population totals. Although the sample fraie was based on the Census from the early 1970s, the household weights have reportedly been adjusted with reference to the 1992 Labor Force Census, so adding up to population should be valid (personal communrication from the Director of the Departamento de Investigaciones Muestrales). 4. See Ravallion, 1994, for a good survey of these issues. 5. In contast, the poor often have myriad income sources, as they try to bnng together revenues from all kinds of activities to be able to meet their subsistence needs. Annex 4: Rural Non-Agricultural Sector 3 The next step in measuring poverty is to relate household-level welfare indicators to some poverty threshold - the poverty line. In constructing a poverty line, many assumptions are generally required, and these can at times become quite contentious. Issues which must be addressed, for example, relate to the nutritional cut- off point to be applied, whether account is to be taken of different requirements between adults and children or between males and females, and what kind of adjustment must be made to allow for non-food items in the basic consumption basket. While various poverty lines have been formulated for El Salvador, their treatment of these issues has not always been very clearly documented (see, IIES-UCA, 1993, and also World Bank, 1994). Nor has the treatment been the same across the different calculations. Yet further issues relate to the partial geographic coverage of many household surveys in El Salvador (missing certain regions, or focussing only on urban areas) and the non-availability of up-to-date popuLation expansion factors which may lead to biased assessments of the distribution of poverty across population sub- groups. Finally, there is an important issue associated with the non-availability of a spatial cost-of-living index. Such an index adjusts for the fact that to reach a given standard of living in Metropolitan San Salvador might cost quite a different amount from other urban areas or rural areas. Combined, these factors issue a strong warning against efforts to provide detailed calculations of poverty rates in El Salvador. Whilst such calculations would undoubtedly be of great value, one might wish to first concentrate on reaching agreement on the appropriate methodology to apply and in securing the requisite data. At the same time, the lack of quantitative poverty measures need not impede unduly one's ability to focus on poverty related questions. If one is prepared to confine one's remarks to broad comparisons of poverty across population sub-groups (say, between rural and urban areas) and to be satisfied in stating that poverty amongst one group is higher or lower than amongst the other (without attempting to state by how much), then methodological differences and uncertainties may be less of a constraint. It is in this spirit that we present in Table 1 some tentative estimates of the incidence of poverty in El Salvador in 1994, based on the consumption expenditure collected by the EHPM. These figures should not be taken literally (ust as all other attempts to measure poverty in El Salvador are likely to be contentious), because they embody a range of strong (and controversial) assumptions. First, a specific methodology was applied to estimate a robust incidence of poverty for the two sub-samples of the EHPM for which highly difference consumption modules were fielded (for a detailed exposition, see Lanjouw and Lanjouw, 1996). Second, it was assumed that the rural cost of living was lower than in urban areas, in proportion to the ratio of the MIPLAN (1995b) urban poverty line to rural poverty line. Third, the poverty line which was taken was simply the one published in MIPLAN, 1995b, for urban areas without any attempt to establish its validity. The purpose of the poverty estimates in Table 1 is to shed light on some of the broad geographic patterns of poverty in El Salvador. At the level of the country as a whole, poverty is much higher in rural areas than in urban areas. This seems to be driven in particular by the relatively low level of poverty in Metropolitan San Salvador. Across the other four broad geographic regions of the country, the evidence is less strong that there exists a clear distinction between rural and urban areas. El Salvador is rather unique in its definition of urban areas in that it counts as urban all municipal centers (cabeceras municipales) without taking into account the actual population of those centers. This is in contrast with many other countries (although it is not unique in Latin America, see Klein, 1993). Poverty in the East; particularly in rural areas, appears to be higher than in the other geographic regions, although the extent to which this is true across different pairwise comparisons varies with the poverty line which is being applied. 6. Note that the cost-of-hving adjustment mentioned above, reduces the gap between Metropolitan San Salvador and the rest of the country. Failure to make such an adjustment would have resulted in an even wider gap. 4 Annex 4: Rural Non-Agricultural Sector m. The Non-Agricultural Sector in Rural El Salvador A study of rural non-agncultural employment in Labn America based on census data suggests that m 1975 roughly 20% of the econonically active population in rural El Salvador was employed in the non- agncultural sector (Klein, 1993). This can be compared with figures of 160, 18% and 40% for Honduras, Guatemala and Costa Rica, respectively, in the early to mid-1970s. More recent census figures for El Salvador have not been calculated, no doubt for data-related reasons associated with the political and military instability of the intervening period. In Table 2 we present estimates based on the EHPM household survey. In 1994, 36.4 percent of the economically active rural population was employed in the non-agricultural sector, nearly twice as many as in the mid-1970s. The range of activities in which the rural population is engaged includes both manufacturing and services. Nearly 30% of all rurl non-agricultural employment is engaged in some form of manufauring activity (combining textiles and carpentry with the generic manufacturing entry). Commerce represents nearly an additional 25%, construction about 13%, domestic service just over 10%, and tansport nearly 6%. The remaining activities are largely various service sector activities. As a proportion of the economically active population, women are far more likely to be active in the non- agricultural market than men. 72% of economically active women are employed in the non-agricultural sector, compared to just about 25% of men. This is not to deny, however, that women are far less well-represented within the economically active population. In the EHPM survey, only about 22.5% of all women of working age are counted among the economically active population (compared to 73.6% of men). The activities in which women are heavily engaged include first of all commerce, and are followed by manufacturing and domestic service. For men, commerce and domestic service are of far less significance, but construction, manufacting and also transport are important sectors. There is geographic variation in El Salvador in the significance of the rural non-agricultural sector. In the Cental 1 region (which includes the departments of Chalatenango, La Libertad, San Salvador and Cuscatlan) nearly 50% of the economically active population is employed in the non-agricultural sector. This contrsts with only 23.2% in the East (Usulutan, San Miguel, Morazan and La Union). The spectrum of activities across regions is fairly uniform in terms of shares of employment. For example, while manufacturing appears to be relatively more irnportant in the West than in the other regions: about 30% of all non-agricultural employment occurs in the textile, carpentry and manufacturing sectors in the West (Ahuachapan, Santa Ana and Sonsonate) it is as high as 24-26% in the other regions. In Table 4 we observe that involvement by households in the non-agricultural sector is broadly associated with lower rates of poverty. The highest incidence of rural poverty in the EHPM survey is observed amongst households which engage in both agricultural labor and farming. In fact, agricultural labor appears to be particularly closely linked to rural poverty in that of eight possible household economic status categories, the three associated with highest incidence of poverty include agricultural labor amongst the household economic activities. It has been remarked in the context of rural India that agricultural labor is a "last resort" activity which households participate in only when faced with acute hardship and no alternative sources of income (Dreze, Lanjouw and Stern, 1992). For this reason the likelihood that agricultural labor households are poor is significantly higher than for many other household types. Such a perspective might also apply in rural El Salvador. Of course, there are different types of agricultural labor. The category applied in Table 4 encompasses both casual daily wage labor and long-term, permanent employment on a farm, plantation or ranch. As a result, not al agricultural labor is likely to be unattractive as a source of income. This is possibly reflected in Table 4 in that households which engage in both agricultural labor as wel as non-agricultural labor run only about an average "risk" of poverty (35% - see Table 1). Annex 4: Rural Non-Agricultural Sector 5 Non-agricultural labor, we have already seen, is also not homogeneous. In Table 4 one household category engaging m non-agricultural employment is relatively highly exposed to poverty (households simultaneously engaging in farming, agricultural labor and non-agricultural labor). However, households reliant only on non-agricultural labor are significantly less itkely to be poor than all other rural households. This observation illustrates an important point. The non-agricultural sector typicatly comprises two distinct sets of activities. On the one hand there is a set of activities which are reasonably productive, relatively well-paid, and which have the appealing feature of being comparatively less exposed than agriculture to climatic variations and uncertainties. On the other hand there are a group of activities undertaken by persons who are unable even to secure an agricultural laboring position; persons who are perhaps old or disabled, or who may be prohibited by custom from participating in the agricultural labor market (for example, women and children). This second set of non-agricultural jobs plays a very different role to the first set. One way to perceive these two is to regard the first set as a source of upward mobility - a route out of poverty, and the second as a type of "safety net" which helps to prevent poor persons from falling into even greater destitution. Both sets of non-agricultural jobs have a very important role to play in reducing, or relieving, poverty. But the types of policies which can be pursued to help realize their potential are quite different. In the Indian state of Maharashtra, the state government supports a large public-works program offering employment at a wage below the prevailing agncultural wage rate to anyone who presents himself at the worksite (see Dreze, 1990 and Datt and Ravaltion, 1994). This program provides non-agricultural employment to poor persons for whom agricultural wage employment is not an option (perhaps because of cultural restrictions, or because climatic conditions have sharply reduced demand for labor). It therefore exploits the "safety-net" function of the non- agricultural sector. The attraction of this approach is that the poor are "self-targeted", i.e. only those who have no other viable employment alternative present themselves at the worksite. Thereby the costly administration of a government targeting scheme is avoided. Policies aimed to expand access of the poor to the high-income non-agricultural jobs, are likely to look very different. In these cases, the emphasis is typicatly on removing constraints and botflenecks to such employment by providing training, supporting infrastructure, etc. In Table 4, while we see evidence of both types of non-agricultural employment (i.e. associated both with a higher risk of poverty and also with a lower risk), the numerical importance of the latter is far greater. Only 2.7% of the rural population falls in the category of engaging simultaneously in fanning, agricultural labor and non-agricultural labor. In contrast, nearly 30% of the raral population is engaged only in non-agricultural activities. And this segment is the least poor of the entire rural population. It seems therefore, that the rural non- agricultural sector in El Salvador functions at present more as a route out of poverty than as a safety net to those who experience acute distress. We turn next to an oft-overlooked segment of the non-agricultural sector. Table 5 examines the FUSADES data to consider the role of small enterprises in rural areas. While the EHPM survey did not enquire specifically into household enterprises the FUSADES rural survey included a separate questionmaire on such activities. In Table 5 we see that of the 300 workers active in rural enterprises, about 40% were family members. Over 50% of all rural enterprises covered in the FUSADES survey were home-based. Commerce was by far the most common form of rural enterprise, although on average such enterprises were smaller in terns of employment per firm than pottery and brick-making enterprises. It is interesting to note that only about 5% of all rural enterprises covered in the FUSADES survey reported having received training. Textile enterprises stand out among the more common enterprises in that they draw particularly heavily on family labor and are most frequently home-based. Nearly a quarter of these enterprises are engaged in a relationship with some larger firm, in which they receive inputs fronm the larger firm, assemble them, and then re- 6 Annex 4: Rural Non-Agricultural Sector sell their output to the same contractor. Such sub-contracting arrangements has been observed elsewhere in Latin America (see for example Lanjouw, 1996), and are argued by Hayamni (1995) to have been common in nual areas of East Asia during the earlier stages of economic development. Hayami (1995) argues that these arrangements are useful to both parties in that they provide to the contractor access to cheap labor, while the home-based firms are able to choose how and when to allocate their family labor, and do not have to concern themselves with bringing the final goods to the market (which, in the case of clothing or shoes for example, might be very far away). The range of both non-agricultural employment as well as rural enterprise activities which are engaged in provides some clues as to the relationship between the non-agriculture and the agriculture sector in rural areas. This broader relationship has received considerable attention in the literature. Mellor and Lele (1972), Mellor (1976) and Johnston and Kilby (1975) have argued that a virtuous cycle between agricultural intensification and non-agricultural activity can emerge on the basis of production and consumption linkages. Production linkages emerge for example when demand of agriculturalists for inputs such as plows and machinery repair stimulate non-agricultural activity via "backward" linkages or where agricultural goods require processing in spinning, milling or canning factories ("forward" linkages). Consumption linkages emerge as rising agricultural incomes feed primarily into increased demands for goods and services produced in nearby towns and villages. While it is difficult to test the strength of such linkages with the available data sources, the fact that a large fraction of non- agricultural activities center around commerce, food processing, transport, and repair activities, suggests that these linkages are certainly present in the El Salvador case.7 However, it is also interesting to note the importance of rural manufacturing and the existence of sub- contracting arrangements between rural home-based enterprises and large, perhaps urban-based, supplier companies. The existence of such rural non-agricultural activities can be very important to the rural economy because they introduce a source of rural income which is less closely linked to agricultural fluctuations. This is in contrast to the activities which are directly linked to agricultural production and incomes. In rural areas, insurance and credit markets often do not operate well, or are missing altogether. This means that, in order to avoid finding themselves in a position where they might need to take consumption loans, farmers' production decisions are often aimed at cropping patterns which minimize the risk of harvest failure, but which have lower- value expected yields (Murdoch, 1995). Access of certain family members to non-cyclical sources of income from manufacturing activities might help to encourage higher-value agricultural production decisions. IV. Correlates of Non-Agricultural Employment and Earnings. We turn now to a closer examination of the correlates of non-agricultural employment in El Salvador by presenting, in Table 6, results from a probit model considering the likelihood of non-agricultural employment for the working age rural population. In light of the discussion in the previous section we look not only at all non- agricultural jobs together, but also distinguish between non-agricultural jobs which can be considered as "low- productivity" jobs and those which are "high-productivity". The distinction is based on whether hourly earnings from these jobs are lower, or higher, than average hourly earnings from agricultuml labor, respectively. In the first column of regression results in Table 6, we find that women are significantly less likely than men to find employment in the non-agricultural sector. We shall return to this finding below. As a person get older, he or she is significantly less likely to be employed in the non-agricultural sector. Compared to the non- 7 De Janvry and Sadoulet (1993) suggest that such linkages might not be so important in Latin America as a whole. With a highly skewed distribution of land and income, a few landowners benefit from the bulk of the income effects of agricultural growth, and these landowners are often absentee and therefore do not demand locally produced goods. This pont indicates that one of the additional benefits of land reform may be the stimulus that is given to the local non-agricultura economy. Annex 4: Rural Non-Agricultural Sector 7 educated, all persons who have been educated are significantly more likely to find employment in the non- agncultural sector. There appears to be a strengthening of the effect of education on the probability of employment as education levels improve. Households with larger per-capita landholdings are less likely to be employed in the non-agricultural sector. This is not the case in all contexts, because where highly desirable non-agricultural jobs are rationed, it is the wealthier households (those with more land) who might be better placed to secure such jobs. In El Salvador, while some family members of the larger landowning households are indeed likely to be working in the non- agricultural sector, they probably reside in San Salvador and therefore do not feature in the FUSADES sample. Cultivating households are also likely to have family members employed in the non-agricultural sector. For these households, the first claim on family members' labor is apparently for assistance in the fields rather than non- agricultural sources of income. The proximity of a household to a paved road significantly improves the likelihood that a family member will be engaged in non-agricultural employment. A similar, but insignificant, influence is observed for distance to nearest secondary school - intended to proxy distance to the nearest town or settlement. Whether the household has a power comnection is strongly significant in increasing the likelihood that a family member will be engaged in some form of non-agncultural employment. Certainly, home-based activities such as tailoring, food preparation or carpentry are much more attractive if the household is connected to the electricity grid. In the FUSADES sample, dummy variables for different departments in El Salvador are significant only in the case of the department of Sonsonate, La Libertad, San Salvador, La Paz and San Miguel. In these departments the probability of non-agricultural employment is significantly higher than in the departnent of Morazan (in the east of the country). For all other departments (with the exception of Ahuachapan) the point estimate is positive, indicating a particularly low probability of non-agncultural employment in Morazan. However, these point estimates are not significantly different from zero. In the second and third columns of Table 6 we consider the same specification against a binary dependent variable indicating, in tum, whether the non-agncultural job is a high productivity one or one which yields a return below the average agricultural wage. The negative impact of age is not significant for high productivity jobs, while for low productivity jobs, it seems that household size is one factor increasing the likelihood that a family member will seek an outside job. For high productivity jobs, the effect of higher levels of education is particularly strong and positive, while for low-productivity jobs, the education variables are all insignificant. The per capita land variable and cultivation dummy both remain negative and significant in the two respective cases. The access to infiastucture variables are broadly similar, although in the high productivity case, the distance variables are not significant, while in low productivity case, it is distance to the nearest secondary school which is significant. Connection to the electricity grid is strongly significant for both high and low-productivity employment. Of the regional dummies, Chalatenango and San Salvador are the two departments in which high- productivity jobs are concentrated. Relative to Morazan, low productivity jobs are more common in Sonsonate, Chalatenango, La Libertad, San Salvador, Cabaias, Usulutan, San Miguel and La Union. A somewhat puzzling finding from Table 6 was the observation that females were significantly less likely to be employed in the non-agricultural sector than men. This finding is due partly to the fact that in Table 6 the relevant domain was taken to include all persons of working age in rural areas. We have already noted that women are far less likely to be "economically-active" than men, as it is common practice to not include non- remunerated domestic activities amongst "economic" activities. In Table 7 we confine our attention to the "economically-active" population, sticking with the FUSADES data for the time being. Now, women are no 8 Annex 4: Rural Non-Agricultural Sector longer significantly less likely to be employed in non-agricultural jobs. In fact, for low-productivity jobs, women are significantly more likely to be employed in such occupations. In Table 8, we retain our focus on the economically active population but apply the significantly larger EHPM dataset to roughly the same specification. The only difference is that we are unable to include the same infrastructure access variables. In this data set, women are significantly more likely to be employed in the non- agricultural sector irrespective of whether the jobs are high or low-productivity ones. Age is positively associated with non-agricultural employment, particularly for the high-productivity jobs. As in the previous tables, education is positively associated with high-productivity non-agricultural jobs, but not with low-productivity jobs. The effect of the land-ownership variable and cultivating dummy is unchanged. Among the regional dummies, Sonsonate, Chalatenango, La Libertad, San Salvador and La Paz are strongly associated with high-productivity jobs (relative to Morazan) while low productivity jobs appear particularly numerous in Santa Ana, Sonsonate, Cuscatlan, La Paz, San Miguel and La Union. The combination of findings from these regressions suggest that non-agricultural employment opportunities are clustered principally in the departments of Sonsonate, Chalatenango, La Libertad, San Salvador and La Paz. The more remote departments, particularly those in the northeast and the far west appear to be less well-served. In Table 9 we turn to the correlates of eamings from non-agricultural employment, basing our analysis once again on the FUSADES data. We present results for two OLS regressions; the first which includes an adjustment for sample selection (Heckman, 1979), and the second which makes no adjustment. The results across the two models are virtually unchanged. Females earn significantly less than men. A women can expect to earn 28% less than a man holding all other characteristics constant.' Earmings are sharply higher with higher education levels. While a person with primary schooling would expect to earn about 17% more than a non-educated person from non-agricultural work, a person with middle schooling would receive about 45% more, a person with high-school level education would receive 79% more, and a person with education at the tertiary level would receive about 223% more. Although the coefficient is not significant, in this regression, a person from a wealthier household would expect to earn more from non-agricultural employment than a person from a household with less per-capita land. If the household is a cultivating one, then a family member employed in some non-agricultural job would be eaning about 47% less than one from non-cultivating household. While the infrastructure variables influenced significantly whether a person was likely to be employed in the non-farm sector, these variables do not seem to influence earnings significantly. Earnings, given employment in a non-agricultural occupation, are also not significantly influenced by geographic location. Before ending this section, we consider briefly some of the factors which may influence the establishment of rural non-agricultural enterprises. In Table 10 we focus on the 101 rural enterprises covered in the FUSADES survey to enquire into their access to infiastructure, as well as any difficulties they report with infrastructure services. It is often argued that the key to rural development is not so much the provision of state-of-the-art components of infrastructure as ensuring access to a basic package which combines simple, or even rudimentary, communications, transport, power and water services (World Bank 1994b). About two thirds of rural enterprises have access to electricity. In El Salvador, very few households without connection appear to provide for their electricity requirements via the purchase of generators. In the s A coefficient c multiplying a dummy variable can be interpreted as a percent change in the endogenous variable only as long as c is close to zero. For larger values, in absolute tenns, the percent change in the endogenous variable is given by lOO[exp(c)-l]. Annex 4: Rural Non-Agricultural Sector 9 FUSADES survey no rural enterprises without an electricity connection reported such private provision of electricity. Water provision was most common via private sources such as a well or river, although nearly 47% of enterprises were catered to by the public network. Of all rural enterprises, nearly 19% reported shortages of water during at least part of the year. Very few rural enterprises have a telephone connection, and 35% of enterprises reported transport-related difficulties associated with the poor state of road infiastructure. Thus, for most inifastructure sectors a sizeable fraction of rural enterprises report not having access to the services, or having problerns with the service. While the situation in rural El Salvador is perhaps not as critical as in other, larger and less densely populated countries, it does appear that access to rural infrastructure is far from perfect. Per capita costs of rural infrastructure provision, in a country such as El Salvador, are presumably not nearly as high as they would be in other countries. Table 11 indicates that the vast majority of rural non-agriculture enterprises obtained their start-up capital from personal savings. In fact, only 7% of enterprises were originally financed through formal sector credit sources. It might appear therefore that the contribution of formal credit institutions to the emergence of this sub-sector of the rural non-agricultural sector has been, at best, modest. In fact, however, this assessment might not be altogether complete. It has been argued, for example, that in rural areas financial markets are so imperfect that there exist no real alternatives to investing one's savings in one own enterprise (Vijverberg, 1988). Hence, the very fact that formal credit institutions are imperfect, might lead to more rather thal less investment in small home-based enterprises. Reform of financial markets in rural areas should thus not only concentrate on facilitating the flow of funds for investrnent purposes but also focus on mobilizing savings and offering real returns on such savings. The net impact of such reforns on the number of rural non-agricultural firms is not clear, but average productivity levels would be expected to rise. V. Conclusions The preceding analysis has demonstrated that the non-farm sector in El Salvador, while not the dominant sector in rural areas, is important in terms of both employment rates as well as incomes. Roughly 36% of the economically active population in rural El Salvador is employed in the non-agricultural sector. Women are particularly highly represented amongst the economically active population employed in this sector; nearly three quarters of all such women are employed in the non-agricultural sector. Of course, the proportion of women in the conventionally defined "econoincally-active" population is only about a third that of men. The range of activities in which the rual population is engaged extends from manufacturing to services. A substantial number of non-farm activities are linked to agriculture in the sense that they feed into agricultural production, transport or transform agricultural goods, or provide goods and services which agricultural households are likely to be purchasing. However, there is also a sizeable manufacturing component which is not obviously linked to agricultural production. Here, household firms are engaged in sub-contracting relationships with large, probably urban-based, firms. The incomes derived from such activities may introduce an important source of liquidity into rural areas which does not covary with agricultural incomes, much the same as remittance income from abroad. The least poor in rural areas are households which are heavily engaged in the non-agricultural sector. In this way, non-agricultural employment may be seen to offer a route out of poverty. The challenge is thus to increase access of the poor to non-agricultural activities which yield high and stable incomes. While data constraints preclude making strong statements about causality, the analysis in this paper points to a number of areas to which further attention may be directed when addressing the design of policies. First, there is strong evidence to suggest that higher income non-agricultural jobs go to those with higher levels of education. While it is not the case that the uneducated are unable to secure any non-agricultuIral 10 Annex 4: Rural Non-Agricultural Sector employment, the kind of jobs obtained tend to be low-productivity activities, which while no-doubt of great value in relieving poverty somewhat, do not, in all likelihood, raise the standard of living of these households markedly. As individuals become better educated, their earnings rise sharply. Having said this, it is important to point out that at a given level of education, women tend to earn significantly less then men from non-agricultural employment. hfrastructure services appear to exercise significant influence on the likelihood of finding non- agricultural employment. Persons residing in remote, inaccessible areas are far less likely to find employment in the non-agricultural sector. Households connected to the electricity network are far more likely to have family members engaged in the non-agricultural sector. It is not certain whether the electricity services themselves stimulate non-agricultural activities, or whether connection to the electricity network acts as a proxy for location near a large(ish) conurbation, where non-agncultural jobs are likely to be relatively more common. Rural enterprises in El Salvador, are fairly poorly serviced by infrastructure services. In the FUSADES rural survey, only 7% of firms had a telephone connection, more than one third of all firms had no access to electricity, more than a third of all firms reported transport-related difficulties, and more than half were reliant on their own sources of water (and a sizeable fraction reported periods of non-availability). While the cost of infrastructure provision in rural areas is higher than in urban areas, in El Salvador the rural population is fairly dense which would help to reduce per-capita costs. The focus of infrastructure providers should not necessarily be on the introduction of highest quality (but high-cost) infrastructure, but rather on a basic package of simple infrastructure services. Regionally, the distribution of non-agricultural occupations is not uniform. The corridor of Sonsonate, La Libertad, Chalatenango, San Salvador and La Paz appear to contain the highest concentration of non- agricultural activities, including in particular the high-productivity jobs which represent the greatest source of upward mobility. The remote departments of Morazan and Ahuachapan appear to be particularly poorly endowed with non-agricultural activities. It is an important priority to try to understand better what factors can help to induce the emergence of more non-agricultural activities in these relatively under-served departments. A very low percentage of rural enterprises report having obtained financing from fonnal sector credit sources in setting up their enterprises. The overwhelmingly dominant source of finance came from personal services. This implies, first of all, that the poorest in rural areas are not those most likely to set up rural enterprises (although, of course, they may find employment in such firms). It might also tell us something about the availability of rural savings facilities. The emphasis in rural financial reform is often on increasing the availability of credit in rural areas. The foregoing comments indicate that attention might also be usefully focussed on improving rural savings mobilization. It is interesting to note tiat in rural El Salvador, very few enterprises report having benefitted from special training in conducting their affairs. This is despite there being a substantial experience in El Salvador with trauning and special programs for micro-enterprises. It thus seems important to consider focussing additional efforts to provide such assistance to firms in rural areas. The observation that at least a fraction of the rural non-agricultural activity involves sub-contracting arrangements between home-based firms and larger firms merits further study. There is already a sizeable experience with "Maquila" industries in El Salvador. These are generally large-scale undertakings in which an industrial park is created, providing a whole package of ifriastructural and logistical services, and foreign investors are then invited to locate their assembly plants there. The smaller, more localized "sub-contracting" model resembles this approach in spirit, although it is of course at a much smaller scale. Important outstanding questions surround just what minimal package of inastructure, training, and finance is necessawy to give further Annex 4: Rural Non-Agricultural Sector 11 smulus to tius approach. Clearly, one un iskable pre-requisite is a measure of security and stability in rural areas. 12 Annex 4: Rural Non-Agrcultural Sector REFERENCES Aoki, M., Murdock, K. and Okuno-Fujiwara, M.(1995) 'Beyondthe East Asian Miracle: Introducing the Market- Enhancing View' in Aoki, M., Kini, H. and Fujiwara, M. (eds) (1995) The Role of Govermment in East Asian Economic Development: Comparative Institutional Analysis, manuscript, Economic Development Institute, the World Bank. Datt, G. and Ravallion, M. (1993) Transfer Benefits fiom Public Works Employment: Benefits for Rural India', Economic Journal, Vol 104, No. 427. De Janvry, A. and Sadoulet, E. (1993) 'Rural Development in Latin America: Relinldng Poverty Reduction to Growth', in Lipton, M. and van der Gaag, J. (eds) Including the Poor (Washington D.C.: the World Bank). Dreze, J.P. (1990) 'Famine Prevention in Idia', in Dreze, J. and Sen, A.K.(eds) The Political Economy of Hunger Vol 2: Famine Prevention, (Oxford: Oxford University Press). Dreze, J.P., Lanjouw, P. and Stern, N.H. (1992) 'Economic Mobility and Agricultural Labour in Rural India: a Case Study, Indfan Economic Review, (Special Number in Memory of Sukhamoy Chakravarh), Vp; XXVII. FUSADES (1993) 'Empleo, Ingreso y Pobreza Rural', Boletin Economico y Social, No. 91. Hayani, Y.(1995) 'n Search of Rural Entrepreneurship in Asia: Concept and Approach', paper presented at a World Bank Conference on Rural Development in Latin America, September, 1995. IIES-UCA (1993) la Reduccion de la Pobreza: Comentaios al Informe de Miplan', mimeo, lIES-UCA. Jonmston, B. and Kilby, P. (1975) Agriculture and Structural Transformation: Economic Strategies in Late Developing Countries (London: Oxford University Press). Klein, E. (1993) 'El Empleo Rural No-Agricola en America Latina', in CEPLAES, Latinoamerica Agraria Hacia el Siglo XX, (Quito: Ecuador). Lanjouw, J.O. and Lanjouw, P. (1996) 'Poverty Comparisons with Non-Compatible Data: Theory and Illustrations', mimeo, Policy Research Departnent, the World Bank. Annex 4: Rural Non-Agricultural Sector 13 Lopez, R (1996) 'Rural Poverty in El Salvador: a Quantitative Analysis', mimeo, Dept. of Agricultural and Resource Ecornonics, University of Maryland at College Park. Mellor, J. and Lele, U. (1973) 'Growth Linkages of the New Food Grain Technologies' Indian Journal of Agricultural Economics, 18(1). Mellor, J. (1976) The New Economics of Growth: a Strate2v for India and the Developin World (Itaca, N.Y.: Comell University Press). MIPLAN (1995a) El Salvador: Evaluacion de los Factores de Persistencia de La Pobreza en Los Hogares Pobres', mimeo, MIPLAN, San Salvador. MIPLAN, 1995b "Principales Resultados: Encuesta de Hogares de Propositos Multiples", mimeo, MIPLAN and Miisterio de Relaciones Exteriores, San Salvador. Murdoch, J. (1995) Income Smoothing and Consumption Smoothing' Journal of Economic Perspecfives, Vol 9, No.3. Ravallion, M. (1994) Poverty Compansons (Chur: Harwood Press). Vijverbr, W. (1988) 'Profits from Self-Employment' LSMS Working Paper No. 43, the World Bank. World Bank, (1994a) 'El Salvador: The Challenge of Poverty Alleviation', Report No. 12315-ES, the World Bank. World Bank, (1994b) Irastructure fbr Development, World Development Report, (Washington D.C.: the World Bank). 14 Annex 4: Rural Non-Agricultural Sector Table 1 The Incidence of Poverty in El Salvador High Poverty Line Low Poverty Line Incidence Persons Incidence Persons of Poverty Poor of Poverty Poor West Urban 0.68 312,343 0.28 127,241 Rural 0.75 449,050 0.38 223,726 All 0.72 761,393 0.33 350,967 Central 1 Urban 0.74 329,117 0.31 138,726 Rural 0.76 560,313 0.33 240,980 All 0.75 889,430 0.32 379,706 Central 2 Urban 0.70 131,827 0.36 66,983 Rural 0.79 237,550 0.32 96,925 All 0.76 369,377 0.34 163,908 East Urban 0.67 287,074 0.30 129,960 Rural 0.79 513,079 0.38 247,231 All 0.74 800,153 0.35 377,191 Metropolitan San Salvador 0.40 518,926 0.08 100,751 National Urban 0.56 1,579,287 0.20 563,661 Rural 0.77 1,759,992 0.35 808,862 Total 0.66 3,339,279 0.27 1,372,523 Source: Encuesta de Hogares de Propositos Multiples, 1994-Ill. Notes: 1-. The Encuesta de Hogares de Propositos Multiples 1994-111 yields potentially problematic consumption figures as two sharply divergent consumption questionnaires were fielded to two non- overlapping sub-samples of total sample. The methodology developed in Lanjouw and Lanjouw (1996) was implemented in order to ensure comparability. For more details, see Lanjouw and Lanjouw (1996). 2. The "high" poverty line refers to a monthly per capita expenditure figure of 667 Colones (approx US$75), and the "low" poverty line refers to a monthly per capita expenditure figure of 334 Colones (approx US$36). These poverty lines have been calculated from the data based on the per capita cost of a basic food bundle (for urban areas) as calculated by the Ministerio de Planificacion in San Salvador, and adding a non-food expenditure allowance in accordance with the average amount spent on non-food items by households with food expenditures equal in value to the MIPLAN food basket. The "low" poverty line is simply 50% of the "high" line. 3. Rural expenditures have been inflated by a factor of 60% to account for MIPLAN's calculation that the basic urban food basket costs 60% more than a basic rural food basket. 15 Annex 4: Rural Non-Agricultural Sector Table 2 Non-Farm Activities in Rural El Salvador (Percentage of Persons Aged Above 10 Years Engaged In Remunerated Labor) Percentage of Population With Primary Occupation in: Male Female TOTAL Fishing 0.7 (2.8) 0.2 (0.3) 0.6 (1.6) Manufacture 4.2 (17.0) 10.0 (13.8) 5.6 (15.4) Textiles/Garments 1.0 (4.0) 8.0 (11.1) 2.7 (7.4) Wood/Straw/Leatherware 1.4 (5.7) 3.7 (5.1) 1.9 (5.2) Utilities 0.3 (1.2) 0.0 (0.0) 0.3 (0.8) Construction 6.1 (24.7) 0.4 (0.6) 4.7 (12.9) Commerce 2.1 (8.5) 28.1 (38.9) 8.4(23.1) Restaurant/Hotel 0.2 (0.8) 1.8 (2.5) 0.6 (1.6) Transport 2.8 (11.3) 0.0 (0.0) 2.1 (5.8) Finance 0.1 (0.4) 0.5 (0.7) 0.2 (0.5) Administration 0.2 (0.8) 0.2 (0.3) 0.2 (0.5) Teaching 0.4 (1.6) 1.6 (2.2) 0.7 (1.9) Health 0.0 (0.0) 1.3 (1.7) 0.4 (1.1) Domestic Service 0.9 (3.6) 12.4 (17.2) 3.7 (10.2) Other Service 4.3 (17.4) 4.1 (5.7) 4.3 (11.8) Total 24.7 (100.0) 72.3 (100.0) 36.4 (100.0) Farming 54.4 11.4 44.0 Agricultural Labor 20.8 16.1 19.6 Note: 1. Column percentages provided in brackets. Source: Republica de El Salvador: Encuesta de Hogares de Propositos Multiples, 1994-EI. 16 Annex 4: Rural Non-Agricultural Sector Table 3 Non-Farm Activities in Rural El Salvador (Percentage of Persons Aged Above 10 Years Engaged in Remunerated Labor) Percentage of Population With Primary Occupation: West Central 1 Central 2 East Fishing 0.0 (0.0) 0.9 (1.9) 1.6 (4.7) 0.3 (1.3) Manufacture 6.3 (18.8) 6.5 (13.4) 5.3 (15.7) 3.5 (15.1) Textiles/Garments 2.0 (6.0) 3.8 (7.8) 3.3 (10.0) 1.5 (6.5) Wood/Straw/Leatherware 2.3 (6.8) 3.1 (6.4) 0.5 (1.5) 0.6 (2.6) Utilities 0.4 (1.2) 0.4 (0.8) 0.2 (0.6) 0.0 (0.0) Construction 4.1 (12.2) 6.5 (13.4) 4.9 (14.5) 2.8 (12.1) Commerce 7.1 (21.2) 9.9 (20.4) 9.7 (28.8) 7.1 (30.6) Restaurant/Hotel 0.1 (0.3) 1.3 (2.7) 0.6 (1.8) 0.0 (0.0) Transport 2.2 (6.5) 2.8 (5.8) 0.6 (1.8) 1.9 (8.2) Finance 0.2 (0.6) 0.2 (0.4) 0.3 (0.9) 0.2 (0.9) Administration 0.5 (1.5) 0.1 (0.2) 0.1 (0.3) 0.1 (0.4) Teaching 0.8 (2.4) 0.4 (0.8) 0.9 (2.7) 0.9 (3.9) Health 0.5 (1.5) 0.5 (1.0) 0.0 (0.0) 0.2 (0.9) Domestic Service 3.0 (8.9) 5.8 (12.0) 2.4 (7.1) 2.0 (8.6) Other Services 4.1 (12.2) 6.3 (13.0) 3.3 (9.8) 2.1 (9.1) Total 33.6 (100.0) 48.5 (100.0) 33.7 (100.0) 23.2 (100.0) Farming 43.7 33.1 50.1 56.6 Agricultural Labor 22.7 18.3 16.1 20.2 Note: 1. Column percentages provided in brackets. Source: Republica de El Salvador: Encuesta de Hogares de Propositos Multiples, 1994-EI. 17 Annex 4: Rural Non-Agricultural Sector Table 4 Poverty and Rural Household Activities Household Percent of Average Per Incidence of Characteristics Population Capita Income Etreme Poverty Agricultural Labor 5.0% 166.33 54.7% And Farming Agricultural 9.6% 212.41 48.7% Labor Only Agricultural Labor 2.7% 272.72 43.9% Farming and Non-Farm Employment Farming Only 26.1% 214.84 41.5% Farming And Non- 19.9% 316.65 35.9% Farm Employment Agricultural Labor 9.1% 274.93 35.2% and Non-Farm Employment Non-Farm Employment 26.1 % 334.46 20.3% Only Non-Farm Income 1.6% n/a 16.3% From Non-Wage Sources Source: Encuesta de Hogares de Propositos Multiples, 1994-III. Notes: 1. Agricultural Labor households are defined as such if at least one household member is employed as a salaried or casual wage laborer in agriculture. Farming households refer to those households where at least one household member is engaged in cultivation. Non-Farm households correspond to those households where at least one household member is employed in a non-agricultural occupation. 2. Extreme poverty is associated with per-capita consumption levels falling below the "low" poverty line. 18 Annex 4: Rural Non-Agricultural Sector Table 5 Rural Enterprises in El Salvador Sector Number Number Percent Percent Percent Perrcent of of Family Home-based With Supplying Firms Workers Members Training Contractor Transport 1 1 100% 0% 0% 0% Other Services 3 3 100% 33% 0% 0% Other Industry 5 6 100% 100% 0% 20% Repair Shop 6 16 44% 33% 0% 0% Restaurant/Bar 5 19 16% 20% 0% 0% Textiles 13 25 73% 92% 8% 23% Wood/Work 10 37 22% 60% 10% 30% Food Proc. 13 53 28% 54% 8% 0% Pottery/Bricks 14 63 13% 7% 7% 7% Commerce 31 77 64% 58% 3% 0% Total 101 300 40% 52% 5% 8% Source: Rural Survey, FUSADES, 1996 19 Annex 4: Rural Non-Agricultural Sector Table 6 Probability of Non-Agricultural Employment as a Primary Occupation (Rural Survey: FUSADES 1996) Probit Model Domain: Entire Rural Population Aged Above 14 Any Non-Agricultural High-Productivity Low-Productivity Occupation Occupation Occupation Obs: 2738 Obs: 2738 Obs: 2738 at 1: 481 at 1: 331 at 1: 150 at 0: 2257 at 0: 2407 at 0: 2588 Variable Estimate Prob Value Estimate Prob Value Estimate Prob Value Intercept -0.862 0.0001 -1.213 0.0001 -1.803 0.0001 Household Size -0.002 0.8785 -0.020 0.1244 0.027 0.0790 Female -0.846 0.0001 -0.866 0.0001 -0.365 0.0001 Age (years) -0.005 0.0137 0.002 0.3555 -0.017 0.0001 Education (Highest level reached) Primary 0.313 0.0002 0.358 0.0003 0.137 0.2287 Middle School 0.282 0.0054 0.412 0.0003 -0.014 0.9194 ligh School 0.663 0.0001 0.900 0.0001 -0.072 0.6280 Tertiary Level 0.883 0.0030 1.289 0.0001 -5.209 0.9993 Per Capita Land -0.279 0.0001 -0.239 0.0017 -0.229 0.0359 Cultivating HH. -0.748 0.0001 -0.716 0.0001 -0.393 0.0016 Distance to Road -0.008 0.0979 -0.007 0.2067 -0.008 0.2976 Distance to School -0.007 0.3041 0.005 0.4751 -0.029 0.0080 Electricty Connec. 0.210 0.0021 0.161 0.0359 0.157 0.0956 Ahuachapan -0.055 0.7940 -0.205 0.3698 0.560 0.1894 Santa Ana 0.244 0.2218 -0.040 0.8521 0.941 0.2261 Sonsonate 0.565 0.0041 0.311 0.1382 1.046 0.0102 Chalatenango 0.037 0.8722 0.562 0.0061 0.822 0.0591 La Libertad 0.491 0.0102 0.311 0.1247 0.904 0.0258 San Salvador 0.843 0.0001 0.562 0.0061 1.183 0.0035 Cuscatlan 0.267 0.2295 0.196 0.4090 0.540 0.2259 La Paz 0.374 0.0744 0.354 0.1067 0.376 0.4100 Cabafias 0.344 0.1448 -0.219 0.4513 1.170 0.0068 San Vicente 0.139 0.5800 -0.063 0.8242 0.706 0.1206 Usulutan 0.300 0.1469 0.136 0.5379 0.809 0.0558 San Miguel 0.4S1 0.0226 0.328 0.1187 0.767 0.0655 La Union 0.344 0.0893 0.056 0.7963 1.016 0.0148 Log Likelihood Model -1058.84 -833.34 -515.92 Constant -1272.55 -1009.49 -581.47 LR Test (Model) 427.42 352.30 131.10 Degrees of Freedom 25 25 25 Critical XI 37.65 37.65 37.65 Source: Rural Survey, FUSADES, 1996 20 Annex 4: Rural Non-Agricultural Sector Table 7 Probability of Non-Agricultural Employment as a Primary Occupation (Rural Survey: FUJSADES, 1996) Probit Model Domain: Rural Population Aged Above 14 and Engaged in Remunerated Work Any Non-Agricultural High-Productivity Low-Productivity Occupation Occupation Occupation Obs: 1592 Obs: 1592 Obs: 1592 at 1: 481 at 1: 331 at 1: 150 at 0: 1111 at 0: 1261 at 0: 1442 Variable Estimate Prob Value Estimate Prob Value Estimate Prob Value Intercept -0.662 0.0153 -1.018 0.0003 -1.695 0.0013 Household Size 0.007 0.6109 -0.017 0.2507 0.037 0.0397 Fernale 0.097 0.3025 -0.136 0.1720 0.338 0.0025 Age (years) -0.007 0.0087 0.003 0.3593 -0.020 0.0001 Education (Highest kvel reached) Primary 0.313 0.0018 0.335 0.0022 0.114 0.3841 Mfiddle School 0.292 0.0147 0.400 0.0020 -0.037 0.8135 High School 0.912 0.0001 1.077 0.0001 -0.068 0.6875 Tertiary Level 1.555 0.0007 1.961 0.0001 -5.505 0.9995 Per Capita Land -0.167 0.0456 -0.132 0.1223 -0.124 0.3090 Cultivating HH. -1.146 0.0001 -1.041 0.0001 -0.648 0.0001 Distance to Road -0.009 0.1322 -0.008 0.2524 -0.007 0.4263 Disance to School -0.005 0.5729 0.009 0.2675 -0.030 0.0148 Electricty Connec. 0.303 0.0002 0.200 0.0220 0.203 0.0588 Ahuachapan -0.220 0.3839 -0.327 0.2163 0.538 0.2917 Santa Ana 0.022 0.9287 -0.246 0.3270 0.915 0.0651 Sonsonate 0.385 0.1048 0.109 0.6580 1.009 0.0390 Chalatenango -0.057 0.8423 -0.395 0.2014 0.968 0.0652 La Libertad 0.366 0.1149 0.167 0.4851 0.902 0.0640 San Salvador 0.896 0.0002 0.494 0.0418 1.209 0.0134 Cusatlan 0.104 0.7013 0.038 0.8907 0.513 0.3355 La Paz 0.376 0.1456 0.373 0.1559 0.398 0.4606 Cabainas 0.452 0.1079 -0.227 0.4893 1.405 0.0068 San Vicente 0.283 0.3452 0.007 0.9827 0.933 0.0877 Usulutan 0.276 0.2751 0.086 0.7429 0.887 0.0802 San Miguel 0.607 0.0124 0.416 0.0947 0.917 0.0670 La Union 0.514 0.0396 0.116 0.6541 1.252 0.0130 Log Likelihood Model -739.82 -649.27 -414.86 Consant -975.36 -813.80 -497.02 LR Test (Model) 471.08 329.06 164.32 Degrees of Freedom 25 25 25 Critical X2 37.65 37.65 37.65 Source: Rural Survey, FUSADES, 1996. 21 Annex 4: Rural Non-Agricultural Sector Table 8 Probability of Non-Agricultural Employment as a Primary Occupation (National Survey: Encuesta de Hogares, 1994-III, MIPLAN) Probit Model Domain: Rural Population Aged Above 14 and Engaged in Remunerated Work Any Non-Agricultural High-Productivity Low-Productivity Occupation Occupation Occupation Obs: 2914 Obs: 2914 Obs: 2914 at 1: 1035 at 1: 544 at 1: 491 at 0: 1879 at 0: 2370 at 0: 2423 Variable Estimate Prob Value Estimate Prob Value Estimate Prob Value Intercept -1.570 0.0001 -2.010 0.0001 -1.641 0.0001 Household Size -0.01 0.0990 0.003 0.9737 -0.024 0.0508 Female 1.338 0.0001 0.290 0.0001 1.269 0.0001 Age (years) 0.004 0.0300 0.005 0.0190 0.001 0.5350 Education (Highest level reached) Primary 0.383 0.0001 0.483 0.0001 0.057 0.3813 Middle School 1.320 0.0001 1.393 0.0001 0.036 0.8133 Hligh School 1.655 0.0001 1.778 0.0001 -0.435 0.0476 Tertiary Level 7.953 0.9981 3.349 0.0001 -0.972 0.0292 Per Capita Land -0.002 0.0070 -0.001 0.1848 -0.001 0.0858 Cultivating HH. -0.328 0.0001 -0.272 0.4363 -0.173 0.0439 Ahuachapan 0.371 0.0718 0.237 0.3416 0.361 0.1193 Santa Ana 0.533 0.0071 0.281 0.2431 0.499 0.0258 Sonsonate 0.884 0.0001 0.588 0.0137 0.706 0.0016 Chalatenango 0.811 0.0001 0.629 0.0112 0.518 0.0293 La Libertad 0.692 0.0005 0.699 0.0032 0.315 0.1634 San Salvador 1.326 0.0001 1.213 0.0001 0.401 0.0766 Cuscatlan 0.682 0.0018 0.463 0.0742 0.532 0.0294 La Paz 0.904 0.0001 0.798 0.0007 0.453 0.0436 Cabanas 0.046 0.8414 -0.010 0.9718 -0.002 0.9944 San Vicente 0.539 0.0096 0.454 0.0697 0.352 0.1361 Usulutan 0.143 0.5035 0.112 0.6636 0.149 0.5418 San Miguel 0.405 0.0564 0.122 0.6416 0.469 0.0478 La Union 0.514 0.0158 0.222 0.3939 0.497 0.0366 Log Likelihood Model -1401.40 -1080.48 -1131.35 Constant -1895.83 -1402.78 -1321.48 LR Test (Model) 988.86 644.60 380.26 Degrees of Freedom 22 22 22 Critical X' 33.92 33.92 33.92 Source: Encuesta de Hogares de Proposilos Mudtiples, MIPLAN, 1994-JIL 22 Annex 4: Rural Non-Agricultural Sector Table 9 Non Agricultural Labor Earnings (Rural Survey: FUSADES 1996) OLS Modd Dependent Variable: (Log) Annual Non-Agricultural Labor lnome Domain: AU Persons Aged 14 and Above ws4th Non-Agricdtral Employment With Adjustment for Sample Selection No Adjustment (MUills Ratio Not Icluded) Variables Estimate Prob. Value Estimate Prob. Value Constant 9.073 0.0001 9.075 0.0001 Household Size -0.009 0.5116 -0.009 0.4998 Female -0.327 0.0001 -0.328 0.0001 Age (in years) 0.007 0.0239 0.007 0.0241 Education (Highest Level Achieved) Primary 0.159 0.134 0.158 0.1346 Middle School 0.373 0.0018 0.373 0.0018 High School 0.580 0.0001 0.582 0.0001 Tertiary 1.172 0.0001 1.172 0.0001 Per Capita Land 0.074 0.4169 0.073 0.4207 Cultivating Household (dummy) -0.628 0.0001 -0.628 0.0001 Distance from Paved Road -0.002 0.7430 -0.002 0.7411 Distance from Secondary School 0.005 0.5896 0.005 0.5881 Electricity Connection -0.081 0.3228 -0.080 0.3257 Ahuachapan -0.326 0.2686 -0.323 0.2675 Santa Ana -0.234 0.3835 -0.234 0.3826 Sonsonate -0.220 0.4028 -0.221 0.4014 Chalantenango -0.122 0.7032 -0.122 0.7028 La Libertad -0.125 0.6288 -0.125 0.6280 San Salvador -0.161 0.5348 -0.157 0.5424 Cuscatlan -0.143 0.6306 -0.143 0.3506 La Paz -0.258 0.3522 -0.258 0.3506 Cabafias -0.289 0.3684 -0.286 0.3686 San Vicente -0.022 0.9481 -0.023 0.9477 Usulutan -0.359 0.1969 -0.359 0.1967 San Miguel 0.120 0.6536 0.120 0.6529 La Union -0.113 0.6810 -0.359 0.1967 AdjustedR2 0.1471 0.1489 Number of Observations 481 481 Source: Rural Survey, FUSADES, 1996. 23 Annex 4: Rural Non-Agricultural Sector Table 10 Rural Enterprises and Infrastructure Services SECTOR OF ACTIVITY Infrastructure Transport Commerce Other Other Repairs Restaurant Textiles Wood- Food Pottery/ Total Sector Serv. Ind. and Hotel work Proc. Brick 1..Road..................................................................................................................................................................................................................... Firms Using Vehicle 1 0 2 0 1 1 6 2 7 6 25 Firms Reporting 0 2 1 2 1 2 6 4 3 14 35 Transportation Difficulties 2. Telephone Firms With Phone 0 0 0 0 2 0 0 1 2 2 7 Connection Firms Reporting Problem 0 0 0 0 0 0 0 0 0 0 0 with Phone Line 3. Electricity Finns With Public 0 2 1 4 4 8 7 7 10 23 66 Connection Firms Without Electricity 1 1 4 2 1 5 3 6 4 8 35 4. Water Firms With Public Supply 1 1 1 3 5 8 5 3 4 16 47 Firms With Private Supply 0 2 4 3 0 5 5 5 10 15 54 Firms Reporting Water 0 1 0 0 3 1 4 4 0 6 19 Shortages Firms Reporting Waste- 0 0 0 1 0 0 1 2 1 4 9 Water Difficulties 5. Waste Removal Firms with Public Waste 0 0 0 1 4 1 1 0 1 5 13 Removal Firms with Private Waste 0 0 0 0 0 0 2 0 2 3 7 Removal Firms With No Effective 1 3 5 5 1 12 7 13 11 23 81 Waste Removal 13 11 23 81 Total Number of Firms 1 3 5 6 5 13 10 13 14 31 101 Source: Rural Survey, FUSADES, 1996 24 Annex 4: Rural Non-Agricultural Sector Table 11 Rural Enterprises and Start-Up Finance Principal Source of Start-Up Finance (percent of firms) Sector Number Number Personal Friends Informal Formal of of Savings and Sources Sources Firms Workers Relatives Transport 1 1 100% 0% 0% 0% Other Services 3 3 67% 33% 0% 0% Other Industry 5 6 100% 0% 0% 0% Repair Shop 6 16 100% 0% 0% 0% Restaurant/Bar 5 19 60% 20% 0% 20% Textiles 13 25 85% 0% 15% 0% Wood/Work 10 37 30% 30% 40% 0% Food Proc. 13 53 46% 8% 23% 23% Pottery/Bricks 14 63 85% 7% 7% 0% Commerce 31 77 68% 16% 6% 10% Total 101 300 70% 11% 12% 7% Source: Rural Survey, FUSADES, 1996 †.....* ~.*........* .. ......:. ..:: : . . .. ....... WHO USES BASIC SERVICES IN RURAL EL SALVADOR? I. Introduction Recent policy changes in El Salvador have laid the basis for sustained growth. However, poverty is still widespread and inequality appears to be increasing. To address this situation, the poor need to have access to the benefits of growth. In particular, public investments in human capital and basic infrastructure need to be targeted to the poor in disadvantaged areas. This paper examines the use of basic services in rural El Salvador. A Benefit Incidence Analysis of public services was not elaborated because of the lack of informnation on per unit public spending within each sector. The distribution of public spending across different socioeconomic groups and regions can be analyzed when two main sources of information are brought together. First, utilization patterns of public services. Second, the allocation of government resources within each sector. The determinants of the use of basic services are not analyzed since the household surveys did not contain information on the quality and price of basic services. The 1994-III Encuesta de Hogares de Prop6sitos Multiples (EHPM), was used to contrast utilization patterns of public services in rural areas with those in urban areas.1 EHPM is utilized to compare the use of basic services by the rural poor and non-poor in contrast to the poor and non-poor in San Salvador and to the poor and non-poor in other urban areas.2 Thus, it provides information on the regional distribution of disadvantaged groups across El Salvador. The 1996 Encuesta Rural, was used to analyze the degree of targeting of basic services across income groups within rural areas.3'4 Hence, it provides the basis for investigating inequality in the distribution of basic services in rural areas.5 This paper looks at educational indicators in Section II. Health indicators are explored in Section _II. Section IV looks at basic needs. In-kind transfers are analyzed in Section V. The last Section discusses the most relevant policy implications coming from this study. II. Education - More than one-third of the rural population cannot read and write More than one in three rural inhabitants in El Salvador older than ten years old cannot read and write, compared to less than one in ten in San Salvador and less than one in five in other urban areas (Figure 1). The poor in rural areas fare even worse, more than 40% of them are illiterate. I EHPM is a nationally representative sample of El Salvador containing 4,220 households and 19,914 individuals. 2 The poverty lines used for EIIPM are 0325 per person per month for those households covered by the long consumption questionnaire, and 0286 per person per month for those covered by the short questionnaire. The welfare measure for ranking individuals in EBPM is total household expenditure per capita. The national poor are the poorest 40% of all individuals who fill below these poverty lines, percentage of the population that Lanjouw (1996) reports as being poor. 3 The Encuesta Rural is a representative sample of rural areas in El Salvador containing 738 households and 4,349 individuals. 4 We use income and expenditure interchangeably throughout the paper, but note that the welfare measure used for ranking individuals is total household expenditure per capita for EHPM and total household income per capita for the Encuesta Rural. 5 Inequality in rural areas is analyzed by creating income quintiles by ranking every individual from the poorest to the richest and then dividing the population into five groups each containing exactly 20% of all individuals. Thus, the poorest income quintile contains the poorest one-fifth of the population while the richest income quintile contains the richest one-fifth. The welfare measure for ranking individuals in the Encuesta Rural is total household income per capita per annum. The Encuesta Rural did not have a household consumption expenditure module. When referring to the rural poor, we refer to the poorest two quintiles or poorest 40% of all individuals. The annual level of income per person at this poverty cut-off is 0 1975. 2 Annex 5. Basic Services Figure 1: Illiteracy rate for poor and non-poor, 1994 by area of residence (%) 45 40 e~35 30 25+ 20 4 0 Rural Otjzer San El Urban Salvador Salvador *Poor ENao-poor MAU I Source: Table A.1 (EHPM, 1994-111) Almost half of the poor population in the Eastern region cannot read and write (Figure 2). This group has the highest illiteracy rate in El Salvador. Among all regional poor and non-poor inhabitants in El Salvador, illiteracy rates are highest in the Eastern and the Western regions, 33 and 29% respectively. Literacy programs need to target the rural population. particularly the poor in Eastern El Salvador. Figure 2 Illiteracy rate for poor and non-poor, 1994 by region (%) 45 --- I 4 e35-. I 30 + .~25 - 20 r 15T b10-- *-5+ 0 [ lE~~~~~Pa., U Non-poorUI Source: Table A.2 (EHPM, 1994-11l) * Females in rural areas have had the highest increase in years of schooling, but both males and females in rural areas have the lowest average overall The gender gap in mean years of schooling has almost closed in rural areas. Females in rural areas aged between 15 and 34 years old have attended school for 3.6 years, on average, compared to 3.8 years attended by males 'm the same age group (Figure 3). Gender disparities for the older age group, those aged 35 years or more, are the widest in rural areas. 6 This paper uses the five regional division with the following departments: Western = Ahuachapan, Santa Ana and Sonsonate; Central I = Chalatenango and La Libertad; Central II = Cuscatlan. La Paz, Cabanas, San Vicente; Eastem = Usulutan, San Miguel, Morazan and La uJnion; and, San Salvador Annex 5: Basic Services 3 Overall, the younger age group in rural areas has achieved the highest increase in mean years of schooling compared to San Salvador and to other urban areas. Those in the younger age group have attended almost four years of school compared to about a year and a half among those in the older age group. However, the average number of years in school for the younger group in rural areas lags behind by three and by five years compared to the younger groups in other urban areas and in San Salvador, respectively. Thus, it is crucial to put in place incentives for retention in rural primary schools. Figure 3: Mean years of schooling for adults in two age groups, 1994 by area of residence and gender 10 9.2 .~8 6.9 7 --6.4 - 6 - - (Males 5ros4 e r Females 1%. 3.8 93 4 36 3 1.9 2 1. 1 0 35+ 15-34 35+ 15-34 35+ 15-34 Rural Otlier urban San Salv-ador Area of residence and age group Source: EHPM (1994-111) Primary school students older than thieappropriate age group prevail across all income groups and regions Gross enrollment rates at the primary level in El Salvador are considerably higher than net enrollment rates across all income groups and regions (Table 1). This indicates that many of the school places are filled by students who are older than the appropriate age group. Thus, the primary education systemn in El Salvador needs to improve internal efficiency by promoting early enrollment and decreasing repetition rates across the country, but focusing particularly on the rural poor. Attending primary school at an older age has a particular perverse impact on the poor by decreasing completion rates in this educational level and lowering enrollments in the following educational cycle. High gross enrollment rates in primary school are a disincentive for the poor because poor children are often needed to help to support the household. In El Salvador, net and gross enrollment rates in secondary school are considerably lower among the poor compared to the non-poor (Table 1). In addition, gross enrollment rates in secondary education are higher among non-poor students and students from better-off regions. This indicates that non-poor students proceed to higher educational levels even at an older age than the appropriate age group. 4 Annex 5: Basic Services Table 1: Net and gross enrollment rates in pre-school, primary and secondary, 1994 for poor and non-poor by area of residence (%) Area of Net Enrollment Rates Gross Enrollment Rates residence Pre-school Primary Secondary Pre-school Primary Secondary Rural poor 12 70 15 12 96 25 non-poor 20 83 34 20 111 54 all 14 74 22 14 101 35 Other urban poor 17 80 30 17 105 50 non-poor 37 89 52 37 114 85 all 29 86 45 29 111 74 San Salvador poor 32 86 23 32 120 46 non-poor 52 90 56 52 110 83 all 50 89 53 50 111 79 El Salvador poor 14 73 19 14 99 31 non-poor 38 88 48 38 112 75 all 26 81 35 26 106 55 Source: EHPM (1994-111) Notes: (a) Net enrollment rates in pre-school, primary and secondary: 4-6, 7-12 and 13.15 year olds enrolled In pre- school, primary and secondary, respectively, as percentage of 4-6. 7-12 and 13-15 year old population. (b) Gross enrollment rates In pre-school, primary and secondary: All children enrolled in pre-school, primary and secondary as percentage of 4-6, 7-12 and 13-15 year old population. (c) Net and gross enrollment rates in pre-school are the same because EHPM did not report any children younger than 4 or older than 6 years attending pre-school. -The poorest girls in rural areas are more likely to attend primary school than boys but least likely to continue into secondary education Poor girls in rural areas are more likely to attend primary school than poor boys. The net enrollment rate for the poorest girls in primary education is ten percentage points higher than the rate for the poorest boys (Table 2). However, poor girls in rural areas are least likely to continue into secondary education than poor boys. This is an indication that girls are more likely than boys to drop out of school as they get older. Thus, it is crucial to put in place incentives for early primary school enrollment for all school-age children in rural areas, but particularly pay attention to the retention of the poorest girls. Table 2: Net enrollment rates in primary and secondary, 1996 by rural quintile and gender (%) Rural Primary Secondary quintile Girls Boys Girls Boys Poorest 77 67 13 18 11 69 64 13 16 [if 77 75 21 21 IV 78 79 39 23 Richest 79 93 43 22 All Rural 77 74 25 23 Source: Encuesta Rural (1996) * School enrollments are strongly pro-poor only at the primary level only The distribution of enrollments in 1996 by educational levels and across income quintiles is shown in Figure 4 together with the Lorenz distribution of household income. The diagonal line (or 450 line) is also known as the line of absolute equality since it goes through those points where the cumulative share of the population equals the cumulative share of household income. Annex 5: Basic Services S Figure 4: Distribution of school enrollments and household income, 1996 by rural quintile 100 80 P~~~~*nI. ,, 60Prar 40 ~ ~~~~ .A' 40 ' E Quintiles Source: Encuesta Rural (1996) In 1996, enrollments at the primary educational level are strongly pro-poor because their distribution is above the diagonal line and the Lorenz curve; which means that the poor's share of enrollments is larger than their population share. Enrollments at the pre-school, secondary and high-school educational levels are weakly pro-poor because their distributions are in between the diagonal line and the Lorenz curve; therefore, the poor's share of enrollments is smaller than their population share but relatively larger than their household income share. The distribution of enrollments in university education is considerably more inequitable than the Lorenz distribution of total household income. E EDUCO schools reach the poor but not the poorest rural children A larger proportion of poor children attend EDUCO schools compared to non-poor children. However, only 8% of the poorest children attend EDUCO schools compared to sixteen percent of children in the second quintile (see Figure 5). Improved targeting of the EDUCO program implies correcting undercoverage of poor children and simultaneously cutting down the leakage to non-poor children. 6 Annex 5: Basic Services Figure 5: Primary school children attending EDUCO schools, 1996 by rural quintile (%) } 20 ; 18 16 e 16 ;14 12 12 10 10 10 - EG E 8 8 a. 6 4 42 E 2 Quintiles Source: Table A.3 Note: Percentage share of all primary school enrollments in each rural quintile * EDUCO schools need to increase enrollments among the poorest rural children, particularly in the Central II and the Eastern regions The Central II and the Eastern regions have the highest poverty rates for the rural population as a whole and also the highest poverty rates for school age children (Figure 6). Almost two-thirds of rural children aged 7 to 10 years old are poor in the Central II region and this proportion is more than half in the Eastern region. In addition, together these two provinces contain about 60% of all poor rural children in the primary school age group (Appendix Figure A. 1). Figure 6: Poverty rates for all and for school-age children by rural region, 1996 (% of population and children 7-10 years old in each rural region who are poor) AUlrural = 49 | Western 3J -. Central I 44 Central __ 64 Eastern 42 54 Rural.S. S. S. 0 20 40 60 80 100 Poverty rate (percentage) |-Al U Children Source: Encuesta Rural, 1996 Annex 5: Basic Services 7 EDUCO schools capture only 7% of primary school children in the Eastern region and 11% in the Central II region (Figure 7). Targeting these regions can improve the poverty focus of the EDUCO program, but it will also be necessary to geographically identify poor areas within these provinces. Figure 7: Primary school children attending EDUCO schools, 1996 by rural region (%) AU rural __ Western10 I Central H I I gt Central ll 1 E:astern _ _~ Rural S.S. 0 0 5 10 15 20 EDUCO enrollments (percentage) Source: Table A.4 Note: Percentage share of all primary school enrollments in each rural quintile * More than one-third of rural poor primary school-aged children are out-of-school Twice as many poor children aged between 7 and 10 years old in rural areas are out-of-school as compared to non-poor children (Figure 8). In addition, primary school-aged children in rural areas are most likely to be out-of-school in contrast to those in any other area of residence in El Salvador. One in every three poor primary school-aged children is out-of-school in rural areas compared to two out of ten in other urban areas and less than two out of ten in San Salvador. Figure 8: Poor and non-poor children 7-10 years old who are out-of-school, 1994 by area of residence (%) e 35- 4' 30 ~,25 20 t ~~~~~~Urban Salpador Salvador |O 10oor 1 | Source: Table A.5 (EHPM, 1994-Oe1) 8 Annex 5: Basic Services * Almost half of rural poor primary school-aged children who are out-of-school report that school expenses are too high The most important reason for being out-of-school for both poor and non-poor children in rural areas is that school expenses are too high1. However, almost half of rural poor primary school-aged children who are out-of-school report that school expenses are too high (Table 3). Improving primary school attendance among the poor requires actions on both the demand for and the supply of education. Thus, easing the financial constraints on the demand for education faced by poor households could significantly increase their children's enrollments at the primary level. Uniforms and stationary expenses for children attending primary schools in El Salvador are seven times as large as school fees for the poor and three times as large for the non-poor. Thus, an alternative for reducing the burden of direct costs of primary education on families with primary school-aged children, targeted particularly to the poor, can be to provide school materials to and to ease requirements on wearing uniformns in disadvantaged rural areas of the country. Table 3: Reason for being out-of-school for poor and non-poor children7-10 years old in rural areas, 1994 (as % of non-enrolled in population group) Reason for being out-of-school Poor Non-poor Needed to work 0 0 School expenses too high 45 39 School is too far 14 9 No teacher at school 0 0 School has closed 0 0 Repeated too many times 1 0 It is not worthy 4 2 Age is inappropriate 15 28 No night school available 0 0 Completed education 0 0 Required at home 6 12 No subsequent grades available 0 0 Other 15 10 Total non-enrolled 100 100 ................................................................................................................................................................ Non-enrolled as percentage of 28 8 population group Source: EHPM (1994-l1l) Note: Sample is restricted to children 7 to 10 years old living in rural areas III. Health * Public health services are not targeted to thepoor or to disadvantaged areas About one-fourth of the population in El Salvador uses public health services when falling ill (Table 4). However, poor and non-poor residents alike have about the same utilization rates of these services within rural areas, within other urban areas and in San Salvador. Thus, public health services are not targeted to the poor. Evenmore, utilization rates are lowest for rural areas, second lowest for other urban areas and highest for San Salvador. Rural residents, poorer than residents from other urban areas and from San Salvador, have the lowest utilization rates of public health services. 7 The second most important reason for being out-of-school for both poor and non-poor children in rural areas is that the age of the child is considered inappropriate. This is an issue to be further explored in rural areas because the sample over which we analyzed the reasons for being out-of-school is already restricted to children 7 to 10 years old, who are of primary school age. Annex 5: Basic Services 9 Table 4: Illness rate and place of consultation, 1994, for poor and non-poor by area of residence (%) Area of Illness Place of consultation residence rate None Private Public Other All places Rural poor 33 10 7 21 62 100 non-poor 36 6 16 22 57 100 all 34 9 10 21 60 100 Other Urban poor 28 11 6 27 57 100 non-poor 31 14 20 26 50 100 all 30 6 16 26 52 100 San Salvador poor 26 5 5 30 61 100 non-poor 25 2 26 30 42 100 all 25 2 24 30 43 100 El Salvador poor 32 10 7 23 61 100 non-poor 30 4 21 26 50 100 all 31 7 15 24 54 100 Source: EHPM (1994-Ill) Notes: (a) Public = public health centers and public hospitals; private = pnvate hospital/clinic, private doctor and nurse; and, other - pharmacy, healer and self-medication. (b) Illness rate is for last 30 days. In rural areas the illness rate is the highest compared to other urban, second highest, and San Salvador, the lowest rate. Although, more than one-third of the rural population reported some type of illness, more than one-third of those falling ill do not use modem health care. About one in every ten rural residents that reports falling ill does not use any type of medical care and six out of ten use pharmacies, healers or self-medication. The Central II and the Eastern regions, the two poorest in El Salvador, have the highest rates of illness (Appendix Table A.6). Close to 40% and about 33% of the residents in these regions, respectively, report some type of illness. Targeting mechanisms of public health services towards the poor and rural areas need to be put in place. In addition to having the highest illness rates, these population groups are severely undercoveraged by modem health services. O Poor rural children are three times more likely to be ill than the non-poor while social security is almost as unequal as household income Rural children in the poorest income quintile are three times more likely to be ill than children in the highest income quintile. Close to 30% of the rural poorest children under 1O years old report an illness that put them in bed for at least a week in 1995 compared to less than 10% of the richest children. Thus, the illness rate for rural children is strongly pro-poor (Figure 9). This means that the poor's share of illness is larger than their population share by income quintiles. Graphically, the distribution of the illness rate by quintiles is above the diagonal line and the Lorenz distribution of household income.8 The illness rate for the rural population as a whole is also strongly pro-poor. 9 The diagonal line (or 450 line) is also known as the line of absolute equality since it goes through those points where the cumulative share of the population equals the cumulative share of household income. 10 Annex 5: Basic Services Figure 9: Distribution of health indicators for all ages and children under 10 years old and household income by rural quintile, 1996 100 N. y I I a 80 Illness allI. Illness ; 60 children / ~40 I Househ, 20 / - . ,' income 20 O ( . ' Social security Quintiles Source: Encuesta Rural (1996) By contrast, access to social security is weakly pro-poor. Only one in every ten rural residents in the poorest income quintile has social security compared to four in every ten for the population in the highest quintile. Thus, the poor's share of social security is smaller than their population share but relatively larger gma their household income share. Graphically, the distribution of access to social security is in between the diagonal line and the Lorenz curve. High illness rates in rural and poor areas of El Salvador are compounded by lack of targeting of public health facilities and by the scarcity of health care financing mechanisms. The findings on illness rates and social security in rural areas by income groups together with the patterns of health facility use indicate that the burden of financing health care is higher on the household budgets of the poor than the non-poor. There is an urgent need in rural El Salvador to increase the access of affordable health care services for poor children. IV. Basic Needs * 2he rural poor in El Salvador have the worst basic needs indicators T'he rural poor in El Salvador have the worst basic needs indicators:9 More than one-third of school-age children are out-of-school; three-fourths of poor rural residents live in overcrowded homes; only 15% of them have access to piped water (including cornmunal sources); only 2% have access to modern sanitation; and, only about one-third have access to electricity (Table 5). 1 9 Although basic needs indicators involve a certain degree of subjectivity in deterrnining adequate levels of access to services, they allow thie identification of more permanent characteristics of poverty as well as the need for basic infrastructure. 10 Children 7 to 10 who are out-of-school is measured at the individual level, the rest of the basic needs indicators are measured for individuals living in households wvith the characteristic. Annex 5: Basic Services 11 Table 5: Basic needs for poor and non-poor, 1994, by area of residence (%)' Access to Children 7-10 who Overcrowding Access to piped modern Access to Area of are out-of-school water sanitation electricity residence Non- Non- Non- Non- Non- Poor poor Poor poor Poor poor Poor poor Poor poor Rural 32 15 75 42 14 28 2 8 35 61 Otherurban 21 7 71 28 35 69 15 53 77 95 San Salvador 15 7 70 18 44 87 41 82 80 98 El Salvador 28 8 74 28 20 65 7 53 46 87 Notes:(a) Overcrowding occurs in households with more than three people per bedroom. (b) Piped water is either inside or outside the home or piped to r common faucet. (c) Modem sanitation is private or shared toilet connected either to the public sewerage system or to a septic fAnk. Children 7-10 who are out-of-school is measured at the individual level, the rest of the basic needs indicators are measured for lndMduals IMng In households with the characteristic. Source: EHPM (1994-111) * All ruralfamilies in El Salvador are severely underserved by basic infrastructure All rural residents are severely underserved by basic infrastructure services. The low degree of inequality in the coverage of these services to rural households does not indicate that these services are weakly pro-poor, but that families from all income groups are underserved (Figure 10). Overall, only about 60% of non-poor residents in rural areas have access to electricity, less than 10 percent of them have access to modem sanitation and one-third of them have access to piped water. By contrast, the proportion of children 7 to 10 who are out-of school and the proportion of residents living in overcrowded homes are highly unequal in rural areas. Non-poor rural families fare considerably better than poor rural families in keeping their primary-school age children enrolled and living in non-overcrowded homes. Figure 10: Distribution of basic needs indicators and household income, 1996 by rural quintile 100 N Chtildren 0 #0 - out-of-school / ~J 80 X 60 - Overcrwin '60 *'-; 40 4 . / //~4. - \, Hosell, 416 20 / sS t t. income f 20- D: ae 2 .- ' aitation C~~~ a Quintiles Note: See notes and footnote for Table x. Source: Encuesta Rural (1996) * Rural areas in the poorest regions of El Salvador, Central II and Eastern, have the worst coverage of basic infrastructure Consistently, rural areas in the poorest regions of El Salvador, Central 11 and Eastern, have the worst coverage of basic infrastructure services (Table 6). Only about half of the rural residents in these regions have access to electricity, less than one-tenth of them have access to modem sanitation and only a fifth of them have access to piped water. In addition, the average distance to paved roads is highest in the Eastern region, more than 9 km, and second highest in the Central II region, more than 6 km. Rural areas 12 Annex 5: Basic Services in general in El Salvador urgently need improved infrastructure, but particularly those hit hardest by the civil war are lagging considerably behind. Table 6: Basic needs by rural region, 1996 (%)a Rural Region 13asic Needs Indicator Western Central I Central u1 Eastern Rural San Salvador Children 7-10 out-of-school 25 11 16 25 22 Overcrowding 55 67 66 66 58 Acoes to piped water 30 31 22 21 27 Access to modem sanitation 14 6 8 6 11 Access to electricity 56 63 48 50 65 Average distance to paved road (km) 4.5 6.1 6.4 9.2 3.1 Notes:(s) Overcrowding occurs in households with more than three people per bedroom. (b) Piped water is either inside or outside the home or piped to a common faucet. (c) Modem sanitation is prvate or shared toilet connected either to the public sewerage systm or to a septic tank. *Children 7-10 who are out-of-school is measured at the individual level, the rest of the basic needs indicators re measured for individuals living in households with the characteristic. Surce: Encuesta Rural (1996) V. In-Kind Transfers T Targeting of publiefood programs to the poor is almost non-existent Food programs at public schools, at public primary schools and at public health facilities need to improve on their targeting to the poor. Leakages to non-poor households prevail in every area of residence and in every public food program, the proportion of non-poor recipients is more than half the proportion of poor recipients (Table 7). The proportions of poor and non-poor families receiving food transfers at public primary schools in rural areas are slightly higher than in other areas of El Salvador, but targeting to the poor is almost non-existent. Public food transfers at public primary schools is not a universal program either. Thus, improved targeting of public food programs, without increasing current transfers, implies correcting for undercoverage of poor households and cutting down on leakage to non-poor families. Table 7: Food recipients by type of public institution, 1994, for poor and non-poor by area of residence (%) Area of Food recipients residence At public schools At public primary schools At public health facilities Poor Non-poor Poor Non-poor Poor Non-poor Rural 40 32 43 36 20 12 Otherurban 11 5 14 14 15 11 San Salvador 26 11 31 16 0. 6 El Salvador 32 18 36 23 17 10 Source: EHPM (1994-111) Note: n' 10 households * Almost half of all in-kind transfers to rural areas are infood and of these more than half go through the schoolfeeding program, but this program is not targeted to the poor Almost half of all in-kind transfers to rural areas are in food and of these more than half go through the school feeding program (Figure 1I). However, poor households receive a smaller share of these transfers than non-poor households. Less than one-fourth of all beneficiaries are from the poorest two quintiles (Figure 12). The school feeding program is by large a primary-school level program, but the inequality in enrollments at this educational level does not explain the inequality in the distribution of food transfers I l. School enrollments are strongly pro-poor at the primary level (Figure 4). Thus, improved targeting of the 11 The distribution of households benefitting from the school feeding program is: 81.8% have primary school students, 6.8% have pre-school students and 11.3% are in neither groups. Annex 5: Basic Services 13 school feeding program in rural El Salvador implies either creating a system of means tests or targeting schools in poor areas. Figure 11: Food transfers by type of institution/program in rural El Salvador, 1996 (% of total by type of institution/program) SCiool feeding _ r a m~~~~poga Otder 12% FIS Private 31% Source: Table A.7 Figure 12: Distribution of transfers of the school feeding program, 1996 by rural quintile (% of total by quintile) 30 9z30 27 27 25 23 20 14 10 Quintiles Source: Table A.8 Less than one-tenth of beneficiaries of the school feeding program are in the Central II,the poorest region, but almost haylfare in the Eastern, the second poorest region T'he poorest region in El Salvador, the Central 11, receives less than one-tenth of the trnsfers from the school feeding programn. However, the second poorest region, the Eastemn, receives almost half of these transfers. Expanding and strengthening the rural school feeding program in the Central 11 region, particularly to benefit primary school children in poor communities, can improve its poverty focus. 14 Annex 5: Basic Services Figure 13: Distribution of transfers of the school feeding program, 1996, by rural region (% of total by rural region) Western 32 Central I |14 Central 11 9 Eastern 6 Rural S.S. 0 0 10 20 30 40 50 In-kind transfer (%/. of total) Source: Table A.9 VI. Policy Implications Education - More than one-third of the rural population cannot read and write. Regionally, the Eastern and the Western regions have the highest illiteracy rates. Thus, literacy programs need to be targeted to the rural population, particularly the poor in these regions. - The primary school system in El Salvador needs to improve internal efficiency by promoting early enrollment and decreasing repetition rates across the country, focusing particularly on the rural poor. Primary school students older than the appropriate age group are common across all income groups and regions. * Mean years of schooling in rural areas is the lowest in El Salvador. It is crucial to put in place incentives for retention in rural primary schools. In particular, poor girls are more likely to drop out of school as they get older. Thus, retention incentives at the primary school level should target the poorest girls. * EDUCO schools reach the poor but not the poorest rural children. Improved targeting of the EDUCO program implies correcting undercoverage of poor children and simultaneously cutting down leakage to non-poor children. * Less than one-fourth of EDUCO children are in the two poorest regions. Targeting these regions can improve the poverty focus of the EDUCO program, but it will also be necessary to geographically identify poor areas within these provinces. * One in every three poor primary school-aged children is out-of-school in rural areas. Almost half of rural poor primary school-aged children who are out-of-school report that school expenses are too high. An alternative for reducing the burden of direct costs of primary education on families with primary school-aged children, targeted particularly to the poor, can be to provide school materials to and to ease requirements on wearing uniforms in disadvantaged rural areas of the country. Health * Public health services are not targeted to the poor or to disadvantaged areas. Rural residents have the lowest utilization rates of public health services and the highest illness rates. In addition this Annex 5: Basic Services 15 population group is severely undercoveraged by modem health services. Targeting mechanisms of public health services towards the poor and rural areas need to be put in place, particularly for the residents of the Central II and the Eastern regions. * Poor rural children are three times more likely to be ill than non-poor children. High illness rates are compounded by lack of targeting of public health facilities and by the scarcity of health care financing mechanisms. There is an urgent need in rural El Salvador to increase the access to affordable health care services for poor children. Basic Needs * The rural poor in El Salvador have the worst basic needs indicators. Non-poor rural families fare considerably better than poor rural families in keeping their primary-school age children enrolled and living in non-overcrowded homes, but all rural families are severely underserved by basic infrastructure services. Regionally, the poorest regions, rural Central II and rural Eastern, have the worst coverage of basic infrastructure. Rural areas in general in El Salvador urgently need improved infrastructure, but particularly those hit hardest by the civil war are lagging considerably behind. In-Kind Transfers * Targeting of food programs at public schools, at public primary schools and at public health facilities is almost non-existent. Public food transfers at public primary schools is not a universal program either. Thus, improved targeting of public food programs, without increasing current transfers, implies correcting for undercoverage of poor households and cutting down on leakage to non-poor families. * Almost half of all in-kind transfers to rural areas are in food and of these more than half go through the school feeding prograrn, but this program is not targeted to the poor. Less than one-fourth of all beneficiaries are among the poor. Thus, improved targeting of the school feeding program in rural El Salvador implies either creating a system of means tests or targeting schools in poor areas. * The poorest region in El Salvador, Central II, receives less than one-tenth of the transfers from the school feeding program. Expanding and strengthening the program in the Central 11 region, particularly to benefit primary school children in poor communities, can improve its poverty focus. 16 Annex 5: Basic Services REFERENCES Encuesta de Hogares de Propositos Multiples (EHPM). 1994-111. Departamento de Investigaciones Muestrales. Ministerio de Planificacion y Coordinacion del Desarollo Economico y Social. Republica de El Salvador. Encuesta Rural. 1996. Fundacion Salvadorena para el Desarrollo Economico y Social (FUSADES). Republica de El Salvador. Lanjouw, P. 1996. Towards a Poverty Profile for El Salvador: Preliminary Results from the 1994 Encuesta de Hogares. Policy Research Department, The World Bank (mimeo). Lockheed, M., A. Verspoor and associates. 1991. Improving Primary Education in Developing Countries. New York: Oxford University Press. Published for The World Bank. Lopez, R. 1996. Rural Poverty in El Salvador: A Quantitative Analysis. Department of Agricultural and Resource Economics, University of Maryland at College Park (mimeo). The World Bank. 1994 El Salvador: The Challenge of Poverty Alleviation. Report No. 12315-ES. Country Department II, Human Resources Operations Division, Latin America and the Caribbean Regional Office. Washington, D.C. Annex 5: Basic Services 17 APPENDIX Table A.1: Illiteracy rate for poor and non-poor, 1994 by area of residence (%) Area of residence Poor Non-poor All Rural 41 26 35 Other Urban 28 14 17 San Salvador 24 8 9 Total 37 15 23 Source: EHPM (1994-111) Note: Population older than 10 years of age who cannot read and write Table A.2: Illiteracy rate for poor and non-poor, 1994 by region (%) Region Poor Non-poor All Western 35 16 24 Central 1 35 17 25 Central 2 35 22 29 Eastem 4f 23 33 San Salvador 24 8 9 El Salvador 37 15 23 Source: EHPM (1994-l1l) Table A.3: Type of primary school attended,1996 by rural quintile (%) Rural quintile Public EDUCO Private Other Poorest 88 8 2 1 11 80 16 4 0 III 88 1 2 1 0 IV 86 10 4 0 Richest 92 2 4 2 All Rural 86 10 3 1 Source: Encuesta Rural (1996) Note: Private includes private schools as well as community and religious schools Table A.4: Type of primary school attended, 1996, by rural region (%) Type of Rural Region school Western Central 1 Central 2 Eastern San Salvador All Rural Public 85 81 86 90 89 86 EDUCO 10 18 11 7 0 10 Private 6 1 2 2 7 3 Other 0 0 1 1 4 Source: Encuesta Rural (1996) Note: Private Includes private schools as well as community and religious schools 18 Annex 5: Basic Services Figure A.1: Rural regional poverty shares for school-age children, 1996 (% of poor children 7-10 years old in rural El Salvador by region) San Salh'dor 6% Western Eastern 30% Central I 17% Central 11 29% - _ _ _ Source: Encuesta Rural, 1996 Table A.5: Poor and non-poor children 7-10 years old who are out-of-school, 1994 by area of residence (%) Area of residence Poor Non-poor All Rural 32 15 27 Other Urban 21 7 10 San Salvador 15 7 8 El Salvador 28 8 18 Source: EHPM (1994-lli) Table A.6: Illness rate and place of consultation, 1994, by region (%) Area of Illness Place of consultation residence rate None Private Public Other All places Western poor 31 10 10 18 62 100 non-poor 32 6 20 22 53 100 all 32 8 15 20 57 100 Central I poor 29 11 4 22 63 100 non-poor 33 5 16 25 55 100 all 31 8 10 24 58 100 Central II poor 39 10 4 24 62 100 non-poor 38 5 15 24 57 100 all 39 7 9 24 60 100 Eastern poor 31 11 7 25 57 100 non-poor 34 5 20 25 51 100 all 33 8 14 25 54 100 San Salvador poor 26 5 5 30 61 100 non-poor 25 2 26 30 42 100 all 25 2 24 30 43 100 Source: EHPM (1994-Ill) Notes: (a) Public = public health centers and public hospitals; private = private hospital/clinic, private doctor and nurse; and, other = pharmacy, healer and self-medication. (b) Illness rate is for last 30 days. Annex 5: Basic Services 19 Table A.7: Transfers in-kind by type of institution in rural El Salvador, 1996, (% by type of institution) Type of Type of assistance institution Food Clothing Health Education Housing Construction Other Private 31 54 72 69 47 30 53 Public 6 0 0 8 0 20 7 FIS 0 0 10 15 7 30 7 School feeding program 51 15 9 8 0 0 0 Other 12 31 9 0 46 20 33 Total 100 100 100 100 100 100 100 Source: Encuesta Rural (1996) Table A.8: Distribution of in-kind transfers, 1996, by rural quintile (% by quintile) Rural Type of transfer quintile Private Public FIS School feeding Other All program transfers Poorest 13 25 0 14 27 16 11 17 0 22 9 31 16 III 15 25 33 23 12 18 IV 24 38 33 27 8 23 Richest 32 13 11 27 23 26 Total 100 100 100 100 100 100 Source: Encuesta Rural (1996) Note: n < 10 households Table A.9: Distribution of in-kind transfers, 1996, by rural region (% by rural region) Rural Type of Transfer region Private Public FIS School feeding Other All transfers program Westem 30 0 33 32 23 28 Central 1 32 25 22 14 15 22 Central 2 13 38 0 9 15 13 Eastem 19 13 33 46 35 31 San Salvador 7 25 11 0 12 7 Total 100 100 100 100 100 100 Source: Encuesta Rural (1996) Note: n < 10 households Table A.10: Rural population by region, 1996, (%) Rural region Poor Non-poor Total Western 8.4 16.3 24.7 Central 1 5.8 10.7 16.5 Central 2 10.5 9.0 19.5 Eastern 13.1 17.7 30.9 San Salvador 2.1 6.4 8.4 All Rural 39.9 60.1 100.0 Source: Encuesta Rural, 1996 20 Annex 5: Basic Services Figure A.2: Rural regional poverty shares, 1996 (% of poor in rural El Salvador) San Salvador 5% Western Eastern 33% Central I VI ~~~15% Central lJ 26% Source: Annex Table A.1 (Encuesta Rural, 1996) Table A.11: Rural population 7 to 10 years old by region, 1996, (%) Rural region Poor Non-poor Total Western 8.6 14.3 22.9 Central 1 8.2 10.5 18.7 Central 2 14.5 8.0 22.5 Eastem 14.8 12.7 27.5 San Salvador 2.7 5.7 8.4 All Rural 48.8 51.2 100.0 Source: Encuesta Rural, 1996 Table A.12: Share of in-kind transfers to household income by rural quintile (% share) Rural quintile Share of income Poorest 10.3 11 3.9 III 5.2 IV 2.9 Richest 5.8 All Rural 5.4 Source: Encuesta Rural (1996) Table A.13: Share of in-kind transfers to household income, 1996, by rural region (% share) Rural region Share of income Western 4.7 Central 1 5.7 Central 2 2.6 Eastern 5.9 San Salvador 10.4 All Rural 5.4 Source: Encuesta Rural (1996) Annex 5: Basic Services 21 Table A.14: Population San Salvador Other Urban Rural Total Poor 2.0 8.8 29.1 39.9 Nonpoor 23.3 21.0 15.8 60.1 All 25.3 29.8 44.9 100.0 Source: EHPM (1994-111) Table A.15: Population by region Western Centrall Central2 Eastern San Salvador Total Poor 10.0 11.6 5.1 11.2 2.0 39.9 Nonpoor 10.7 11.6 4.5 10.0 23.3 60.1 All 20.7 23.2 9.6 21.2 25.3 100.0 Source: EHPM (1994-111) Table A.16: Population 7 to 10 years old San Salvador Other Urban Rural Total Poor 2.5 10.2 36.1 48.8 Nonpoor 18.5 18.0 14.7 51.2 All 21.0 28.2 50.8 100.0 Source: EHPM (1994-111) Table A.17: Population 7 to 10 years old by region Westem Centrall Central2 Eastern San Salvador Total Poor 12.0 14.2 5.9 14.1 2.5 48.8 Nonpoor 8.6 10.2 3.9 10.0 18.5 51.2 All 20.6 24.4 9.8 24.1 21.0 100.0 Source: EHPM (1994-111) Table A.18: Rural population by quintile and gender Quintile Male Female Total 1 9.9 10.0 19.8 2 10.3 9.8 20.1 3 10.5 9.6 20.1 4 10.1 9.9 20.0 5 10.4 9.6 20.0 All 51.2 48.8 100.0 Source: Encuesta Rural (1996) Table A.19: Rural population by region and gender Region Male Female Total Westem 12.8 11.9 24.7 Central 1 8.4 8.1 16.5 Central 2 9.8 9.7 19.5 Eastem 16.2 14.7 30.9 San Salvador 4.0 4.4 8.4 All 51.2 48.8 100.0 Source: Encuesta Rural (1996) 22 Annex 5: Basic Services Table A.20: Rural population by region Region Poor Non-poor Total Westem 8.4 16.3 24.7 Central 1 5.8 10.7 16.5 Central 2 10.5 9.0 19.5 Eastem 13.1 17.7 30.9 San Salvador 2.1 6.4 8.4 All 39.9 60.1 100.0 Source: Encuesta Rural (1996) Table A.21: Rural population 7-10 years old by quintile and gender Quintile Male Female Total 1 9.8 11.7 21.5 2 14.1 13.1 27.2 3 10.9 9.4 20.3 4 10.3 9.6 19.9 5 5.7 5.3 11.1 All 50.8 49.2 100.0 Source: Encuesta Rural (1996) Table A.22: Rural population 7-10 years old by region and gender Region Male Female Total Westem 11.1 11.9 23.0 Central 1 9.6 9.0 18.6 Central 2 11.3 11.3 22.5 Eastem 15.4 12.1 27.5 San Salvador 3.5 4.9 8.4 All 50.8 49.2 100.0 Source: Encucsta Rural (1996) RURAL LAND MARKETS' Introduction 1. Traditionally, problems concerning rural poverty in El Salvador have focused on land issues-land scarcity, tenure insecurity, and low productivity. Government-sponsored land distribution programs in the 1980s, and later under the 1992 Peace Accords, did not achieve their poverty reduction objectives and created considerable distortions in land markets. As a result, land has not been allocated efficiently neither for agricultural use, nor for housing or urbanization. How can land markets contribute to reduce rural poverty, environmental degradation, and lack of agricultural competitiveness? This will require recognizing more explicitly land's multiple functions and uses, and the need for a flexible land market to improve land allocation into these competing functions and uses. Land needs to be viewed not only as generating income from agriculture but also for rentals, a more dynamic use as a store of value for household savings (collateralization) and shelter, and as an input which can have productive uses well beyond agriculture. 2. Land transfers under the agrarian reform programs, have created a group-the agrarian reform beneficiaries (including those from the Peace Accord-mandated land transfer program, PTT)-which hold land under special status (i.e., agrarian reform tenure which includes collective, joint and individual ownership, with characteristic strategic default2 and association incentives. These groups suffer from special market constraints, often restricting land flows to the landless, which the reforms was meant to help. This study analyzes briefly-as a special case-the agrarian reform cooperatives, as opposed to the individual beneficiaries, because cooperativization under collective ownership constituted a key element of the 1980s agrarian reform-led development strategy. These cooperatives have benefited from more land per capita, better land, and more credit, and now hold the equivalent of 20 O/o3 of the best agricultural land in the country, and leave 6% of it idle. 3. Section A describes how changes in land tenure patterns in the last century led to dual land markets, and the impact of Government interventions. Section B reviews and updates policy changes being implemented by the Government to create a more efficient land market in El Salvador. Section C summarizes findings and recommendations on improving land markets with the objective of: (i) reducing rural poverty; (ii) developing a dynamic and flexible land market to enhance the competitiveness of the sector; and (iii) creating incentives to reduce resource degradation. The paper does not address regional or municipal planning, although it should be given some consideration. 4. Attachments present theoretical underpinnings and literature reviews on related land market issues. Appendix A discusses the conceptual framework of common property vs. private tenure regimes. Appendix B reviews the literature on urban vs. rural land allocation, and the role of government in the land market due to market failures. Appendix C presents agrarian reforms experiences. Appendix D presents the legal regime differences between agrarian reform and non-agrarian reform land. l This Annex was written by Cora Shaw (LCSES). Comments were received from Alberto Valdes, Ian Bannon and Louise Cord. FUSADES and MAG assisted in gathering information. 2Strategic default is when a borrower, even able to repay, decides not to do so because of lack of consequences, e.g., repossession or denial of further credit. The expected value of potential default losses is less than the debt value. FUSADES, Diagnostico del Sector Agropecuario, 1996. The 20% refers to cooperative actual holding of soil classes I- IV land. When taking into account all land held by cooperatives, they hold 30% equivalent of land under soil classes I-IV. 2 Annex 6: Rural Land Markets A. Two Land Markets 5. As discussed in the review of Salvadoran land distribution experience (Appendices C and D), land markets interventions over the past 16 years in El Salvador were comprehensive (about 50 % of land under crops), and created a reformed sector which has a major impact on overall agricultural productivity, and on the dynamics of the land market. During the agrarian reforrn process, the Government placed restrictions on land transactions for reform beneficiaries. The principal reason was that supporters feared that distributed land, if allowed to be sold, would eventually be reconcentrated in the hands of large land- owners, through force or coercion. Another reason was that agrarian reform beneficiaries would sell the land for consumption purposes. Therefore, each of the resulting tenure regimes in the reformed sector had limitations on sale, rental, or parcelization (these restrictions are sumarized in Appendix D). These restrictions have created a segmented land market-a severely restricted one confined to agricultural land, irrespective of soil potential, and a largely unregulated one generating chaotic urbanization patterns. Table 1: Agrarian Reform vs. Non-Reform Land (ha) Beneficiaries Land Market Segment Hectares (Families) or Owners Ha per Beneficiary Total Soil Classes I-VIII 2,045,900 I. Land Distribution A. Phase I 215,000 37,000 5.81 B. Phase III 80,000 47,000 1.7 Subtotal Agranan Reform 295,000 84,000 C. Peace Accords (Programa de Transferencia de 78,000 30,000 2.6 Tierras) Sub-total Land distribution 373,000 104,000 II. Non-Reform Market 1,227,157 N/A Ill. Total 1+11 1,600,157 IV. Unaccounted for (Water bodies, mountains tops, 444,843 roads, cities) V. Total 2,,045,000 of which (partial): Renters/ Sharecrop 36,372 Land Under Cuftivation 736,400 Land under soil classes l-IV 696,700 Urban encroachment since 1974 (estimated at 0.8% of 86,000 territory in 1974, and 5% in 1995) Sources: Based on Thiesenhusen (1993), MAG/GOPA (1996), MAG (1989), and Appendix 1. 6. Intervention in El Salvador was through policies and practices that affected both land agrarian and non-reform market sectors, although differently. This section will discuss issues that affected both segments differently, i.e., access to land, debt, land security, agrarian institutions, cooperative management and access to social services, and other developments in land prices, urbanization, and availability of long- term financing. Not all interventions are unjustified, however. Land markets cannot be totally liberalized due to public good concerns, including the environment, irreversibility of rural-to-urban land conversion, protection of critical land or water sites. In El Salvador, however, these public goods regulations mostly exist but are not enforced. (1) Access to Land and Rural Poverty 7. Clearly, agrarian reform in El Salvador is a controversial socio-economic and political issue, a subject which for many transcends a strictly economic analysis. Furthermore, for El Salvador today, with a substantial 4From Appendix C, Table 1 and current CNRIUE urban/total area estimates. In the Western part of El Salvador, the urban to total area ratios are: Sonsonate 48/1239 (5%), San Salvador 156.4/1042 (15%), La Libertad 5711600 (4%); and Ahuachapan 41.6/1239 (3%). Annex 6: Rural Land Markets 3 history of land reform policy in the recent past, the link between long-term rural growth and land tenure structure is a question which cannot be elucidated with rigor based on the available analysis. However, in a modest way, the analysis undertaken under the Alleviating Rural Poverty Chapter (VII) of the Main report and Annexes 3 and 4, provide a relevant reference to identify some fundamental parameters. Box. .I I ,Oove'rnment intervention Justified in the Land Market? 'Th -rationaIe for government intervention rests on market failures. However, regulations to internalize the costs of mmu,r,eue,d private decisions depend on both the adequate existence of markets for negotiating contracts and government capacity. . enforce, hese regulations. Food eurity. So,me countrie a cernedo eternal vulnerability in eth event of food scarcity, d onver .'riltural tio,:urban use'may'negatively affect ,fd supply. However, market liberalization as-result nd ,c,,eapr .a.vailability of food. Exter'nlities.Ub ....growth can provide rationalefor intervention in: nd m,arkets, particularly i tpen externalities addressed are, pollution.and noise imposed by urban dwellers.on rural dwellers;or .chemical pollution, .odr,'or c s om farm activitie to urban'dw.ellrs.. Positive externalities:inciude o'penspace associated with rura activitie. 'Ennronment. Expansion of cities into prime agricultural land (mostoften the case), may lead to use of less q ' .id ivolv.ig deforestation and soil erosion due to.more intensive cultivation or expansion of the. agricultural frontier. Asymmetrnc lnformation.. Because speculation distorts land. allocation and is detrimental to the functioning of land marke, itcan wors,en land distribution. This affects mostly peri-urban areas, where land prices rise the fastest. There are sound `re6asons fritervention in this case, When- groups can gain from privileged access to information on public investments or ther .c.hanges that imay suddenly affect land prices in a given area. But the impermanence syndrome is not the result-of pecla'tion; ratherj, the expected rental value of land increases due to expansion of the urban sector, as higher value- uses (e{p.g suburban housing.m.ay-compete with agricultural use. irrteesibili't.: It is toocostly. to revet. urban land to rural use, therefore. rurA :to urban use must be carefully planned.. r .s.'o,.,meho onstrained toavoid.errors in conversion. - si ' fn'tio, of land". any. agrarian refonns were based on the belief that land is not a commodity, but is invested with a.s.o.a :function'tat 'ttranscends production of goods:and services. This agrarian school of .thought was mostly based on 'wester ideooges, -and did,not-address the-rituistidreligious underpinnings of land's role-found in indigenous cosmologies, nthi.s,sc,hoolof ,thought, land's "social" function was to provide each.citizen with a means of subsistence; therefore,land distribution.was equivalent to distribution of social justice. 8. On the one hand, access to land appears to be one of the most important determninants of household income. El Salvador has a relatively high laborAand ratio, higher than most of Latin America, partly the result of being a land scarce country. However, findings in Chapter VII suggest that: * The least poor in rural areas are those households which are heavily involved in nonfarm employment, and thus a route out of poverty is to increase access of the poor to non-farm activities; * Among the poorest households, farmers and landless agricultural workers are equally poor. On average, landless rural workers in agriculture are the poorest segment of the rural population. In general, farmers are better off than landless workers. However, the percentage of farmers in extreme poverty is quite similar to that of landless, at 27% and 30% respectively. Adjusting household income for adult equivalence, the incidence of extreme poverty among landless agricultural workers is lower than that of farmers. * To surpass the poverty line, farm size would have to be expanded from 2 ha to about 12.5 ha, or from 2 ha to 5 ha to reach the extreme poverty line. Thus, more land certainly raises household income, but a scale (average) of less than 12 ha is unlikely to lift the typical household above the poverty line. Similarly with landless workers in agriculture, these parameters provide a rough order of magnitude of the land resource needs to lift them out of poverty. From Bhadra and Salazar Brandao, Urbanization, Agricultural Development and Land Allocation, World Bank Discussion Paper # 201, Washington, 1993. 4 Annex 6: Rural Land Markets 9. El Salvador has an availability of only 1.1 ha of Table 2: Members in Agrarian Refonn arable land6 (Soil Class I-IV) per person actively Cooperatives (Direct Beneficiaries) engaged in agriculture. The total number of land Survey Year No. of Associates 1980/81 31,183 distribution beneficiaries varies due to changes in status 1981/82 22,194 overtime, and classification categories and 1982/83 26,205 1983/84 28,410 methodologies. It is estimated that in 1992, Phase I 1984/85 27,436 beneficiaries reached 37,000 families, Phase III 47,000,7 1985/86 27,174 and PTT 30,0008 (1996). The total number of the 1986/87 30,268 and PT1r000 196.Tettl ubro h 1987/88 34,149 agricultural labor force is estimated at 640,000, of which 1988/89 36,558 40% are landless, land poor, seasonal workers or 1989/90 33,096 are ~~~~~~~ ~ ~~~~~~~~~~~~1990/91 32,380 unemployed.9 1991/92 31,137 1992/93 no census 1993/94 28,893 10. Therefore, considering the scarcity of farm land Souce: MAG/PERA. in El Salvador, these estimates underline the unreasonableness of relying primarily on land redistribution to alleviate poverty among the rural poor, and the importance of non-land factors (i.e., education, access to infrastructure and credit, technology, economic incentives) in the poverty equation. (2) Debt Overhang 11. Both land market sectors were affected by the conflict, and incurred heavy losses because of it. The agrarian reforn cooperatives did not reimburse their debt. The commercial sector most affected by war and technology failures (mostly the cotton growers in the Eastern part of the country) incurred debts as well. In May 1996, the Government condoned 70% of commercial and agrarian debt for prompt payment. Repossession remains an issue, and many banks will not lend against small farmers' land because of the political and administrative barriers to repossession. 12. International experience warns that agricultural sector retrenchment and accompanying debt are a fact of adjustment. Where country policies eliminate distortions that promoted land hoarding, oversized farms will shed land to more efficient farmers. In New Zealand, for example, the Government's role was to resist calls for further protection and help instead ease out farmers through streamlined exit procedures: repossession by banks, and "exit bonuses" for small farms. 6 Area under Soil Classes I-IV is 696,700 ha, under Soil Classes I-VI is 933,600 ha (1.46 ha per agricultural worker), and area under cultivation is 746,400 ha. 7 Thiesenhusen (1993). The number for Phase I beneficiaries is higher than the PERA number in the table because it includes about 6,000 beneficiaries of colonization programs prior to 1980. 8 Hevia, H. , MAG/GOPA Consultant Report, Dec. 1996. 9Seligson et.al. (1993)EI SalvadorAgricultural Policy Analysis. Land Tenure Study. USAID Report, September 1993. Annex 6: Rural Land Markets 5 13. In agrarian reforn debt, Table 3. El Salvador: Source and Status of Farm Debt maturities and interest rates for Ag. Reform all debt differ, even within Land Debt Reprogrammed Conmaercial Observations For instance, Worldng Capital Debt categories. For istance, Debt agrarian reform beneficiaries' Agrarian From Piase I, ISTA-BFA Commercial Debt of up to debt for land was mostly at 30 Reform 11 and III or Public ol6,665 was Cooperatives Fronm P1T Banks forgivem. years, 6%, with 4 years of and other Agrarian waiver (dispensa).13 For some Beneficiaries Reforn coops allowed to cooperatives, however, the auction land to interest rate was 9%. Since few pay reminder pad t e c itra o ..................................for...........3............. 3 '~ paid, the criteria f oters - - ------------ Commnercial differentiation was not an issue, or Public and remains unclear. By 1994, ... Decree ja Banks only 1% (in real terms) of ISTA Condonnation No. 699" 699" Decree iaw debt and 5% of FINATA debt with Prompt No.698'2 had been repaid. At the Payment beginning of 1996, the Land Bank had not even calculated the repayment schedules of its beneficiaries, nor was clear on the terms of the loan.14 ISTA debt became repayable only when agrarian reforn cooperatives received titles, in many cases after the departure of the co-gestor.15 14. Commercial farmers incurred commercial debts during war and later as profitability declined due to removal of protection in the late 1980s and early 1990s. Decrees 292 and later 698 were approved by the Legislature to provide relief to these farmners. The largest farmn debts are concentrated in the ex-cotton sector in the Eastern part of the country. With the advent of pest resistance and lack of technology, cotton growers' cost became untenable, in addition to the environmental cost of high pesticide applications."6 The elimination of COPAL's monopsony power and liberalization of the cotton market, plus the cost issue literally eliminated cotton production. No cotton has been produced there for five years. 15. In May 1996 the Legislature approved Decrees 698 and 699 condoning much of agrarian and commercial debt. They reduce 70% of debt value for prompt payment, and foregoes the last p5,000. Essentially, if debt is up to 021,665, it is forgiven. A key element of the legislative package also includes Decree 719,17 which allows cooperatives to sell part of their land to pay the remaining 30%, with the only proviso that they would have to have at least a land/member ratio of 7 ha. This was included to avoid rewarding cooperative members where exodus was large, although this ratio was three times the maximum allocation given to individual beneficiaries by parcelizing cooperatives. As stated above, non-agrarian reform land has no such restriction on sales. These decrees allow land under both land market segments to eliminate of debt and become available to the land market. IOLey del Regimen Especial de la Tierra en Propiedad de las Asociaciones Cooperativas Comunales y Comunitarias Campesinas y Beneficiarios de la Refofma Agraria, later extended to PTT beneficiaries. Ley de Reestructuracion de la Deuda Agraria, approved May 9, 1996. 12 Ley de Apoyo a la Reactivacion del Sector Agropecuario, approved May 9, 1996. 3 waiver means no payment is required for either principal or interest. Grace means only interest payments are due, no principal. According to the titles, debtors were to obtain ISTA's accord for dispenisa, but in practice most do not pay. For instance, the titles mentioned 4 year grace period instead of dispensa, although the Peace Accords stipulated "Agrarian Reform terms". 5 e.g., in La Carrera, debt was due as of 1994: 14 years after the farm was intervened and transferred to the cooperative management, and that an ISTA co-gestor was installed. 16 Anecdotal evidence suggests as many as 58 applications. 7 Ley del Regimen Especial de la Tierra en Propiedad de las Asociaciones Cooperativas Comunales y Comunitarias Campesinas y Beneficiarias de la Reforma Agraria, approved May 30, 1996. 6 Annex 6. Rural Land Markets Boi 2: Fan Debt Overhang and dSecor :Cisis: :Howdid New Zealand Handle It?. X .... The ai rs thIayc dIIevelopin odeveloped, toqdebt-rlte agriultural crises are sector polwcies and farmer perspectives.: Ths box explores ithe experience:of agncuitural sectors of New Zeland, te circumaepolicies ad ..rket p s Whichled -hem to sucht riss, andthe adjustments the coupynyato- recov er ... :Beore984d the stateh thtraditoaly endirectlyinvolved wit thepolicy ;decisions of:fa sheepanbeef.Government etablishedaban ofprotec tion tforf, arme, Mc ude set prices (SMPs) and other subsidies like. m eting ad ' credit assistance, as as a vaiety tax brks.hi 1984 as Xthe nwyelet: L came to power with a 'moremariket" platform, and quickly removed state :susidi t : ceaiwnsectors of thie economy.i For:anagricultural: sector sed&to State protection, the effects of Govement'slibel. econmicpoliciesintheshort-ermwdinp heigned excagerate, increasd inrest ra d an: .in.ailitto competeithim oports. : -d . e, Thcb frmrfs' respo,.nses tto A Gov.ernment policy and the country's new economic realities occ.urredquicy,d hey'drastical.yhangedthefiagricultual setor.Th most noticeableresponses were: a considerabledrp in br modificatios in, arm, :.management, including a1decrease in wage. labor. (in favor offamily labr) and.. apiy ei capital nd inputs,' 'lfarm owners souht' to decrease 'expnditures vernment facilitated t the insectrtroutheNewStartGrantPro ,whichgavefamers intenblesituations (malywi e' Gin e due to.-dclimate-iducdlcopfailue) tehance-to escapetheir bank debt and eiet' teagicutua setr Te.nval 'are'. lan.,d.'' was. sod' andthus trnsere toaets who coud- prouc moeefcety adhdacs.o rdt ti Xanr h privat tagricltual bank wiSth:-bad prfolios did: not necessar I:need to -go under those .amr in un,vible buiess.iru.taces coul .Xescape thi e:eindividual 0debt: crises,: and- the size and -compegtitiveness of :New Zealad.'s. agrcaultures,ector bec aime 'morearationalized. ..' Zea..la,i,,ad.', fa fod that.the: wayout of thae crisis, althoug diffiTcul, as to "rigtsize" their fars, becme oetitive, ior elge get ,out o-fthedbusiness.f ' ;di:: :E 16. Decrees 698 and 699 provide up to June 30, 1997 for beneficiaries to apply for relief. Agrarian reform institutions have been slow in calculating balances so that paymnent can be effected. An extension of the deadline will probably have to be considered. (3) Lack of Land Tenure Security 17. Lack of tenure in security is considered a land market imperfection. The effects of land titling and registration (registration publishes land rights and provides conveyances public faith before third parties) is likely to enhance farmers productivity and income by providing incentives to invest, notably in attached investments, through credit if land is used as collateral, and also improves land market fluidity by reducing the transaction costs of land conveyances or long-term rentals (short-term rentals rarely have written registrable contracts). Another reason for land insecurity has been the perverse effects of the 1979 Rental Law (D.L. 207), which enabled land transfers to renters-essentially colonos who clained renter status. Although the results of this land transfer were positive for its 47,000 beneficiaries, it was not so for the colonos who were expelled before FINATA could act. A lingering effect of this is the insecurity of rentals for unregistered owners who may fear not being able to recover their land if rented, particularly older fanners or single women without adult children who cannot till the land. is. In El Salvador, land security is hamnpered by inefficiencies in land adjudication and registration and landowners' perceptions. Land security means different things to different people, who adjust their investment decisions based on these perceptions. 19. Land security means recognition by society of bona fide ownership. The point is that, while title registration measures are needed, these do not replace society's perception of wrongly acquired land. Measures forestalling land invasions, respect for long and short-term rental agreemen and authorizing agrarian reform beneficiaries (6% of cooperative land is idle) to rent land openly, must complement land securitization . Annex 6: Rural Land Markets 7 20. Salvador's land registry and Table 4: Tenure Security Implications for Each Land Market Segment Concept Amrarian Reform Free Market cadastral system suffers from out-of- tenure loss of land, reversion of land invasions date land records. Land records insecu- agrarian reform risk to investents in (registry and cadastre) are updated by rity capital, iition, the National Registry Center (CNR). crn-ierr'create impression of normal written Only about 40% of El Salvador's rentals idle/available land in the contracting procedures cooperative, that it has surplus adequate properties are registered, and of those, land less than a third have cadastral ...short-.term some rentals.allowt.opaynd.ebIts nomatl so'rnai coontract's........... references. Many..properties with validrentals to input suppliers, adequate, but for very references. Many properties with valid equivalent to bartering land use short periods (ofen one registration cannot be located. against inputs. crop season) Property owners with title certificates that are not registered cannot use their titles as security for mortgages or other collateral. Some properties are registered by municipalities, but this practice does not guarantee their legal rights before third parties. Registration is not obligatory in most of the country, and despite the legal security it provides, substantial transaction costs discouraged land registration. 21. Deficient land records affect everybody, but are particularly notable for Government-adjudicated land. Beneficiaries got their title only after debt was paid. The agrarian reform cooperatives got their provisional title only after co-gestion, with 30-year repayment period starting then. Therefore, there was no incentive to push for title. Their registration was also problematic because of deficient land records of previous owners. This was a particular problem for PTT adjudications because of agreed timetables for the implementation. 22. Many anomalies occur with ISTA titles. ISTA held the titles until co-gestion was over or debt was repaid (as for many other aspects of the Agrarian Reform, there were either no rules or no systematic application of rules). The titles were listed in the ISTA President's special volume, instead of the notaries' protocol as in private conveyances must be.'8 Once the titles were granted, the beneficiaries did not know they had to register the title in the Land Registry and were left on their own to do so. Another anomaly is that the titles are listed in a volume belonging to the President of ISTA, unlike the regular deeds that must be listed in the protocol of the issuing notary, protocol which is presented to the Supreme Court for inspection and quality control yearly, as per Decree 678.'9 23. The Government has progressed on land registry and cadastre services to support the transparency and lower transaction costs in the land market. The Bank-financed Land Administration Project supports the establishment of the National Registry Center and financing for a process to systematically regularize title and deed registration of the national territory. The three key features of this operation are the concept of a single land registry and cadastre institution, the obligation to register by notaries in regularized areas, and in-field adjudication with alternative dispute resolution mechanisms (Box 3). 18 Notaries must present their protocols to the Supreme Court yearly. 9 Ley Especial para la Transferencia e Inmuebles e Inscripciones de Titulos de Propiedad Expedidos por el ISTA, approved Nov. 14, 1991. An additional anomaly is that the notaries in the ISTA legal services charged twice the going rate for deeds (if an individual beneficiary from D.L. 747 sold to another), but was able to offer services on the spot and without giving the buyer and seller the opportunity of quality control of original title by a non-ISTA notary. 8 Annex 6: Rural Land Markets Ba 3 Lson fo Wrl an~'uorrted Lad Adminiistration.. Projects... - : r dp ' ii d t ' f e st4 rmland.1 d i i;td ............ t C:' it000l'|00t04'tt:' t': :ti0 (:0 : oehantw ecadet has been, inolved in land reform andland itenure issues under its pov enediua deelpen trtgy n 1975 th,e' Ban ulshdalndpfr apr vluto fprjcsinTaln ta-nd Kren'' and;Ea.a 0+'tt00 recent exerience suggeSt that.land-:reform programsare most cost-effectiveand contribute .to equity objectives a whenntheyare raide' 'nd cobinedwith agle-conflict-resolution mechnisms.The.project showed tha ladtitlin isf anipratse in i''crea,sin rralpsroutivit i areas wh ere (a) atn dresite arehigh; (bn fe is a it, ) itn systes ola'nd--allocation-anddidspute. adjudicaion have broken down; and (d) f:ormal credit systgems are willing and':able -to- lend- to -small ifarmers 0against the 0security iof land titles. fExcept for the credit :linkc, these: conditions daccurately describ2e thte isituation in EIl-Salvad.or. Thereis sigificant syng between titling and crdit,; an titling: lan cad edt poutvt inceasesonmbetteraccessto formal creditstV: e ',.'A 0r.eview of World -BankLe experenc withE rua Ititln project ithrugou teworld Lreporte tat, excet.for th Thalan Land TitingI. ad.-Hprojects',4* the12 opraion's sure,yedpefomd pooly.' The0 ' mai prblm were ':lac ofS poliica s,upport conlicin ,bureau,crati.c: prortis;lack tof intittonal capacty or .support; an,d .complex 0 multipl.e objetive o' which,.d ti..tlg. .wasnl a adjunct.:By con'tr ffateTlandprojectss were reltivyscssfld d to fu'll poitl disti suppor G'enmn ''''''t-tc itmenXt0 0 to '''" adqaerscs; singl t itlin ob'ctve an eaivl aorbelndpl eenvtironment Thie Emutual .cooperati:on between .the: Land ;Agency and thfe' Word :Bank 0must Walso: be considered. a Efaco of the suess soV fa. t nec a The Hank-supotd Sadr LndAmin istation projecti dr'as lessons-:fromthe be gin ofim a t piotpoetdrteArcl Secterrok eRef andAdinito ratndp-roje t,wch ove ed:- 5,800 n th oate D1eparmen mnicipalities of Santa Catarina; Mazahaut and Santo zDomingo Guzman in 1995. This first phase showed ithat te: regisratonpogram ,w,as high,ly-:p,op.ular,:ad nancd- w,hen adequate- local pubrlicity wvas given before visits by field tea,ms. SIt: also. demosrated thf,ef effiectiveness iof the imapper/notary: field verification: teams, :but showed up institutional weaknesses 0that: ledt,o-slo"wdata enry,:unaderscorng theneedtostreamine and computerize thefCNR. It also provided the first set of harddata .o'n-the" matude. of theproblem and target group. Basedondatta reomt othese dmunicipities, the! Goveret estimates tha most, about40%ofpropertiesnationwide arere Kstered,iand iesstliithan.atird i ae dca. stt tref nces; ofte uer: pacl, otwresal 7%o the 'rba wer beot 0 2ad9%o terrlwr eo h) hsdt if t poverty reduct.io.n. potentil gof theproposed project. Futermore, thepilot oexpe pdenceincols exmn a yield lowe costs pe hetaefor thie proposed projec thanrecn simiar exprinesin Eat .Asia.-:0 0;i:t:;X:00 :T:tE:00:it:000::00-:::0:f .registr recorc.:s. Once te CNR was included in t:he land: tran sfer- effort, itwo transitory derees. were pzrepaed to -allowV stealie procedures,for- adverse po-ssession an.d.other pre.ven-tive annotations to facilitate:registration. :i00i:t: : 24. One caveat in the "free" land market is that there are still niches where there is insufficient regulation for private low-income housing development. A glaring gap is the lack of protection of leasors or arrendatarios con promesa de yenta. Under this scheme, a developer (lotificador) enters into an agreement with a landowner (agrarian cooperatives are excluded) to "develop" his/her land. The developer then parcels the plot into small housing lots without services (average 250mn2), organizes a leasor group, and sells to this group typically at 10 years and low monthly payments (around ¢60-70). Because the contract between the developer and the landowner does not involve a real right's transaction, it does not constitute a lien, and need not be registered according to present legislation. Therefore, the landowner is not prevented to sell by a registered lien, or mortgage the land, creating opportunities for fraud. The Governmnent is aware of this gap in the law and is preparing legislation to protect leasees. As of early 1996, the largest developer (Argoz-Catarnaran) had created about 130,000 housing plots under this scheme, in 460 locations. This scheme has been written about favorably, ignoring the downside described here.22 In stark contrast, although almost all agrarian reform cooperatives provide housing lots to their members, only 6,000 housing units (new and remodeled) were created since 1981 *23 25. With respect to rentals, Phase HII transfers of rented land clouded the rental market, creating insecurity for formal long-term rentals and therefore reducing the timespan of rentals, often to one season. 201{ans P. Binswanger and Klaus Deininger, World Bank Land Policy: Evolution and current Challenges. This paper was prepared for the World Bank Agricultural Sector Symposium, 1994. 21 Environment Department, Research and Policy Division Working Paper No. 1992-35, March 1992. 22 Straasma, "sMaking Land Markets Viable": Two Successful Approaches to Collecting Loans Made to Small Farmers in Central America and the Caribbean", V. Wisconsin, Dept. Of Agricultural Economics, 1990. 23 MAG/PERA, Evaluacion del Proceso de la Reforma Agraria XI, San Salvador, El Salvador, 1992. Annex 6: Rural Land Markets 9 Another drawback is that it deprives of a secure and stable source of income to farmers no longer able to till their soil (older farmers or single women without adult children), who will tend to leave their land idle. About 45,000 beneficiaries got land though Phase III of the agrarian reform, but many renters were dislodged before FINATA could proceed, increasing landlessness. It is estimated that the number of landless, land-poor or unemployed have remained the same in absolute numbers as at the onset of the conflict and the agrarian reform (about 255,000, or 40% of the 1992 agricultural labor force).24 (4) Agrarian Institutions 26. Agrarian reform beneficiaries depend of agrarian reform institutions (Appendix C). Both non- agrarian reform and agrarian reform beneficiaries in rural areas suffer from inefficient public services, but the agrarian reform institutions added a layer of paternalistic, inefficient and sometimes corrupt bureaucracy. Agrarian reform institutions have their own objectives, including self-perpetuation-this has been the experience throughout the world. In El Salvador there were three such institutions: ISTA, FINATA and the Land Bank. FINATA was closed in 1995. The Land Bank was created in the early 1990s to provide financing for land purchases and later put in charge of implementing the PTT. 27. In addition to the titling issues mentioned above, agrarian reform beneficiaries had ISTA co- gestores and required institutional approval for many cooperative decisions. Over time, cooperative management reached a symbiosis with the ISTA co-gestor, who functioned as a link with government services, obtained FIS and SRN grants and acted as contractor to the cooperative (the "NGO"). But the ISTA co-gestor did not collect debt payments. This system did not promote self-sufficiency or entrepreneurial spirit. (5) Cooperative Management's Accountability and Access to Social Services 28. Agrarian Reform Cooperatives held high hopes of reducing rural poverty (Appendix A). Land's "social function" was supposed to provide the fruits of ownership of the factors of production to the workers. To land was added initial working capital credit, and later technology and managerial capacity (through the co-gestores) to palliate the lack of experience of agrarian reform beneficiaries in the management of large farms. 29. Pervasive inadequate incentives and the war context adversely affected the economic performance of most cooperatives. Lack of investments stalled productivity improvements, diversification and innovation. Most cooperatives were not run as revenue-earning enterprises, and savings were used for dividend distribution, social services to members or other uses. Population living on cooperative land falls into two categories: members and ex-colonos (squatters who were either excluded from membership at the outset, not incorporated since, or later squatters who are tolerated by the cooperatives). Members are only 20% of the population living on cooperative land. Although the wage jobs per cooperative have remained stable at about 28,000 wages/year per cooperative), the ratio wage/permanent jobs decreased from 4 to 1 to 2 to 1 (1988-94). This evolution means that more members have permanent jobs, replacing temporary labor by other members and squatters (at a ratio of 255 daily wage jobs per single permanent job). Therefore, only 60% of the cooperatives' active population works in the cooperatives, of which 42% in temporary wages. Women were particularly affected as they were the first to be let go when wage jobs dwindled. 30. Why have Salvadoran agrarian cooperatives survived? What did they do, as opposed to their counterparts in Peru or other countries where they became cooperatives in name only and all production 24 Seligson, op.cit. 10 Annex 6: Rural Land Markets core of political cooperative leaders able to successfully jostle for political power. The platform for these leaders has been facilitated by Govemrnment inability to refuse political demands from agrarian reform cooperatives. Therefore, these cooperatives have been more successful than average rural dwellers in facilitating access to social services for their members, even though they are not the poorest. These services include education, health, pensions, life insurance, and housing. Table 5: Social Services Provided by Cooperatives and PTT Communities, 1994 No. of Average ISTA PTT Total families Families/Servi Cooperative (beneficiari ce es) With Schools 194 22 216 84,500 391 With Adult Literacy 99 15 114 84,500 741 No. Housing Units 637 437 1074 84,500 79 Providing Housing and Housing Lots 178 17 195 84,500 433 With Health Clinics 49 4 53 84,500 1594 With Free Health Services 149 10 159 531 With Health Insurance 154 6 160 528 With Pension Payments 107 107 84,500 790 Sick pay 237 13 250 84,500 338 Paid vacations 18 18 84,500 4694 Paid Funeral Services 262 24 286 84,500 295 Others (life insurance, family funeral 105 NIA. 105 84,500 805 services) ..................................................................................................................................................................... Number Surveyed 264 22 286 84,500 Source: MAG, PERA X, 1996, pp. 168-178; also table in page 180. National averages from 1992 Census: 678 for schools and 2708 for health sevices. 31. Although disappointing in terms of contributing to agricultural growth, agrarian reform cooperatives have contributed as channels for access to social services. Because they are legal entities, they have access to govemment and other programs. The issue is how to overcome the deficiencies in production while preserving the benefits of organization. It seems that the cooperative organization's main problem is govemance, as management has been in many cases wielding their privileged access to the information and influence that comes through membership in political organizations (asociaci6n y confederaci6n) and with the handling of the cooperatives' business end. Therefore, community, self- selected and solidarity organizations that can manage rural enterprises need to be nurtured and their access to services and inputs facilitated, improving their governance over that of the present cooperative model. This can provide one of the answers to the looming voluntary parcelization that could result from both implementation of D.L. 719 and from P.T.Tproindivisos. (6) Other Distortions Affecting the Salvadoran Land Market 32. In addition to the direct land interventions described above, other distortions affect the land market. The mechanism through which these distortions work is through land prices: (a) subsidization of capital intensity (e.g., agricultural machinery), which encourages larger farms, even if these are not more efficient, or by preferential exchange, interest or tariffs rates for inputs favoring mechanization or substituting for labor (these inputs have a 1% tariff level, while maize has 10%); (b) border/trade protection to land-extensive crops or livestock; (c) undertaxing/not collecting tax on land (including idle) or on agricultural income, therefore converting the sector in a tax heaven. Fiscal reform has increased collection, but contribution of agriculture-despite its 9% contribution to GDP and 50% to exports, remains marginal at 0.03% in 1995; and Annex 6: Rural Land Markets 11 (d) better access to credit-particularly when it is subsidized-through real collateralization (therefore denying credit to those without titles-mainly the smaller parcel owners) but not repossessing in case of default. 33. Other factors include land price increases due to Governnent demand or urbanization, as this adds a substantial premium on land prices, and availability of long term credit. 34. What can the Government do? Common Governnent interventions are agricultural zoning; public purchase of private development rigthts estate tax relief and rebates, and taxes or subsidies to internalize environmental externalities. Box 4: How Policy-Distortions Promote Inefficiencies in the Land Sales Markets25 International experience has demonstrated that for a given technology, factor prices, land quality and: faming skills, there -will be an optimal <'operational: holding size" for a country's farms, at which the disincentive costs of adding rmore. workers :fully.offsets theeconomies of scale for purchasing inputs, accessingcr.edit, management-skills, etc.. Mounting evidence:and particular.examples continue to illustrate that the small, private "family" farm is the most efficient form of economic operation. However, in most countries social and political distortions, along with ill-conceived incentives, often prevent the land market- from establishing an optimal holding pattern. Land markets and land tenure have historically been dictated by societal power relations. Fromrthe traditional feudal.systems. of Europe, to the landlord estates of China, Japan, -and Korea, powerful ruling classes maintained control over land, and thus. the economy, through coercion, keeping peasants indentured through tenancy contracts or usufructary rights. :Over time, peasants (slaves, serfs) gradually became free, so landlords generated new, economic distortions to main,tain peasant labor, -in,cluding: (i). lnd monopolization, confining peasants to small, unproductive plots by buying up productive lands and establishing vagrancy laws; (ii) differential taxation, establishing oppressive taxes and tributes for peasants, while .larger estates receive tax breaks and privileges; and (iii) establishing and monopolizing agricultural product markets in order to restrict market access for the pesants. While many of these power methods and land market distortions may seem antiquated with the advent 'of miodern economies, they remain. a: part of many countries' land markets. Many countries continue to undertax rural landholdings,' land thus .becomes a tax haven and there is incentive to accumulate and consolidate. hi many cases, large landowners receive credit subs,idies, and-in-almost, all cases receive better terms than small owners who have no collateral, no credit history, or are too widely dispersed and remotely located to have economies of scale in transaction costs. Legal restrictions over' certain forms of land holdings can also i'nhibit a would-be seller or long-term renter of land from entering the market. The: power that land represents and provides, has historically led to concentration in the hands of a few. The policy distortions that the ruling classes utilized to consolidate their power in the past are reflected in the distortions of today. If incentives and: legal restrictions on tenancy make the options to sell or rent long-term difficult, uneconomical- or risky, then the land sales market will be unable to move toward amore optimal distribution of ownership-operational holdings, and the sector will be under-productive. 35. Since the signing of the Peace Accords in 1992, the dynamism of the land sales market has been defined primarily by four principal factors: (a) the post-peace upward trend of land prices toward prewar levels; (b) the conversion of rural land for housing (subdivision or lotificaci6n); (c) remittances, which finance housing; and (d) increased price levels established due to Govemmet obligation to purchase properties for PTT and the advent of Peace. 36. Although no hard data is available, it is estimated that the driving force determining land prices have been increased supply during the first years of adjustment due to the decline of agriculture's profitability following elimination of subsidies and protection for the sector, the inability to use land productively for agriculture in some areas directly affected by war, as well as outmigration from the rural areas due to agriculture's stagnation. 25 Source: From Binswanger, Hans P., Klaus Deininger and Gershon Feder, Power, Distortions, Revolt, and reforn in Agricultural Land Relations, World Bank, WPS No. 1164, July, 1993 12 Annex 6: Rural Land Markets (a) Evolution of LandPrices. 37. Since the signing of Peace in 1992, in some Table 6: Land Price Ratios rural areas land prices (in dollar equivalent prices) have Year Minimum Maximum Ratios incraseddueto increased demand in turn spurred by values (0IJha) Values (0IJha) Max/Min increased due to increased demand in turn spurred by 1992/93 2,860 11,430 8.5 population growth and remittances from emigrants, the 1993/94 2,413 24,300 8.89 Government's land purchases program as part of the 1994/95 2,150 30,000 13.95 pent-up for construction h by ~~~~~~~~~~~~~~ ~~1995/96 2,860 38,000 13.29 PTT, and pent-up demand for construction unleashed by Source: MAG/GOPA and OCTA, 1996 peace. In fact, price ratios between most and least expensive Government-purchased land jumped from about 5 after Peace to more than 13 in 1996,26 reflecting its higher opportunity cost in alternative (housing) uses. (b) Conversion of Rural Land to Urban Uses. 38. El Salvador is the most densely populated country in the continental Western hemisphere (about 250 inhab/km2). Not surprisingly, an influential factor in the price increase has been the rise in demand for housing land, and the phenomenon of subdivision. The high population density of El Salvador and the expanse of housing has meant that land historically utilized for agriculture is being sold for housing lots. Because of the higher price for housing properties, the cost of agricultural land is also being influenced, rising above the net present value of its productive value overtime, and therefore no longer worthwhile allocating for agricultural purposes. .oi5:-0-Why . Good Arcultua :Land Converte to Urba .nXUse?27000 t :400Xi-X:l- .0-0:.0:0-000000-00-00g0 . .Atssueis.. t. lossof prime"a iu ua land .(Cl ss an Isoiqualit..................... ............a.. l o r a an gee..... r.. ........ T t0endwa oserved in the UK and- the 0US.t The Eeeconomic picplsta exliitaefudj th hisor of o- itie- (0mosl in uoe.0 Citites -meredt asiceters lfor mreing ofarcltualpouts,tospl w inpts to fcltatethe ind.us.r revoon.7 Thereisahighcorrlaton be e entheloionofpri meland nd.hun s s. T... ot communiacations and otherinfrastructure mght have played an iportiant role in detennrining thle location of .earlycities.s rr Thi ypothesiswas test empirically in t dfidigs idte ta"hle1.1i% o t n is n SilClassuejs eand W2Iofral.i ti 50 0eof ih00 t r.ai; t r Soil Coneration Sevc estiatesbs on s n967- ata) prime agu alladisco toura set t wc th r of land in general and three times the rate of non-prime ra=.l land.. . Ts.is.notgexplain-ed y migration patterns, buCtrather by the uenerl icrease .,::.[ in oplaion and . oveal tda I"sityd It developing countries, the urbanization trend has been different, and, coXn-tra to developed countre s_ large cities, i th. de.velopingworlcitiesarestillexperiencing high positive rates of i. U n as e farmlands. Together ithX - populati growth,$ urbanization creates high pressure on land tin both urban and ruralsectors. Demandor!ts sae forhousing-and uban activities,Obid ding upthe price ol cpai ti of rent the-ln cunfeny produc;es c. The expectationof higher ztureza rentn oe w c rr thric' 39. Urban encroachment has been significant in El Salvador. No national data are available, but it is estimated as about 86,000 ha since 1974. Recent data and IGN estimates indicate a current average of about 5% of the national territory is under urban use; in 1974, this was about 0.8%. More specifically, in Western part of El Salvador, the urban to total area percentages are: Sonsonate 48/1239 (5%/), San Salvador 156.4/1042 (15%), La Libertad 57/1600 (4%/O); and Ahuachapan 4r1.6/1239 (3% a). 40. Many low-income rural housing parcels have stemmed on this newly urban land. Argoz- Catamaran, the largest (but not the only) development company, has sold 130,000 housing plots in rural areas in the last 17 years (para. 24). The owner has created a rural Housing Bank, to raise long-term capital to finance more housing developments. 26 OCTA, 1996. s.Bhadra and Brandao, Urbanization, Agricultural Development and Land Allocaton, World Bank Discussion Paper 201, 1993. Annex 6: Rural LandMarkets 13 Box 6: Direct (physical) and 3idirect Effects-of Conversion of Rural to Urban Land Physical effects result.from the building: of investment and infrastructure, change in the density of use as a result of urban pressureandoverall populationgrowth. Indirect effects which result from mixing land uses (residential, manufacturing/industrial, and agricultural). These include regulatory, technical, speculative andmarket effects. Research results inNewlersey2M indicate hat all the indirect effects. below have a negative effect on output supply and input, demand: vegetables was. -the only subsector to benefit, livestock was the.most adversely affected, and demand for labor-increased. (a) Rgulatory ef,fcts respond to the.. ba-ization's demand for farmers to intemalize externalities related to the use of pesticides and chemical fertilizers. (b) Technical effects related to the imcreased costs of agricut prduction close to urbanreas; these generate csts due to vandalism, and either positive or negative impact on farnm mnagement, depending on whether proxity it costlier to changes in agricultural production organization due to proximity. (c)YSpeculative effects-can be more:substantial, as farmers are reluctant to invest in fixed farm-rea:ted inprovements because of the increasid opportunity cost of land as a result of urbanization demand. Land becomes: a financial rather than a productive asset. This is caled the "impermanence syndrome". These speculative effects may become very important.im countries with high densityof populatiomn (d) Market effects occur because suburbanization reduces:farm trnsport costs. (c) Long-term Financing for Land Purchases 41. An additional problem to access land is the lack of long-term credit for land purchases. Only for housing is there a mechanism for 10-15 year financing, at market rates. This averaged 19.5% in February 1996 and is made possible by the issuance of up to 3-year bonds exempt from encaje bancario. No such mechanism exists for agricultural uses of land. B. Toward a New Paradigm for Land. A New Agrarian Policy 42. Recognizing the need to improve agricultural policy, combat poverty, and increase employment and incomes through better land use and secure tenure, the Government formulated a new agrarian policy in 1994. The current Administration has made big strides in redressing the interventions and tackling the distortions discussed in the previous section. 43. The 1994 agrarian policy sought to: (i) finalize, as quickly and efficiently as possible, the land transfers agreed in the Peace Accords29 and complete the parcelizations requested by agrarian reform cooperatives; (ii) ensure security of tenure to promote investment and enable transfer of land; (iii) streamline state agrarian reform institutions; (iv) consolidate and restructure agrarian debt; (v) create a freer market for land to facilitate access and create employment opportunities for the rural poor; and (vi) design and install an efficient rural finance system. (1) Finalizing Land Transfers. 44. The two ongoing land transfer programs have suffered delays. In 1994, Peace Accord land transfers were speeded up after being bogged down for lack of land registry and cadastral data impeding the clear definition of sellers' rights to holdings. In November 1994 the CNR was assigned to help accelerate titling and registration, and by December 1996 the process of financing and adjudicating land had been practically completed. Outstanding titling, registry, and cadastral activities are mostly due to deficiencies in sellers' land titles or lack of interest in the beneficiary pool. The next step is to parcel, title, 28 Lopez, Adelja, and Andrews, "The Effects of Sururbanization on Agriculture", American Journal of Agricultural Economics 1991, 70(2), cited in Badra, op.cit. 29. The 1992 Peace Accords committed the Govemnment to distribute 166,000 ha to 47,500 excombatants and squatters by December 1995. These targets were later adjusted to explicit demands. 14 Annex 6: Rural Land Markets and register the proindivisos-with subsidies to those who demand it, and finalize the same work for cooperatives-at their cost and request. Decree 719 approved the streamlining of agrarian reform cooperative's voluntary parcelization. (2) Ensuring Tenure Security. 45. The Land Administration Project supports the establishment of the CNR and finances a national process to systematically regularize title and deed registration. It involves finalizing the regularization of the national territory, and strengthen CNR's systems so that it is able to keep them updated while maintaining financial self-sufficiency. This will also improve security of rentals (see (5) below). (3) Streamlining State Agrarian Reform Institutions. 46. The Government is consolidating the three agrarian reform agencies into ISTA and the Land Bank. Since land transfers are practically finalized, these should be closed. (4) Consolidating and Restructuring Agrarian Debt. 47. ISTA and Land Bank's functions of calculating debt and ascertaining tenure status must be completed. So far, 50% of individual agrarian reform beneficiaries have become debt free. Decrees 698 and 699 approved in May 1996 condoned 70% of agrarian beneficiaries' debt for prompt payment. Individual beneficiaries received an additional relief of 05,000, so that debt under 021,665 was condoned. Decree 719 streamlines procedures for agrarian cooperatives to sell land to pay remaining debt. (5) Creating a Freer Market for Land. 48. The Government, supported by the Land Administration Project, is carrying out a national land regularization process for this purpose. All agrarian reform beneficiaries, be they individuals, in cooperatives or proindivisos, will have their land regularized by this, process-the ones without fully condoned or paid debt will be recorded with a mortgage. On rentals, land owners may feel more protected while renting registered land, and for longer periods, thereby activating a formal, longer-term rental market more conducive to better management and planning horizons. A clear policy announcement of the security of private rentals may be needed, providing access to land to (successful) farmers who want to extend their area without outright purchase, as well as a source of income to farmers who do not till their fields. This measure may improve land management, since the survey indicates that renters use land conservation measures more often than owners. 49. Another effect of Decrees 698 and 699 reducing debt and Decree 719 allowing sales and rentals to agrarian reform beneficiaries is to reduce their barriers to convey. It requires cooperative land to be sold at public auction, as the most transparent manner to achieve sales and protect cooperative members. C. Findings and Recommendations (1) Findings 50. Land policy in the recent past focused on administered land distribution, and previously on modernization of agriculture toward cash crops (coffee, cotton) in the past century. Cash crop promotion led to skewed land distribution, and distribution in the 1980s and 1990s to quell rural unrest and address poverty. Both instances are believed to have caused rural unrest without substantially reducing rural Annex 6: Rural Land Markets 15 poverty. Both instances are believed to have caused rural unrest without substantially reducing rural poverty. 49. New land policy needs to recognize more explicitly land's multiple functions and uses and the need for a flexible land market to improve land allocation, expanding the role of land not only as generating income from agriculture but also for rentals, a more dynamic use of land as a store of value for household savings mobilization (collateralization) and shelter, and other productive activities. 50. Since the early 1990s, the Government has started firmly in the path to redress the inequities of past policies and reducing rural poverty. Trade liberalization, fiscal reform and rural investment have reduced the distortions which made agriculture, or its adjunct, landowning, a store of rents. It has also made strides to connect the differentiated land market segments, thereby increasing its flexibility to allocate land to its multiple uses, not just agriculture, in addition to complying with its Peace Accord commitments and reducing barriers for agrarian reform land to become more productive and better managed.. 51. The following recommendations are designed to deepen the ongoing measures and expanded in the direction of creating income generating activities in the rural areas. (2) Recommendations 52. Rural poverty alleviation should be aggressively based on diversification of income, where agriculture is but one option. This makes particular sense given El Salvador's land scarcity. This involves promoting diversification of income, including reduction of rural transaction costs, and improvement of rural financial markets and social and human capital. Reduction of transaction costs involves improvement of physical capital in rural areas (infrastructure, transport, information channels, electricity, telephones, market information systems). Improvement in social and human capital includes adult literacy, vocational training, and supporting new rural organizations, non-political producer/entrepreneurial organizations and strengthening their management. 53. Debt relief work must be completed, and the deadlines perhaps extended. The decrees provide up to June 30, 1997 for beneficiaries to apply for relief. Agrarian reform institutions must be helped or the work contracted out for calculating balances so that payment can be effected. 54. The outstanding issues on land tenure security involving land titling and registration, which are not included in the Land Administration regularization activities, refer to ownership of land subject to government land transfer programs, and are: * Farm subdivisions within the cooperatives, currently under cooperative ownership, as allowed under Decree 719. It is suggested that this process should remain a direct responsibility of the cooperatives, who should contract and finance the technicians and developers to process such subdivisions without direct government involvement. * Farm subdivision of PTT units, currently under joint ownership (proindivisos). As part of a government assistance program, it is suggested that a subsidy be offered to PTT beneficiaries to cover the cost of physical subdivision of their land. Rather than offering free government services, thus subsidizing the parceling service (into quality/size equivalent parcels): such a scheme represent a subsidy to the demand for subdivisions, thus encouraging beneficiaries to take the initiative, organize themselves if they so desire and proceed to manage the process themselves. For beneficiaries who left their properties and perhaps live abroad, it is recommended that their land be put on a trust or be sold after due process. Such due process should include a campaign of public announcement in the mass media indicating to owners that they should 16 Annex 6: Rural Land Markets sign the corresponding documents for the sale of the land, and that their right prescribes after a period of sevemi (say, 8-10) months. The Government has drafted legislation to facilitate the land market under proindivisos, where the absentees who have no designated representatives would be represented by the Procurador. 55. Formalize arrendamientos con promesa de venta by moving forward on draft decrees. 56. The other aspect of land tenure security involve the security of rentals. Ongoing and regularization, a policy announcement, and the closing of agrarian institutions may help re-establish confidence in potential landlords. 57. Close ISTA and Banco de Tierras, and transfer remaining functions to manage these to a special office. 58. The challenge with rural development based on cooperative organizations is the improvement of governance. Cooperative management should be subject to controls like any other business enterprise for the benefit of their members as shareholders: external audits, management representativity and professionalism. Decree 719 supports the creation of grass-roots organizations which may provide production incentive-based alternatives to traditional cooperative organization, as well as organizing small farmers and cooperatives into farmer associations providing services (input and output marketing, technology dissemination, bulk purchase of social services like life-insurance). These associations would be better equipped to provide feedback to agents of technology generation and dissemination on appropriate technology, and on how to face diversification and globalization. 59. Land market liberalization should be deepened (particularly with respect to rentals) in order to allow land used now below its potential to be assigned to other users through market mechanisms. Also, other uses of land (other than agricultural production) in rural areas should be recognized: as housing and saving instruments. 60. To encourage the emergence of non-Government long-term credit for land purchases, rules for repossession in case of default should be streamlined. 61. Eliminate other distortions that prevent a level playing field for agriculture, including improving tax collection in agriculture. 62. A final suggestion on a point not covered in the report, is to consider the need for more effective regional/municipal planning to avoid chaotic urbanization. Annex 6: Rural Land Markets 17 BIBLIOGRAPHY Bhadra Salazar Brandao (1993), Urbanization, Agricultural Development and Land Allocation, World Bank Discussion Paper #201, Washington, D.C. Binswanger, Hans P. And Klaus Deininger (1994), World Bank Land Policy: Evolution and Current Challenges. This paper was prepared for the World Bank Agricultural Symposium. Bimswanger, Hans P., Klaus Deininger and Gershon Feder (1993), Power, Distortions, Revolt, and Reform in Agricultural Land Relations, World Bank, WPS No. 1164 Durham, W.H. (1979), Scarcity and Survival in Central America. Ecological Origins of the Soccer War, Stanford University Press, Stanford, Califomia. Lopez, Adelja, and Andrews, "The effects of Sururbanization on Agriculture", American Journal of Agricultural Economics 1991, 70(2), cited in Badra. Seligson et. Al. (1993), El Salvador Agricultural Policy Analysis. Land Tenure Study. USAID Report Simon, Laurence R. And James C. Stephens Jr. (1982), El Salvador Land Reform 1980-1981: Impact Audit, Oxfam America, Boston. Straasma (1990), "Making Land Markets Viable: Two Successful Approaches to Collecting Loans Made to Small Farmers in Central America and the Caribbean", Wisconsin, Dept. Of Agricultural Economics. The World Bank (1995), Guatemala. Land Tenure and Natural ResourcesManagement. ReportNo. 14109-ES. I Appendix A Property Rights and Land Management Conceptual Framework30 The nature and exercise of property rights over natural resources is at the heart of modem natural resource economics and debates over natural resource management policy. Property rights or resource tenure regimes-historically evolved, legally defined (in either national or local customary law), and culturally specific-largely embody the set of incentives facing natural resource users, with important implications for time preferences, investment strategies (i.e. intensivity of relative factor utilization), and approaches to externality issues. A strong linkage between tenure over and management of natural resources is often claimed in the natural resource economics literature. Two principal, somewhat overlapping strands of theory have been elaborated. The first3" has tended to focus on the relationship between tenure security (especially land tenure security) and various measures of investment and productivity. This "property rights" literature has argued that the absence of clearly defined property rights over environmental goods acts as a disincentive to optimal and sustainable resource use. As many critics have pointed out, however, the existence of property rights (usually understood as individual property rights) alone is hardly a sufficient condition for sustainable natural resource management.32 Viewing land as a capital good, property rights advocates argue that tenure security becomes important not only in restraining resource use, but also in inducing maintenance and investments. Feder and Onchan (1987), for example, claim that ownership security affects both investment incentives and the availability of resources to finance investment. On the demand side, property rights proponents argue that "economic agents who cannot be sure of receiving the benefits of their efforts (because of positive externalities) do not have as strong an incentive to work and to invest as they would have in a situation in which all externalities were internalized" (Wachter 1992). On the supply side, access to credit may be facilitated by clear ownership rights over collateral. While there is some empirical evidence to suggest a positive relationship between possession of formal property rights over land (e.g.. title) and level of farm investment and improvements (cf. Feder and Onchan 1987), many are wary of generalizing these results to different socioeconomic contexts. Roth et al. (1989), for example, conclude that while land titling may be an appropriate policy for increasing tenure security, investment, and output in many situations, it is important to recognize that possession of title is not necessarily synonymous with security, title does not confer the same set of rights and security of rights in all situations, and the impacts of increased ownership security depend on technology and the nature of credit and input markets. Similarly, Atwood (1990) argues that in much of Africa, land titling would not have the intended impact, would not be economically justifiable, or would even be counterproductive, due to differences in transactions costs, credit sources, productivity, security and access to formal or informal institutions faced by different groups of people. A second strand of the literature has been largely devoted to an important variation of the natural resource tenure and management relationship: common property regimes. Bromley (1989) defines common 30 From: World Bank, Guatemala. Land Tenure and Natural Resource Management, July 1995. 31 See, for example, Feder (1 987a, 1987b); Feder and Onchan (1987); Feder and Noronha (1987); Chalanwong and Feder (1985); Feder and Feeny (1991). 32 Bromley (1989) notes that there is ample and growing evidence of destructive land-use practices undertaken by private (freehold) owners of land and related natural resources. Appendix A Page 2 of 2 property as a situation in which the management group (the "owners") has a right to exclude non-members, and non-members have a duty to abide by exclusion. Individual members of the management group (the "co-owners") have both rights and duties with respect to use rates and maintenance of the resources. Ostrom (1994) has usefully identified seven "design principles" of what she calls "common-pool resources": clearly defined boundaries; congruence between appropriation and provision rules and local conditions; collective-choice arrangements; monitoring; graduated sanctions; conflict-resolution mechanisms; and minimal recognition of rights to organize. Typically, common property regimes involve common ownership of the resource system, but provide the members with individual use rights (Atwood 1990). Many problems of traditional common property regimes are related to the breakdown of the mechanisms of internal governance, the loss of effectiveness of external access controls (leading to encroachments and invasions), and lack of access to credit and other services due to collective property's often ambiguous legal status in both civil and common law systems (Morse and Woodman 1988). Although many of these systems have been flexible enough to adapt their rules and mechanisms of internal governance to changing socioeconomic conditions, recent pressures, such as population growth, increasing commercialization of agriculture, and changes in traditional norms and values have sometimes undermined the working of the traditional rules and mechanisms of internal governance or have made them obsolete, requiring new or adapted rules (Wachter 1992). Thus, contrary to those who claim that tragedy must always come to the commons, many common property theorists argue that degradation is not the fault of the property regime, but rather of the breakdown of the incentive mechanisms necessary for the concept of property to mean anything (Bromley 1989). In this view, it is the dissolution and not the existence of local- level common property management institutional arrangements that has led to resource degradation in the extensive economic margin of many developing countries (Bromley and Cernea 1989). Appendix B Rural vs. Urban Land Allocation El Salvador's population is growing and urbanizing rapidly. The 1992 population census found that the urban population grew from 25% percent in 1979 to about 50% in 1992. How does population growth and urbanization affect the land market? How does population growth affect the land market? Most rural land fragmentation originates in population growth, including inheritances. Minimum size parcels are generally perverse because deprive some members of the family (generally the daughters) from access to inheritance value when other assets are absent. Indivisibility is also perverse because it creates barriers to buying. selling or renting without the explicit consent of all owners, often spread out. How does urbanization affect the land market? A recent World Bank publication33 surveyed the literature on the interactions between urban and agricultural development and the implications for the conversion of land from rural to urban users. This conversion is considered irreversible in most cases. The review concluded that there are no general equilibrium models addressing these implications. Most models analyze from the perspectives of either urban or rural economists, with dual economy models incorporating the agricultural sector in a highly stylized fashion, ignoring distance, soil fertility and land reallocation. From the work of von Thuenen, however, we can extract insights into interesting hypothesis about the effects of proximity to urban centers of profitability of farming. Another conclusion of the paper is that there does not seem to exist a systematic discussion of market failures in land markets, although most governments intervene heavily in this market. Government intervention takes different forms, and address direct and indirect effects of land conversion. Is Government intervention justified in the rural land market? Typical Government Interventions. The rationale for Government intervention rests on market failures. However, regulation to cause cost internalization of unregulated private decisions depend on both the adequate existence of markets for negotiating contracts and Government's capacity to coerce and/or enforce these regulations. (See text Box 1) 33 Bhadra, D., and Brandao, A. S. P.,(1993) Urbanization, AgriculturalDevelopment, and Land Allocation, World Bank Discussion papers No. 201. I Appendix C Agrarian Reform Experience The problems concerning land in El Salvador-land scarcity for the poor, tenure insecurity, and low productivity-have been principal components of socio-political contention throughout much of the nation's history. With a population of well over five million people, and an area of 20,000 square kIn, El Salvador is one of the most densely populated country in Latin America.34 Within a context of poor soil quality and rapid population growth, government intervention has skewed land tenure pattems causing high levels of landlessness and exacerbated natural resource degradation. Table C.1: Land Use Potential Land Use Potential Area (in Ha) 1! % of Area Intensive Agriculture (Classes I-lll) 357,190 17.0 Umited Cultivation (Class IV) 332,861 15.8 Rocky-Umited Agr. (Class V) 45,585 2.2 Perennial and Tree crops (Class VI) 200,996 9.6 Pastures Only (Class VIl) 858,644 40.8 Subtotal 1,795,276 85.4 Not Suitable for Agriculture (Class Villi) 253,587 12.1 water 38,386 1.8 Urban 16,831 0.8 Total 2,104,080 100.00 Source: OAS, El Salvador: Agricultural Zonirtcation Study, 1974 Pre-1980 Government Intervention and Land Tenure Patterns. Why have land tenure patterns traditionally have been skewed? Durham35 argues that skewed land patterns occurred due to State's intervention in favor of coffee exports and cash crop diversification in the late 1800s and early 1900s. He describes these State interventions in three phases: (a) Pre-conquest and Colonial Periods. Prior to the introduction of indigo and cattle, Indians were organized in relatively equitable communalVejidal systems. Colonial powers rarely became involved in direct production of balsam and cacao, the main export crops, and relied basically on rents and tribute from the Indians. With the introduction of indigo and cattle, starting in the second part of the XVI century, Spanish colonialist received land grants to the detriment of Indian population. By the early XIX century the haciendas had appropriated about a third of the colony's land area. To retain a labor pool, a vagrancy law was enacted in 1825, and penalty was imprisonment. (b) The expansion of coffee. Despite these encroachments and land tenure changes, most Indians had managed to preserve this communal/ejidal land into the first days of the Republic, founded in 1839. Most indigenous land was concentrated in the fertile volcanic highlands. Government promoted coffee exports, and in 1847, a landowner planting more than 15,000 coffee trees was exempted from military and public service for himself and his employees. From 1857, public land was granted to anybody who planted at least two/thirds with coffee. In 1881, the Government abolished the traditional communal land system, leading to the almost complete loss of traditional lands to private landowners for coffee production. A large titling program was initialed in 1882, intended to speed the up the growth of coffee production. The 1882 law required all occupants of ejido lands to register their claims and pay the titling fee within six month; 34 More than 250 inhabitants per Km2. El Salvador Natural Resources Management Study, World Bank Report No. 12355-ES, 1994. 35 This sections borrows from W. H. Durham, Scarcity and Survival in Central Anierica. Ecological Origins of the Soccer War, Stanford University Press, Stanford, Califomia, 1979, and from Binswanger, H., Deininger, K., and Feder, G., Power, Distortions, Revolt, and Reform in Agricultural Land Relations, World Bank Working Paper 1164, Washington, 1993. Appendix C Page 2 of 5 otherwise "unclaimed" lands were sold in public auctions. Illiterate Indians were often no aware of these requirements, as opposed to well connected individuals. The reaction was a series of popular uprisings (1872, 1875, 1880, 1885, 1898), culminating in the 1932 massacre in Santa Ana, Ahuachapan and Sonsonate, where about 1% of the population was killed. This created a mass of dispossessed landless wanderer seeking seasonal work, a characteristic of Salvadoran labor throughout this century. To address this problem, in 1896, the Legislature reconfirmed full land ownership to those agriculturalists who still held their communal land plots. (c) Agricultural Diversification. The value of coffee exports dropped after the Great Depression and the 1932 revolt, and well into the Second World War. To promote diversification and lessen reliance on coffee, large areas of the Pacific Lowlands were converted to cotton in the 1950s and 1960s in response to favorable market conditions. Sugar also expanded rapidly. Land owners therefore reduced land available for rentals to colonos, sharecroppers, renters, or squatters, who used it to produce food. Many of the newly landless responded by emigration to the city, to Honduras36, or staying around large farms, which provided seasonal jobs. It is no surprise that the resulting high rate of landlessness, and disequal land tenure instigated widespread social discontent in rural areas, which eventually contributed to the civil war. In 1971, half of El Salvador's 271,000 farm units consisted of less than one hectare and were considered to be of poor soil quality. Ninety percent of all farms were less than five hectares, and six families held more property than the 133,000 smallest-scale farmers. The Government response was the Agrarian Reform of 1980. This was enacted in direct response to the culminating social tensions, in an attempt to avert disaster. However, the Agrarian Reform was implemented under circumstances of violence and virtual chaos, as elements of the military, sponsored by the land owning class opposed to the reform reacted against the reform objectives. Despite the comprehensiveness of the agrarian reform (to be described below), land tenure remains skewed in El Salvador (see table below). Table C.2: Land Distribution Among Land Owners (1989)37 Size of Land Holding (hectares) Form of Tenure 0-2ha 2-5 ha 5-20 ha 20-50 ha 50 ha + Total No. of Land Owners 196,181 39,978 31,822 9,072 3,786 280,839 Percentage of Total 69.9 14.2 11.3 3.2 1.3 100.0 Land Area for 131,291 132,461 300,187 275,671 387,547 1,227,157 Landowners %of Land Area for 10.7 10.8 24.5 22.5 31.6 100.0 Landowners Source: MAG, Land Use and Tenure Survey, 1989 Recent Land Transfer Programs To stem further discontent, and the Militar Junta that took power in the late 1970s decreed a series of land transfer programs. These took place over the decade of the 1980s, under war conditions, and its objectives were mired from agrarian transformation to political jostling. The cooperative movement became politically organized and established as an interest group to be contended with and as a vehicle for political power for cooperative leaders. The number of members, however, declined after 1989 during the last phase of the conflict, and further after Peace in 1992 (See table 2). 36 Durham estimates that 130,000 Salvadoran emigrants returned to El Salvador from Honduras as a result of the soccer war (1969), adding to the landless problem, Scarcity and Survival, op.cit., p. 170. 37 Does not include land renters, sharecroppers, and agriculture cooperatives (reform and non-reform). Appendix C Page 3 of 5 The 1980 Agrarian Reforms The Basic Law of Agrarian Reform (Decree 153) was issued by civilian-military junta under a state of emergency, on March 5th, 1980. The reform as envisioned by the revolutionary junta had three goals: greater income equality, expanded employment opportunities in the rural sector, and increased and diversified agricultural production.38 The reform was originally designed to be implemented in two phases. During Phase I of the reform the Government was to acquire, through voluntary sale or expropriation, all properties that exceeded 500 hectares in size (landowners were allowed to keep 100 to 150 hectares- depending on land quality-for own farming). The Government expropriation would include all machinery and livestock that was found on the properties.39 The original landowners were to be compensated according to the average property values on the owner's 1976 and 1977 tax declarations410, which was below their market value. Government paid in long-term Government bonds. These bonds can be exchanged at about a 30% discount in the secondary market, and be used to pay tax liabilities. The first phase of the reform was a massive transfer of assets to farms' resident workers. Over 400 properties were targeted for expropriation during the initial days of the reform, encompassing 218,000 hectares, or 12% of Class I-VII land in 1974.41 Most of the land expropriated in this first stage of the reform (60%) were used for pasture or were lying fallow at the time, thus was not as politically sensitive. Most of the remaining 40% were large cash crop plantations owned by absentee farmers. The beneficiaries of the reform were the resident permanent workers on the plantations. Most non-resident seasonal laborers farm and plantation workers were excluded from the reform, either because they were not working on the farm at the time of the take-over, or because they would have exceeded the target beneficiaries/land ratio. The new land owners (beneficiaries) were identified for each property expropriated, then organized by ISTA into cooperatives under collective land ownership. Land was titled to the cooperative. This also implies that cooperative membership can change. Cooperative management is elected yearly, and the General Assembly can expel members. An ISTA administrator (co-gestor) would remain on the cooperative to provide technical assistance, and to oversee the election of a directorate to run the agrarian reform cooperative.42 Table I shows the total amount of land involved in the reform and the number of beneficiaries. Phase II of the 1980 reform was to affect properties ranging from 100 and 150 hectares (depending on quality) to 500 hectares, and was generally considered the heart of the reform. The reform would have involved 1800 properties, comprising 23 percent of the nation's best farmland. Nearly 75 percent of the nation's export crops (coffee, sugar, and cotton) were to be affected by this Phase, providing land for as many as 50,000 families.43 The Constitution of 1983 established the property size limit at 245 hectares, virtually nullifying Phase II44, and avoiding much of the political upheaval which would have occurred with 3 Wise, Dr. Michael, (1986)'"El Salvador's Agrarian Reform", USAlD/El Salvador study, p. 13. 39 ibid 40 Carlos Augusto Granados, Director of Consultores Agropecuarios (COAGRO), interview by author, October 16, 1992, San Salvador. 41 Laurence R. Simon and James C. Stephens Jr. (1982),EI Salvador Land Reform 1980-1981: Impact Audit, Oxfam America, Boston, p.9. Oxfam put the number of properties at 376, while Wise put the number at 469 in 1986, although both sources cited the possibility of counting properties twice. The sources did agree on the area involved. 42 ISTA representatives were initially supposed to remain for one year, as ownership and responsibility was gradually turned over to the beneficiaries. However, whether due to social tunnoil or the Government's desire to control the cooperatives, representatives usually remained for several years -an extreme was 14 years in La Carrera in Usulutan and indicates the symbiosis between the co-gestor and cooperative management. 43 Simnon and Stephens, p. 15. 44 Land owners had two years to sell land over 245 ha. Appendix C Page 4 of 5 the expropriation of the aforementioned economically vital properties. Part of the remaining identified farms over 245 ha were adjudicated by ISTA to militants in the cooperative movement (Acuerdo del 3 de Julio de 1994), who were assimilated to the PTT beneficiaries. The final phase of the reform (Phase III) was not originally part of the 1980 agrarian reform. Dubbed the "Land to the Tiller Program", Phase III proclaimed that all campesinos renting small plots of land shall become owners of that land (with a maximum of seven hectares per beneficiary). Despite great fluctuations in the amount of land under rental agreements, provoked by minimum wage laws of the 1960's and 1970's, the level of sharecropping and land held in rental arrangements had surged in the early 1980's. An initial survey of USAID/El Salvador estimated that some 220,000 hectares or 13 percent of the land in farms45 would potentially be affected by Phase III, providing land for 117,000 beneficiaries.46 However, this program did not take into account land use practices of renters, which rotated crops under rentals to, among other reasons, diversify their crops under different microclimates and spread climatic risk . Finally, the potential for transformation was never realized, as the threat of expropriation provoked that about 25,000 renters were preemptively evicted from their holdings, accentuating landlessness on the countryside. By 1990, only 47,000 of the potential 117,000 beneficiaries received land, and only 80,000 ha of the estimated 220,000 ha were actually transferred. Nevertheless, a FINATA study47 concluded that those who did receive land individually had improved productivity vis-a-vis collective farming. B3ox C: EI l Sa..ador's A ai . Observions and Finding (1981) .......... .*. :0:0:0:0tOver -60% of El Salvador's ruralpopulation-th.e:landess ugrnt laborers which c much of l Salvador poorest citizens.-were.not otential beneficiares of thel lad rf program.. * 0The p;eanr Chur chacadecs:an many loc agian expertsweretexuded frmf thd plan ig and Phase I:~~~~~~~~~~~~~~~~~~~~~~~ex q * j -000- This gphaseonloygaffects 1gt5% fof thfe nation's arable fland, and hadl}. virtuall n effect o expor crp produc hd::tion. * Th d senumber of beneficiaries wvas relatively ismall (40,t591),i because onily penrmanent iresident laborers wnere Nincluded. 100;0 l-000-0.0-. The ref.would.ave iolved 1,800propies, cmprisin g23% of.th.e.n tions bestfarmn withnely75% of .t nation--4s export :crops;(coffiee,;sugar -an dcotton).However, Phse U was unpoular wiith oefl ldhlders an was-4--- never iEplme ted .. . E ..... - The -eConstitution of 1983 established the land-olding lmit of 245-ha,- there1y nullii;4ng Phase II Also, Phase H h.s.. ot -- n beeflly ainistered. Te nber of properties exceeding- te Constitutional lmit is cently, uder disute. :1 Proga :placed restictions on beneficiies to- sharecropr ad rent; their land,it thus ig.no. r prevalicultura *l000000t It bougndealy50,000- peasant. families ifor g30 iyears gontot mosty poor pltots of land wiich gdo 0not 0provide subsistence and cann000000iot Rsutin continuous foo-ro prduction.00000000000000 4.---0-00000000000000 ----. --0000 The agrarian reform of 1980 was of unprecedented scope in El Salvador . Table 1 in the main text of the annex illustrates the amount of agriculture in the reform sector. Land distribution programs have transferred the equivalent of 54% of land under soil class I-IV, 40 % of Soil class I-VI, and 51% of land under crops, resulting in a very significant impact on Salvadoran land tenure structure.49 Despite moving toward a more economically viable land distribution, the agricultural sector, and in particular the reform 45 Assuming all land Class I-VII is in farms. 46 Simon and Stephens, op.cit p. 15. 47 This is consistent with findings in other countries. i Laurence R. Simon and James C. Stephens Jr., EI Salvador LandReform ImpactAudit 1980-1981, op.ct. 9 As initially written, the reform of 1980 would have directly affected nearly half the land in El Salvador, and nearly half of the rural poor. Wise, p. 53. Appendix C Page 5 of 5 sector, continues to be constrained by inefficient land market, administration system (registry and cadastre), weak institutions and perverse incentives. In 1992, the Legislature approved D.L. 747, which allowed cooperative General Assemblies to parcel out lots for housing and production to its members. Under this law, members can opt for private ownership of parcels, shares or a mix of both types of tenure. ISTA would parcel and adjudicate. About 60% of agrarian reform cooperatives have opted for these "new" alternative forms of tenure. In May 1996, Law 716 allowed cooperatives to auction part of their land to pay debt. The Peace-Accord Mandated Land Transfer Program (Program de Transferencia de Tierras- PTT). As an integral component of the 1992 Peace Accords, GOES agreed to distribute 166,000 ha of land to some 15,000 former combatants of the armed forces, 7,500 former combatants of FMLN, and 25,000 tenedores, or farmers directly displaced by the war.50 The objective of the PTT, as well as non- agricultural efforts to create employment originating from the Peace Accords, is to provide economic and social opportunities for the participants and those most affected by the war to reinsert themselves in Salvadoran society. Through the course of PTT implementation, the number of beneficiaries has declined as potential participants take advantage of other options. Since 1992, the original scale and number of prospective beneficiaries of the PTT has diminished, as reflected in Table below. Table C.3: Scope and Advance of the Land Transfer Program (PTT) No. original No. Actual Fraction Beneficiaries Beneficiaries 51 Area (ha) FMLN (& Tenedores) 7,500 25,730 60,000 25,000 FAES (Armed Forces) 15,000 7,739 18,000 TOTAL 47,500 33,469 78,000 Source: Banco de Tierras 'Informe Ejecutivo: Avance del Programa de Transferencia de Tierras", 1996. T'he early stages of the land transfer were characterized by delays in implementation, disgruntlement concerning the quality of land to be transferred, high land prices, and not surprisingly, political tension. Delays in the purchase of land for distribution were primarily due to legal problems and were reflective of El Salvador's archaic land registry and cadastre situation, where only a minority of plots have current formal and registered title. Delays in transferring purchased land stemmed from irregularities in the rosters of beneficiaries, principally in the case of the FMLN and tenedores. More than three years since the beginning of the Program52, the land transfer process is nearly completed. The 78,000 ha represents 11 percent of El Salvador's land under crops. 50 Tenedores are displaced people from the regions most affected by the war. They were identified for the PTT by the FMLN. 51 Later information reduces total to about 30,000 52 The official beginning of the PTT was October 30, 1992. By December 1996, 30,000 beneficiaries had received land, including about 250 pending; the process is considered almost closed. Appendix D Recognized Forms of Land Tenure and Restrictions in Salvador's Land Market ........... . .. ...... ... ............. .... ..... . . .. ... R ig ht to : R ig h t to Form Xf T.nure Characteristics DIrectly Relevant Legislation: it :u l/ Sel Rent Restrictionst Agrarian Refonn p op e ..e.................s......................................................................................s............. ...................................................................................................... Individual Individual land -Constitution established Yes, with Yes, For sale, partition, donation (FINATA; Phase IlIl rights (or in regulatory framework for restric- restric- or rental of property, Reform). selected cases, Agrarian Reform properties tions but tions, but recipient must be a reform If family residence, family) which -D.L. 747, Art. 25 regulates land there is no no beneficiary or own 7 ha or rights belong to the appears on the transfers. moni- monitoring less after the transfer. family (as per Family title deed- Deeds toring me me- Code). executed by chanism. chanism. FINATA. Joint Ownership All beneficiary -Decree 839 (as it pertains to Dec. 839, Dec. 839, -Dec. 839 rental and sale (Pro Indiviso): names appear the Civil Code) can sell to yes, to restricted to landless Dec. 839 and Dec. on title deed. -Decree 713 (Programa de landless landless -Dec. 713. 713. Beneficiaries of Transferencia de Tierras). Dec. 713. -Dec. 713- the 1992 Peace Yes, free. Accords Land Transfer Program ....)............................ ..... .......................... ..... .............................. ...................... .......... ....................... .......................... ........................................ ........................................................ Non Refonm Individual Freehold land Constitution of the Republic, Free: No -Holding not to exceed 245 If family residence, rights belong to Art. 105, Inc. 3: 'The land decision by restrictions hectares. rights belong to the the individual owners cited in this article individual (Constituti -Judicial intervention legal family (as per Family whose name (individual holders; Inc. 2) have owner. on, if property is inherited by a Code). appears on the the right to freely transfer, Agrarian minor (16 years). title deed. share, divide and rent their Law of properties. Land belonging to 1979). cooperative associations and communes will be subject do separate regulations.' Also Civil Familiar (Group) Land rights -Family Code (refers to Civil Free; must Free; must Family Code (refers to belong to a group Code, Art. 988; in cases of be decided be decided Constitution Art. 105). or family with succession due to death). by entire by entire collective names -ISTA Law. group group. on title deed. -D.L. 842. registered on title, including 'life companion s", if they are legally registered as such. Collective Land rights -General Law of Cooperative Free; Free; both No limit to size of holding. belong to Associations (refers to Civil decision within the registered Code). made by cooperativ members of the -Law of the Creation of ISTA; consensus e or to traditional and Decree 842, where it of the individuals cooperative (prior pertains to properties regulated General outside the to 1 980s Agrarian by ISTA, taken over from the Assembly cooperativ reform). now defunct ICR. of e. Associates Corvnercial Freehold land Commercial Code (refers to rights belong to Civil Code). the individual whose name appears on the title deed. LAND DEGRADATION PROBLEMS IN EL SALVADOR' Introduction Land degradation is thought to be the most important natural resource management problem in El Salvador. It is commonly thought that '75% of the country's surface is degraded.' Land degradation is thought to lead to reduced agricultural productivity and sedimentation of reservoirs. If land degradation problems are as severe as is thought, they are likely to have a significant impact on agricultural growth, growth potential and on the welfare of a significant proportion of the population-about half of the population of 5.6 million lives in rural areas and a third is employed in agriculture. Although there is a perception that lad degradation problems in El Salvador are well understood, this understanding is in fact very superficial. The commonly-quoted figure of 75% of the country's surface being degraded, for example-in addition to being almost certainly exaggerated-provides no indications of severity of degradation, of its effects, of the areas in which it is most severe, or of its causes. TIhis annex reviews existing data on natural resource problems in El Salvador's rural sector and combines it with new data from a household survey in a strong analytical framework. To the extent possible, it attempts to identify areas (both geographic and thematic) where interventions are both important and feasible. The analysis focuses on effects of degradation on farm productivity rather than on downstream effects such as sedimentation of reservoirs, since this is the critical aspect from the viewpoint of agricultural sector growth. Because of the limitations of the available data, it is difficult to arrive at definitive answers to the problems of land degradation in El Salvador's agriculture. Nevertheless, several themes emerge from the available information: * Land degradation problems in El Salvador, while significant, are much less severe and widespread than is commonly perceived. * Farmers' responses to land degradation problems, while far from universal, are more substantial and widespread than is commonly perceived; where responses are lacking, there appear to be good reasons. These results, although they contradict much of the conventional wisdom about the situation in El Salvador, are consistent with those of recent and broader analyzes of land degradation in Central America (Lutz, Pagiola, and Reiche, 1994). 1 This annex was written by Stefano Pagiola and John Dixon. The authors would like to thank Carlos Restrepo, Jaime Acosta, and Margarita de Sanfeliu of FUSADES, and Claudia Ocafla, of the University of Maryland for their assistance in preparing this report. 2 Annex 7: Land Degradation Evidence on Land degradation This section reviews the available data for clues on the extent and severity of land degradation problems in El Salvador. Much of it is based on the results of a survey of 302 farm households carried out by FUSADES, a Salvadoran NGO, in 1996 (Box 1). Although this survey has limitations for the analysis of land degradation problems, it represents an important and rich new source of information on conditions in rural El Salvador. Box 1: The FUSADES Survey This study draws heavily on data from a household survey carried out by FUSADES, a Salvadoran NGO, with World Bank assistance, in early 1996. The survey included a main sample of 630 rural households randomly selected from all regions of El Salvador. This sample was designed to be representative of the rural population, based on information from a labor force survey carried out in 1992. 192 of the households in this sample were farm households, defined as rural households that farmed more than 0.5 manzanas (0.35ha) of land. In addition, a complementary sample of 110 farm household was also collected, for a total of 302 farm households. Each of these households fanns between 1 and 3 fields, for a total of 472 fields. For each field, a wide range of data were collected, including field characteristics, ownership, crops produced, conservation practices, and other improvements. The data collected on these 472 fields form the main source of information for this study. Detailed information on yields and input use was also collected for each household's main field (that is, for a total of 302 fields). These data are used in the productivity analysis. While these data provide a very rich source of information on many aspects of land use in El Salvador, several caveats should be bome in mind. First and foremost, the sample frame is based on population rather than land. One cannot, therefore, use these data to deternine the proportion of land under different uses that experiences a particular problem, or on which particular practices are used. Land owned and operated by non-rural residents, for example, would not be captured at all. The majority of land degradation problems in El Salvador, however, are generally ascribed to small fanners. Large commercial units generally occupy the better, flatter land. Moreover, the crops often grown on large holdings, including coffee and sugarcane, tend to provide considerable protection to soils. The sample, therefore, should tend to over-represent fields likely to have land degradation problems. It should also be noted that in the case of some departments the estimates are based on a small number of observations. Total Mean Mean Number of Number of Field area number of field size Denartment Households Fields (ha) fields (ha) Ahuachapan 22 31 52.7 1.4 1.7 Santa Ana 25 30 66.5 1.2 2.2 Sonsonate 15 27 21.0 1.8 0.8 Chalatenango 17 26 126.9 1.5 4.9 La Libertad 20 30 48.8 1.5 1.6 San Salvador 14 19 27.3 1.4 1.4 Cuscatlan 15 26 32.8 1.7 1.3 La Paz 16 26 59.0 1.6 2.3 Cabafnas 17 26 47.9 1.5 1.8 San Vicente 13 20 25.1 1.5 1.3 Usulutan 29 47 64.6 1.6 1.4 San Miguel 36 62 144.7 1.7 2.3 MorazAn 23 38 87.9 1.7 2.3 La Union 40 64 116.0 1.6 1.8 Total 302 472 921.3 1.6 2.0 Source: FUSADES Survey Annex 7: Land Degradation 3 Risk factors El Salvador is thought to be particularly vulnerable to land degradation. Much of the terrain is mountainous, and the soils are easily erodible. Rainfall is distributed in two rainy seasons divided by distinct dry seasons, so that rainfall is often heavy when ground cover is poor. Map 1 shows the proportion of land in USDA capability classes VI to VIII, which are generally considered unsuitable for cultivation. About 65% of the country falls within these land capability classes. The proportion of such land is particularly high in the northern and eastern parts of the country. Map 1: Proportion of land classified as unsuitable for cultivation N 0 10 20 30 40 50 CHALATENANGO I I I Kilometers ~~N~&.sONSOwATE -~~~~~~ MORAZAN ~~w ~LA q_~~~~~~~~~~~~~_' lt~~~~~~~~~~~~~~~~u~o' % Land in Classes V-iI.. 30 40 5086070 8090100 ~ P ISULLJTAN . T. Source: Data from 2nd Agncultural Census Because of population pressure, inequalities in access to land, and the effects of the civil war in the 1980s, substantial portions of land classified as unsuitable have been cultivated, usually by small farmers. Much of it has been planted to various combinations of maize, beans, and sorghum (collectively known as granos basicos, basic grains). Common cultivation patterns of these crops often leave the soil bare at the beginning of the rainy seasons. Erosion Considering the long-standing concern over soil erosion in El Salvador, there have been surprisingly few efforts to measure it. Measurements of erosion rates and yield were carried out on run-off plots at Metapan, in Santa Ana, over the period 1975-1980 by Flores Zelaya (1976-1981). The treatments studied include a traditional practice and two different conservation measures: live barriers and bench terraces. Table 1 presents mean soil loss and crop yields measured under each of the practices. Unfortunately, no clear link between practices, erosion rates, and yields can be drawn from these data. Although the experiment measured soil loss and yields over 5 years, it only included a few replications of each practice, a problem that was 4 Annex 7: Land Degradation exacerbated when measurement problems led to data from one of the blocks having to be discarded. Moreover, there was significant variation in weather conditions during the study period, as well as exogenous factors such as pest attacks on the beans crops. Table 1: Erosion rates and basic grains yields under alternative management practices at Metapan, Department of Santa Ana, 1975-80 Traditional practice Live barriers Terraces Soil Yield Soil Yield Soil Yield Rainfall loss Maize Beans loss Maize Beans loss Maize Beans Year (mm) (mt/ha) (mt/ha) (mt/ha) (mt/ha) (mt/ha) (mt/ha) (mt/ha) (mt/ha) (mt/ha) 1975 1,895 137.01 1.88 1.15 129.04 2.65 1.35 58.11 2.62 0.71 1976 1,397 72.17 1.50 1.20 5.10 2.45 0.89 5.95 1.88 0.40 1977 1,192 12.68 3.40 1.33 .. 4.37 1.50 .. 2.62 0.92 1978 1,928 4.50 2.12 1.20 6.89 4.51 0.60 3.25 2.81 0.23 1979 1,716 18.51 2.62 1.33 19.95 3.10 1.51 6.89 2.01 0.07 1980 2.11 0.48 3.15 0.45 1.91 0.17 Source: Fifth and Sixth Reports on Research and Runoff Plots at Metapan Notes: Soils at the site are latosols, reddish-clayey, yellowish-red, and gray forest podzol; average slope is 30% Treatments were replicated 3 times on 1,200 square meter plots; Block III plots omitted due to measurement problems 1978 and 1980 beans yields were affected by severe attacks from slugs (Vaginulus sp.) indicates soil loss was too small to measure One aspect of these data which is worth mentioning is that the most-quoted part is the first-year erosion rate under the traditional practice: 137 metric tonnes per hectare (for example, in World Bank, 1994). Clearly, this single observation is not representative of the observed erosion rates under that practice. Indeed, the similarly high rates observed under all practices in the first year of the experiment suggests this may be an effect of plot establishment. This type of misquoting or misinterpretation of experimental results is one of the reasons the conventional wisdom on land degradation problems in El Salvador is so badly wrong. Table 2 shows the distribution of fields reported as experiencing erosion by farmers in the FUSADES survey. As might be expected, the reported incidence of erosion increases with slope. Only 22% of surveyed fields with mild slopes experience erosion, in contrast to 83% of fields on steep slopes. Map 2 shows the incidence of reported erosion in the different regions of El Salvador. As expected, a relatively high proportion of fields in northern and western departments experience erosion. More unexpected is the high proportion of fields on which erosion is reported in La Libertad. This reflects the high proportion of fields on moderate and steep slopes found in that department. Productivity problems: Evidence from experimental measurements Experimental measurements of the productivity effects of land degradation are even scarcer than measurements of erosion rates. Table 1 provides some information, but it is mostly inconclusive. Beans yields were lower under the bench terracing practice than under either of the other two practices, while corn yields were lower than under the live barriers practice. This is most likely due to the reduction in area cultivated that results from terracing. Corn yields under live barriers appear to be higher than under the other practices, but the results with respect to beans yields are erratic. Annex 7: Land Degradation 5 Table 2: Incidence and effects of erosion problems on fields on undulating or sloping land Mild Moderate Steep slope slope slope fields) ha) fields) ha) fields) ha) No erosion problems 78 59 46 48 17 28 Experience erosion 22 41 54 52 83 72 Effects of erosion on fields affected: No significant problems 64 29 30 24 18 19 Future production will decline: Severe decline 0 0 5 5 16 16 Moderate decline 14 4 14 7 11 16 Minor decline 0 0 2 0 5 2 Will need more fertilizer to compensate 0 0 3 1 5 5 Total 14 4 25 13 37 39 Production was reduced in 1995 7 3 17 21 16 9 Fertilizer is washed away 14 4 32 36 32 43 Need to use more fertilizer to compensate 0 0 8 7 11 15 Source: FUSADES survey Notes: Percentages for effects of erosion do not add to 100 because of multiple responses Map 2: Proportion of fields reported as having erosion problems 0WM ~~~N 0 10 20 30 40 50 Kilometers 6 Annex 7: Land Degradation Productivity problems: Evidence from national statistics This section examines the available evidence on agricultural productivity using MAG national statistics. MAG yield data are based on a survey in which 553 sample points throughout the country are monitored, with 20% of sample points being changed every year. The current sample frame is based on 1976 land use map; a stratified sampling procedure is used, based on 11 land use types. The observations are then expanded to obtain regional-level and national-level estimates. The current sample design was not designed to obtain department-level estimates.2 Although some areas could not be visited during the war and were omitted from the estimates, MAG believes the data should not have been affected since the areas concerned have relatively little agricultural production. Figure 1 shows maize yield trends over the last 30 years.3 Although high-yielding hybrid varieties were introduced in the early 1960s (Walker, 1980), agricultural statistics only began distinguishing between hybrid and local varieties in the 1971-72 crop year. As can be seen, average maize yields have increased substantially, more than doubling in the last 30 years. The early part of this increase reflects the switch to hybrid varieties. Interestingly, while hybrid varieties do not show any significant trend in the last 20 years, local maize varieties show strong and sustained yield increases in the last decade. This probably reflects spill- over effects of improved management and inputs resulting from modern variety use. Figure 1: Maize yield trends in El Salvador, 1960-61 to 1993-94 3 - 7 2 0 1980 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 -61 -63 -65 -67 -69 -71 -73 -75 -77 -79 -81 -83 -85 -87 -89 -91 -93 Data Fitted Data Fitted Data Fitted All varieties -.- - Hybrid -_- --- Local ------ Figure 1 does not suggest that degradation is causing widespread productivity declines. National averages, however, can mask local declines. Figure 2 shows maize yield trends at the highest available level of disaggregation: the regional level. These data show similar trends: little or no yield change among modem varieties and relatively strong and sustained yield increases among local varieties. Unfortunately, this disaggregation is not a very useful one, since each region includes sections of all major agro-ecological zones, 2. A new land use map is under preparation. Once it is complete, a new sample frame will be developed. The number of sample points will be increased to 850. The new design will allow statistically valid estimates to be made at the department level. Even this, however, will fall far short of the degree of detail required for analysis of land degradation problems, as discussed below. 3. Data shown are for the first maize harvest, which typically accounts for over 90% of production. Annex 7: Land Degradation 7 from the mountainous north to the coastal plains. This is a common problem with data collected according to administrative boundaries. Similar time trends are observed for beans and sorghum, the other main subsistence food crops. Figure 2: Maize yield trends in El Salvador by region, 1972-73 to 1993-94 Region I Region 11 (Santa Ana, Sonsonate, Ahuachapan) (Chalatenango, La Libertad, San Salvador, Cuscatlan) 3 -A '~~ '.1j 1972-7S 1977-78 1982-83 1987-88 1992-93 1972-73 1977-78 1982-83 1987-88 1992-93 Region III Region IV (La Paz, Cabanas, San Vicente) (Usulutan, San Miguel, Morazan, La Union) 3 -3 - 2 - 2 /-V -- 0 0-- 1972-73 1977-78 1982-83 1987-88 1992-93 1972-73 1977-78 1982-83 1987-88 1992-93 Data Fitted Data Fitted Data Fitted All varieties -i- - Hybrid -*- - -- Local ------ Interpretation of these results is difficult. Certainly they do not provide unambiguous evidence that productivity is declining. The observed pattern of stable or rising yields does not necessarily imply that degradation-induced productivity changes aren't occurring, however. Yields could be rising, for example, because increases in fertilizer use have been sufficient to offset the damage caused by degradation (Figure 3).4 In this case, degradation would be reflected in rising costs and falling returns rather than in declining yields. The only firm conclusion that can be reached at this point is that if degradation is in fact occurring, it has not been sufficiently important that it could not be offset by increasing input use. 4. Unfortunately, disaggregated data on fertilizer use on food crops over time are not available, so a full productivity analysis is not possible. 8 Annex 7: Land Degradation Figure 3: Rising fertilizer use can offset the effects of degradation Effect of Yield degradation Yield Y2 ------- ---- Mv2\i t5 lt, - - - - - - - - - - - - - - - - - - - - - - - - - * Increasing Inputs o Constant Inputs e______________ FFertilizer Time Fo F1 F2 Application Productivity problems: Evidence from farmer perceptions The FUSADES survey asked farm households what changes they thought yields in their area had experienced in the last 10 years. As can be seen from Map 3, there is a widespread perception that yields have declined (or that declines have only been averted by increasing fertilizer use). Overall, only about 34% of households perceive yields as having either increased or remained the same. The regional distribution of these perceptions is surprising, however, especially the high values for perceptions of yield decline in Ahuachapan, Sonsonate, La Libertad, and San Salvador. Map 3: Proportion of farm households perceiving yields in their areas as having declined in the last 10 years N~~~~~~~~~~~~~~~ Note: Includes those perceiving yields have remained unchanged because of increased fertilizer applications Annex 7: Land Degradation 9 Table 2 shows the nature and severity of problems that farmers expect to experience on fields on undulating and sloping land affected by erosion. As might be expected, the severity of erosion problems increases with slope. On 64% of the fields with mild slopes which experience erosion, it does not lead to significant problems. In contrast, most fields on steep slopes which experience erosion have problems as a result. Future productivity declines are expected on 37% of fields on steep slopes; almost half of these will experience severe productivity declines. The incidence of current problems (loss of fertilizer and of current production) is also higher on moderate and steep slopes. Fields on steep slopes are also more likely to require additional fertilizer to compensate for previous erosion. Fertilizer use has had to be increased to compensate for the effect of past erosion on 1 1% of fields on steep slopes, and 8% of fields on moderate slopes.5 Map 4 shows the regional distribution of fields on which farmers perceive erosion as causing productivity problems caused by erosion on fields; Table 3 provides additional detail on the nature of the reported problems in each department. Overall, farmers do not expect erosion to cause serious problems on almost one third of affected fields. The proportion is particularly high in San Vicente and Cuscatlan. In contrast, about one quarter of fields affected by erosion are expected to experience lower yields in the future. The proportion is particularly high in Chalatenango and Santa Ana and to a lesser degree in Morazan. About one in six fields affected by erosion report that production in the current year was lower because of it, and twice that many report fertilizer being washed away (which would affect current costs but have no long-term effects). Map 4: Proportion of fields on which erosion is reported as causing productivity problems N 0 10 20 30 40 50 SANTA ANA ~~~~~~~~~~~~~~~~Kilometers ',~~~~~~~~~~~~~~~~. - :-- 0 ~ 10r 20 30 401 50 60 70 _UoU120T3N40 . a~~~~~~~~~~~~~~~~~~ 5. These fields account for only 2% of surveyed fields (no fields on mild siopes are reported as requiring higher fertilizer use). In contrast, 10% of households perceive that yields in their area have only been maintained because of higher fertilizer use. 10 Annex 7: Land Degradation Table 3: Regional distribution of problems caused by erosion on fields affected Need to increase Does not cause Future production Productionfell Fertilizer is fertilizer use to serious problems will decline in 1995 washed away maintain production (% (%- (% (%0 (% FT (%7 (% (% (% Department fields) area) fields) area) fields) area) fields) area) fields) area) Ahuachapan 22.2 22.0 22.2 7.8 22.2 39.0 55.6 70.2 11.1 15.6 Santa Ana 22.2 22.9 55.6 48.6 0.0 0.0 22.2 22.9 0.0 0.0 Sonsonate 40.0 29.6 0.0 0.0 40.0 26.0 40.0 26.0 40.0 59.2 Chalatenango 15.4 5.5 61.5 28.1 15.4 22.6 23.1 33.6 0.0 0.0 LaLibertad 53.8 64.1 7.7 3.3 15.4 17.1 15.4 8.8 0.0 0.0 San Salvador 33.3 27.5 0.0 0.0 33.3 13.9 33.3 58.6 0.0 0.0 Cuscatlan 87.5 95.4 12.5 4.6 0.0 0.0 0.0 0.0 0.0 0.0 La Paz 20.0 26.1 20.0 10.9 20.0 4.3 20.0 43.5 20.0 15.2 Cabafias 12.5 7.7 12.5 7.7 12.5 15.4 25.0 46.2 0.0 0.0 San Vicente 85.7 95.2 14.3 4.9 0.0 0.0 0.0 0.0 0.0 0.0 Usulutan 25.0 46.4 25.0 14.3 25.0 17.9 12.5 5.4 12.5 14.3 San Miguel 0.0 0.0 29.4 8.9 0.0 0.0 52.9 29.0 17.6 17.6 Morazdn 23.5 27.0 41.2 33.3 29.4 42.9 52.9 39.7 5.9 3.2 La Uni6n 29.2 21.8 20.8 17.6 25.0 16.3 33.3 55.3 4.2 9.1 Total 30.1 23.3 26.7 17.3 16.4 16.2 30.8 33.2 6.8 7.4 Source: FUSADES survey Notes: Percentages for each department do not sum to 100 because of multiple responses Productivity problems: Evidence from input and output data The FUSADES survey collected detailed information on yields and input use for each household's main field (that is, for a total of 302 fields). Efforts were made to carry out a cross-sectional productivity analysis using these data. The results were very poor, however, and are not reported here. This partly reflects the difficulty of carrying out cross-sectional analysis in an area with as much diversity of conditions as El Salvador. It also reflects the fact that for many of the independent variables, the only available measures were subjective qualitative ones. If the analysis is restricted to subsets of the data originating in relatively similar areas, so as to avoid some of these problems, too few observations are available. Adoption of conservation practices Conservation practices have been promoted by a long succession of projects in recent decades (Box 2). Despite these efforts, the conventional wisdom is that adoption of conservation practices in El Salvador is extremely low. Table 4 shows the extent of use of different conservation measures on fields in the FUSADES survey. Although only a little more than a third of the surveyed fields have some form of conservation, on moderate and steep slopes the proportion increases to over half. The most commonly-used conservation measures are cultural ones, especially minimum tillage and the use of crop residues for soil cover. Use of these measures peaks on moderate slopes and then falls. Among the structural measures, stone walls are most frequently encountered. Use of stone walls increases with slope, with as many as 20% percent of fields on steep slopes being so protected. Map 5 shows the regional distribution of use of conservation measures. Comparison with Maps 2 and 4 indicates a close correspondence between departments in which erosion and erosion-induced productivity problems are common and use of conservation measures. Annex 7: Land Degradation 11 Box 2: Soil conservation projects in El Salvador Since 1955, the government, through the Ministry of Agriculture and Livestock (MAG), has undertaken a variety of conservation projects (Hernrndez Navas and others, 1994; Perdomo Lino, 1990): * META (Improvement of Agricultural Lands) 1962-71. First government program to train soil conservation technicians. Technical assistance to conservation was provided on 47,000 ha. * FAO Project ELS/71/506. Integrated development project on 2,000 ha in Metapan, Santa Ana. Introduced the concept of watershed conservation. Gabions and reforestation were used to reduce the threat of downstream flooding. * FAO Project ELS/73/004. Integrated development project on 25,000 ha in the Rio Tamulasco watershed, Chalatenango. First project which combined work on conservation with efforts to increase productivity. * FAO Project ELS/78/004 (Conservation and Improvement of Renewable Natural Resources in the Northern Watershed of the Cerr6n Grande Reservoir), 1980-84, extended by Project ELS/84/006 (Development of Rural Communities and Watershed Planning), 1985-86. Worked in a 124,000 ha area in Chalatenango. Its major objective was to protect the Cerr6n Grande reservoir from siltation. Initially, simple conservation measures such as contour plowing were promoted; more complex and effective conservation measures were introduced later, including stone and vegetative barriers, bench terraces and individual terraces, diversion ditches (bench- type and trenches), earth ridges, drainage canals, and ditches. Farmers adopting conservation measures were provided with a package of incentives, including agricultural inputs ( seeds, plants, fertilizer, and insecticides), tools, materials for constructing soil conservation works. and technical assistance. About 4,000 ha were conserved. * FAO Project ELS/86/005 (Agroforestry Assistance to Rural Communities with Scarce Resources), 1987-92. Initially implemented in Cabafias, then extended to Usulutan and Morazan in 1989. The conservation practices promoted in this project were similar to those used in the Cerr6n Grande project. Stone barriers were the most commonly used measure, followed by diversion ditches, individual terraces, and alley cropping. Madrecacao, eucalyptus, and fruit tree orchards were also established. A similar set of incentives to those used in the Cerr6n Grande project was offered to participating farmers. Credit on generous terms was also provided, and has proven very popular among farmers. About 1,200 ha were conserved. Some measurements of yields on conserved plots were made at Guacotecti, in Cabainas, but the sample was small and there were neither control plots nor baseline measurements, so their usefulness is extremely limited. * WFP "Food for Work" Program, 1972 to date. The Food for Work program has often sponsored conservation work, but the quality of work has been poor. * Minimum-Tillage Project, 1970 to date. The extension service in Guaymango, Ahuachap6n, has advocated minimum tillage since the 1970s. A package of productivity-improving measures (use of hybrid maize and improved sorghum, use of N and P fertilizers, increased plant density, and application of herbicides and insecticides) and soil conservation measures (use of crop residues as mulch, living and dead barriers, and planting on the contour) was promoted. 2,450 ha are considered as being conserved. MAG estimates 100,605 ha were conserved under these projects, but the true area is probably much smaller, either because conservation measures were destroyed or have decayed after the projects ended, or because they were never built in the first place (Perdomo Lino, 1990). It should also be mentioned that some conservation measures also have other purposes. Some cultural conservation practices, for example, also affect current yields. Stone walls may be built primarily because of the need to use stones removed from fields. Although such measures also have conservation effects, this may not be the primary motivation for constructing them. 12 Annex 7: Land Degradation Table 4: Use of conservation measures Mild Moderate Steep Flat slope slope slope Total (% (% (% (% (% (% (% (% (% (% fields) ha) fields) ha) fields) ha) fields) ha) fields, ha) Cultural practices Crop rotation 2 5 2 1 4 2 4 3 3 3 Minimum tillage 7 11 11 6 20 25 17 13 13 18 Use of crop residue as mulch 10 13 9 5 23 22 11 9 15 16 Cover crops I 1 6 6 7 4 4 3 4 3 Total 16 25 23 14 39 42 30 25 27 32 Hedges 0 0 3 2 6 4 0 0 3 2 Stone walls 2 2 6 10 11 17 20 31 7 14 Ditches 0 0 2 1 2 2 2 1 2 3 Terraces 3 8 3 3 2 1 2 2 1 1 Gully control 0 0 0 0 5 3 2 1 2 2 Any measure 20 33 36 29 54 58 52 57 38 48 More than one measure 6 7 8 4 21 19 13 13 12 13 Source: FUSADES Survey Map 5: Proportion of fields on moderate and steep slopes with some form of conservation +~~~~~~~~~~ t~~~~~~~~~~~~~~01 03 05 Annex 7: Land Degradation 13 Summary Despite the availability of the data from the FUSADES survey, it remains difficult to ascertain the extent and severity of land degradation problems in El Salvador. Nevertheless, it seems clear that catastrophic statements such as '75% of the country's surface is degraded' are substantially exaggerated. * Extent of degradation. A more plausible order of magnitude is that about 50% of fields on moderate slopes and 80% of fields on steep slopes experience erosion, and that about one-third of fields on moderate slopes and two-thirds of fields on steep slopes experience productivity problems. Since fields on moderate and steep slopes account for about 30% and 10% of surveyed fields, respectively, the total area affected is much smaller than 75% of the country's surface.6 This is all the more true when it is recalled that the sample is likely to over-represent fields on steep slopes and less favorable soils. Another way to express this result is that about a quarter of farm households farm fields affected by erosion, and about a fifth farm fields affected by productivity problems. * Severity of degradation. Unfortunately, the available data are insufficient to arrive at even order- of-magnitude estimates of the severity of degradation, except to note that, to date, it seems to have been possible to overcome its effects by increases in input use. Regionally, the areas most affected appear to include those in the northern and eastern part of the country (the 'usual suspects') but also sections of the western part of the country. Causes of land degradation To the extent that land degradation occurs, it does so as a result of land use decisions. These decisions are unlikely to be irrational. There is a need to understand why particular land use decisions are made. Without a better understanding of the causes of problems, appropriate policy prescriptions cannot be made. Many explanations have been proposed for the use of degrading practices, including farrner ignorance, insecure tenure, lack of credit, and poverty. We will attempt to examine each of these possibilities to the extent that available data allows. If we assume farmers are rational, then their land use decisions depend on a comparison of the returns they can obtain under each practice available to them. While many cultivation can degrade the soil, action to slow or arrest degradation through changes in crop and management practices or through the adoption of conservation techniques is likely to be costly, either directly in terms of investment requirements or indirectly in terms of foregone production. The critical question is: do the long-term benefits of reduced degradation make these costs worth bearing? The diagram in Figure 4 depicts a stylized choice between two land use practices: a degrading practice and a conserving one. The top panel shows the flow of returns under each practice. Under the degrading practice, yields and hence returns gradually fall. Under the conserving practice, stable yields can be achieved after an initial investment.7 The bottom panel shows the discounted returns to each practice.8 In this panel, 6. Because the FUSADES sample is based on household sample frame rather than a land sample frame, however, no direct estimate of the proportion of the country's surface affected by these problems is possible (see Box 1). 7. In practice, conservation measures might only slow, rather than arrest, degradation. Conversely, some practices might not only stop degradation but result in improvements in land conditions, thus leading to increasing returns over time. The same reasoning would apply to these cases. 14 Annex 7: Land Degradation the short-term costs and long-term benefits of adopting the conservation practice can be compared directly. If the long-term benefits exceed the short-term losses, we expect the practice to be adopted unless a constraint prevents it. Note that off-site effects such as downstream sedimentation, are not included in this analysis. From the farmers' perspective, this is an externality which they have no incentive to take into account. Flgure 4: Incentives to adopt conserving practices: conceptual framework Conserving practice Returns per hectare time Short-term losses Discounted Conserving practice per hecturns rLong-term benefits per hectare Degrading practice time Table 5 shows that farmers in the FUSADES survey undertake a variety of improvements to their fields. This suggests that farmers do undertake improvements which they perceive to be beneficial. Any theory which attempts to explain why conservation measures are adopted must also explain why these other investments are made. Table 5: Improvements to fields Proportion Number offields Total area undertaken byfarmers Improvement (n) (%) (ha) (%) (n) (%) Wells 41 8.7 138.8 15.1 28 68.3 Fences 282 59.7 738.6 80.2 116 41.1 Fruit trees 138 29.2 336.3 36.5 56 40.6 Timber trees 103 21.8 366.0 39.7 31 30.1 Coffee trees 53 11.2 99.1 10.8 15 28.3 Other trees 53 11.2 180.5 19.6 13 24.5 Source: FUSADES Survey 8. Since the choice of which practice to use is made by farmers in light of their own objectives and constraints, the appropriate discount rate to use is the farmers' own subjective rate of time preference. Annex 7: Land Degradation 15 An important factor to bear in mind when considering the causes of land degradation is that the analysis of the previous section suggests that not all farmers are in fact affected by land degradation. The conditions facing these farmers would be represented, in a diagram such as Figure 4, by essentially constant returns over time under their current practices. Under these conditions, they would have no incentive to adopt conservation measures. Thus, we should not expect universal adoption of conservation measures. Farmer ignorance Farmer ignorance, either of the problem or of possible solutions, is often blamed for the use of degrading practices or the failure to adopt conservation measures. This explanation seems unlikely, since it flies in the face of evidence worldwide on farmer rationality. Certainly the extension services available to Salvadoran farmers, and particularly to poor subsistence farmers, have been far from optimal. Nevertheless, management practices aimed at increasing crop productivity (including use of hybrid seed and agrochemical inputs) diffused widely in El Salvador during the 1970s and 1980s (Walker, 1980). In any case, given the numerous projects which have promoted conservation measures throughout the country in the last decades (Box 2), and the evidence from the FUSADES survey that a wide variety of conservation measures are in fact used, farmers seem unlikely to be, at this point, unaware of the options available to them. Inadequate conservation alternatives If the available conservation practices are not very effective, as illustrated in Figure 5A, or are very costly, the short-term losses to adopting them will exceed the long-term benefits. Information on the effectiveness of different measures does not exist, despite the numerous projects which have, and continue to, promote them. Even information on their costs is difficult to obtain. Research on soil conservation practices in Central America suggests that, except under very particular circumstances, expensive conservation measures such as terraces are unlikely to be cost-effective (Lutz, Pagiola, and Reiche, 1994). Although the data in Table 1 are not conclusive, for example, they do suggest that bench terracing is unlikely to be a cost-effective conservation measure. Beans yields were lower under this practice than under either of the other two practices, while corn yields were lower than under the live barriers practice. This is most likely due to the reduction in area cultivated that results from terracing (a cost which is often forgotten). The reduction of soil loss also does not appear to be significantly higher than with live barriers. When combined with the substantially higher cost of terracing, this makes this conservation practice likely to be unattractive under the conditions found in Metapan. The results with live barriers, which consisted of rows of perennial and densely growing plants planted along the contour, appear to be more promising, at least with regard to corn yields, which are higher than under the other practices. The results with respect to beans yields, however, are erratic. More research is clearly needed to establish the productivity effects of different conservation practices. The case of Guaymango, in which widespread adoption of conservation tillage has been achieved, has been widely cited as a success story for soil conservation in El Salvador. In this area of Ahuachapan and Sonsonate departments, a long-standing effort which began in the early 1970s, finally resulted in practically complete adoption of the recommended package of productivity-increasing and conservation measures by the mid- 1980s. A recent review of this case, however, shows that the conservation measures included in the Guaymango package reduced returns from what they would have been had a similar package of productivity- increasing measures been adopted alone (Sain and Barreto, 1996). That the conservation measures were adopted anyway appears to result from their being presented together with the productivity-improving measures (so that their separate effect was not apparent), from making adoption a precondition to the access to credit, and to the fact that while inclusion of conservation measures lowered returns, they remained 16 Annex 7: Land Degradation positive overall. Perhaps agriculture is indeed more sustainable with these conservation measures than without. But is it really desirable to fool (by hiding conservation measures in a broader package) and force (by tying adoption to credit access) poor farmers into adopting conservation measures that reduce their already-low returns from what they might have been? Macro policy framework Macro policy distortions can affect the relative returns to conservation and can, at times, discourage its adoption. No simple relationship exists between the effects of policy distortions on incentives to conserve. In the example in Figure 5B, distortions that lower output prices reduce the incentive to adopt the conservation practice. This is not a general result, however: policy distortions can also encourage conservation (Pagiola, 1996). Figure 5: Effects of different factors on farmers' incentives to adopt conservation practices Dgrading practice g _i Conserving practice r - /Conserving practice ~~~~~~~~~~~~~~ C / t c. 0X n pran practice time time Short-term losses E 2 e2 A/ ' e Short-term losses Degrading practice Long-term Conserving benefitsLogtr C p!9~onserving practice E / Degrading pra time time A. Effect of ineffective or expensive B. Effect of low prices resulting conservation practices from policy distortions Length of lease _ Conserving practice E grading practice E3 O s / C o / ~~~~~~~~~~~~~~~~Degrading pracM cl ca time time Short-ter n losses o Short-term losses 2~~~ ~Co iserving practice Long-term benefits 0 Conserving practice ' X v o @ v Long-~~~~~~~~~~~~~~~~~Logterm benefits Degrading p actice a Degrading practice time time C. Effect of short leases D. Effect of insecure tenure Annex 7: Land Degradation 17 The extent of policy distortions related to trade pricing and marketing of agricultural products has been reduced by the economic liberalization efforts of the 1990s. Although the real prices of several agricultural products have been declining in recent years, this does not appear to be a direct consequence of policy interventions. Tenure problems Tenure problems are a popular explanation for land degradation problems in El Salvador. In principle, tenure can affect conservation decisions in two main ways: * Leases may be too short for farmers to gain the long-term benefits of adopting conservation practices (Figure SC). * If tenure is insecure, farmers are not sure they will be able to gain the long-term benefits of adopting conservation practices. The possibility that they will be evicted acts in the same way as a discount rate, and reduces the perceived long-tern benefits of adopting conservation practices (Figure 5D). A first cut at determining the likely impact of tenure on land degradation can be made by examining the prevalence of different forms of tenure. In the past, such information was scarce, especially in light of the sensitivity of the problem. Table 6 shows the prevalence of different tenure forms found in the FUSADES survey. Despite the conventional wisdom, the vast majority (76%) of fields operated by farm households are in fact privately owned.9 Maps 6 and 7 show the proportion of fields owned and rented in different departments. The proportion of fields owned by their operators tends to decrease, and the proportion rented or share-cropped to increase, from west to east. Share-cropping is only significant in Cabafias and Cuscatlan. As can be seen from Table 6, rented fields tend to be smaller than owned fields. Rented fields are slightly more likely to be on moderate or steep slopes than owned fields (54% of surveyed fields vs 45%), but the difference is not huge. Table 6: Prevalence of different tenure forms Mild Moderate Steep Flat slope slope slope Total Mean (% (% (% (% (% (% (% (% (% (% size Tenure fields) ha) fields) ha) fields) ha) fields) ha) fields) ha) (ha) Owned 81 81 69 82 76 91 70 87 76 87 2.2 Rented 10 13 27 16 18 7 26 11 17 10 1.2 Sharecropped 2 1 0 0 1 0 2 1 1 1 1.2 Rent-free loan 2 0 2 1 2 0 2 1 2 0 0.5 Squatting 5 3 2 1 2 0 0 0 3 1 0.8 Invasion 1 0 0 0 1 1 0 0 1 0 0.9 No answer 1 0 2 1 1 0 0 0 1 0 0.8 Total 100 100 100 100 100 100 100 100 100 100 2.0 Source: FUSADES Survey Note: Squatting and invasion are similar in that farmers are farmers are occupying land that is not theirs, but squatting does not necessarily imply conflict with the owner. 9. While it is likely that some categories of tenure (especially squatting and invasion) are under-reported, these seem unlikely to change the proportion of owned fields significantly. 18 Annex 7: Land Degradation Map 6: Proportion of owned fields N _ 1 l .- 0 ' l b , , - l0 10 20 30 40 50 Kilometers | CABANS ...... 4- ~~~~ M~ORAZAN UNION 60 65 70 75 80 85 90 95 SLT4 Source: Data from FUSADES Survey Map 7: Proportion of rented fields - r^^ Sharecropping | N 2 ,-hXaREffiX (<5% in all other departments) Source Dt f D Annex 7: Land Degradation 19 Table 7 shows the relationship between tenure and use of conservation measures. As might be expected most conservation measures are found on owned lands. Surprisingly, however, some rented lands also have conservation measures. This is a result that had already been found by McReynolds and others (1994) using data from the late 1980s. Upon reflection, however, this is less surprising than it might seem. Conserved land, being more productive, is likely to command higher rental rates than unconserved, degraded land. If conservation is profitable, therefore, it is in the landlord's interest to carry out. Landlords also have an incentive to require that conservation practices be undertaken, since any degradation occurring during rental will reduce the landlord's future income from renting out a field. However, this does not seem to happen. While 59% of rented fields report restrictions on what crop can be grown, only 9% report other kinds of restrictions. The survey did confirm that most leases are relatively short-term and insecure. Only 37% of rented fields had been rented for more than 3 years. Conversely, only 7% of farmers expected to continue renting the same field for more than 2 additional years, while 81% did not know how much longer they would rent the same field. Table 7: Effect of tenure on use conservation measures Cultural Number practices Hedges Walls Terraces Ditches Totai offields (% (% (% (% (% (% (% (% (% (% (% (% Tenure (n) (ha) fields) ha) fields) ha) fields) ha) fields, ha) fields) ha) fields) ha) Owned 356 795 23 30 3 3 8 15 1 1 3 3 35 48 Rented 78 94 40 41 9 6 1 1 47 46 Sharecropped 9 9 56 52 11 16 67 68 Rent-free loan 8 4 13 8 13 8 Squatting 14 14 29 39 7 15 36 55 Invasion 3 3 33 53 33 53 No answer 4 3 75 78 75 78 Total 472 921 27 32 3 2 7 14 1 1 2 3 38 48 Source: FUSADES Survey Credit availability Both the availability of credit and its lack have, at various times, been blamed for causing degradation. Salinas and others (1993), for example, argue that formal credit encourages degradation because (i) it finances production in fragile areas; (ii) it does not finance conservation; and (iii) it encourages the use of agrochemicals. On the other hand, Van Doom (1992) notes that access to subsidized credit was an important factor motivating farmers to participate in the FAO's soil conservation project. According to the FUSADES survey, none of the farmers have asked for credit for the purpose of financing conservation measures. The most important purpose for seeking credit was to finance production (75%), with land acquisition a distant second (13%). (It is interesting to note that land was used as collateral for 34% of loans.) Given this lack of demand, it seems difficult to blame lack of credit of farmers' failure to conserve. On the contrary, available evidence from the many conservation projects carried out in the country seems to indicate that farmers may be willing to adopt conservation measures for the purpose of obtaining credit. Rather than a benefit of credit, conservation seems to be, for many farmers, a cost of obtaining it. 20 Annex 7: Land Degradation Poverty There has been relatively little analysis of the causal links between poverty and land degradation. In principle, poverty can affect conservation decisions in two main ways: * Subsistence needs cannot be met while adopting the conserving practice. * Poor farmers have higher discount rates and so value long-term benefits less. Pagiola (1995) suggests that, since poor farmers tend to be heavily dependent on agricultural production for their livelihood and have few alternative income sources, it is very much in their interest to conserve their soil resources, given the extremely high long-term costs of failing to do so. If the subsistence requirements are sufficiently high relative to what their production technology can provide, however, it is possible that no sustainable practices exist that would allow a household to meet its subsistence requirements. In this case, the household would have no choice but to adopt unsustainable land use practices. Summary Many farmers do employ a range of conservation practices. But many do not. The available data does not allow definitive answers to the reasons for some farmers' failure to adopt conservation measures. It does seem clear that many of the conservation measures which have been promoted, and particularly the more expensive structural measures, may not be cost-effective for farmers. Conversely, on steep slopes cultural measures, though cheap, may not be sufficiently effective. There may, therefore, be a need for additional research on cost-effective conservation techniques, particularly for farmers on steep slopes. That many conservation measures are not cost-effective from the farmers' perspective has been established in many instances throughout the world (in Central America, for example, by Lutz, Pagiola, and Reiche, 1994). Available data do not allow estimates of how many farmers do have adequate measures available to them and yet fail to adopt them. To the extent that such farmers exist, neither ignorance nor lack of credit seem likely to play important roles in their failure to adopt conservation measures. Rental regimes may well provide low incentives to conserve, but they only affect a relatively small proportion of all fields. Whether, and in what way, poverty affects conservation decisions remains to be established. Even though conservation measures are relatively widely adopted, interventions to increase adoption might be justified if constraints or market failures mean farmers under-invest in soil conservation. In El Salvador the main constraints to farmer adoption of soil conservation are usually identified as farmer ignorance, insecure tenure, lack of credit, and poverty. * Ignorance. Ignorance is unlikely to be a constraint. Many farmers do use various forms of conservation, showing that they are widely known. * Tenure. Farm households own three quarters of the fields and almost 90% of the land they operate, so the impact of any tenure problems is likely to be limited. Although there are reasons to expect rental practices to result in under-investment in conservation, primarily because of the short length of most leases, the survey reveals that conservation measures are in fact used on rented lands; indeed, a greater proportion of rented fields than of owned fields have cultural conservation measures (Table 2). * Credit. None of the farmers in the survey had asked for credit to finance conservation measures, although many sought credit for other purposes, so credit is unlikely to be an important constraint. * Poverty. Available data are insufficient to determine whether poverty is an important constraint to investments in conservation. Annex 7: Land Degradation 21 The available evidence, although insufficient to allow a full cost-benefit analysis of the profitability of conservation measures under different conditions, does suggest that farmers make appropriate conservation decisions given the severity of the threats they face and the cost and effectiveness of different conservation measures. Conclusions and recommendations Given the many weaknesses of the available data, even after the quantum improvement provided by the data from the FUSADES survey, general conclusions must be drawn with caution. With this caveat, available data suggest that: a The extent and severity of land degradation problems in El Salvador is substantially lower than is commonly perceived, affecting between 50% and 80% of fields on moderate and steep slopes and about 20-25% of farm households. Even with these revised estimates, however, land degradation remains an important national problem. * Many farmers do invest in conservation measures. Many that do not are likely to fail to do so because of the inadequacy, from a cost-benefit perspective, of available measures. Of the commonly-advanced reasons for farmers' failure to adopt conservation measures, neither ignorance nor lack of credit appear to play important roles. Rental regimes may discourage conservation on rented lands, but this only affects a small proportion of all fields. The effect of poverty remains to be established. The more nuanced picture of conditions that emerges from the analysis in this paper requires a more targeted approach to interventions designed to address land degradation. Unfortunately, however, available data make it difficult to do so. In particular, collection of data along departmental lines or-even worse-regional lines makes it very difficult to identify hot spots. Targeted Policies. The available evidence does not indicate a need for new broad, national-level policies. Nevertheless, there is likely to be a need for targeted interventions to address problems experienced in specific areas. The nature of these interventions will vary according to the specific problems encountered. To identify the necessary interventions, improved data will be required. Data collection. Despite decades of concern over land degradation and a multitude of projects, data that would allow identification of the extent and severity of land degradation problems remain extremely limited. MAG, for example, collects yield data according to administrative boundaries. Because of the very wide diversity of agro-ecological conditions in each department, such data provide very little information on land degradation problems and offer little guidance in the development of appropriate policy responses. A new sample frame is currently being developed for MAG's production data collection. This frame should be designed to collect data based on agro-ecological conditions rather than administrative boundaries. The need to collect data according to agro-ecological conditions is also a key emerging conclusion of the Land Quality Indicator program being jointly implemented by the World Bank, FAO, UNDP, UNEP, and the CGIAR. On-going national data collection must also be supplemented by targeted research efforts aimed at measuring the linkages between different land use practices and long-term productivity. Farmers are the ultimate decision-makers when it comes to land resource management. The available evidence suggests that they are responding rationally to land degradation problems, in light of their incentives and constraints. Land degradation problems are also highly site-specific, and differ substantially from place to place. Under these conditions, national-level policies are blunt instruments which are unlikely to be cost- effective and may well have perverse effects. Effective policy responses need to be carefully targeted at a sub-national level. 22 Annex 7: Land Degradation References Flores Zelaya, 0. 1976-81. Informe Anual de Parcelas para Investigaci6n del Control de la Erosi6n y Escorrentfa Superficial en el Distrito Forestal de Metapa'n. San Salvador: Servicio de Ordenaci6n de Cuencas HidrogrAficas y Conservaci6n de Suelos. Hemrndez Navas, J.R., R. Lazo Mel6ndez, and J. Bueno Alferes, 1994. "Soil Conservation Projects in El Salvador." In E. Lutz, S. Pagiola, and C. Reiche, eds, Economic and Institutional Analyses of Soil Conservation Projects in CentralAmerica and the Caribbean. Environment Paper No.8. Washington: World Bank. Leonard, H.J. 1987. Natural Resources and Economic Development in Central America: A Regional Environmental Profile. New Brunswick: TransactionBooks. Lutz, E., S. Pagiola, and C. Reiche. 1994. "The Costs and Benefits of Soil Conservation: The Farmers' Viewpoint." World Bank Research Observer, Vol.9 No.8. McReynolds, S., T. Johnson, and C. Geisler. 1994. "Factors Affecting Land Use and Soil Management Practices in El Salvador." In E. Lutz, S. Pagiola, and C. Reiche, eds, Economic and Institutional Analyses of Soil Conservation Projects in Central America and the Caribbean. Environment Paper No.8. Washington: World Bank. Pagiola, S. 1995. "The Effect of Subsistence Requirements on Sustainable Land Use Practices." Presented at the Annual Meeting of the American Agricultural Economics Association. Indianapolis, August 6- 9, 1995. Pagiola, S. 1996. "Price Policy and Returns to Soil Conservation in Semi-Arid Kenya." Environmental and Resource Economics, Vol.8, pp.255-271. Perdomo Lino, F.A. 1990. "El Recurso Suelo en El Salvador." Activo 20-30. San Salvador: Fundacion Ecologica Salvadorefia. Sain, G.E., and H.J. Barreto. 1996. "The Adoption of Soil Conservation Technology in El Salvador: Linking Productivity and Conservation." Journal of Soil and Water Conservation, Vol.51 No.4, pp.313-321. Salinas, A.D.R., A.V. Ceren Duefias, and S. Cardoza Lopez. 1993. "Informe Tecnico: Credito para el Pequenio Agricultor de Escasos Recursos en Areas de Laderas, Fragiles y de Altas Pendientes." Project FAO/TCP/ELS/2251 (A): Formulacion de una Politica Nacional de Desarollo Agricola y Rural Sostenible. San Salvador: FAO. van Doom, J. 1992. "Informe de la Mision de Consultoria en Evaluacion del Impacto del Proyecto." Project PNUD/FAO/ELS/86/005: Apoyo Agroforestal a Comunidades Rurales de Escasos Recursos. San Salvador: FAO. Walker. T.S. 1980. Decision Making by Farmers and by the national Agricultural Research Program on the Adoption and Development of Maize Varieties in El Salvador. PhD Dissertation, Food Research Institute, Stanford University. World Bank. 1994. Natural Resources Management Study. Report No.12355-ES. Washington: World Bank. IMAGING Report No.: 16253 ES co