Report No. 12673-CO Colombia Poverty Assessment Report (In Two Volumes) Volume l: The Main Report August 8, 1994 Country Department III Country Operations Division I Latin America and the Caribbean Region MICROGRAPHICS Document of the World Dank Report No: 12673 CO Type: SEC Table of Contents Currency Equivalents and Glossary of Abbreviations ................ x Prk1ifce ................................... xiv Straegic Overview .................................... ns W-si6n Estratgica ..................................... . xxvu Chapter - Cbombia Povcny Profile I. Introduction ......................................... 1 II. Trends in Poverty ...................................... 2 Poverty in Colombia: 1992 ......... .................. 2 Prgress inPoverty Reduction: 1978-1992 ........... 6 The Index of Unsasfied Lsic Needs (NBI) ................. 8 m. A Profile of the Poor ................................... 11 PrNcipal Characristics of Households in Povety ............. 11 Mm Probabiliy of Being Poor .......................... 14 Poverty and the Urban Labor Market ..................... 18 IV. A Profile of Social Indicators ........................... 21 Chapter 2 - Promodng Income-Eaning Opporunit for the Poor: The Macroeconomic Framework I. Bawgrcd ........................................ 28 Poverty Reduction: Major Macroeconomic Challenges ..... ...... 31 II. Structal Reforms, Shocks, Responses and te Poor: 1990-92 ..... ..... 32 Macroeonomic Performance and the Poor: 1990-92 ............ 32 Income Determination: The Role of Policy Reforms and lonomic Shocks, 1990-92 ................................ .. 36 - iv - m. Macroeconomic Policy Challenges in the Medium Term: Adjusting to Cusiana's Oil Bonanza .............. 44 Colombia's Medium Term Macroeconoraic Perspectives: Adjusting to the Oil Boom .45 Proposed Medium-Term Macroeconomic Policy Framework .48 ChaWter 3 - Human Resource Development I. Introduction .52 II. Education ............................................ 53 Sector Overview ........... ........................ 53 Incidence Analysis: Who Benefits from Public Expenditures in Education? ....................................... 58 Main Consraints and Challenges ......................... 63 Conclusions and Recommendations ........................ 66 m. Health .............................................. 69 Sector Overview ............................. 69 Incidence Analysis: Who Benefits from Public Expenditures in Health? . 73 Main Constraints and Challenges .77 Conclusions and ReCOmmendations .79 Chapter 4 - Providing Energy, Water, Sewerage and Shelter to the Poor 1. Introduction .82 ... Energy 83 Sector Overview .83 Incidence Analysis: Who Benefits from Public Subsidies? .88 Conclusions and Recommendations .93 m. Water and Sewerage .97 Sector Overview .97 Incidence Analysis: Who Benefits from Public Subsidies? .100 Conclusions and Recommendations .104 IV. Housing Policy Framework: A Program to Benefit the Poor .... ......... 106 SectorOverview ................................... 106 InidenceAnalysis ................................. 112 Conclusions and Recommendations ........................ 112 - V . Chapter 5 - Rural Programs atd the foor I. Overview of Main Rural Programs .......................... 117 U. Fondo de Desarrollo Rural Integrado (DR) ....fl9.... ......... 119 DRI Spending . .. I 1.2. Distributive Impact .......................... ., 121 Main Constraints and Challenges .In Conclusions and Recommendations ...! II. Plan Nacional de Rehabilitaci6n (PNR) .124 PNR Spending .125 Distributive Impact ..126 Conclusions and Recommendaticns .12 IV. Instituto Nacional Colombiano de la Reforma Agraria (INCORA). ... 127 INCORA Spending.. 128 Distributive Impact ...129 Main Constrnts and Challenges ..130 Conclusions and Recommendations ...131 V. Agricultural Credit Bank (Caja de Cr6dito Agmrio) .132 Lending to Small Farmers. 132 Distributive Impact .133 Main Constraints and Challenges .134 Conclusions and Recommendations .135 VI. Overall Conclusions and Recommendations.. 135 Chapter 6 - Poverty Reduction in a Decentralized Setting: Equity, Efficiency and Accountability L Inroduction .138 I. Recent Evolution of Institutional Arrangements for Poverty Reduction .139 m. A Framework for Institutional Development: Efficiency, Equity and Accountability for Poverty Reduction in a Decentalized System .148 IV. A Strategy for Institutional Development: Strengthening Local Capacity. 156 - vi - Chapter 7 - The Social Safety Net 1. Introduction .................................... 160 II. Colombian nstitute of Family Welfare (ICBF) ..................... 160 lmtitutionl Armngements and Coverge .................... 162 Effectiveness of the Program ........... ................ 63 Policy Recommendations .............................. 168 m. Programs of the Office of the President .169 Evaluation of the Programs .169 Conclusions and Recommendations .170 IV. Social Security .171 Recent Reform of the Social Security System: Law 100 of 1993 .173 Assessment and Recommendations .175 v. Proposa for New Anti-Poverty Programs ........... ............. 176 VI. Overall Recommendations ................................... 178 Bibliography ........................... 181 Map (IBRD No. 18370RI) ........... ................ 191 - vi - - List of Tables, Figures, and Boxes Tables 1.1 Poverty in Colombia - 1992 ................................. 3 1.2 International Poverty Comparisons Incidence of Poverty. 5 1.3 NBIs and the Urban Poor, 1991... 9 1.4 NBIs and the Rural Poor, 1991. 9 1.5 Characteristics of the Poor According to Alternative Definitions, 1991 .11 1.6 Characteristics of the Average Poor Household, 1992 .12 1.7 Incidence of Poverty by Household Group, 1992 .13 1.8 The Probability of Being Poor, Marginal Effects, 1992 .14 1.9 Education and the Probability of Being Poor, 1992 .16 1.10 Household Poverty Over the Life-Cycle, 1992 .17 1.11 Education - Income Differentials, 1990 .20 1.12 Evolution of Education Indicators ............................. 21 1.13 Global andChronic Malnutridon, 1989 .27 2.1 AnnualGrowthRates, 1970-90. 29 2.2 Economic Growtiandits Sources, 1990-92 .34 2.3 The Labor Market, 1990-92 .................. ; ....... 33 2.4 Estimated Income Effects of Policy Reforms and Economic Shocks, 1990 92 ... 37 3.1 Public Expenditure in Education, Selected Years, 1970-92 .4 3.2 Distrbution of Public Educational Expenditure, Selected Years, 1975-90 S 3 3.3 Spending per Student by Level, Selected Countries .56 3.4 Enrollmentby Level, 1992. 58 3.5 School Attendance by Income Level andAge. 59 3.6 Subsidies in Education, 1992 .60 3.7 Distribution of Subsidies to Public Primary Education, 1992. 66 3.8 Distribution of Subsidies to Public Secondary Education, 1992 .61 3.9 Distribution of Subsidies to Public Higher Education, 1992 .62 3.10 Distribution of Subsidies by SENA, 1992 ..62 3.11 Total Educadon Subsidies, 1992 .62 3.12 Growth in Public Speding in Health ........................... 70 3.13 Non-Access to Health Services ........... ............. 73 3.14 Distribution of PHS Services .......... .............. 74 3.15 Distribution of ISS Services .......................7.....4.. 74 3.16 PHS and SSS Subsidy Distribution ....................... 75 3.17 PHS Subsidies by Income Level, 1992 .......................... 76 3.18 Distribution of ISS Subsidies, 1992 ......... ................... 77 4.1 Subsidies in Electricity Tariffs as a Percentage of Cost ................ 85 4.2 Residential Natural Gas Tariffs as a Percentage of Cost, December 1991 ..... 85 4.3 Monthly Energy Used for Cooldng per Lower-Income Household .... ..... 87 4.4 Coverage of Electricity by Income Group, 1970, 1974, and 1992 .... ...... 89 4.5 Distribution of Electricity Subsidies ............................ 90 - vili - Tables 4.6 Housel.,ld Stratification and Income Distribution, 1992 ................ 92 4.7 Water and Sewerage Tariff by Socioeconomic Strata .................. 98 4.8 Coverage of Water by Income Gror.p, 1970,1974, and 1992 ............. 101 4.9 Coverage of Sewerage by Income Group, 1970, 1974, and 1992 ... ....... 102 4.10 Distribution of Water and Sewerage Subsidies ............... 103 4.11 Ihe Housing Problem in Colombia, 1991 ........................ 107 4.12 Program Targets and Observed Performance (Number of Housiag Solutions) ... 109 4.13 Subsidy for a 2 SM Household and the Estimated Required Paymenu for a Housing Solution ....................... 114 Investment Spending by Selected Rural Programs .118 S.'. Distribution of Subsidies by Selected Rural Programs, 1991 .118 5.3 Contribution of Selected Rural Programs to Total Subsidies . .119 5.4 DRI - Program Spending, 1988-92 ..120 5.5 DRI - Distribution of Progra Subsidies, 1991.. 122 5.6 PNR Program Speing, 1988-92 ............................. 125 5.7 DPNR Distribution of Prgram Subsidies, 1991 ..126 5.8 INCORA, Program Spending, 1987-91 ..129 5.9 INCORA - Distribution of Program Subsidies, 1991 . . .129 5.10 Caja Agraria - Lending to Small Farmers, 1988-91 . . .133 5.11 Caja Agraria - Distribution of Program Subsidies, 1991 .133 6.1 Local Govemnmen and Population .. . 142 6.2 Local Govemnien Spending, Seleced Years, 1990-94 ......... ....... 145 6.3 Oranizing Principles Decentralization and Poverty Reduction .... ........ 150 6.4 Rural Program Population and Expenditure, 1992 .................... 153 6.6 Administradve Load for Targeting Poverty . . ..................... 153 7.1 ICBF Expenditure, 1989 and 1993 . . .162 7.2 HCB and CAIP Costs Under Different Assumptions of Program Coverage .... 163 7.3 Daycare Use, 1992 ..165 7.4 HCB Programs among Household with Child Age 2-6 . .166 7.5 Presidential Programs, Budgets and Expenditures, 1990-94 .170 F1gUres 1.1 Poverty Incidence, 1992..................................... 3 1.2 Regional Distnbution of the Poor, 1992 ......................... 4 1.3 Evolution of the Inidence of Extreme Poverty ..................... 6 1.4 Distribution of the Poor According to NBIs, 1991 ................... 10 1.5 Probability of Being Poor and Number of Children in Household, 1992 ...... 15 1.6 Gender and Poverty, 1992 ................................. 15 - ix - Figures 1.7 Regional Difference in Poverty (Head Count Index), 1992 ............... 17 1.8 Labor Force Participation, 1990 .............................. 18 1.9 Labor Force Participation Rate, 1990 ....... ................ ... 19 1.10 Life Expectancy at Birth for Men and Women .............2........ n 1.11 Colombia - Infant Mortality Rates .......................... 22 1.12 Evolution of Fertility Rates by Zone ........... ................ 22 1.13 Under-Five Mortality Rates ................................. 24 1.14 Life Expectancy, by Region ................................. 25 1.15 Infant Mortality Rates ................................... 25 1.16 Fertility Rates by Level of Education and Zone, 1987-90 .26 2.1 GDP Growth Rates at Constance Prices, 1970-90 .29 2.2 Nominal Interest Rates, 1990-93 .33 4.1 Monthly Electricity Charges, Costs and Subsidies per Legal Urban Household, by Strata, 1992 .91 4.2 Monthly Gas Charges, Costs and Subsidies per Legal Urban Household, by Strata, 1992 .92 4.3 Monthly Water and Sewerage Charges, Costs, and Subsidies per Legal Urban Household, by Strata, 1992 .98 4.4 Supply of Housing Solutions, Collective and Individual .111 6.1 Municipal Property Tax Income .141 6.2 Local Government Revenues, 1992-2000 .141 6.3 Situado Fiscal per capita, by Region .145 Boxes 3.1 Escuela Nueva - A Successful Program .,erving the Rural Poor .57 4.1 Switching from Cocinol to LPG .87 5.1 Rural Development Investment Project, A Decentalized, Demand Driven Program of Rural Investments .121 6.1 Study of Regional Practices in Grant Design. 146 6.2 Demand Verification and Pricing Signals in Mexico's PRONASOL .149 6.3 Recommendations for Improved Efficiency .151 6.4 Recommendations for Improved Equity ....... .................. 152 6.5 Recommendations for Improved Accountability .154 7.1 Characteristics of Principal ICBF Programs ........................ 161 7.2 Factors Affecting Use of HCB Services .............. .......... 166 7.3 Estimated ISS Deficits and Obligations, 1992 .173 7.4 Estimating the Costs of a Social Safety Net .179 Currency Equivalents (As of April 14, 1994) Currency Unit = Peso (Col$) Col$1 = US$0.00121 US$1 = Col$827.96 Fiscal Year January 1 to December 31 Glossary of Abbreviations AGETIP Senegal's Public Employment Program (L'Agence d'execution de travaux d'interest public contre le sous-emploi) Apertura Opening of the economy BATUTA Music Education Program, Office of the President Bienestina Fortified flour made and distributed by ICBF CAIP Day Care Centers (Centro de Atenci6n Inegrl al Preescolar) Caja de CrEdito Agraria Agricultural Credit Bank CAJANAL Social Security for Governent enployees (Caja de Compensaci6n Nacional) Caprecom Social Security for Employees in Communications (Caja de Previsi6n Social de Comunicaciones) Caprecundi Social security for employees of Cundinamarca Dqment (Caja de Previsi6n del Departmento de Cundiamma) CASEN Household Survey CCI Community Child Centers (Centros Comunitarios pam la Infancia) CCRP Regional Population Research Ceter [NGOJ(Corporaci6n Centro Regional de Poblaci6n) CDRs Municipal Councils for Rural Developmem (Consejos de Desarrollo Rural Municipal) CELGAC Cundinamarca's Electricity Utility Company CEPAL Economic Commission for Latin America (Comisi6n Econ6mica para Latinoam6rica) CERT Tax credit certificate (Certificado de Reembolso Tributario) Cesantis Redundancy rights Cocinol Gasoline for cooking purposes COLGAS LPG Distribution Program CONPES National Council for Economic and Social Policies (Consejo Nacional de Politica Econ6mica y Social) - xi - CORPES Regional Council for Economic and Social Policies (Consejo Regional de Politica Econ6mica y Social) Corporan6nimas Social security for corporate employees (Caja de Previsi6n de las Corporaciones y Sociedades An6nimas) DANE National Statistics Department (Departamento Administrativo Nacional de Estadistica) DAPRE Administrative Department, Office of the President (Departamento Administrativo de la Oficina de la Presidencia) DBRS Division of Basic Rural Sanitation, INPES DFIET Division of Institutional Strengthening, DNP DNP Department of National Planning (Departamento Nacional de Planeaci6n) DRI Integned Rural Development Program (Desarrollo Rural IntegraL..)) DTF Coloznoian .ierest Rate Index (Dep6sitos a T6mino Fijo) ECOPETROL Colombian Oil Enterpri'e (Empresa Colombiana de P6troleos) EDP Essential Drugs Pi gram EMP Economic Modernization Program EMPOS Water and Sanitation Companies (Empresas de Obras Sanitarias) Escuela Nueva New School Program ESMAP Energy Strategy Management Program (World Bank) FAG Agriculture Guarantee Fund (Fondo Agricola de Garantia) FAMI Program for the Family, Women and Infants (Familia, Mujer e Infancia) FAO Food Agricultural Organization FECODE National Teachers' Union FEDESARROLLO Foundation for Higher Education and Development (Fundacion Para la Educaci6n Superior y Desarrollo) FES Foundation for Higher Education (NGO) FFAP Fondo Financiero Agropecuario FINAGRO Fund for the Financing of the Agricultural Sector (Fondo para el Financiamiento del Sector Agropecuario) FINDETER Municipal Development Bank (Financiera de Desarrollo Territorial) FIS Social Investment Fund (Fondo de Inversion Social) FNC National Coffee Fund (Fondo Nacional del Caf6) FNCV National Fund for Local Roads (Fondo Nacional de Caminos Vecinales) Fondo DRI Integrated Rural Development Fund (Fondo para el Desarrollo Rural Integrado) FSES See PPFSES GIS jeographical Information Sy. cm HBF Family Welfare Prograan (Hagares de Bienestar Familiar) HCB Community [cnild care] welfare center (Hogar Comunitaio de - xii - Bienestar) HIMAT Colombian Institute for Hydrology, Meterology and Land Preparation (Instituto Colombiano de Hidrologfa, Meteorologia y Aderuaci6n de Tierras) ICA Colombian Agricultural Institute (Instituto Colombiano de Agricultura) ICBF Family Welfare Ihstitute (Instituto Colombiano de Bienestar Familiar) ICFES National College Entrance Examination ICT National Construction anid Mortgage Company (Instituto de Cr6dito Territorial) IDB Irnter-American Development Bank IDEMA Agricultural Marketing Institute (Instituto de Mercadeo Agropecuario) IFS International Financial Statistics, IMF IMF International Monetary Fund INCORA Colombian Institute for Agrarian Reform (Instituto Colombiano de Reforma Agrria) INDERENA National Institute of Renewable Resources and the Environment (Instituto Nacional de Recursos Renovables y del Medio Ambiente) INEA National Institute for Alternative Energy (nstituto Nacional de Energias Alternas) INPES Natonal Institute for Specia Health Programs (nsttuto Nacional Para Programas Especiales de Salud) INSFOPAL Institute for Municipal Development (Instituto de Fomento Municipal) INURBE National Institute of Urban Reform and Low-income Housing (Instituto Nacional para la Reforma Urbana y La Vivienda de Interds Social) IPC Community Development Program anstituto de Programas Comunitarios) ISS Institute of Social Security (Instituto de Seguridad Social) 1MP See PPJMF JNT National Tarriff Board (Junta Nacional de Tarifas) kBtu Thousand British Thermal Units kWh Kilowatts per hour LPG Uquified petroleum gas MAM See PPMAM Misi6n Social Special commission in arport of dalization and targetui of social services (Misi6n de Apoyo a la Descentralizacifn y la Focalizaci6n de los Servicios Sociales) MOA Ministry of Agriculture MOH Ministry of Health - xiii -. MOPT Ministry of Public Works NBI Index of Unsatisfied Basic Needs ODEPLAN Chile's National Planning Office PDI Program for Municipal Institutional Development (Programa de Desarrollo Institucional Municipal) PDIC Integrated Campesino Development Plan (Plan de Desarrollo Integral Campesino) PHS Public Health System Plan DIA Targeted Nutrition Assistance Program (Plan de Desarrollo Integ.al Alimenticio) PNR National Rehabilitation Plan (Plan Nacional de Rehabilitaci6n) PNUD Umted Nations Development Program (Programa de las Naciones Unidas Para el Desamrollo) PPFSES Fund for solidarity and social emergencies (Fondo de Solidaridad y Emergencia Social) PPJMF Presidential Program for the Youth, Women and Family (Programa Juventud, Mujer y Familia) PPMAM Medellin Metropolitan Area Presidential Program (Progma Presidencial para Medellin y su Area Metropolitana) PPP Purchasing Power Parity ProFamilia Family Planning Institution PRONASOL Mexico's National Solidarity Program (Programa Nacional Para la Solidaridad en Mexico) Revolucift Pacfifca 1990-94 Economic Development Plan SENA National Vocational Training Service SIF Social Investment Fund SINTAP National System for Transfer of Agricultre Technology (Sistema Nacional de Transfeencia de Tecnologia Agropecuaria) SlTC Standard Intemational Trade Classification SSPD Regulatory Agency for Public Services (Superintendencia de Servicios PNblicos Domiciliarios) Tcal Tera-Calories (Tera-calorias) UAP Unit of Land for the Rural Household (Unidad Agricola Familar) UAP DNP unit in charge of water and sanitation (Unidad de Agua Potable y Saneamiento) UDA DNP unit in charge of agarian development (DNP, Unidad de Desarrollo Agrario) UDT DNP unit in charge of inter-governental relations (DNP, Unidad de Desarrollo Territorial) UNDP United Nations Development Program Preface Sustainable poverty reduction is the World Bank's main objecdve. The Bank's approach to poverty reduction has evolved, reflectng both the progression of development thinking and the lessons of experience. Two recent Bank publications promoted the policies and practices found to work best. Based on cross-country analysis, World Developnen Report 1990 concluded that the countries most successful in reducing poverty have followed a two- part approach. They have promoted policies tat emphasize efficient growth and make use of labor, the poor's most abundant asset. And they have promoted expenditures and institutions that provide equitable access to social services, including education and health care. World Developmert Report 1990 also emphaized the importance of a social safety net to protect society's most vulnerable groups. The 1991 policy paper, Assistance Strategies to Reduce Poveny, established guidelines on how the Bank should promote the two-part approach in its policy dialogue, analytical work, and lending. Asstance Strateges recommended that poverty assessments be prepared for all active borrower countries to provide policy advice and help design Bank-pported programs. In its simplest form, the porty assessment addresses ree questons: Who is poor? Why are they poor? And what should policymaker do about it? This report presents the poverty asessment for Colombia. The report has three parts: stategic overview, main text, and anexes. hne strategic overiw identifies areas where the government and the Bank need to focus to ensure future progress in poverty reduction. It offers a context for the assessment's different elements and a sense of priorty among policy options. The main report has swven chapters detaiing the analysis and recommendatons of both Bank and government teams regarding: the nature and extent of poverty in Colombia, the effectiveness of economic management, the adequacy of govement efforts to provide basic social and infrastructural services to the poor, the effectiveness of the instutional framework, and the availablity of a scia safety net. The annexes provide methodological and informat support for the report. The World Bank and Government of Colombia took the oppornity to work together in desgning this study. While the Bank was seeig to produce a poverty assessment, the Misn Socal in DNP warted to conduct a study on the incidence of government expenditure. Juan Luis Londof o (then Deputy Director of DNP) and Emnesto May (task manager for the poverty assessment report) agreed to set up joint Bank/DNP teams to work on both studies. An agreement to produce two paralle publications was formaized In April 1993. The Government of Colombia hired a group of consultants to provide background reports on educa, health, pubic housing, childcae, public utility services, rural programs, and the programs of the Office of the Presdent. The Bank assigned sector specists to work closely with the consultants. DNP also appointed counterrs to collaborate with the Bane's spedalist Carlos Eduardo Velez Edcavarfa was hired to head the Miln Social's effort, and a special module an socl was included in the national household survey of September 1992. - Xv - The Bank mission visited Colombia on May 9-21, 1993. The mission was led by Emeto May and integrated by Ariel Fisibein (deputy mision leader), Jacques van der OWag, Wilam McGreevey, Tim Campbell, Eleanor Shreiber, Isabele Girardot-Berg, Anne- Marie del Castillo, Jayme PortCarreiro, nakoor Persaud, and Guilherme Sedlacek. Mbisson members worked closely with DNPIMisI6n Social counterparts as well as with tIhe consultants. See tie list of joint World Bank/DNP worldng teams. The poverty report was prepared by tie Bank team on the basis of worldng papers drafted by the Colombian consultants and the outcome of the mission. In February 1994, the first draft of the report was discused wift all those involved. Ten eams prePared background repot for the studies. Frai. sco lasso and Hnando Moreno Guerrero produced the statscal informadon in the povrty profile. Teams from FEDSARROLLO prepared the reports on educafion (Carlos Geardo Molina, Doris Polanfa, and Mariclo Alviar), health (Carlos Gerardo Molina, Ursula Giedidn, Maria Clara Rueda, id Maurico Alviar) and macr cnmic policy (duardo Lora and Ana Maria Herrera). ECONOMETRIA (Alvaro Reyes Posada and Marfa Gloria Cano) prepaed the reports on eltity and gas and on water and sanitation. CENAC (Fabio Giraldo IsaLa and Jorge Enrique Torres) ptepared the report on housing. Juan Jos6 Perfetti and Marfa del Rosano Guerfa de Mesa wrote about nrral progams. Carmen Elisa Florez and Regina M6ndez prepaed tie report about ICBF, and INVAL (Alvaro Toledo Romero, Lucero Zamudo Cirdenas, and lbardo Sarmdento Anzola) prepared te report on the progams of the Ofaice of the Presdent The following people also contributed significntly: Demnis de Tray, Darnie Morrow, Scott Quehl, Helena Rbe (pee reviewer), Jeffrey Hammer (pe reviewer), Mada Elena Maldonado, Francisca Castro, Ana Mara Llorente, Juan Manuel Rojas, and Carlos AIberto Media At various stages, the teams benefitted gray firom the comments and advice of many people: Juan Luis Londoflo, Hernando Jos6 Gomez, Cecilia Marfa Velez, Tarsico Castalleda, Hugo L4pez Castalo, Nelson Ortiz, Aida Soler, Mguel Urrutia, and Kristine lbegt. Bruce Ross-Larson provided the final editng, working with Heather Cochran and Paul Holtz. Rafael E. Flores and Elena B. Rodriguez provided excelent seruarial asusitne. - xvi. Joint Word BankIDNP Woridng Teams Taut WeddsBan Cr,euadftat DNP/MIsId SeWd mow Ernesto May (Mission Leader) Carlos Eduardo V6elz (General Coordinator) Povaty Proftie AnsI FRszbeln IDeputy Mission Ledr) Fracsco Lasso Carbs Albato Medina (Technical Assisnt) Gaurav Datt H4emndo Moreno Gurrro Macroeconomico Framework Emesto May FEDESARROLLO, Eduawdo Lois Cas Esteban Posada Ana Made lererH Hemando Zuteta' Labao Marlts Guhrm Sedlacek H"el Jacques van der Gasg FEDESARROLLO: Cados Garafdo Mot Clara Guzmrn - Eleanor Schebe Mats aaCr Rueda Hectr Fablo Medina Education Jacqs van der Geag FEDESARRDLLO: Carlos Gerad Molha Camn Helena Vergara- Mcbal Poetshnik P.rgait Peii* Eetity end Gas Jayrne Porto-Carelro ECONOMETRIA: Alvaro Ree Posada Eva Mari Udibe* Mrt Gltoda Cano Sonib Rojas- Water and Santtion Thakoot Persud ECONOMETRtIA: Alato Reya Posada Eva Mar Urbe^ Maria Glori Cano Ruben Datto Avendallo* Houing Thakoor Peraud CENAC: Pablo Gircido Isaz Dego Fernfndezo Jorp Enrique Tonfes Ramnrez Luz Stella Gonzhlet Andres Escobar Urb Rul Progam abolbs Girardot-Beri Juan Jos4 Pentett del Conal Cara Gonzilez Anne Mari del Cstillo Consult Made del Roscuo Guenra de Mesa tecenhtIon Thn Campbel Sol Safet Not Willam MeGrevy ICIIF Carmen Esa F16tez beetr Pbza- Regia Mnt" Mirim Gutlne Progas of th Presidncy INVAL: Atvao Toedo Romero Nancy Tirado* Lcero Zamudio Ckdenm Libardo Sermkent Anzoa *Pe Reviewers ased by OW Strategic Overview Povery Reduction Strategy: Steering for measures that will enhance rural Colombia toward Growth with Equity incomes. This need is accentuated by the recent deterioration of rural incomnes and The recent Cusiana and Cupiagua oil dis- the economy's anticipated adjustment to edpoverty coveries give Colombia an excellent the oil boom. A proactive policy frame- redon strategy opportunity to move onto a higher work must ensure that the rural sector needs tofocus in growth path. Colombia has the resources and the poor share the oil windfall gains. three priority to pursue a coordinated poverty reduction Work on a coherent rural develop- areas: 1) rural strategy. It now needs effective and effi- ment strategy, notably absent in the gov- development cient human resource policies and target- emient's 1990-94 Development Plan, 2) social and ed poverty programs-in short, a strate- has just been initiated. Previously, the sttural gy. In devising and implementing that authorities focused on agricultural poli- servcsran strategy, policymakers must give priority cies and rural expenditure programs. The serwcest and 3) to three areas: (1) rural development, (2) Ministry of Agriculture attended to agri- decentralizatio social and infrastructural services, and cultural policy. And the implementing capacity buil8ing. (3) decentralization, capacity building, agencies of the rural expenditure pro- and institutional and institutional strengthening. It will grams (Integrated Rural Development strngthening also be important to improve the moni- Program, National Rehabilitation Plan, toring and evaluation systems for all Colombian Institute for Agrarian Reform, poverty reduction programs. and Agricultural Credit Bank) worked The pace of institutional reform in directly with local governments-but Colombia's development strategy is a without adequate coordination to develop major challenge. The policies are gener- an institutional framework conducive to ally well conceived but need further defi- broader rural development. nition to become fully operational. So far The first steps have recently been the government has seriously underesti- taken. The functions of the implementing mated the difficulties of implementation agencies for rural expenditure programs and institutional change. The government have been streamlined and specialized, must now specify the implementation making them more complementary. details of its social sector development Placing one institudon, the Integrated program. Rural Development Fund, in charge of financing infrastructure and services for Rural Development the productive rural sector assures greater consistency in formulating programs and Extreme poverty in Colombia is mainly a implementing actions in the field. By rural problem. Seventy percent of the. bringing together municipalities. rural people with incomes below subsistence communities, and public officials to coor- live in ural areas, underscoring the need dinate local rural development projects in ' xviii - the Municipal Councils for Rural Rural) to support rural capital expendi- Develop-ment, the government has framed tures by the private sector. a more integrated vision of rural develop- The trade protection and domestic ment. Appropriate consideration can now support mechanisms in agriculture, in be given to the links among national, addition to hindering competitiveness, regional, and local rural projects and are inefficient instruments for protecting between social and productive sectors. the rural poor. To reduce poverty, the Consistent with a more comprehen- government should reverse the recent The basisfor rural sive and coordinated rural development backsliding in trade policies, eliminate development is strategy, the government needs to pro- agriculture price and investment subsi- vide adequate incentives for the agricul- dies, and redirect resources to targeted afficgemagrwdth ture sector, invest in economically attrac- rurl programs. growth tive rural infrastructure projects and ser- To aid the poor during the current vices to support production, and elimi- period of stress and to begin a transition- nate the urban bias in the delivery of al scheme for reallocating agricultural social services. But improving the living resources, the proposed Emergency Rural conditions of the rural poor is not all that Employment Program should be put in can be done. Because migration from the place, evaluated, and expanded as appro- rural to urban areas is another mecha- priate. As proposed, the pilot employ- nism for increasing living standards, the ment scheme is to be coordinated by the government should ensure the mobility Ministry of Agriculture and managed by of labor. It should eliminate distortions in the Integrated Rural Development the labor market and provide high-quality Program. Particular care should be taken basic rural education and health care. to ensure efficient labor-intensive tech- niques are used in the proposed rural Incentive Frameworkfor Agriculture infrastructure projects and to set wages in the program below market rates. This The basis for rural development is effi- will ensure targeting through self-selec- cient agricultural growth, possible only tion and give an incentive to leave the by removing today's price distortions. program as soon as alternative work The governmen, made significant opportunities develop. To maximize the progress in liberalizing the sector, dis- impact of the employment program, it mantling the Agricultural Marketing should be executed by community Institute (IDEMA) and relaxing agricul- groups (such as Juntas de Acci6n ture supports. But these advances were Communal, Empresas Solidarias, reversed in 1993. In response to the women's groups) with funds from 1991-92 crisis in agriculture, the govern- municipalities, the Integrated Rural ment increased its trade protection, reini- Development Fund, the National tiated market interventions to support the Rehabilitation Plan, or the Fund for purchase of important crops-IDEMA's Solidarity and Social Emergencies. purchase of agriculture products doubled between 1992 and 1993-and extended Rural Infrasructre and Services the date for bringing interest rates to small farmers closer to market levels. Is Increasing rural productivity through also created an investment subsidy economically attractive rural infrastruc- scheme (Incentivo a la Capitalizacidn ture projects is one of the most effective - xix means to reduce poverty in Colombia. have been impressive in improving the Programs to improve rural infrastructure, school environment to increase learning services, and land consolidation are and community involvement. If the important instruments of poverty reduc- Escuela Nueva is to continue to function tion. The challenge is to identify, design, on a national scale, it will require and implement projects with high eco- renewed commitment from senior policy- nomic and social rates of retum, while makers. In health, the clinic-based care encouraging private sector participation. system appears poorly suited to the rural In the short to medium term, the most poor. A model based more on out-reach Resources should important constraint in rural project services should be considered. In addi- be channeedfrom delivery is low implementation capacity, tion, a reasonably attractve career pack- exising subsides a constraint that should c- -cled imnme- age for a permanent, rural-oriented pub- eg husing diately. As local and depa aiental opera- lic service needs to be developed. for energy. houstng tional capacity is enhanced, additional and agricultural resources could be effectively channeled Socal and Infastructural Services price support through direct transfers to rural munici- systems toward palities or through the cofinancing sys- Although Colombia has made impressive basic education, tem. With the added resources, special gains during the last two decades in health cae, na safeguards and incentives will be expanding services, much remains to be and required to ensure that national poverty done to improve the access and quality of reduction goals are reflected in planning services for the poor in both urban and most to and budgeting. In this context, the rural areas. Many poverty programs have Integrated Rural Development Fund pro- mainly benefited the better-off, and elim- grams can play a special role to channel inating these subsidies would free up additional funds, develop capacity at the considerable resources to fmance more local level, influence the allocation of poverty reduction. In particular, local resources, and promote specific resources should be channeled from projects in support of the poor. existing subsidies for energy, housing, and agricultural price support systems Delivery of Socil Services in RuralAreas toward basic education, health care, rural infrastructure, and social assistance to the The design and delivery of social ser- most vulnerable. vices are biased toward urban centers. Correcting this bias requires tilting the Education balance of social spending toward rural areas and-more important-modifying The education system is at a crossads. the delivery modes for such services as The decentralization effort will pose education and health so that they more opportunities and challenges and requires closely respond to the special circum- a genuine commitment to improve edu- stances of the rural population. cation, reflected in govemment priorities Education is the only sector that has and expressed in a Minister of Education made a concerted effort to develop a who embodies these principles. With an rumral model for the delivery of services. overall coverage rate of roughly 86 per- The results of the program, Escuela cent at the primary level and nearly 50 Nuev, are very positive. Achievemenu peren at the secondary level, Colombia -xU- is still far from the constitutional man- expenditures. A centerpiece of the gov- date of universal basic education to age emment's new strategy in the sector, as 15. Low coverage in basic education presented in the recent Social Security remains a problem among the poor and in Law, is the universal provision of a stan- certain areas (most notably, the Atlantic dard package of health services, begin- Building local region). Furthermore, the overall quality ning with primary services. The goal is to of basic public education is extremely deliver a basic health package to all politic al comftlnent low. Colombians, regardless of their income, to education is as The projected growth in central gov- by subsidizing the poor. But the imple- critical as building einment transfers to departments and mentation details of the law are incom- capacity municipalities should substantially plete or evolving, and they promise to be increase the resources available to difficult to institutionalize. regional and local govemments for edu- A commitment to continuity and flex- cation. Although these expanded ibility in pursuing universal basic health resources should theoretically lead to coverage over the medium to long term increased access to education, achieving will be required from the next adminis- this objective will depend upon local tration and from future governments. The government commitment. Building local Minstry of Health needs to be strength- political commitment to education is as ened to aggressively guide the system critical as building capacity. National, and set quality and efficiency control departmental, and municipal authorities standards, particularly important as the must enhance the quality of public edu- system moves to integrate public and pri- cation by raising the educational require- vate health providers under the new ments of new teachers. They must move framework. toward a system that rewards perfor- Notably absent from the govern- mance rather than tenure. They must ment's plans for basic services are verti- improve the availability of instructional cal programs for immunization and materials in schools and provide ade- endemic disease control, which are espe- quate facilities to limit overcrowding, cially important for the poor. Vector- and ultimately eliminate the three-shift borne diseases are still prevalent in system in public secondary schools. urban slums and coastal and river valley areas. Malaria and dengue fever control Health programs have declined in both efficien- cy and effectiveness in recent years. Most of Colombia's poor rely on the pub- Mass immunization campaigns have lic health system. Present service use and proved effective but not yet achieved expenditure outcomes indicate sufficient universal coverage. The Ministry of quality services are not reaching this Health needs to retain service responsi- group. Effective care through the public bilities for these programs and actively system reaches only 35 percent of the pursue their implementation. To enhance population, compared with the target of their effectiveness, the rural out-reach 60 percent. Overall expenditure is insuf- strategy should prove critical. In the ficiently targeted toward the poor. The decentralized provision of health ser- richest 40 percent of the households ben- vices, local governments need to be dis- efit from almost one-third of the couraged from overinvesting in health I xxi I facilities whose recurrent costs may not liquid petroleum gas and kerosene appli- be sustainable. Critical elements in the ances available through Community search for better quality will be holding Councils with subsidization. Network local officials accountable for the effec- extensions of natural gas could be subsi- tiveness of services and strengthening dized in poor neighborhoods through low user demand. hook-up charges and efficient two-burner stoves. For rural households, extending Energy the electricity network may be too expen- sive, and isolated generating units could Cridcal elements in For many years, the strategy for setting provide basic lighting and communica- the search for beter electricity tariffs was anchored in the tion services. Small communities can be quality will be belief that by maintaining low prices, served by installing generating equip- qa will be poor households would have access to the ment such as diesel motors. hogkalls services. Evidence indicates, however, accountabe for the that the distribution of the massive subsi- Water and Sewerage effectiveness of dies (representing 1.4 percent of GDP) services and has been slightly regressive and thus not Lack of access to water and sewerage is strengthening user primarily beneficial to the poor. Although mostly a problem of the poor in demad the government recently increased elec- Colombia. For both services, coverage of tricity tariffs a%d reduced stfatification (a the poorest fifth of the population in poorly designed pricing scheme that 1992 was well below the countly aver- allegedly charges residential users on the age: for water, it was 65 percent (com- basis of income), the process needs to be pared with the average of 83 percent) and accelerated. The government should fur- for sewerage it was 37 percent (com- ther simplify or eliminate stratification, pared with 67 percent). This lack of cov- establish a more general residentil tariff erage of the poor is not explained by an based on marginal cost pricing, and com- unwillingness to pay. In the main cities, plement it with targeted social tariffs for households in the lowest income decile low income users. The resources attained spend up to 9 percent of their income on through changes in public pricing could water and sewerage, while the population reduce regressive taxes or finance more at large spends 1 to 2 percent. targeted interventions. Increasing coverage in water and Complementary measures will be sewerage goes beyond the need to follow required to accompany the tariff adjust- the pricing policies for cost recovery and ments. With up to 11 percent of poor targeted subsidies recommended in the urban household income spent on energy, report It requires meeting the needs of economical alternatives for household scatered rural populaons or of residents energy are critical. Natural gas, liquid in inaccessible, unserviced, periurban petroleum gas, and kerosene are alterna- areas. In the urban centers, this must be tive fuels to be promoted along with the done without crating incentives for oth- electricity tariff reforms. Market forces ers to settle in even more dificult tin. can accomplish the switch from electrici- Apart fiom technological constraints, the ty, and the govemment should make wre cost per household of traditional services these brces are at play. For low-incom in these areas is prohibitive. hcoseholds, the. govenment could make This ,drings two issues into focus. -xlii - Intermediate technologies should be used for very poor and vulnerable groups, and for residents in inaccessible or very poor developing such a mechanism must be a areas, whether in peripheral settlements priority. near urban centers or in dispersed rural Resources should be made available zones. And housing policies should pro- by eliminating subsidies (such as those vide more access to basic services and for public utilities) or other programs that avoid further proliferation of unautho- mainly benefit the wealthy, or by assign- rized urban setdements. Tie government ing a small fraction of the expected oil Lack of coverage should promote such altematives as com- revenues. Experience elsewhere suggests of e poorfor munity standpipes, cisterns, tanker that, if efficient programs can be trucks, and water purification tablets as designed, an appropriate level of spend- energs wategI solutions for today's unserviced areas. It ing for social protection (leaving aside and sewerage is should explore ways of upgrading the traditional responsibilities for education not exPlained by quality of service provided under such and health services) would range from 1 an unwlngness alternatives by encouraging private sec- to 2 percent of GDP. In assigning to pay tor participation through a bidding resources to the very poor and vulnera- process that would assign both lucrative ble, the government must ensure tAat and nonlucrative neighborhoods to be resources are not wasted on ineffective serviced by the bidders. A subsidy of the projects. Bureaucratic red tape must be same magrnitude as that provided to low- kept to a minimum, and the bulk of social income households for conventional ser- spending must reach those in greatest vices could be awarded to ensure service need. in the poorest neighborhoods. Within the Restructured and refocused, the pro- existing housing policy of direct subsi- grams run by the Family Welfare dies to the poor-to purchase a new Institute (ICBF), the Office of the house, improve an existing one, title President, and the Social Security System land, or access sanitation services-the could form a social safety net providing government should direct more resources more efficient and equitable assistance to to finance urbanized and urbanizable lots those in greatest need, substantially alle- and water and sanitation works. viating poverty. New options for poverty programs worth considering are Plan Social Assistance to the Most Vulnerable DIA (a food stamp program) and the Groups Emergency Rural Employment Program In designing these programs, the govern- The poverty reduction strategy should ment observed five rules: direct most of ensure that the poor gain from-and con- the subsidy to beneficiaries, minimize the tribute to-gowth. But some poor peo- need for additional staff or management, ple, especially the old and the infirm, will induce self-targeting and establish auto- not be able to respond to new oppomtni- matic exit points for recipients, promote ties. And others, although benefiting, will complementarities between services, and still have incomes inadequate for basic test the program as a pilot project and necessities, or be vulnerable to income- evaluate results before implementation. reducing shocks such as drought or loss As the government develops its social of the family breadwinner. Colombia safety net, it must follow these guiding does not yet have an effective safety net principles. Only after careful evaluation - xxiii - of the benefits and costs of such pilot governments, organized interest groups, projects should any program be institu- and private sector-can reach a profi- tionalized. ciency in decisionmaking and manage- ment commnensurate with the sweeping Decentralization, Capacity Building, expansion of subnational government and Institutional Strengthening responsibilities under the new constitu- tion and legislation. The proposed strate- The effectiveness of Colombia's social gy incorporates required changes at all expenditures and poverty reduction levels of government: central, depart- Wuhdece6 a&4dM efforts depends on a precarious balance mental, and municipal. One critical com- poverty reduction of power between central and local gov- ponent is to transform departments into and insttutional ernments. With the decentralization of its the intermediate links between the central deont hav public sector, Colombia is shifting to government and municipalities. A second deelopmenthave local governments the responsibility for is to open public administration to the become inseparabk delivering social programs and many private sector, NGOs, and organized infrastructural services to the poor. Local communities in planning, financing, pro- governments have eagerly embraced the gram implementation, training, manage- new powers and autonomy in spending ment, and evaluation. A third is to conveyed in recent legislation, particular- enhance human resources at all levels of ly 1993's Ley de Competencias y government. Recursos (on local government functions Given the magnitude of the task, the and financing). But the effectiveness and implementation of the strategy should efficiency of social spending and man- start with pilot projects in each area-to agement are seriously in question in all make more effective use of limited insti- but a handful of cities and departments. tutional capacity. The projects should To induce effective spending, the central continuously test intergovernmental government is developing an incentive coordination-and include evaluations at framework for local authorities. Properly each stage as well as strategies to expand defining this framework is critical for the application to more recipients. success of decentralized services in reaching the poor. Define an Incentive Frameworkfor Local With decentralization, poverty reduc- Governments tion and institutional development have become inseparable. The "War on Essential programs in health, education, Poverty" clearly requires a strategy for water, sanitation, and social assistance institutional development an overarching will be delivered by municipal govern- vision, explicit long-term objectives, a ments pursuant to centrally mandated central body responsible for husbanding dictates, but strongly influenced by local subnatiornal institudons, greater capacity priorities. Making these arrangements among local governments to formulate work depends on effectively deploying programs of local public spending and to the instruments of decentralization. The evaluate their poverty impact, and atten- incentive framework now being designed tion to professionalism. This vision calls combines both fnancial incentives and for an entirely new level of technical administraive controls: intergovermmen- capacity so that local communities- tal revenue transfers, matching grants, - xxiV - and regulatory provisions. By them- Open Public Administration to Private selves, however, the fiscal instruments Providers used to decentralize public services do not ensure the poor wili benefit from Given their limited capacity, local gov- social spending-that resources will be ernments should start using the private well spent in response to shifting local sector, NGOs, and organized communi- needs. ties to support their activities in plan- The government should ensure that ning, finance, pronam implementation, Spedal attention is services are efficient, roughly equivalent training, management, and evaluation. needed n, hunwnfor everyone entitled to services, and pro- This option is more realistic now. Some vided in a way that responds to the needs departments and municipalities have ministries, of the poor. The government's success in already taken the lead and can provide man tries, inducing effective spending on social pri- positive examples. The central govern- deartment, and orities will depend on its ability to moni- ment has created a positive climate for muncpalities tor outcomes and make adjustments to public-private sector interaction during the incentive framework. One priority the past few years. The new agencies for area for change relates to the terms of the regulation and control provide a fresh transfer formulas. The government opportunity for private sector participa- should consider including social expendi- tion in monitoring and evaluating decen- ture performance indicators in the formu- tralized programs. And the country has las-such as infant mortality or academic NGOs capable of assuming a leadership performance scores-to compensate role. achievements. Enhace Hum Resources atAUl Levels Trasorn Departments Special attention is needed in human The passivity of departments in the resources in sectoral ministries, depart- decentralization process must be over- ments, and municipalities, and the come, and their role as intermediaries required skill mix must be adjusted between the central government and through training or staff composition municipalities facilitated. The depart- changes. The areas of technical expertise ments need to create or strengthen their in need of emphasis are: sectoral techni- own technical assice units for munici- cal knowledge; targeting techniques; palities and participate in the central gov- impact evaluation; contemporary man- ernent's pilot projects. They also need agement training; information systems; to be part of the new intergovernmental identification, design, and implementa- information syster, bringing local and ton of projects; and public administra- regional resource allocation priorities to tion, planning, and budgeting. Training the attention of central government agen- and evaluation progams should include cies. And they need to train departmental operational plans and start with the most staff to sufport municipalities whenever advanced deparm ts and cities, gradu- the municipalities cannot fully comply ally expanding to the rest of departments with the new responsibilities. and cities. - xxv-^ It is iso inportant to create a pool of with little connection to lower levels of professional managers and public offi- government. A much more robust man- cials at the subnational levels of govern- agement information system similar to ment. Colombia needs to establish pro- Chile's or Argentina's could serve as a fessional qualifications for local civil ser- single source of reliable information for vice with salary schedules and education- local and central goverments to assess al institutions equipped to produce quali- progress, gauge needs, and manage pro- fied upper and middle-rank personnel. grams. To be effective, the system must be linked to the budget and planning A limiation in Monitoring and Evaluation Systems process at all levels of government. assessing the To implement the poverty reduction strat- Social Assistance Programs eectveness of egy, the government must continuously and monitor advances and modify action The absence of any credible estimates of prOgrams has been plans. For this, the government needs to the benefits and identification of recipi- the lack ofpre. enhance its monitoring and evaluation ents of social assistance programs seri- established systems. A clear limitation in assessing ously undermines efforts to recommend monitoring and the effectiveness of expenditures and expansion, contraction,or elimination. At evaluaon sysems programs has been the lack of pre-estab- best, programs provide information about to uwasure the lished monitoring and evaluation systems program costs and unit costs per recipi- to measure the impact and cost-effective- ent, but cost information cannot be a u c ness of poverty reduction programs. So. guide to assessment without knowledge gffec'i'ess lf developing adequate information systems of benefits. For example, within ICBF poverty rduction should be among the top priorities. there have been no comparatve longitu- programs dinal evaluations of the community child Decentralization Effort care centers (HCB) and day care pro- grams (CAIP). The decision to increase To support closer links among govern- ICBF financing to the HCB program, and ment levels, two types of information to increase parents' payments to the systems will be required. Colombia CAIP program while seeking alternative should develop performance criteria to modes to finance it, was based on the montor progress in institutional develop- govemnment's decision to target limited ment. The Department of National resources to the poor. But the effective- Planning Unit for Intergovernmental ness of ICBF's current expenditure mix Relations (DNP-UDT) has the responsi- is impossible to determine. bility to monitor and evaluate-at arm's A sound analytical base to assess the length-the implementation and impact benefits and cost-effectiveness of each of the institutional development strategy. program is urgently needed to determine Even so, given its limited experience future resource allocations. Evaluation with implementation issues, it will need modules should be incorporated in all support. social assistance programs. Resource An intergovernmental information allocation decisions should be made only system is also required. The information with this information in hand. The Office systems for resource allocation for social of the Presidency, which should become spending have so far been mainly centraL more active as an incubator of poverty - xxvi - programs, can be effective exploring development, expanded social and infra- innovative approaches only if it concen- structural services, decentralization and trates on timely evaluation of the results. institution-building, and monitoring and To function as pilot projects, the evaluation systems-must be pursued Programs of the Presidency need to together. Adequate coordination will be incorporate evaluation modules from the required across ministries and among dif- start. On the basis of the information ferent levels of government, between the generated, assessment can be made on public and private sectors, and among With the expected whether a pilot program should be users and providers of social services. oil revenuesfrom expanded and transferred to a line min- The change of administration in August Cusiana and istry-or discarded. 1994 provides an excellent opportunity to incorporate poverty reduction as a core Cupiagua. the time Poverty Progress Reports element in the new four-year develop- is ripe for ment plan. The recent reversal in the Colombia to To ensure adequate attention to the trend toward poverty reduction- embark on a implementation of the poverty reduction observed in declining rural incomes- renewed effort of strategy, and improve the approaches and the level of violence in the country- poverty reduction being followed, there should be a period- which contributes to social disintegra- ic progress report on poverty. A National tion-make the task critical. With the Commission for Poverty Reduction expected oil revenues from Cusiana and should be established to produce this Cupiagua, the time is ripe for Colombia annual or biannual report. This commis- to embark on a renewed effort of poverty sion should be a pluralistic body with reduction. representatives of central, departmental, The challenge is for the government and municipal governments, NOOs and to start implementing the proposed local communities. As an input to this poverty reduction strategy. By systemati- report, national and regional councils for cally pursuing each element in the strate- economic and social policies and NGOs gy, it will be possible to continuously should prepare background reports on monitor advances, make required modifi- governmental and nongovernmental cations, and further elaborate the specific efforts. All levels of government should details of the corresponding action plans. be required to include poverty reducdon The Bank stands ready to support this strategies in their annual planning and strategy through economic and sector budgeting exercises. work, lending operadons, and technical assistance. Patience and perseverance are Conclusion required, however, for the fight against poverty is a long-term endeavor. The The proposed poverty reduction strategy rewards will come-the most vulnerable is an integrated framework to tackle in society will attain a higher standard of poverty systematically in Colombia. All living. four elements of the strategy-rural Vision Estrategica Estrategia para la Redu-ci6n de la Por ello, la tarea actual involucra la Pobreza: c6mo encaminar a Colombia especificaci6n detallada del mecanismo hacia un crecimiento con equidad de ejecuci6n para el programa de desa- rrollo social. Los recientes descubri mientos de petr6leo en Cusiana y Cupiagua brindan Desarrollo rural La estrtgia de a Colombia una excelente oportunidad rdccOn de la de acelerar su ritmo de crecimiento. La pobreza extrema de Colombia es, Pobreza debe cen- Colombia cuenta con los recursos nece- fundamentalmente, un problema rural. trarse en tres aeas sarios para emprender una estrategia El setenta por ciento de la poblaci6n p)ioas: IJ coordinada de reducci6n de la pobreza. con ingresos por debajo del nivel de desarrovio is Lo que precisa ahora son polfticas efi- subsistencia vive en zonas rurales, lo 2) servcos caces y eficientes en el Area de recursos que demuestra la necesidad de adoptar sociakes y de humanos, y programas focalizados de medidas para aumentar los ingresos en iYs'y reducci6n de la pobreza: en pocas pa- ese sector. Esta necesidad se ve acentua- 3) descent ra- labras, una estrategia. En el disefto y da por el reciente deterioro de los ingre-e capacidafomepeo aplicaci6n de la misma, las autoridades sos rurales y el ajuste previsto de la raCiva ope- deben conceder prioridad a tres areas: 1) economfa a la bonanza petrolera Para rakvatyfonta- desarrollo rural, 2) servicios sociales y que el sector rural y los pobres puedanl km U- de infraestructura y 3) descentra- compartir los beneficios de dicha cional lizaci6n, fomento de capacidad operati- bonanza es menester contar con un va y fortalecimiento institucional. Seri marco de polfticas adecuadas. tambien necesario mejorar los sistemas La formulaci6n de una estrategia de seguimiento y evaluaci6n en todos coherente de desarrollo rural, sorpren- los programas de reducci6n de la dentemente olvidada en el Plan de pobreza. Desarrollo de 1990-94, solo se ha El ritmo de las reformas institu- empezado a elaborar recientemente. cionales constituye uno de los grandes Hasta ahora, los organismos compe- desaffos de la estrategia de desarrollo de tentes se ocupaban dnicamente de las Colombia. Por lo general, las polfticss polfticas agrarias y de los programas de aplicadas estan bien concebidas pero, gastos en zonas rurales. El Ministerio de para que s . plenamente operativas, Agricultura se encargaba de la polftica necesitan mayor concreci6n. El gobier- agraria mientras que los organismos de no, sin embargo, ha subestimado las ejecuci6n de los programas de gasto dificultades asociadas con la puesta en mrales (Programa de Desarrollo Rural marcha de las reformas institucionales. Integrado, Plan Naciondl de - xxviii - Rehabilitaci6n, Instituto Colombiano de condiciones de vida de los pobres que Reforma Agraria y Banco de Crddito viven en zonas rurales. Otro mecanismo Agrario) se relacionaban directamente para mejorar los niveles de vida de la con los gobiernos locales, pero sin la poblaci6n es la migraci6n del campo a coordinaci6n necesaria para establecer la ciudad, por lo que el gobiemo debe un marco institucional conducente a un garantizar la movilidad de la mano de desarrollo rural mas amplio. obra, eliminar las distorsiones del mer- Recientemente se han dado los cado de trabajo y proporcionar servicios La base del primeros pasos en este sentido. Se han de educaci6n y de salud basicos de alta desarrollo rural es agilizado y especializado las funciones calidad en las zonas rurales. un crecimiento de los organismos responsables de los agricola eficiente programas de gastos en el sector rural, Marco de incentivos para el sector mejorando asf su grado de complemen- agricola tariedad. Al asignar a una sola institu- ci6n, el Fondo de Desarrollo Rural La base del desarrollo rural es un creci- Integrado, a cargo del financiamiento de miento agrfcola eficiente, y 6ste s6lo es infraestructura y servicios para el sector posible si se eliminan las actuales dis- rural productivo se consigue una mayor torsiones de precios. El gobierno logr6 coh.erencia en la formulaci6n de progra- avances notables en la liberalizaci6n del mas y en la realizaci6n de actividades sector, desmantelando al Instituto de en el lugar mismo de implementaci6n. Mercadeo Agropecuario (IDEMA) y Agrupando a las autoridades munici- reduciendo los subsidios a la agricul- pales, representantes de las comu- tura. Pero estos avances se revirtieron nidades rurales y funcionarios de las en el afto 1993. En respuesta a la crisis entidades pdblicas en los Consejos agricola de 1991-92, el gobierno inten- Municipales para el Desarrollo Rural sific6 el proteccionismo comercial, rein- con el fin de coordinar los proyectos trodujo el apoyo a la adquisici6n de locales, el gobierno ha concebido una importantes cultivos-la adquisici6n de visi6n mas integrada del desarrollo productos agrfcolas por el IDEMA se rural. Ahora serb posible tener en cuen- duplic6 entre 1992 y 1993-y aplaz6 la ta, como es debido, las relaciones entre fecha prevista para establecer las tasas los proyectos rurales de alcance de inter6s de los pequeftos agricultores -;cional, regional y local y entre los en niveles mas pr6ximos a los del mer- sectores sociales y productivos. cado. Asimismo formul6 un programa De acuerdo con una estrategia mas de subsidios a la inversi6n (Incentivo a integral y coordinada del desarrollo la Capitalizaci6n Rural) con el fin de rural, el gobierno debe ofrecer incen- estimular las inversiones privadas en el tivos adecuados al sectot agrfcola, sector agrfcola. invertir en proyectos de infraestructura Los mecanismos de proteccionismo rural de altos rendimientos econ6micos comercial y de apoyo a la agricultura y sociales, y eliminar la tendencia a nacional, ademias de frenar la competi- favorecer a las zonas urbanas en la tividad del sector, son poco eficaces prestaci6n de los servicios pdblicos. como instrumentos de protecci6n para Pero no basta con mejorar las. Ia poblaci6n rural pobre. Para reducir la - xxix . pobreza, el gobierno deber§ revertir las rural que sean econ6micamente atrac- polfticas de proteccionismo comercial tivos es uno de los medios mas eficaces recientes, eliminar los subsidios a la de reducir la pobreza en Colombia. Los inversi6n y a los precios agrfcolas y programas de mejora de la infraestruc- reorientar dichos recursos hacia progra- turm y los servicios rurales y de consoli- mas rurales focalizados a los grupos daci6n de tierras son instrumentos m4s vulnerables. importantes para la reducci6n de la Para ayudar a los pobres en estos pobreza. El reto consiste en identificar, momentos de dificultad e iniciar un pro- disefiar e implementar proyectos con ceso de reasignaci6n de los recursos gran rentabilidad econ6mica y social, al agrfcolas, habra que poner en marcha, mismo tiempo que se alienta la partici- evaluar y ampliar, segdn convenga, el paci6n del sector privado. Programa de Empleo Rural de A corto y mediano plazo, el Emergencia propuesto por las autori- obstaculo mas importante en la rea- dades. De acuerdo con la propuesta, el lizaci6n de proyectos rurales es la limi- plan piloto de empleo sera coordinado tada capacidad operativa de los gobier- por ei Ministerio de Agricultura y dirigi- nos locales, que se debera fortalecer a la do por el Programa de Desarrollo Rural mayor brevedad posible. A medida que Integrado. Se debera poner atenci6n mejore la capacidad operativa a nivel especial en asegurar que los proyectos municipal y departamental, se podra de infraestructura nural que formen parte canalizar eficazmente un mayor volu- del programa sean intensivos en mano men de recursos al campo atraves de de obra y quu los sueldos que se paguen transferencias directas a los municipios sean inferiores a los del mercado. Esto rurales o mediante el sistema de cofi- iltimo es esencial para establecer un nanciamiento. Estos recursos adi- mecanismo autoinducido de selecci6n de cionales requerirhn salvaguardias espe- participantes y ofrecer un incentivo para ciales e incentivos para garandzar que abandonar el programa en el momento en el proceso de planificaci6n y progra- en que se presenten oportunidades labo- maci6n se tome en cuenta el objetivo de rales alternativas. Para que el impacto reducci6n de la pobreza nacional. En del programa de empleo sea mayor, este contexto, el Fondo de Desarrollo debera ser ejecutado por grupos comuni- Rural Integrado puede contribuir en tarios (atrav6s de las Juntas de Acci6n forma significativa a canalizar nuevos Comunal, Empresas Sohldarias o grupos fondos, fomentar la capacidad operativa organizados de mujeres) con fondos de a nivel municipal, influir en la asig- los municipios, el Fondo de Desarrollo naci6n de recursos locales y promover Rural Integrado, el Plan Nacional de proyectos especificos en apoyo de los Rehabilitaci6n o el Fondo de Solidaridad pobres. y Emergencias Sociales. Prestacidn de servicios sociales en las Infraestructura y servicios rurales zonas rurales El aumento de la productividad rural El diseflo y la prestaci6n de servicios mediante proyectos de infraestructura sociales tienden a favorecer a los centros - xxx - urbanos. Pam corregir este sesgo habra subsidios liberarfa recursos considera- que inclinar la balanza de los gastos bles para financiar otras iniciativas para sociales hacia los sectores rurales y. reducci6n de la pobreza. En especial, sobre todo, modificar las modalidades los recursos destinados para subsidiar de prestaci6n de los servicios de edu- energfa, vivienda y precios agrfcolas caci6n y salud de manera que respon- deberan orientarse hacia ensefianza dan mnas fielmente a las circunstancias basica, atenci6n de salud, infraestruc- Los recursos de,ti- especiales de la poblaci6n rural. tura rural y asistencia social a los mas nados hasta ahora La educaci6n es el Gnico sector que vulnerables. para subsidiar ha realizado un esfuerzo concertado por ene#ga, vivenday desarrollar un modelo rural de Educaci6n precios agricolas prestaci6n de servicios. Los resultados deberdn orientarse del programa Escuela Nueva son muy El sistema educativo se encuentra en hacia enseflanm positivos. El mejoramiento significativo una encrucijada. El esfuerzo de descen- bdsica, atencidn de del entorno escolar ha permitido tralizaci6n presentara oportunidades y salu Infraestruc- aumentar el aprendizaje y la partici- desaffos y requiere un compromiso tua nuralyasisten- paci6n comur.itaria. Pero para que el autdntico por mejorar la educaci6n, cia social a los mds programa Escuela Nueva continde fun- compromiso que se debe ver reflejado vuderables cionando a escala nacional se requiere en las prioridades gubernamentales y el compromiso constante de las altas representado por un Ministro de instancias de la administraci6n. En Educaci6n que encarne estos principios. materia de salud, el sistema de atenci6n Con una tasa de cobertura total de clfnico-curativo no parece id6neo para aproximadamente el 86 por ciento en el la poblaci6n rural pobre. Convendrfa nivel de ensefianza primaria y de casi el pensar en un modelo mas orientado 50 por ciento en el de la secundaria, hacia los servicios de extensi6n comu- Colombia esta todavfa muy lejos del nitaria-vfa agentes de atenci6n pri- mandato constitucional de ensefianza maria. Ademas, hay que ofrecer universal bisica. La baja cobertura de la opciones atractivas para hacer carrera ensefianza basica (hasta los 15 afios) dentro del servicio pdblico permanente continda siendo un problema para la orientado a la poblaci6n rural. poblaci6n pobre y para ciertas regiones del pafs (en particular de la regi6n atlan- Servidos sociales y de Infraestructura tica). Ademas, la calidad general de la ensefianza pdblica de nivel basico es Aunque en las dos dltimas decadas sumamente baja. Colombia ha conseguido ampliar signi- El crecimuiento previsto de las trans- ficativamente sus servicios, es mucho lo ferencias del gobierno central a los que queda por hacer para facilitar el departamentos y municipios debera acceso y mejorar la calidad de los servi- incrementar notablemente los recursos cios que reciben los pobres tanto en el disponibles de los gobiernos regionales medio rural como en las ciudades. y locales para educaci6n. Aunque estos Muchos programas contra la pobreza nuevos recursos deberfan significar, en han beneficiado sobre todo a los menos teorfa, un mayor acceso a la educaci6n, necesitados, y la eliminaci6n de estos la consecuci6n de este objetivo - xxxi - dependera del empefio de los gobiernos detalles de aplicaci6n de la ley son locales. El lograr un comproniiso polfti- incompletos o estan en gestaci6n, y su co local por la enseflanza es tan o mas institucionalizaci6n promete ser diffcil. importante que el construir nuevas Los pr6ximos gobiernos deberan dar escuelas. Las autoridades nacionales, claras pruebas de continuidad y flexibi- departamentales y municipales deben lidad en la bisqueda de la prestacion mejorar la calidad de la educaci6n publi- universal de servicios basicos de salud a ca elevando el nivel de preparacion mediano y largo plazo. Hay que reforzar exigido al nuevo personal docente, y el Ministerio de Salud para que pueda El lograr un corn- avanzar hacia un sistema que premie la orientar el sistema de una manera agre- pmriso poltkco calidad del desempefio docente, mas que siva y fijar normas de control de calidad localpor la la antiguiedad en el cargo. Ademas, y de eficiencia, especialmente impor- ensefanza es tan o deberan mejorar la disponibilidad de tantes a medida que el sistema pase a mds inpor que material didactico en las escuelas y cons- incorporar a los proveedores de salud el constrir nuevas truir instalaciones suficientes para evitar publica y privada dentro de su nuevo escueas la sobrepoblaci6n escolar y, con el tiem- marco de desarrollo. po, eliminar el sistema de triple tumo en Dentro de los planes gubernamen- las escuelas secundarias publicas. tales de servicios bAsicos cabe seflalar la ausencia de programas verticales de Salud inmunizaci6n y lucha contra las enfer- medades end6micas, que son particu- La mayor parte de la poblaci6n pobre de larmente importantes para los pobres. Colombia s6lo puede contar con el sis- Las enfermedades transmitidas por tema de salud publica. Segun los datos mosquitos y otros vectores son todavia actuales sobre utilizaci6n y costos de los frecuentes en los barrios de tugurios de servicios, sin embargo, este grupo no las ciudades, en los valles fluviales y esta recibiendo atenci6n adecuada. El en la costa. Los programas contra el sistema publico s6lo estA atendiendo al paludismo y el dengue han perdido efi- 35 por ciento de la poblaci6n, mientras ciencia y eficacia en los dltimos aflos. la meta es del 60 por ciento. El total de Las campaflas masivas de inmu- gastos no esta suficientemente foca- nizaci6n han demostrado su eficacia lizado hacia los pobres. Los dos quintiles pero no han alcanzado todavia cobertu- mas altos de ingreso de la poblaci6n ra universal. El Ministerio de Salud reciben casi una terceza parte de los gas- debe responsabilizarse de estos progra- tos. Un elemento fundamental de la mas y promover activamente su ejecu- nueva estrategia gubernamental en este ci6n. Un elemento critico en este senti- sector, presentada en la reciente Ley de do es la estrategia de servicios de Seguridad Social, es la prestaci6n uni- extensi6n comunitaria rural. En un sis- versal de servicios de salud, que tema descentralizado de prestaci6n de empezard por los servicios primarios. El servicios de salud, es menester disuadir objetivo es ofrecer un conjunto basico de a los gobiernos locales en evitar inver- servicios de salud a todos los colom- siones excesivas en instalaciones de bianos, cualquiera que sea su ingreso, salud cuyos gastos operativos y de subsidiando a los pobres. Pero los mantenimiento tal vez no scan - xxxii - sostenibles. Otros elementos cffticos petr6leo y el keros6n son combustibles para lograr la mejorfa en la calidad de alternativos que deberAn promoverse al los servicios es la asignaci6n clara de mismo tiempo que se ajustan las tarifas esta responsabilidad a los funcionarios electricas. Las fuerzas del mercado locales y el fortalecimiento de la pueden llevar a la sustituci6n de la elec- demanda de los usuarios. tricidad por otras fuentes de energfa, y el gobierno debera garantizar la libre Energfa accion de dichas fuerzas. En relaci6n con Elementos crftcos los hogares de ingreso bajo, el gobierno para lograr la Durante muchos aflos, la estrategia para podrfa distribuir y subsidiar aparatos de mejorta en la cali- fijar las tarifas electricas se basaba en la gas licuado de petr6leo o de kerosen dad de los servicios convicci6n de que si se mantenfan bajos atrav6s de los consejos municipales. La de salud es la asig- los precios, las familias pobres tendrfan ampliaci6n de la red de gas en los barrios naci6n clara de acceso a los servicios. La realidad pobres podrfa subsidiarse reduciendo los esta responsabili- demuestra, sin embargo, que la distribu- costos de conexi6n y utilizando estufas dad a los urn- ci6n de los subsidios masivos del sector de dos quemadores con bajo consumo de cionarios locales y (nada menos oue el 1,4 por ciento del energfa. En el caso de las viviendas elfortalecimiento PIB) ha sido ligeramente regresiva y, por rurales, donde la ampliaci6n de la red de la demanda de lo tanto, no ha beneficiado primordial- el6ctrica podrfa resultar demasiado cos- los usuarios mente a los pobres. Aunque el gobiemo tosa, se podrfan atender los servicios ha subido recientemente las tarifas el6c- basicos de iluminaci6n y comunicaci6n tricas y reducido la estratificaci6n (un atrav6s de generadores aislados. En las sistema de precios mal disefiado que, en comunidades pequefnas los generadores teorfa, cobra a los usuarios residenciales podrfan ser, por ejemplo, motores diese. de acuerdo con sus ingresos), es preciso acelerar el proceso. El gobiemo deberfa Agua potable y alcantarilado simplificar todavfa mas o eliminar por completo la estratiflcaci6n, establecer En Colombia, la falta de agua potable y una tarifa residencial basada en los cos- alcantarillado es un problema casi exclu- tos marginales y acompafiarla con tarifas sivo de los pobres. Para ambos servicios, sociales que favorezcan especfficamente en 1992 la cobertura del quintil mas a los usuarios de ingreso bajo. Los recur- pobre de la poblaci6n era muy inferior a sos liberados con los cambios en las tari- la media nacional: en el caso de agua, el fas permitirfan reducir impuestos regre- 65 por ciento (frente a una media del 83 sivos o financiar intervenciones mas por ciento); en el de alcantarillado, el 37 focalizadas. por ciento (frente al 67 por ciento). Esta Asimismo, se requeriran medidas falta de cobertura de los pobres no se complementarias que acompafien los explica simplemente porque no est6n dis- ajustes de tarifas. Habida cuenta de que puestos a pagar. En las principales ciu- hasta un 11 por ciento de los ingresos de dades las familias del decil de ingresos las familias pobres urbanas se gasta en mAs bajos gastan hasta el 9 por ciento de energfa, es fundamental proveer altemnati- sus ingresos en agua y alcantarillado, vas que permiitan sustituir el uso de clec- mientras que la poblaci6n en general solo tricidad. El gas natural, el gas licuado de gasta entre el 1 y el 2 por ciento. -%xxiii- Para aumentar la cobertura de los concede subsidios directos a los pobres servicios de agua potable y alcantarillado -para comprar una nueva casa, mejorar no bastard con aplicar polfticas de pre- una ya construida, registrar la propiedad cios basadas en recuperaci6n de costos y de la tierra o tener acceso a servicios de subsidios focalizados como se recomien- saneamiento-, el gobiemo debera desti- da en este informe. Habra que atender las nar mas recursos a financiar parcelas necesidades de poblaciones rurales dis- urbanizadas y urbanizables y obras de persas o de residentes en zonas periur- abastecimiento de agua y saneaniento. La cobertura banas inaccesibles y desatendidas. En los linitada de servi- centros urbanos, habra que hacerlo sin Asistencia social a los grupos mds vui- cios de energ(a, crear incentivos para que otros se asien- nerables agua potab y ten en lugares todavfa mAs inaccesibles. alcantarillado no se Aparte de los obstaculos tecnol6gicos, el La estrategia de reducci6n de pobreza explica simpk- costo de los servicios tradicionales por debe procurar que los pobres se benefi- mente por lafalta familia en estas zonas es prohibitivo. cien del crecimiento econ6mico, y con- de disponibi Todo ello obliga a considerar dos tribuyan a 61. Pero algunos de ellos, de pago de la asuntos. En el caso de los residentes en sobre todo los ancianos e invalidos, no pob goid lugares inaccesibles, tanto en asen- podran aprovechar las nuevas oportu- timientos perif6ricos pr6ximos a centros nidades. Otros, adn obteniendo algunos urbanos como en zonas rurales dispersas, beneficios, no tendran los ingresos sufi- deberan utilizarse tecnologfas interme- cientes para atender sus necesidades dias. Por otra parte, las politicas de basicas o correran peligro de perder sus vivienda deberan ofrecer mayor acceso a ingresos, por eventos imprevistos tales los servicios pdblicos basicos y evitar la como el de una sequfa o la p6rdida del ulterior proliferaci6n de asentamientos sost6n de la familia. Colombia, que urbanos no autorizados. El gobierno carece todavfa de una red de protecci6n deberi promover alternativas como las social eficaz para los grupos mds pobres tomas de agua comunitarias, las cister- y vulnerables, deberfa otorgar a la nas, los camiones cisterna y las tabletas creaci6n de tal red un caracter priori- de purificaci6n del agua como soluciones tario. para las actuales zonas desatendidas. Los recursos podran conseguirse Debera estudiar la manera de mejorar la eliminando subsidios (por ejemplo, los calidad de los servicios pres!ados con destinados a los servicios pdblicos) u tales alternativas alentando la partici- otros programas que benefician funda- paci6n del sector privado mediante un mentalmente a las personas de ingresos proceso de licitaci6n que asigne altos, o asignando a tal fin una pequefia simltfineamente a los licitantes barrios parte de los ingresos previstos de la lucrativos y otros menos rentables. Pam bonanza petrolera. La experiencia de garantizar el servicio en los barrios m4s otros pafses revela que, si se formulan pobres podrfa conceder un subsidio de la programas eficientes y bien focalizados, misma magnitud que el ofrecido a las un nivel adecuado de gasto en protec- familias de ingreso bajo para los servi- ci6n social (aparte de las responsabili- cios convencionales. En conformidad dades tradicionales en los servicios de con la actual polftica de vivienda, que educaci6n y salud) oscilarfa entre el 1 y - xxxiv - el 2 por ciento del PIB. Al asignar Descentralfzaci6n, fomento de capaci- recursos para los grupos m4s pobres y dad operativa y fortalecimiento inst. vulnerables, el gobierno debe evitar tucional todo despilfarro en proyectos ineficaces. Los costos administrativos deben La eficacia de los gastos sociales y de reducirse al mInimo, y el grueso del los esfuerzos de reducci6n de la pobreza gasto social debe beneficiar a los mas en Colombia depende de un equilibrio necesitados. precario de fuerzas entre e' gobierno Como consecuencia Una vez reestructurados y reorienta- central y los gobiernos locales. Con la de la descentra- dos, los programas dirigidos por el descentralizaci6n del sector pdblico, Izaci&4 la reduc- Instituto Colombiano de Bienestar Colombia esta transfiriendo a los gobier- cn de lapobreza y Familiar (ICBF), el Despacho nos locales la responsabilidad de los el desarroUlo int- Presidencial y el Sistema de Seguridad programas sociales y de muchos servi- tucional son ahora Social podrfan constituir una red de cios de infraestructura destinados a los nseparables protecci6n social, que ofirecerfa asisten- pobres. Los gobiernos locales han cia mas eficiente y equitativa a quienes recibido con entusiasmo su nueva autori- mas la necesiten y reducirIa considera- dad y la autonomia de gasto conferidas blemente la pobreza. Entre los posibles por la legislaci6n reciente, en especial la programas de lucha contra la pobreza Ley de Competencias y Recursos de podrfan considerarse el Plan DIA (un 1993 (sobre las funciones y finan- programa de estampillas para la compra ciamiento de los gobiernos locales). de alimentos) y el Programa de Empleo Pero la eficacia y eficiencia del gasto Rural de Emergencia. En la formu- social y la administraci6n pdblica est.n laci6n de estos programas, el gobierno en serios problemas en casi todas las se atuvo a estas cinco normas: dirigir la ciudades y departamentos. Para lograr mayor parte del subsidio a los benefi- una mayor eficacia del gasto, el gobier- carios, reducir la necesidad de personal no ceniral estA disefando un sistema de y los costos administrativos, establecer incentivos para influir en las decisiones un mecanismo de selecci6n autoinduci- de las autoridades locales. La definici6n do de beneficiarios y definir criterios adecuada de ese sistema es impre- para la exclusi6n automatica de aque- scindible para que los servicios descen- llos receptores que ya no necesitan el tralizados lleguen de verdad a los programa, promover la complemen- pobres. tariedad entre los servicios y ensayar el Como consecuencia de la descentra- programa como un proyecto experimen- lizaci6n, la reducci6n de la pobreza y el tal para evaluar sus resultados antes de desarrollo institucional son ahora insepa- la implementaci6n generalizada. Al rables. La "Guerra contra la pobreza" establecer la red de protecci6n social, el requiere claramente una estrategia de gobierno debe seguir esos principios desarrollo institucional: una visi6n glo- rectores. Ningfin programa debe institu- bal, objetivos explfcitos a largo plazo, un cionalizarse sin antes haberse evaluado organismo central encargado de super- detenidamente los beneficios y costos visar una administaci6n eficiente de las de dichos proyectos experimentales. instituciones subnacionales, mayor capacidad de los gobiernos locales para - xxxv - formular programas de gasto pdblico y educaci6n, abastecimiento de agua, pan evaluar su efecto sobre la pobreza, saneamiento y asistencia social en con- y atenci6n al profesionalismo del per- formidad con las disposiciones del go- sonal. Para ello se requiere un nivel bierno central per- teniendo muy en nuevo de capacidad t6cnica que permita cuenta las prioridades locales. Para ser a todas las comunidades locales-go- exitoso el nuevo sistema se require del biemnos, gnupos de interes organizados y uso adecuado de los instrumentos de sector privado-decidir y administrar descentralizaci6n. El sistema de incen- con la competencia requerida la enorme tivos que ahora se estA preparando com- expansi6n de las responsabilidades de pagina los estfmulos financieros con los los gobiernos subnacionales prevista en controles administrativos: transferencia la nueva constituci6n y legislaci6n. La de ingresos entre gobiernos, donaciones estrategia propuesta incorpora los cam- de contrapartida y ckisposiciones regu- bios requeridos en todos los niveles de ladoras. No obstante, esos mismos gobierno: central, departamental y instrumentos fiscales disefiados para municipal. El primer requisito funda- descentralizar los servicios publicos no mental es convertir a los departamentos garantizan por sf solos que los gastos en intermediarios entre el gobierno cen- sociales redunden en beneficio de los tal y los municipios. El segundo serfa pobres-que los recursos se gasten abrir la admmistraci6n pdblica al sector correctamente en respuesta a los cam- privado, ONGs y comunidades organi- bios de las necesidades locales. zadas en areas como la planificaci6n, El gobierno debe comprobar que los $inanciamiento. implementaci6n de pro- servicios son eficientes, que todos los gramas, capacitaci6n, administraci6n y autorizados a recibirlos se benefician en evaluaci6n. El tercer requisito serfa proporci6n semejante y que se prestan mejorar los recursos humanos en todos en forma acorde con las necesidades de los niveles de gobiemo. los pobres. El exito del gobierno en la Dada la magnitud de la tarea, la eje- promoci6n de una mayor eficacia de los cuci6n de la estrategia debera comenzar gastos en prioridades sociales depen- con proyectos piloto en cada Area, para dera de su capacidad de supervisar los hacer una utilizaci6n mAs eficaz de la resultados y continuamente ajustar el limitada capacidad institucional. Los sistema de incentivos. Una esfera prio- proyectos deberan comprobar en todo ritaria para posibles cambios son los momento la coordinaci6n interguberna- componentes de las f6rmulas de trans- mental y prover evaluaciones en cada ferencias. El gobierno deberfa exami- fase asf como estrategias para ampliar nar, por ejemplo, la posibilidad de la aplicaci6n a un mayor ndmero de introducir en estas f6rmulas indicadores beneficiarios. de eficacia de los gastos sociales-por ejemplo, cambios en la mortalidad Formulacidn de un sistema de incentivos infantil o en el nivel de rendimientos para los gobiernos locales academicos. Esto permitirfa retribuir a los departamentos y municipios que Los gobiernos municipales realizaran mAs avanzen en el area social. los programas esenciales de salud, - XXIVi - Transformaci6n de los departamentos mente, el pafs cuenta con ONGs capaces de tomar el liderazgo. Hay que superar la pasividad de los departamentos en el pr-oceso de descen- Mejoramiento de los recursos humanos tralizaci6n y facilitar su funci6n de a todos los niveles intermediarios entre el gobierno central y los municipios. Los departamentos Se require prestar atenci6n especial a los deben crear o fortalecer sus propias recursos humanos de los ministerios sec- Se requiere prestar dependencias de asistencia t6cnica para toriales, departamentos y municipios y atencidn especial los municipios y participar en los modificar la composici6n de especiali- a los recursos proyectos piloto del gobierno central. dades disponibles en el personal me- hImos de Deben tawnbien formar parte del nuevo diante actividades de capacitaci6n o cam- los minsterios sistema de informaci6n interguberna- bios de personal. Las especialidades sectoriales, mental e indicar a los organismos pdbli- tecnicas que necesitan mayor atenci6n departamentos y cos centrales las prioridades locales y son las siguientes: conocimientos t6cni- municipios regionales de asignaci6n de recursos. cos sectoriales, t6cnicas de selecci6n de Asimismo, deben preparar al personal beneficiarios, evaluaci6n de impacto de de los departamentos para que ayuden a programas, capacitaci6n en tecnicas los municipios cuando 6stos no puedan modemas de adrministraci6n, sistemas de desempefhar debidamente sus nuevas informaci6n, identificaci6n, disefho e responsabilidades. implementaci6n de proyoctos y adminis- traci6n pdblica, planificaci6n y pre- Apertura de la admniseracidn piblica al supuesto. Los programas de capacitaci6n sector privado y evaluaci6n deberAn incluir planes ope- rativos y comenzar con los departa- Dadas 'as limitaciones en su capacidad mentos y ciudades mas avanzados, para operativa, los gobiernos locales deberan ampliarse posteriormente a todos los recurrir inicialmente al sector privado, demas gobiemos locales. ONGs y comunidades organizadas para Es tambi6n importante crear un que apoyen sus actividades de planifi- equipo de administradores profesionales caci6n, financiamiento, implementaci6n y funcionarios pdblicos para los gobier- de programas, capacitaci6n, adminis- nos departamentales y municipales. traci6n y evaluaci6n. Esta opci6n es Colombia debe fijar requisitos profesio- ahora mas factible que antes. Algunos nales para los funcionarios locales y departamentos y municipios han dado ya establecer escalas de sueldos e institu- los primeros pasos en este sentido y ciones de ensefianza que permitan contar pueden mostrar resultados positivos. con personal competente de rango medio AdemiAs, el gobierno central ha creado un y superior. clhina favorable a la interacci6n del sector pdblico y privado durante los dItimos Sistemas de seguhmiento y evaluac6n afios. Los nuevos organismos de regu- laci6n y control ofiecen nuevas posibili- Para poner en practica la estrategia de dades de participaci6n del sector privado reducci6n de la pobreza, el gobiemo debe en el seguimiento y evaluaci6n de los supervisar en todo momento los avances programas descentralizados. Y, final- conseguidos y modificar los planes de xxxvii - acci6n. Para ello, debe reforzar sus sis- proceso de planificaci6n y presupuesto temas de seguimiento y evaluaci6n. Un de todos los niveles de gobierno. obstculo que impide evaluar la eficacia de los gastos y programas ha sido la falta Programas de asistencia social de sistemas preestablecidos de seguimiento y evaluaci6n que permitan La ausencia de estimaciones verosfmnles cuantificar el impacto y la eficacia en sobre los beneficios y de identificaci6n funci6n de los costos y beneficios de los de los beneficiarios de los programas de U obsdcuo que programas de reducci6n de la pobreza. asistencia social representa un grave Unapide evaluar la Por ello, una de las prioridades ha de ser obstaculo para los intentos de recomen- eficde evgas- la creaci6n de sistemas de informaci6n dar la expansi6n, contracci6n o elimi- tos y programas ha adecuados. naci6n de los programas. En el mejor de yp lga fd d a los casos, los programas suministran temspreestadbkci- Esfuerzo de descentralizaci6n informaci6n sobre los costos de los pro emos de seguiieto gramas y los costos unitarios por benefi- y evaluacin que Para promover vfnculos mas estrechos ciario, pero la informaci6n sobre los cos- permitan cuan- entre los distintos niveles de gobiemo, se tos no puede orientar la evaluaci6n si no fcar el imacto y requeririn dos tipos de sistemas de infor- se conocen los beneficios. Por ejemplo, I e y maci6n. Colombia debe formular crite- en el ICBF no se ha realizado ninguna efa e nf rios de desempefjo para supervisar los evaluaci6n longitudinal comparativa stos progresos del desarrollo institucional. La sobre los hogares comunitarios de bie- de Unidad de Desarrollo Territorial del nestar (HCB) y los programas de aten- programas de Deparamento Nacional de Planeaci6n es ci6n infantil (CAIP). La decisi6n de mducc6f de la la encargada de supervisar y evaluar- aumentar el fmanciamiento del ICBF con porea mparcialmente-la implementaci6n y el destino al programa HCB y de elevar las avance de la estrategia de desarrollo cuotas pagadas por los padres en el pro- institucional. Atin asf, dada su experien- grama CAIP, al mismo tiempo que se cia limitada en problemas relacionados buscaban modalidades alternativas de con implementaci6n, necesitara alguna financiamiento, se bas6 en la decisi6n formi de apoyo adicional. gubernamental de destinar a la poblacion Se requiere tambi6n un sistema de pobre los recursos disponibles. Pero en informaci6d intergubernamental. Los este momento es imposible detminar la sistemas informativos sobre la asig- efectividad de la asignaci6n del gasto de naci6n de recursos para gastos sociales ICBF. han sido hasta ahora de caracter central, Se necesita urgentemente una base con poca relaci6n con los niveles de analftica s6lida para evaluar los benefi- gobierno subnacionales. Un sistema cios y efectividad en costos de cada pro- mucho mas s6lido de informaci6n sobre grama, a fin de deterninar la asignaci6n la gesti6n, semejante al de Chile o de recursos en el futuro. Deberin incor- Argentina, podrfa funcionar como fuente porarse m6dulos de evaluaci6n a todos dnica de informaci6n confiable y permi- los programas de asistencia social. Las tir al gobierno local y central evaluar los decisiones sobre la asignaci6n de recur- avances, determinar las necesidades y sos s6lo podran tomarse teniendo en enc&iuzr los programas. Para ser eficaz, cuenta la informaci6n que estos generen. este sistema debe estar vinculado al El Despacho Presidencial, que deberfa - xxxviii - dedicarse a incubar programas de lucha mentos de la estrategia-desarrollo rural, contra la pobreza, dnicamente podra ampliaci6n de los servicios sociales y de explorar con eficacia nuevos plantea- infraestructura, descentralizaci6n, fomen- mientos si se concentra en la evaluaci6n to de capacidad operativa y fortale- oportuna de resultados. Pam que puedan cimiento institucional y sistemas de funcionar como proyectos piloto, los seguimiento y evaluaci6n-deben Programas de la Presidencia deben seguirse en forma simultAnea. Se requiere incorporar m6duios de evaluaci6n desde una coordinaci6n adecuada entre los Los igresos pre- su inicio. Basandose en la informaci6n ministerios y entre los diferentes niveles vistos de los recur- generada, se podri evaluar si un proyec- de gobierno, entre el sector publico y el sos petroleros de to piloto debe ampliarse y transferirse a pnvado y entre los usuarios y los pro- Cusiana y un ministerio sectorial, o desecharse. veedores de servicios sociales. El carabio Cupiagua ofrecen a de gobiemo en Agosto de 1994 ofrece Colombia ua Informes periodicos sabre las acividades una magnifica oporbunidad para hacer de oportwidad exce- contra la pobreza la reducci6n de la pobreza un elemento lente de emprender central del nuevo plan de desarrollo. El an esfuerzo renova- Pam que la estrategia de reducci6n de la retroceso reciente de la tendencia hacia la do de reduccidn de pobreza reciba la atenci6n que se merece reducci6n de la pobreza (manifiesta en el lapobreza y continuamente se mejoren los progra- descenso de los ingresos rurales) y el mas existentes, deberan prepararse nivel de violencia en el pais (que causa informes peri6dicos sobre las actividades desintegraci6n social) dan a esta tarea contra la pobreza. Para la elaboraci6n de caracter prioritario. Los ingresos pre- este informe anual o bianual se debera vistos de los recursos petroleros de nombrar una Comisi6n Nacional pam la Cusiana y Cupiagua ofrecen a Colombia Reducci6n de la Pobreza. Esta debera ser una oportunidad excelente de emprender un organismo pluralista, con represen- un esfuerzo renovado de reducci6n de la tantes de los gobiernos centrales, departa- pobreza. mentales y municipales, ONGs y comu- El desaffo esta ahora en que el go- nidades locales. Como base para dicho biemo decida implementar la estrategia informe, los consejos nacionales y propuesta. Siguiendo sistematicamente regionales de polfticas econ6micas y cada uno de sus elementos, sera posible sociales y las ONGs deberan preparar supervisar en todo momento los avances documentos infonnativos sobre las diver- conseguidos, introducir las modifica- sas iniciativas gubernamentales y no ciones necesarias en forma oportuna y gubernamentales. Para esto, todos los continuamente afinar los detalles de los niveles de gobierno deberan incluir sus planes de acci6n correspondientes. El estrategias de reducci6n de la pobreza en Banco Mundial estA dispuesto a apoyar los planes y presupuestos anuales. esta estrategia con estudios econ6micos y sociales, prestamos y asistencia tMcnica. Conclusl6n No obstante, se necesitara paciencia y perseverancia, porque la lucha contra la La estrategia propuesta de reducci6n de pobreza es un reto de largo plazo. Tarde o la pobreza representa un planteamiento temprano, la recompensa llegar& los gru- global de acciones sistematicas contra la pos mas vulnerables de la sociedad alcan- pobreza en Colombia. Los cuatro ele- zarfn un nivel de vida mas alto. Chapter 1 Colombi: Poverly Profile During the past few decades the living conditions of the average Colombian have improved substantially. Poverty has declined steadily from an estimated 50 percent of the population in 1964 to 19 percent in 1992. Since the early 1950s, life expectancy at birth-now at 69 years- has risen almost two decades, while infant mortality rates-now 30 per thousand live births-have been slashed by a factor of four. Indeed, Colombia has better social indicators than those expected for a country at its level of development. While Colombia's overall poverty picture has generally improved, striking disparities in income, wealth, and living standards between regions and socioeconomic groups remain. Poverty incidence is more than three times higher in rural than in urban areas. Infant mortality rates in Choco, the department with the worst social indicators, are similar to the national average twenty years ago. The incidence of child malnutrition is 50 percent higher in households in which the mother has no education than in those in which she has pnrimary education. Overall, the sustained improvement in the main social indicators of the last decades should not divert attention from those Colombians still experiencing conditions that the country, as a whole, left behind ten or twenty years ago. The experience of other countries in Latin America suggests that faster rates of improvement in iving conditions can be achieved. L. Inroduclon 1.1 Colombia has made substantial progress in improving the living sandards of its population. But the challenge of poverty reduction remains significant. In 1992, six million Colombians had incomes below a commonly accepted minimum subsistence level. Child malnutrition and infant mortality remain high despite continuous progress for several years. Regional differences in most indicators suggest that the efforts required to surmount gaps in income and social well-being are enormous for some areas. 1.2 This chapter analyzes the nature and extent of poverty in Colombia, and the distinguishing chaacteristics of the poor. It provides points of reference for the rest of the poverty assessment report. Section II starts with a brief description of poverty in Colombia in 1992. Through different poverty lines the level and regional incidence of poverty are examined. To provide some perspective, the incidence of poverty in Colombia is compared with other Latin American countries. The section then presents, in addiftion to 1992, estmates of the incidence and severity of poverty for 1978, 1988, and 1991. Yn doing so, it provides the basis to compare poverty outcomes across time. The section concludes with an overview of the evolution of the index of "Unsatisfied Basic Needs" (NBI)-a widely used - 2 - measure of poverty in Colombia-and an assessment of its accuracy and usefulness as an indicator of poverty. 1.3 To characterize the poor in Colombia, Section m uses a probability model in which the chances of falling below the poverty line are linked to individual and family factors, like education, number of dependents in the family, or geographical location of the household. With labor income representing the pnncipal component of total income of poor households, special attention is given to the labor market characteristics of the poor and to the main determinants of labor income. Through simulation exercises the section concludes by presenting potential gains in poverty reduction associated with future education efforts. Finally, Section IV complements the above review by looking at other social indicators, like net primary school enrollment or under-five mortality rates, to assess improvements in standard of living of Colombia's poor. This assessment ends by comparing Colombia's record with that of other countries in Latin America and the world. It. Trends in Povefty Povely in Colombia: 1992 1.4 More than six million Colombians-18.8 percent of the popuation-have incomes below a commonly accepted subsistence level to buy a nutritionally adequate basket of food. Seventy percent of those individuals live in the rural areas of the country. 1 As can be seen in Table 1.1, the incidence of such an extreme form of poverty is more than three times higher in rural than in urban areas. The rural poor are poorer than the urban poor. On average, their incomes are 43.3 percent below the subsistence level. The same deficit is only 31.3 percent for the urban poor. 1.5 Given the arbitrariness in defining any poverty line, it is important to consider several poverty lines in doing the incidence analysis. The poverty line used so far is exetme-it considers only persons with incomes inadequate to meet their food energy requirements. Individuals with higher incomes may still be poor. Figure 1.1 shows, for rural and urban areas and for the nation as a whole, how the incidence of poverty increases as multiples of the original poverty lines are considered. A doubling of the poverty line2 would 1 These findings are based on normati fom the household suvey condted by the Dpamto Admbami Nacoal de Estdic (DANM) in Septmb. 199Q (ENH77). Tb urban infomation rdIs to thirty dee ciies/metpolHtan areas, while - Mr areas hae been surveyed in four regions of te counny. Unfortunaly, t vey had no coverage of the Ten*odos Nadoaw (nationa terories). me niden of povery was esmtd uSig the "indigence line" calculatd by DANE. Ffteen Ilnes wer used (fourteen for urban areas and one for e rual aea) in order to reflect differences in cost of livig, whicb ranged in value bftween US$25 and US$33 per capit per month. Ihese lines represent approxmy 50 percent of aveae monthly per capina income in the rual areas, and an average of 25 percent in e ub areas. For a detaed descrwon of the household survs used for the rport, as we as the methodologies folowed in the generadon of information to be presented in dds chapter, die reader shold reOr to Anmex 1. 2 In addition to the 'indigence line" being used ere, DANE also calculates a poverty line' which is twice the indigence line in ud area and 2.3 times that tnurban areas. The conderation of a higher poverty line which is twice the value Of an *exeme povert Ile" is the apprach folowed by CEPAL in its multple sdes of povert in Latin America. -3- Table 1.1: Poverty In Colombia, 1992 Nad_a IIi- RIt Number of People (millions) 6.2 1.9 4.3 Incidence (percent)a 18.8 9.9 31.2 Poverty Deficit (percent)b 39.4 31.3 43.3 Poverty (Gap (percent)C 7.4 3.1 13.5 a. Proportion of population falling below the poverty line. b. Percentage by which average income of the poor is below the poverty line. c. Indicates depth of poverty as a proportion of the poverty line. It is equal to the product of incidence and poverty deficit. Source: Estimates based on household surveys (EN177). Fire 1.1: Poverty Incidence, 1992 100% 90% 80% 70% 60% Edae Urban 50% :W 40% t National 30% 20% 10%~ 0% 0.5 1 1.5 2 3 5 6 Multiple Of base povwart" ie Source: Estumates based on household surveys (1NH77). indicat that the incidence of poverty in utban areas is 36 percent and in rural areas 65 percent, while that for the country as a whole is 48 percent. A large segment of the popuation appears to be concentrated close to the povrty line. This is particularly true in urban areas where more than 13 percent of the population have incomes between 1 and 1.5 times the poverty line. By doubling the poverty fine, one observes a doubling of the incidence of poverty in nural areas, and almost a fourfold increase in urban areas. Legitimate coneen for the "extremely poor'-mosdy rural-therefore should not result in the neglect of the urban poor. The following paragraph provides a more detailed picture on this particul issue. -4 - 1.6 Regional Distibution of the Poor.3 Figure 1.2 shows the distribution of the poor is sensitive to the poverty line used. For the "extremely poor,' as mentioned before, poverty is mostly a rural problem. Within rural areas, the extremejy poor are somewhat more concentrated in the Oriental and Atlantic regions, even though the incidence of poverty in the Pacific is as high as in the Oriental region. In urban areas, the extremely poor are found almost in equal numbers in large and medium size cities as in small ones, although the incidence of poverty is clearly higher in small cities. For the regional distribution of the "poor" (with two times the base poverty line), the picture is quite different. The proportion living in nual areas drops sharply, with all regions other than Central experiencing significant reductions. In this case, 44 percent of the poor are in cities. As a consequence, as one moves toward a less stringent definition of poverty, the issues related to urban poverty become more important and clearly deserve the attention of policymakers in Colombia. Figure 1.2: Regional Distribution of the Poor, 1992 (incidence of poverty in parenthesis) Extremely Poor Flulal Rlelfb Medium (11%) 16% 15 < FljZ ~~~~~~ura Sma (I a% Pe Tl r16 O Its! (83%)21% Rusral cmtra 31)8 (23%) 18% Poor Rtml R ifi Udan larg. nd (70%)11% Matium (33%) RuralOa Ant 25% (85%)15%/ \6) FW Or1 6td Utdn (42%) (60%) I6% _Wr At tS% MSe%) 14% Sowe: Esdmats based on household suveys (ENH77). 3 A Hlet of Ib deparsaet nlcued tl each region can be found iL Annex 1, Table 1.1. -5- 1.7 International Comparisons. Through the use of a common poverty line, the incidence of poverty in Colombia can be compared with that in other countries in Latin America.4 The results, presented in Table 1.2, indicate that the incidence of poverty in Colombia is similar to Jamaica, Venezuela, and Chile. Poverty in Colombia appears moderate when compared with other countries in the region.' It is noticeable, however, that in Colombia the incidence of rural poverty relative to that of urban poverty (3.6) is significantly higher than the rest of the region (except Jamaica). This underscores the need for measures that will enhance rural incomes and narrow the gap between rural and urban regions. Table 1.2: International Poverty Comparisons Incidence of Poverty (percent) Suy location, year National Urban Ruld Rwwl Urban This Study: Colombia, 1992 14.5 7.0 24.9 3.6 PADDLE Study: Colombia, 1989 n.a. 8.0 n.a n.a. Brazil, 1989 40.9 33.2 63.1 1.9 Chile, 1989 10.0 9.9 10.4 1.1 Costa Rica, 1989 3.4 3.5 3.2 0.9 Guatemala, 1989 67.0 50.9 76.5 1.5 Honduras, 1989 70.5 54.4 82.6 1.5 Jamaica, 1989 12.1 4.4 18.3 4.2 l Panama, 1989 31.8 26.9 36.8 1.4 Venezuela, 1989 12.9 10.8 23.5 2.2 Source: Psacharopoulos and others (1992). 4 Consistent with a recent study tbat estimated poverty in a large group of Latn Amerin counties (PsachY los and othr (1992)), the incidence of poverty in Colombia for 1992 is estmated using a poverty line equal to the equivalent it Colombian pesos of US$60 per capita per month in 1985 PPP dollars. The lines used ranged in value between US$21 and US$M per capita per month. 5 The difference between this report's estimate of utban poverty and that of Psacharopoulos and others (1992) is explained by the diferent methods by which the original survey information is adjusted to account for income under-repoting and the number of urban areas covered. The 1992 survey includes smaller cities which experience hwher ince of poverty. See Annex 1 for details on the methodology foLowed in this study to adjust for income under-teporting. -6 - Progress in Poveny Reducion: 1978-1992' 1.8 Between 1978 and 1992, Colombia reduced extreme poverty by six percentage points-from 24 to 18 percent of the population. Similar reductions can be observed under different poverty measures.7 But given a population increase of approximately 30 percent between 1978 and 1992, the number of individuals below the poverty line was not reduced-5.9 million Colombians in 1978 compared with 5.8 million in 1992.8 It is worth noting, however, that these gains occurred in a period when most other Latin American countries experienced increases in both the incidence and number of the poor. Between 1980 and 1989 the incidence of poverty in the region increased from 27 to 32 percent of the population.9 Colombia was one of the few forttnate exceptions. Figure 1.3: Evolution of the Incidence of Extreme Povert 40.0% 8s.4% 33.1% 31.2% 30.0% 29.0%1 23.6% l° National 20.0%- .8.7%16.0% 17.7% U Urban 10.0% _ % 8.0% 0.0% 1978 1988 1991 1992 Source: Estimates based on household smveys (ENH19, ENH61, ENH74, BNH77). 1.9 The decline in national poverty was concentrated in 1978-91. Poverty increased in 1992. This reversal in the trend toward poverty reduction was the result of a sharp increase in nrual poverty between 1991 and 1992, associated with the crisis in the 6 To ensure consecy between the differen houshold srveys the urbn sector is Hinmtd to seven large and medium size chies Dannquila Bogot, BucaraManga. Cai, Manizales, MedeUli and Past) which are incde in al of them. As a consqne , the restg numbers are not direcy comparable with those discussed in the previous section which have a broader covemge. 7 Indeed, the anaysis shows an unambigus delie in poverty-naoa, rural and urban-betveen 1978 and 1992. That is, the decine is observed regardess of the choice of poverty meues and pover lines. 8 As explahied in footnote 6, the incidence of urban povety was estdma on the basis of iofmtion corresondn to seven adwr than bthy tbrce cities. 9 See Pa _opoulos and others, op. cit. -7 - agriculture sector and the corresponding decline in rural incomes during 1992 (Figure 1.3). This decline resulted from a number of factors: a very severe drought, falling international prices for both exportable and importable products (especially coffee), an intensification in rural violence, the dismantling of the system of support prices, and the short-term difficulties faced by agriculture in adjusting to the relative price correction from the liberalization of trade.'0 1.10 In the sub-period 1978-88, Colombia experienced a significant reduction in both rural (from 38.4 percent to 33.1 percent) and urban poverty (from 12.1 percent to 8.0 percent), each contributing evenly to the decline in national poverty. During this period real per capita income (in terms of the cost of the food basket for low income families) of the bottom 20 percent of the income distribution grew at an annual rate of 1.3 percent in rural areas and 2.2 percent in urban areas. During this same period average income for all groups fell by 1.6 percent in rural areas, and grew at an annual rate of only 0.7 percent in urban areas. This resulted in a sharp reduction in income inequality (both urban and rural) between 1978 and 1988, which is the main factor explaining poverty reduction during those years." Intersectoral population shifts, on the other hand, had a very limited effect on poverty. The distribution of population did not experience significant changes, and the share of the rural sector decreased by only two percentage points during the period under review. 1.11 Since 1988, however, changes in poverty have followed different patterns in the 2 sectors. The incidence of urban poverty has remained virtually unchanged-about 8 percent-between 1988 and 1992. Rural poverty, on the other hand, fell from 33 to 29 percent between 1988 and 1991, before increasing by approXimately tWo percent in 1992.12 Overall, the evidence reviewed indicates that, since 1988, changes in national poverty are mainly associated with trends in the rural sector. 10 Chapter 2 presents a detaied analysis of the extent to which each of the mentioned facto. contbud to the reducdo in rural incomes. 11 See Annex I for the resls of die decomposition analysis of changes in poverty between 1978 and 1988, which identifies the relative contbution of growdh and inequalty. Rsearch indicates t reallocation of capital to enhamce agricultue productivity and humn ass was the main factor explaining the teduction of income inequalty in Colombia during the 1970s and 1980s. Furthermore, variations in returns to human capital were the main channel by whkh labor income inequalities were ameliorated durig Colombia's stuct transformation (see Londoflo (1990)). 12 Real per capita income of the 20 percent poorest indiviuals m urban areas experienced only a minor increase ben 1988 and 1992 (inceasing 1.8 percent up to 1991 and then falling 1.1 percent in 1992). The trend is very different in rural areas, where real incomes of the poorest 20 percent increased at an annual rate of S percent between 1988 and 1991. and fel by 13 peren in 1992. Ihese values are estmated on the basis of the household survey and using the cost of the food basket for low income groups as a delator. -8- The Index of Unsatisfed Basic Needs (NMI)13 1.12 So far the analysis and discussion have focused on poverty as measured by income insufficiency. But since 1987 the Goverunent of Colombia has used the index of "Unsatisfied Basic Needs" (NBI) to measure poverty. Based originally on the results of the 1985 census, which did not include questions on household incomes or expenditures, the NBI methodology has defined poverty in terms of five indicators: (a) inadequacy of housing (judged by types of materials used for walls, roofs, and floors); (b) lack of access to public utilities such as electricity, potable water, and sanitation; (c) high household occupation density (more than three people per room being the threshold); (d) school-age children not attending school; and (e) large number of dependents per wage earner per household (more than three being the threshold). A household is defined as living in poverty if it meets any of the five criteria. More recent studies have excluded the last indicator ("high dependency") given the availability of nationwide information on household income for which it was supposed to be a proxy. 14 1.13 During the last few years the NBI index has been given a prominent role as a poverty indicator in Colombia, and it has been frequently used by the government to target public programs for poverty reduction. However, a straightforward comparison with the income-based indicator used in this report shows some potential problems with the NBI index as it is currently estimated. Although the poor (defined on the basis of three alternative poverty lines) experience higher incidence of the four NBIs than the population as a whole, as shown in Tables 1.3 and 1.4, the incidence of unsatisfied basic needs (as measured by the NBI index) among the income-poor is far from overwhelming. In the urban areas, for example, only one in three individuals in the extremely poor group is poor under the NBI criteria. To the extent that policymakers wish to reach families with incomes below the poverty line, the use of the NBI index would involve large errors of exclusion and inclusion. 1.14 First, by targeting specific unsatisfied needs, most of the poor are not reached. For example, more than 90 percent of the poor in urban areas and 80 percent of those in rumral areas are not identified by the indicator of inadequate housing. By targeting the composite index a larger proportion of the poor can be reached, particularly in rual areas, but at a much higher cost. In the rural areas, for example, the targeted population would be 44 percent of the areas' total population. Second, using the NBI index (or its components) for targeting purposes implies that a large number of non-poor would be included in poverty reduction programs. For example, 80 percent of those classified as poor according to the NBI index m urban areas have incomes above the lowest poverty line. The diagnosis does not improve much even if one is stricter about the use of the NBIs, for example, by focusing on the population that has three or four unsatsfied basic needs instead of only one (Figure 1.4). While the group with three or four NBIs represents only 4 percent of those with NBIs in urban areas and 15 percent of those in rural areas, even among this reduced group more than 60 percent of the individuals are not income-poor. 13 Aemx I mviews the ind of each of te four NBIs in both the rurl and uban areas thgh time, usir4 informaion for 1973.1985,1988, and 1991. 14 PNUD (1991). -9 - Table 1.3: NBIs and the Urban Pbor, 1991a Inadeqwte Inadequate Crowded Schol Populton Group Housing Serices Households Absence Nlw Incidence in population (percent) 2.2 1.0 8.3 2.7 11.5 Incidence among poor (percent) 1 Poverty Line 7.3 2.8 26.1 7.9 33.0 1.5 Poverty Lines 5.9 2.6 21.6 6.4 28.2 2 Poverty Lines 4.9 2.2 18.2 5.4 24.0 Percentage of non-poor among 1 Poverty Line 73.8 78.7 75.4 77.1 77.6 1.5 Poverty Lines 48.4 52.0 50.5 55.1 53.4 2 Poverty Lines 29.9 31.5 31.7 37.6 35.1 a. The September 1992 survey does not include the information needed to calculate the NBI index. 1991 survey information (ENH74) was used instead. Urban areas include only seven large and medium-size cities. Source: Esimates based on household surveys (ENH74). Table 1.4: NBIs and the Rural Poor, 1991 Inadequate Inadequate Crowded School Populdation Group Housing serIces Houeholds Absence NM Incidence in population (percent) 14.2 24.1 20.0 12.8 44.1 Incidence among poor (perent) 1 Poverty Line 19.6 35.1 32.4 16.8 61.6 1.5 Poverty Lines 17.2 31.1 28.6 15.6 55.9 2 Poverty Lines 17.2 29.4 26.2 15.2 53.8 Percenage of non-poor among 1 Poverty Line 60.0 57.9 53.1 61.7 59.5 1.5 Poverty Lines 42.9 39.2 32.7 42.3 40.1 2 Poverty Lines 24.3 23.8 18.1 25.7 23.7 Source: Esdmates based on household surveys (ENH74). - 10- Figure 1.4: Distribution of the Poor According to NBIs, 1991 FJURAL AFAS UREBAN AREAS 3 % 4NE 3 4 NED 1) 2 NE 2 N8I 1 NBI 1 NBI Source: Estimates based on household surveys (ENH74). 1.15 The choice of poverty indicators can have important policy implications.'5 In identifying the poor, differences in definitions do matter. Table 1.5 presents some basic characteristics of the poor according to six alternative definitions of poverty. In addition to the three poverty lines and the NBI index, the table also includes individuals with unsatisfied basic needs and incomes below one poverty line, and those who are either below two poverty lines or with unsatisfied basic needs. The NBI poor have significantly higher levels of income than those identified by poverty-line criteria. Average household income per capita among the former is almost three times higher than those earning less than one poverty line. An aspect in which NBI and the "extremely poor" criteria do not differ is in the location of the poor-according to both criteria three out of every four poor individuals live in rural areas. The regional distribution of the mrual poor differs significantly between definitions. Under the poverty line criteria, rural poverty is more heavily concentrated in the Oriental region, while under the NBI criteria the Atlantic region has a larger share of the rural poor. 1.16 The Govemmeit of Colombia, through the Misi6n Social, is currently revising the NBI metiodology. An alternative poverty index that would incorporate information on housing, services, income, employment, human capital, and other factors affecting standards of living is being considered."6 Using a broader set of information than the current NBI index, the new index will measure the severity of poverty. iS See Glcwwe and van der Gagu (1990). 16 See Velez and Guenero (1994). - 11 - Table 1.5: Characteristics of the Poor According to Alternative Defnitions, 1991 1 2 Povery 1 1.5 2 Povety Line & Poverty Poverty Poverty Lines Chaacteri&* NBI li'i Lines Lines or NBI NBI Popu&xtion Monthly income per capita (pesos) 8,388 8,863 13,334 17,752 22,304 25,631 56,721 Activity ratio (percent) 31 30 31 33 35 38 41 Head's age (years) 45 46 45 45 45 46 46 Female head (percent) 15 16 16 17 17 16 18 Head's schooling (years) 2.7 3.5 3.8 4.2 4.2 3.3 6.2 Living in rural areas (percent) 84 74 65 60 61 74 43 Atlantic Region 29 23 24 25 26 32 24 Oriental Region 34 33 29 28 27 27 26 Central Region 16 20 24 26 27 24 31 Pac fic Region 20 24 23 21 20 17 19 So ce: Esdmates based on household surveys (ENH74). III. A Prople of the Poor 1.17 Households in extreme poverty differ from the average household in several identifiable ways."7 In many cases, those differences are poverty-generating factors. For this reason, a profile of the typical household in poverty can be a powerful tool in the design of poverty reduction programs. This section defines such a profile for the poor in Colombia. Principal CharacteAstics of Households in Poverty 1.18 Table 1.6 presents a comparison of the demographics of poor and nonpoor households, both in urban and rural areas. The first striking difference between poor and nonpoor relates to household size and number of children.18 Poor households have, on average, twice the number of children as nonpoor ones. This is an important factor explaining low per capita incomes resulting from high dependency ratios. The other element 17 Tne results discussed in tbis sectton correspond to households with icome per capita lower than the base poverty line, the group referred to as the exremely poor. 18 Me analysis refers to chdren under the age of 10. - 12 - is the nmber of people employed. Note that in urban areas the average number of people employed (per household) is lower among the poor. Households in poverty are also relatively young and poorly educated. Both the head and the spouse in poor households are, on average, two years younger than those in non-poor ones. The difference in schooling between poor and non-poor households is approximately three years in urban areas, while somewhat lower in rural areas. The combination of young parents-and more children-with low education results in low earning capacity for these families. While earning capacity is expected to increase over time, low education is likely to be a handicap over the family's entire lifecycle. Finally, a female head-without other considerations to family composition-does not appear to be a characteristic that distinguishably identifies poor households. Table 1.6: Characteristics of the Average Poor Household, 1992 Urban Rwul Charwcisd Poor Non-poor Poor Non-poor Household size 5.2 4.4 5.6 4.5 Number of children 1.7 0.8 1.8 1.0 Number of people employed 1.1 1.8 1.7 1.9 Head's age (years) 42.0 45.0 46.0 48.0 Spouse's age (years) 36.0 39.0 39.0 40.0 Head's schooling (years) 4.6 7.5 2.6 4.0 Spouse's schooling (years) 4.9 7.4 2.9 4.3 Female head (percent) 27.0 23.0 18.0 18.0 Source: Estimates batsed on household surveys (ENH77). 1.19 The way in which each of the factors affects poverty outcomes can be seen more clearly by comparing the incidence of poverty among groups of households with similar charactristcs, as in Table 1.7. For example, the incidence of poverty among households with four to six members is twice that of households with less than four members. Similarly, the head count index increases with the number of children-there is a fourfold increase in the incidence of poverty that occurs in urban areas when one compares households with one and two children. As expected, the incidence of poverty diminishes as the head's level of schooling as well as his/her age rises, except that there seems to be higher incidence of poverty among households whose head is over 65. Finally, and consistent with the - 13 - information summarized in Table 1.6, female-headed households have only a marginally higher incidence of poverty, and only in urban areas.19 Table 1.7: Indidence of Poverty by Household Group, 1992 Urba Rwl Household factor Gnnpmng (percent) (percent) Number of members Less dtan 4 4.1 14.5 4-6 9.5 28.6 Number of children 1 6.7 22.4 2 23.3 33.6 3 26.6 44.8 4 or more 39.2 62.7 Head's age 26-35 years 11.0 30.3 3645 years 8.5 31.2 46-55 years 6.3 21.8 56-65 years 6.3 21.8 66 or older 7.3 33.5 Head's education level 0 years 19.6 33.8 14 years 13.2 31.0 5-10 years 8.7 19.6 11 years or more 2.1 4.9 Household head Male 8.0 26.6 Pemale 9.7 26.4 Source: Estmates based on household surveys (ENH77). V9 The insignificant effet on poverty of female headsbip, apparent In Tables 1.6 and 1.7, could be somewhat surprising given sanlod p _smVdom. However, once one brigs into consideration distwtions related to family composition the expected die corlaona appea. In pacular, it is important to disinguish beween fema-headed households: (a) with and without childn, and (b) bein sigle and two-pab households. Whie those houolds without chidrn have a rdlavely low ncidence of povety (5 percem in uban es and 15 peren in nrua areas), those with children experience much higher incidence of povety (appmnrael_y 17 per in urban areas anid 40 percent in nal aeas). On the other han single parm households-of which 84 per have a female hehave a significantly higher incidene of poverty (twice dtat of two-parent households in urban areas). Fuorder discusion on the ise of female-headed households is presetd in the tae, as part of the analysis of the probability of bein poor. - 14 - The Probability of Being Poor 1.20 To establish the relative role of the different factors in the configuration of the poverty profile in Colombia, it is necessary to isolate their individual effects.20 One of the benefits of such analysis is to be able to assess the impact which a change in a particular factor would have on the probability of an individual being poor, if all other factors are kept constant. These results are presented in Table 1.8. For example, it is estimated that an additional year of education, for a household head, decreases the probability of the household being poor by 3.1 percentage points in rural areas and by 1.2 percentage points in urban areas. The overall results can be easily summarized. The risk of a household being poor diminishes as the head's (and the spouse's) level of education rises, and to a lesser extent as the head's age rises. It increases, on the other hand, with the number of children and with female headship. Table 1.8: The Probability of Being Poor, Marginal Effects, I9M2k (percent) Household Chwwracteric Urban Rural Number of children 4.6 9.7 Head's age -0.5 -0.2 Head's education -1.2 -3.1 Spouse's literacy -4.6 -2.0 Female headshp 4.0 10.5 a. The marginal effects should be interpreted as the percentage change on poverty associated with a unit change in the explanatory variable. Source: Esmates based on household surveys (ENH77). 1.21 So far the results have been presented in the context of marginal effects. A more revealing approach is to assign differing values to target characteristics and simulate the resulting probability of being poor while mainaining all other variables at their national mean values. In this context, it is possible to talk about the probability of being poor for given factors, and comparisons can then be made across characterisdcs. Examples are presented in Figures 1.5 and 1.6, where the probabilities of being poor are quantified for households with a given mnuber of children and female headship. The simulation approach is probably most fruitful in analyzing characteristics that allow for a high degree of differentiation-like education or age-or those for which the control of other factors is essential in the analysis-like spatial location. 20 The probability of an indiviual being poor is a fundtion of the varlous personal characteristcs so far discussed, such as age, gender, and years of schooling. Since the factors being exmin are limied dependent variables, tL, way to technically handle their relatonship is by properly fiing a probit model. A detailed description of the model and the estimaon results is presented in Annex 1. Urban and rurl areas are considered separately. - 15 - FIgure 1.5: Probability of Being Poor and Number of Children In the Household, 1992 Oy ~~i_dCh Sowre: Estimates based on household surveys (ENH77). FIge 1.6: Gender and Poverty, 1992 RdzOIIY of Ddn Pba x 10_ 30 2hadJ Uzbai Sow=e: Esimates based on household smuveys (ENH77). 1.22 EducatIon Among the various charcristics consdered, education allows for a substmanl degree of dif on with respect to the probabilit of being poor. As Table 1.9 idIcates, this ranges on average from 40 percent for those with no education in the rural areas to 4 percen for those with some university (or post sw ondary) education. For any give level of education-either for the head or the spouse of the household-the probabilt of being poor in the rural areas is about double that in the urban areas. Although not presenotd in Table 1.9, a rural household in wbich neither the head nor the spouse have any - 16 - education has a 50 percent probability of being poor. That probability decreases to 22 percent for a mral household where both the head and the spouse have primary education. Table 1.9: Education and the Probability of Being Poor, 1992 Edaucatdon Rural Areas Uran Areas Household Member Level (percent) (pereent) Head 0 years 39.6 21.2 5 years 22.9 11.4 8 years 15.2 7.4 12 years 7.9 3.9 Spouse Illiterate 35.9 16.9 Primary 26.0 9.8 Secondary 14.5 6.7 Source: Esimates based on household surveys (ENNH77). 1.23 Age. As bas been pointed out before, the probability of being poor diminishes with age. This observed relationship can be explained, a priori, both because: (a) incomes tend to increase with age; and (b) the composition of the family changes in time. By distinguishing the independent effect of each of these factors, the simulation analysis allows a better understanding of their corresponding interaction. Table 1.10 portrays the dynamic changes in poverty risks experienced by a typical Colombian family through its lifecycle. As shown, the probability of being poor increases as a young couple has children (between ages 25 and 40), and later diminishes, as those children grow up and leave the household. This pattern is more intense than it would be if one would simply consider changes in age, disregarding changes in family composition. 1.24 Spatial Location. Large differences in the incidence of poverty among mral regions remain even after differences in household characteristics are considered. Figure 1.7 presents the predicted head count index for the four mrual regions. These values are estimated considering households with averge national characteristics. In other words, the values control for the fact that households in the four regions differ in their level of schooling, age, and family composition. The main finding is that households with similar characteristics experience a higher risk of poverty in the Oriental and Pacific regions than in the Adantic and Cental regions. 1.25 Comparison with the actual head count indices (also included in Figure 1.7) indicates an interesftng contrast between the Atlantic and Oriental regions. The incidence of poverty in these two regions is roughly similar-28.2 percent of the households in the Atlantic and 29.7 percent in the Oriental region. Once differences in household - 17 - Table 1.10: Household Poverty Over the Lifecycle, 1992a ProbabIy of Being Poor Number of Household (percent) Head's Age Children Size I 25 0 2 18.6 8.9 30 1 3 26.4 12.5 40 2 4 34.4 15.3 50 1 4 22.5 7.4 60 0 2 13.4 3.5 a. These probabilities were estimated for a male-headed household with average characteristics. Source: Estimates based on household surveys (ENH77). Fgure 1.7: Regonal Difference in Poverty (Head Count Index), 1992 celC Aa Antlanc 0 5 10 15 20 25 30 35 40 HNo Count 1win_ Smoce: Estimts based on household surveys (ENH77). characteristics (education, age, and family composition) are controlled &lr, the incidence of poverty in the Atlantic region is estimated to be almost 10 percentage points lower than in the Oriental region. The difference between predicted values reflects the distinct effects of other fators like in as ure in each of tbese regions. But another aspect of this comparison needs to be considered. In the Adantic region the predicted incidence of poverty is lower than the actual. It is the only region in which this occurs. The explanation relates to the direct and negative effects of observed family characteristics in the incidence of poverty in the Atlantic region. Among all regions, the Atlantic presents-on average-the worst set of - 18 - family characteristics within its population. To mention a few, the rural Atlantic region fares poorly both in terms of fertility and education. It experiences, by far, the highest fertility rate in the country (5.4 compared with the national average of 2.9). Its performance in education is also very inadequate: six out of the seven departments of the Atlantic region have the lowest rates of coverage for primaty education in the country." Povenly and the Urban Labor Mawet 1.26 Low-income households derive most of their income from labor market sources, either through wage-employment or self-employment. Only about 5 percent of the income of the lowest two deciles is derived from other sources. As a result, the risk of poverty is closely associated with the extent to which a household participates in the labor market and the way in which the market remunerates its labor. This section analyzes these questions within the urban areas. 2 1.27 Labor Market Participation. The low labor market participation of poor households should concern policy makers. The participation rate declines from 65 percent in the ninth decile to 35 percent in the first decile. Figure 1.8 shows that this pattern is more pronounced for women. Only 22 percent of the adult females in the lowest two deciles of the distibution participate in the labor force. Even when they do participate, women in the lowest deciles tend to work fewer hours. Working women in the first decile work four hours fewer a week than the average woman. Fre 1.8: Labor Force Participation, 1990 40 lgt 0.n 2a Om. rdDoeo 4t Doec 5Uh VW OtIh NO 7tm0.cf O 8fI0. Oft. Si-as Top 5% Sore: Estmates based on housthold surveys (ENH68). 21 lhe informan is shown in Annex 1, Table 1.28. 2 Ae analysis is based on dte June 1990 Household Surey, which includes a special module for de inormal labor met. The detailed result are presened in Annex 1. Tbis analysis is based exchsivey on camings (that is, it excludes other incomes) and does not corre for income under-reporting. us, it is not direcdy comparable with the Itset discssed in other parts of the report - 19 - 1.28 Educated Colombian women are more likely to work. Mothers are less likely to work.23 For a poorly-educated mother, the potential benefits of working outside the household can be easily outweighed by the corresponding costs, particularly when transportation costs are taken into account. Limited access to affordable child care facilities is another deterrent. 1.29 Age also affects labor force participation significantly. Labor force participation varies significantly over the lifecycle (Figure 1.9). Among men it reaches 90 percent at 30 years of age and remains above that level until the age of 50. Among women, participation rates reach 50 percent around 25 years of age, remain near that level up to 40 years of age, and then decline rapidly. Participation rates of older individuals remain high, particularly among men. The participation rate among men at age 60 is 70 percent, and remains at 40 percent for those 70 years of age. Lower-income men are more likely to continue working than better-off men. This may reflect limited access to social security, which forces a significant proportion of the elderly to keep working.' Figure 1.9: Labor Force Participation Rate, 1990 100 so 80 70 -'so60 f 4 | - Men .50 -- Women 40- 30 20 10 so eD _ . * o co a o t c o 0 ( 0O _- _ _ CS N e e e 0 0 * X - ID IC IC e 0X Age Sorce: Esdmates based on household surveys (ENH68). 23 A female with incompee prniay education has an 11 percent probabilty of participation compan:d to a 34 percent probabilty for one with secondary education. Mme presence of children reduces the probability of participation from 25 to 20 percent See Velez and Wimter (1991). 24 Only 18 percent of the utban population older than 60 years receives pensions from the social securty system, wbile 45 percent of that same growp does not receive any tpe of income and thus depends eniely on the support of their Smilies. See Castailo (1992). - 20 - 1.30 Wage Differentials. Two other important market characteristics are worth mentioning: observed gender and education wage differentials. Female workers earn, on average, 25 percent less than male workers. Half of this difference is because women work fewer hours than men. Thus, the average gender differential in hourly earnings is only 12 percent. This differential is small compared with other Latin American countries, suggesting a good degree of female integration in the urban labor market.25 The barriers experienced by Colombian women mostly relate to difficulty finding work (particularly in the case of poorly- educated mothers) and not to unequal pay. 2 Income differentials, on the other hand, do show that the Colombian labor market rewards education. As shown in Table 1.11, men's returns for 5 (primary), 11 (secondary), and 16 years (university) of education are 33 percent, 84 percent, and 169 percent, respectively. There are no significant differences in returns for men and women.27 Table 1.11: Education - Income Differentials, 1990 (percent) &lucaaon Level Males Females I year 14 23 S years 33 36 11 years 84 91 16 years 169 165 Source: Esumates based on household surveys (ENH68). 1.31 The estimates of the earnings equations presented in Annex 1 indicate that the retrns to primary education are not evenly distributed by years of schooling. The first year, particularly among females, shows a very large return. To a lesser extent this is also true of the last year of primary schooling." The high reurns to the first year of education may demonstrate the high payoff from basic socializing skills a child learns when sarting school. The low reus to intermediate years of primary schooling may reflect that a complete primay education is a threshold in the labor market. 1.32 As the probability analysis already indicated, education can be a powerful tool to reduce poverty. Giving primary education to an illiterate should have a large impact on his/her earning capacity. Based on the estimated returns to education and given current levels 25 aPsmaopolous and Tos (1991) fon that the average male pay advantage was 30 percent in a sample of 15 countries in tie region. 26 Ihe observed gender wage differentials are not assocated with differenes m endowmes or employment charcts, but with the reward males and females recerve for those. Annex I presents the reut of such decompostion anaysis. 27 These rates of reurn were estmated controlling for sectoral choice m employment. When rates were estmated without controls for secor of employment, females showed higher returns to education than males Te resuts of altenatve regresions are icde in Annex 1. 28 Mhe fln and fifh years ogedher explain 25 and 29 pementag points of remn for maes and femals, respecvly. - 21 - of education in urban areas, however, the direct poverty reduction effect of a "basic education for all" strategy would not be very large. If every worker with less than five years of schooling was given exactly that level of education, the proportion of the families with labor earnings below the poverty line would fall by one percentage point. But it would still have a significant effect on urban poverty through its indirect effect on fertility and family size. When the potential impact of changes in education on expected family size is considered, the same intervention would reduce poverty by four percentage points. A strategy that gives five years of education to everyone also increases the chances for further schooling. Thus, these estimates underestimate the poverty reduction effects of such a strategy. Improvement in the quality of primary education and expansion in the coverage of secondary education would probably have a significantly larger effect. If every worker with less than eleven years of schooling was given exactly that level of education, the proportion of families below the poverty line would drop 50 percent. IV. A Profile of Social Indicators 1.33 During the past four decades the living conditions of the average Colombian have improved significLntly. Basic social indicators show persistent p- zgress. Life expectancy at birth has increased by eighteen years since the early 1950s, while infant mortality rates fell from 123 per thousand live births in 1955 to recent estimates of 30 per thousand live births in 1990 (Figures 1.10 and 1.11). The prevalence of global child malnutrition has fallen from 22.7 percent in 1966 to 12.5 percent in 1986.29 1.34 A rapid reduction in population growth, currently at 1.7 percent per year, and significant expansions in education levels have fostered Colombia's social progress. The slowdown in population growth since the early 1960s is the result of one of the world's fastest fertility transitions. Fertility rates have fallen from seven to three children per women in a period of 25 years.' Interestingly, fertility has declined both in urban and nmral areas, with the latter falling faster in the 1980s (Figure 1.12). During the same period, as can be seen in Table 1.12, the Colombian population became more educated. Illiteracy fell more than 50 percent, and average levels of schooling doubled since 1965. Table 1.12: Evolution of Education Indicators Indkator 1965 1975 1985 Iliteracy (perent) 26.5 17.7 12.2 I Average Schooling (yea) 2.3 3.0 4-3 |Primuy E&dcaion (percent) 54.7 79.4 87.8 Source: DANE and DNP. 29 DNP (1991). 30 The rictiaon i fervels res was of etuc of a accessfw anadoy pnomgnteducionsPeandr ctilden Pe o ua g In a( en1 of inreastng levels of educafton and corwomitmt teductions Ithe de emnaW for duldnen. See Rofhan (1992). - 22 - Flgure 1.10: Life Expectancy at Birth for Men and Women 765 70 60 ~ ~ ~ ~ ~~~~~~~--men i i _ _ _ _ _ _ _ _ _ _ _ 0 W olme 46 . _ _ . . - _ - Source: DAM igure 1.11: Colombia - Infant Mortafit Rates 140 120 100 so '60 40 20 So&uce: DANE. igure 1.12: Evolution of Fertility Rates by Zone 7 P"bd ~ ~ ~ ~ ud Source: DANE. - 23 - 1.35 Today, Colombia has better social indicators than those expected for a country at its level of income. Based on the analysis of information corresponding to ninety countries, the under-five mortality rate in Colombia is 12 percent lower than expected for 1989. Life expectancy is two years higher.3' There was significant progress between 1960 and 1989. According to the estimates, Colombia's under-five mortality rate is 7 percent above its expected level in 1960, and 12 percent below the corresponding level for 1989.32 1.36 The pace of progress in Colombia needs to be put in perspective. As compared with the performance of other countries in Latin America and the Caribbean, the Colombian experience does not appear exceptional. Witness the reduction in the under-five mortality rate. This finding can be clearly presented graphically. Figure 1.13 shows the divergences (expressed in percentage terms) from expected rates of mortality in 1960 and 1989 for the 24 countries under consideration. In Colombia the divergences are 7 and -12 percent, respectively.33 The distance from the 45-degree line is an indicator of the relative progress that a country has made in reducing the under-five mortality rate between 1960 and 1989. The further below the line a country is, the larger its progress in reducing mortality rates. The fact that Colombia falls below the line is an indication that it reduced the mortality rate beyond the level that could be expected-given its growth performance during the period. However, it is easy to notice that countries like Chile, Nicaragua, Trinidad & Tobago, Costa Rica, and Jamaica are further below the line, suggesting a relatively faster rate of progress in the reduction of under-five mortality rates. Furthermore, while some of those countries (such as Chile and Nicaragua) started with expected mortality rates higher than Colombia, others (such as Costa Rica and Jamaica) already had low mortality rates in 1960. 1.37 Recent general improvements in Colombia's overall standards of living mask large remaining differences among regional and social groups (Figures 1.14, 1.15, and 1.16). The Pacific region shows the worst indicators. Life expectancy is two years below the national average. Infant mortality is 20 percent higher than the national rate. Significant differences also exist within regions, and specific departments can be identified as clear outliers34 Three departments-Choc6, Nariflo, and Caqueta-present particularly worrisome health conditions. In the Choc6, the department with the worst social indicators, life expectancy is sixteen years lower than in Bogota (fourteen years lower for men and eighteen for women). Infant mortality rates in Choc6-at 82.3 per thousand live births-are similar to the national average twenty years ago. In CaquetA, they are similar to the national average ten years ago. In terms of education, the problem of low primary coverage seems concentrated in the Atlantic region. Except for La Guajira department, it shows the lowest indices in the country. 31 The expected values were esdimated econometrically on the basis of a suggested relatonship between social indicators and GDP per capita. Anmex 1 has details on the information and medhodology used. 32 The proportional divergence from expected vahles for 24 countries in the Latin America and the Caribbean region is shown in Table 1.31 in Annex 1. 33 Countuies in the soudteastern quadrant, for example, had higher-than-expected mortality rates in 1960 and lower than-expected in 1989. 34 See Annex 1, Tables 1.24, 1.25, and 1.28. R ~~~~~~Divergence 1988 W~~~ °o * *o * o * r0 e a e a a e a e a 0 * 5- *~~~~~~~~~~~~~~ g~~~ * F.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - 25 - Figue 1.14: Life Expectancy, by region AtlFi Bogd a Central OWomen Pacfic= Naino 1 Caquet (C Chco (PI 45 50 55 60 65 70 75 Yew. Source: DANE. Figure 1.15: Infant Mortalt Rates 0 1 0 20 30 40 50 s0 70 80a 90 Value Source: DANE. - 26 - F%mr 1.16C: Ferfility Rae by Level of Educatio and SZane, 1987-90 SuMeior0 - Frimary .-i:X ;aUn No Education i O 1 2 3 4 5 6 L ~~~~~~~~~~~~Rake Source: Enusau de Prevalena Demogr4fica y Sud(EPDS 1990]. 1.38 R=Me infonnathon on the midence of duhld mainunon alo iat lage differences between sociwmic group8. The VWide=e, presenfted in Taible 1.13, clearly sbows Inutins owidon with the economir and social mcWisnce of the family. C)f particiar inWest is the impotat role played by the mote's eahcationl level. This also mluences the probability of child nulnutrton dugh its effea on family size. The iniec of global child manptoXfor examp, -is SO peret lower in hbeolds where the motber bas primary education tfian in thos in whmich she bas no etcation In addition, as women withi prmary education have significantly lower fertility rates thanhose with no education (igme 1. 16), a fiuter reduction in tbe incklene of uonassociate d with family size should be expected. 1.39 The susXined imrvmn in social indicat in the last deas shotld not divert anention bom tbose Colombians still experiewing condkitins tbat the couty, as a whole, left bebinld ten or twny years ago. The expiw of otbor couse in the region- as discussed above--suggst tha fasteras of mxprouvement in the main social indicators can be achieved. - 27 - Table 1.13: Global and Chronic Malnutrition, 1989 G(hba MalnuUieon Chmawic MauUNNOR Charactefisdc (percent) (percewt) Size of Family Less than 4 members 9.6 15.9 More than 6 members 15.1 24.2 Mother's Level of Education None 20.8 Primary 10.8 24.0 Secondary 9.8 12.7 University 5.8 5.8 Monthly Family Income One Minimum Salary 16.2 28.1 From 1.1 to 2.5 Minimum Salaries 10.5 15.2 From 3.6 to 4.5 Minimum Salaries 8.1 11.0 More than 4.6 Minimum Salaries 2.5 6.5 Basic Sanitation Water Supply 9.6 15.4 No Water Supply 17.5 27.7 Sanitation Service 9.8 14.7 No Sanitation Service 18.3 30.3. Water Treatment 11.2 19.6 No Water Treatment 15.7 21.8 a. Refers to weightage relationship. b. Refers to height/age relationship. Source: DNP (1991b). Chapter 2 Promoting Income-Earning Opportunities for the Poor: The Macroeconomic Framework Colombia's excellent record ir: macroeconomic management has provided a solid basis for continuous gains in poverty redruction. The 1991-92 crisis in the agriculture sector-related to both internal and externat shocks-led to important reductions in rural incomes and a corresponding increase in poverty indicators. Colombia must adopt corrective measures to get back on the path to poverty reduction. The oil discoveries in Cusiana provide an excellent opportunity to attain a higher growth path and devote resources to much needed infrastructure and social services that will directly benefit the poor. Good macroeconomic policy is essential if Colombia is to realize these potential benefits. The appropriate macroeconomicpolicyframework should be based on prudent fiscal policy that ensures continuous gains in the fight against inflation and investment in projects with high economic and social returns. I. Background 2.1 Colombia's competent macroeconomic management, based on the govermentis continuous commitment to correcting internal and external imbalances, is best evidenced by its enviible growdt record during the 1980s (Figure 2.1 and Table 2.1). During the "lost decade" in Latin America, Colombia maintained an average annual growth rate of 3.5 percent while adjusting to a worldwide recession and coffee and oil price swings. Between 1980 and 1985 Colombia's economy grew at an average 2.2 percent annually. Subsequent macroeconomic adjustment raised the average annual growth rate to 4.5 percent between 1985 and 1990. Economic growth was positive in every year during the decade. Although Colombia's economic growth in the 1980s was not as rapid as in the previous decade, maintaining stability and avoiding a recession provided the basis for continued success in reducing both poverty and ineqWalit. 2.2 The relatively smooth growth of the Colombian economy masked important structural deficiencies. A closer look at economic performance indicators suggests a lack of dynamism in the economy since the mid-1970s. The contribution of manufacturing to GDP remained relatively constant. Non-traditional exrort growth was not diversified. Manufacturing remained oriented toward the domestic market. Low and frequendy negative productivity growth in most sectors suggests that economic growth depended on ireasing I As Modey (1992) indicat Colombia's sowdown in economic gfowd dring th flirt pan of th 1980s-whlch led to iile capacky in the economy-prvided te baS for an expansiay adjusmnt and sbiaton effot In 1986-90. Combining a deprecin of the ral exchange rame of over 50 pecn and improvements in the fiscal positionte authrii wer aucceaf in d;eltg the balance of payms deficitt while both emplyme and real otp grwth _ ad. Tte shp 8 in ra wages ananed m 1980-84 were mainaed and povery reducton througho the decade was achievd. 1970 -~~~~ ;* 1971 ~~~ [ ~~~~~1972 1 1973 1974 1975 1977 B ~~~~19787 !A%O P C.60 o17 1979 1980 I - ~~~~~~~1981 1982 1983 1984a I. ~~~~~~1987 1989 11990 - 30 - capital inputs rather than efficiency gains.2 To advance the gains in poverty alleviation, Colombia required a change in its growth strategy. 2.3 Recognizing that macroeconomic measures alone would be insufficient to raise productivity and output growth, and to promote long-term employment and earning opportunities, the Barco Administration anounced in Febuary 1990 its Economic Modernization Program (EMP) to improve efficiency of resource allocation and use. The EMP contained a set of structural reforms and accompanying macroeconomic policies designed to raise economic growth to 5 percent anmally, bring inflation below 20 percent, and reduce poverty. The centerpiece was trade refonn. It envisaged the gradual elimination of quantitative restrictions (QRs) on manufactured imports and scheduled reductions in tariffs to reach an average tariff rate of 15 percent in 1994. Taking advantage of a strong balance of payments in 1990-91, the Gaviria Administration advanced the timetable for implementation. The trade reform was put in place in 18 months, not the five years originally envisaged. The government also went beyond the original scope of the EMP by extending trade liberalization to agriculture. The state monopoly on key agricultura! imports was eliminated, and a variable tanff scheme implemented for some commodities. Although progress has been slow in completing a regional trade agreement under the Andean Pact, Colombia and Venezuela have freed their bilateral trade and adopted common external tariffs. In December 1993, Colombia signed a bilateral agreement with Chile that will allow for the gradual elimination of tariffs over a ten-year period. A free-trade pact between Colombia, Mexico, and Venezuela has also recently been completed. 2.4 The Gaviria Administradon has also pursued other structural reforms. To improve resource mobility and facilitate the supply response to the new trade regime, fmancial sector reforms authorized ownership of fiancial institutions by foreign investors and free entry into all segments of the market, subject only to pmdential requirements. Access to foreign exchange was improved, and the government liberalized direct foreign investment. The labor regime was modified to reduce labor rigidities and facilitate industry restructuring. Reforms in the public sector were to improve the efficiency and focus of public expenditures. Public monopolies were eliminated in sectors critical to trade, including railways, ports, shipping, and agricultural marketing. Most of the govermment's non-oil industrial holdings were divested, as were the five banks nationalized during the banking crisis of 1982-85. Further divestment is underway. The intention to reorient the state in a liberalized economy was spelled out in the government's 1990-94 Development Plan (the Revollucidn Pacffica), emphasizing social sectors, infrastrucre, and enviromnental protection. 2.5 Structural reform during the last three years was hampered by significant shocks to the economy. For poverty reduction, the most important ones are those that 2 Since the nid-1970s. toal facor productivty gsowtb fell both in absolute terms and as a propotion of ouWut growth. 4 major coributing actor was the decling produc of capital, which was also rfed i e nease in the capitaouu rado of the economy. See Wold Bank (1991), pp. 16.21. - 31 - directly affected agriculture: the severe drought during 1992, sharp reductions of the external price of coffee and other commodities, and intensification of guerrilla and drug-related violence in rural areas. These shocks were exacerbated by the real appreciation of the exchange rate-primarily the result of large external capital inflows-and the restraint of agricultural credit due to the Caja Agrania financial crisis.3 Poverty Reducton: M4jor Macroeconomic Challenges 2.6 The recent decline in rural incomes in Colombia led, as shown in Chapter 1, to an increase in poverty indicators in 1992. Policy debate therefore focused on reversing the rural deterioration. The policy response is embedded in an emergency policy package presented by the Ministry of Agriculture, and the recently approved Law 101 of 1993.4 The measures include some trade and credit policy revisions, rural work programs, and an investment subsidy scheme. The policy debate has led some govenmment officials to question the direction and pace of structural reform in Colombia. They argue that reforms have had a negatve impact on employment and caused hardship to low-income groups in bGth rual and urban areas. Even if the reforms are maintained and consolidated-the most likely and desirable case-the government may need to take compensatory measures, including reallocation of public expendiures toward the poor and development of adequate safety nets. 2.7 To determine the best way to deal with rural stress and promote income- earning opportmities for the poor, it is essential to identify how structural reforms and major shocks interacted to produce increased poverty in 1992. Only on this basis can one assess whether corrective measures are required, or if there is a need to redefine the current macroeconomic policy framework. Section II analyzes the effects of each of the major reforms and shocks on the income of the poor and defines a strategy to regain the path for poverty reduction in Colombia. 2.8 In the medium term, the major macroeconomic policy challenge in Colombia is adjustment to the expected oil boom that will result from recent discoveries in Cusiana and Cupiagua, which have more than doubled Colombia's proven oil reserves. The expected oil revenues provide Colombia with an opportunity to attain a higher growth path and to devote resources to infrastructure and social services that will directly benefit the poor. Good macroeconomic management, however, is essential to realize these potential benefits. The experience of countries like Venezuela and Nigeria, which have lived through similar shocks, proves caujion is required. Section m explores Colombia's medium-term macroeconomic perspectives as the country adjusts to the expected oil bonanza. In doing so, it incorporates recently approved social security reform. It then recommends a macroeconomic policy framework aimed at steering Colombia to a course of higher sustinable growth with equity. 3 See Chapter for a descriton of Ca3a Agmrias aciviies and c dfficuldes. 4 See Secion L, pargraphs 2.27-2.32 for details of the policy tesponsc and dth corespondi discssions. - 32 - II. Structural Reforms, Shocks, Responses and the Poor: 1990-2 2.9 This section assesses the effectiveness of macroeconomic management in Colombia during 1990-92 in terms of poverty reduction.' Based on a computable general equilibrium model, the analysis explores the role of each of the major reforms. shocks, and corresponding policy responses in explaining observed changes in the income of the poor. The model simulates the behavior of the economy between 1990 and 1992. By alternately including or excluding individual reforms or shocks, the model generates counterfactuals for each event. The effect of each event is therefore identified. The purpose is to determine how disadvantaged households fared from macroeconomic changes in 1990-92. The section begins with an overview of macroeconomic performance of the economy during this period and its consequences for the poor. It concludes with a summary assessment of the factors behind the income variations observed during 1990-92, and a set of resulting policy recommendations.6 Macroeconomic Performance and the Poor: 1990-92 2.10 During 1990-92 policymakers faced the difficult task of simultaneously implementing the structural reforms of the EMP and reducing inflation, which had reached a 15-year peak of 32 percent in December 1990. The government's 1991 macroeconomic program thus called for tight fiscal and credit policies. As envisaged, the deficit of the non- financial public sector was reduced from 0.6 percent of GDP in 1990 to 0.2 percent in 1991, and domestic interest rates rose sharply (Figure 2.2). But this rise-combined with falling global interest rates and a slower depreciation of the nominal exchange rate-widened the differential between the returns on peso and dollar assets, encouraging strong capital inflows fed partly by drug money. The capital inflows undennined the tight monetary policy, strengthened the balance of payments, and led to the appreciation of the real exchange rate. In response, the govermnent advanced the timetable of trade reform to reduce the external imbaiance. Success was limited: the current account surplus in 1991 ended above ; percent of GDP. Gross international reserves increased to over nine months of imports of goods and services. The uncertainties related to the stability of the exchange rate and the pace of implementation of the structural reforms caused private expenditures to stagnate. Private consumption grew only 0.9 percent during 1991 (-0.7 percent in per capita terms), private S eased autenion is now being given to die role of macro conomic policy in povery reduction-fthe Ranw$s Poverty Rert (1990) is a good example. Clealy, effective adjustm to negative external shocks is essental if one is to minimiz their effcts on the poor. Even structurw reforms aimed at improving te effciency of resource use and generatng incom-earning oppomuies for the poor thrugh incrxased labor demand cant in the short rn, lead to deteriortion in poverty. Impetfecdons in the labor market can. duting die adjustmnt process. result in higher unemployment and conesponding declines in labor income. Accompanying demnupd-reduig measures may hurt the consumption of the poor and the non poor aike. 6 The radter interested in the details of the analytical work in this section should tefer to Annex 2. Section I of the Annex oudines the main fetues of the general equilibrium mode developed by FEDESARROLLO. Section n then presents a brief description of the nmjor reforms and shocks experienced by the economy. a analyzes the role dhat each played in expaining changes in the income of the poor. The analysis is ceterd aound the criis in the agricuture sector and three main policy issues which have imporan income distribuon implications: (a) trade liberalization; (b) tax reform; and (c) minimum wage policy. - 33 - Figure 2.2: Nominal Interest Rates, 1990-93 (percent) 45 40 20 30~~~~~~~~~~~~~~~~~~~~~~~~~ 19S0 1991 8 Yma 1995 a. Weighted average of weekly rates on three-month bills. b. Average monthly rates based on survey of financial intermediaries. Source: Banco de la Rtpublica and IMF staff estimates. investment fell 14.1 percent, and total imports dropped 3.9 percent (Table 2.2). The reslting slow economic growth of 2.1 percent was generated mainly through the expansion of the agriculture sector (5.2 percent) and infrastructure (3.4 percent). Manufacturing gIew just 1.1 percent and construction contracted 0.7 percent. Even then, the 12-month inflation rate only fell to 27 percent by year end. 2.11 The 1992 macroeconomic program was designed to consolidate trade liberalization, further tighten fiscal policy through tax reform, and modify monetary policy to lower interest rates, capital inflows, and real exchange rate appreciation. Early in 1992, a pronounced slowdown in the combined net placement of open-market instauments in local and foreign currency helped lower the rates on three-month open market bills from 36 percent in late 1991 to 20 percent in May-September 1992. As a result, the surge in capital inflows in 1991 and the first half of 1992 subsided, as did the pace of reserve accumulation by the Central Bank. For the year as a whole, real appreciation of the peso was 2 percent. Imports grew sharply while exports remained stable, and the current account surplus feli to about 2 percent of GDP. By December 1992, interest rates stabilized around 27 percent and inflation fell to 25.1 percent. Greater confidence and stability spurred private expenditure: private consumption grew 4 percent, private investment 27.5 percent, and imports 29.4 percent. Expenditure was further stimulated by the government, as public consumption - 34 - Table 2.2: Economic Growth and its Sources, 1990-92 (1975 pesos) Anna Change (percent) Percentage of G4P Sector 1 1992 1999-92 1999 1991 1992 Demand Private Consumption 0.9 4.0 2.4 68.0 67.2 67.5 Government Consumption 2.3 12.4 7.2 11.1 11.1 12.1 Private Investment -14.1 27.5 4.7 8.0 6.7 8.3 Government Investment -1.8 7.4 2.7 6.1 5.8 6.0 Inventory Accumulation 63.8 3.2 30.0 1.7 2.7 2.7 Exports 3.6 6.7 5.1 20.7 21.0 21.7 Imports -3.9 29.4 11.3 15.5 14.6 18.3 GDP 2.1 3.5 2.8 100.0 100.0 100.0 Supply Agriculture, Fishing, and Forestry 5.2 -0.9 2.1 22.4 23.1 22.2 Coffee 15.0 -0.4 7.0 2.8 3.1 3.0 Odter Agricultue 4.2 -2.6 0.7 9.2 9.4 8.9 Mining 0.7 0.7 0.7 4.8 4.7 4.6 Manu6fcturing 1.1 4.6 2.8 22.0 21.8 22.1 Coffee Processing -7.6 30.7 9.9 2.9 2.7 3.4 Other Manufactuing 2.4 0.9 1.7 19.1 19.1 18.7 Infr.5#cuIe 3.4 1.4 2.4 10.0 10.2 10.0 Construction -0.7 13.6 6.2 3.1 3.0 3.3 Services 1.3 4.5 2.9 37.6 37.3 37.8 a. Tranortaton, storage, communications, electricity, gas, and water. Sorce: DANE. increased 12.4 percent and public investment 7.4 percent. Growdt picked up to 3.5 percent and came mainly from the non-tradeable sectors. Construction expanded 13.6 percent and services 4.5 percent. Agriculture (excluding coffee) contracted 2.6 percent. Manufactuing (excluding coffee processing) grew just 0.9 percent. This resource allocation responded to sharp changes in relative prices due to trade liberalization, the drop in international commodity prices (the price of coffee, for example, fell from US$0.95 a pound in 1990 to US$0.69 a pound in 1992), and the 11 percent real appreciation of the exchange rate during 1990-92. Prices in the a ture sector were firiher affected by the dismantling of support prices and purchase agreements which, unil 1990, were managed by IDEMA to reduce price uncertainty and sustain profitability of the main crops. Overall, in relation to the GDP deflator, agriculture prices dropped 13.3 percent between 1990 and 1992. 2.12 Despite the relatively slow economic growth during 1990-92 of 2.8 percent a year, employment expansion averaged 5 percent a year, and the unemployment rate fell from 8 percent in 1990 to 7 percen in 1992. Urban employment expansion occurred in all segments of the labor market (Table 2.3). Employment of skilled workers grew 9.3 percent a year, while for unskilled workers it increased by 5 perce per year. Although contruction - 35 - Table 2.3: The Labor Market, 199092 Numnber of People (housands) Vanaions in Employment Vataon in (percent) Real Wagesa Locadon 1990 1991 1992 1991/90 1992/91 1992/90 9 ercenf) Main Urban Areas Labor force 4,816.30 5,112.20 5,327.80 6.14 4.22 10.62 Employed 4,324.70 4,610.60 4,841.20 6.61 5.00 11.94 Unskilled formal 1,999.70 n.a. 2,244.80 12.26 0.34 Unskilled informal 1,515.80 n.a. 1,629.50 7.50 2.92 Skilled 809.10 n.a. 966.80 19.48 -3.88 Unemployed 491.60 501.60 486.60 2.04 -2.99 -1.02 Unemployment rate 10.21 9.81 9.13 (percent) Other Urban Areas Labor force 2,819.40 2,812.90 2,960.30 -0.23 5.24 5.00 Employed 2,560.60 2,565.90 2,720.70 0.21 6.03 6.25 Unemployed 258.80 247.00 239.60 -4.57 -2.99 -7.42 Unemployment rate 9.18 8.78 8.09 (percent) Rural Areas Labor force 5,191.50 5,836.70 5,628.90 12.43 -3.56 8.43 Employed 4,905.80 5,589.50 5,382.80 13.94 -3.70 9.72 -22.39 Unemployed 285.60 247.20 246.10 -13.47 -0.43 -13.84 Unemployment rawe 5.50 4.23 4.37 (percent) Total Labor force 2,827.20 13,761.80 3,917.00 7.29 1.13 8.50 Employed 1,791.10 2,766.00 2,944.70 8.27 1.40 9.78 Unemployed 1,036.10 995.80 972.30 -3.89 -2.35 -6.15 Unemployment rate 8.08 7.24 6.99 (pcent) LII I I I a. Results of the 1990-92 simulaton. Source: DANE household suveys, Sepmber 1990,1991, and 1992; FEDESARROLLDO 1990-92 imulaton. - 36 - and services grew faster than manufacturing, formal employment expanded at a higher rate than informal occupations-6.3 percent a year compared with 3.7 percent a year. In the rural areas, employment also increased significantly-4.7 percent per year-with the unemployment rate falling from 5.5 percent in 1990 to 4.4 percent in 1992. Given the sectoral composition of economic growth during the period, however, the expansion of employment in different segments of the labor market reflects substantial differences in the behavior of real wages. As estimated by the 1990-92 simulations, real wages in the fornal sector remained basically stagnant, while those in the informal sector increased by 3 percent between 1990 and 1992. Real wages for skilled workers dropped 4 percent while those for nual areas fell 22 percent. 2.13 With labor representing the principal source of income for poor households, employment and wage behavior during 1990-92 led to a sharp income redistribution between urban and rural areas. During this period, as estimated by the 1990-92 simulations, the average household income of the four lowest decdiles in the rural area dropped 14.7 percent in real terms, while those in the urban areas experienced a 10.6 percent increase. This led to the deterioration of 1992 poverty indicators described in Chapter 1. Income Dedenninm on: The Role of Policy Reforms and Economic Shocks, 1990-92 2.14 Table 2.4 summarizes the role different reforms and shocks had in the behavior of the income of the poor in Colombia between 1990 and 1992. As presented in Annex 2, Section II, the simulation results indicate that trade liberalization has benefited both rural and urban populations. Tax reform has extracted income from both groups. In these two cases, however, the rumral population has been relatively protected. But that is as far as protection goes. Minimum wage policy has transferred income from the rural poor to the urban poor. The change in international agriculture prices has benefited the urban sector and reduced living standards of the rural population. And the relaxation of agricultre supports (related to the restructuring of IDEMA) has diminished transfer of resources from the urban to the rumral sector. In aggregate, the rural poor have been negatively affected by developments between 1990 and 1992. 2.15 Trade Reform. As presented in Annex 2, Table 2.9, trade liberalization has improved both mral and urban incomes. Indeed, the rurl poor have benefited the most. Trade liberalization has increased the income of the lowest two rural quintiles by 3.4 percent. Even though prices have generally declined in response to liberalization, relative prices have moved strongly in favor of agriculture-with the jump in modem agriculture prices leading to an aggregate 2.7 percent increase in agriculture relative prices (with respect to the GDP deflator). This has generated demand for rural labor and a 7.4 percent increase in rural wages. In urban areas, expansion of the internediate manufactured goods sector (intensive in formal unsldlled labor) has brought unskilled workers from the informal to the formal sector. The final outcome has been a reduction of 0.4 percent in employment in the infonnal sector, an increase of 1 percent in the number of formal jobs, and a drop of 0.4 percent in the - 37 - Table 2.4: Estimated Income Effects of Policy Reforms and Economic Shocks, 1990-92 Complete Simulaon Compared with Counterfactual (percent change) (percent change) l Ag,iculture Agricuure Area . Trade Tax Minimum International Domestic 1990-92 Reform Refonn Wages Pnces Prices Urban income Lower quintile 10.58 0.76 -2.40 0.18 0.36 2.86 Upper quintile 11.71 0.84 -3.10 -0.36 0.55 3.05 Gini coefficient 0.57 0.03 -0.34 -0.23 0.08 0.06 Rural income Lower two quintiles -14.72 3.39 -0.86 -2.08 -5.51 -4.27 Upper quintile -14.71 2.85 -0.82 -2.06 -5.69 -4.08 Gini coefficient 0.01 -0.38 0.02 0.02 -0.11 0.12 Source: Annex 2, Section H. unemployment rate. The reduction of the informal work force, combined with expanded labor demand, has led to a 0.7 percent increase in real wages in the informal sector. Trade reform thus has caused an even 0.8 percent increase in real urban incomes. 2.16 These results contradict the view of some Colombian policymakers, who blame trade reform for the recent deterioration in rural incomes. Indeed, trade liberalization has improved income levels and income distribution in the rumral areas. Furthermore, improvement of real incomes in the lowest deciles of both rural and urban areas shows trade liberalization has enhanced poverty alleviation in Colombia. 2.17 Tax Reforms. The simulation presented in Annex 2, Table 2.12 indicates that tax reforms have contracted both urban and rural incomes. Urban production decline and increased direct taxaton of corporate income have reduced capital income by an estimated 1.1 percent. Urban labor incomes have differed depending on specific segments of fte labor market. Unskilled formal workers protected through the minimum wage policy have experienced increased layoffs as frms have accommodated production cuts. Elnployment of unsldlled workers in the formal sector has declined 3.7 percent, leading to a similar drop in labor income. Informal unskilled workers have encountered increased competition from those displaced in dte formal market, but have been able to expand employment through wage concessions. Thus, while wages in the informal labor market have fallen 2.9 percent, employment has increased 1.2 percent and the resulting real labor income has declined only 1.8 percent. Unemployment has increased by 1.3 percent. Reduced labor demand, in turn, has led to a decline of 2.0 percent in real wages for skilled workers and to a corresponding drop of 2.4 percent in real labor income. In all, the induced reduction in different types of urban income seems to have been progressive. Total real income of the upper and lower quintiles in the urban population bas declined 3.1 percent and 2.4 percent, respectively. The - 38 - urban Gini coefficient has been reduced by 0.3 percent. In mrual areas, the reduction of total incomes has affected labor more severely than capital incomes (1.1 percent compared to 0.6 percent), but without important effects on income distribution. All income groups have experienced income reductions of about 0.8 percent. 2.18 In conclusion, recent tax reforms in Colombia have had a contractionary effect on real incomes, with a somewhat positive effect on income distribution. Not only has urban income distribution improved as a result of the modifications in the tax stmcture, but the rural sector has been somewhat protected from the accompanying income contraction.7 2.19 Minimum Wages.8 Between 1990 and 1992, minimum wage policy is estimated to have increased wages of unskilled formal workers by 5.3 percent, while reducing employment in this segment of the market by 5.1 percent (see Annex 2, Table 2.14). With real wages for unskilled informal workers having increased by only 1.3 percent, jobs have increased 1.8 percent, with many workers remaining unemployed. The unemployment rate has been 1.8 percent above the level of frictional unemployment. Even then, the urban poor bave been better off because of the minimum wage policy: real urban income of the lower quintile has increased 0.2 percent. The contractionary effect of minimum wages has implied a 2.9 percent reduction of real urban capital income. This explains the estimated 0.4 percent fall in income of the upper quintle of the urban sector and the corresponding improvement in urban income distribution (with the Gini coefficient dropping 0.2 percentage points). Effects on rural areas have been quite different. The simulation indicates price increases of agriculure goods have been mild compared with those of urban goods and services. The contract;onary effects of minimum wages have reduced demand for mral products, leading to a decline of 1.8 percent in agriculture relative prices (with respect to the GDP deflator). This has caused real rural incomes to decline about 2 percent for all income groups. The effect on real wages and capital income has been of the same order of magnitude. 2.20 In conclusion, minimum wage policy has been detrimental to poverty reduction in Colombia. Protection of unskilled workers in the formal sector has been maintained through the transfer of income from nrual to urban groups. The loss from such policies appears to be substantial-almost 1 percent of GDP growth-and, against some beliefs, not conducive to overall poverty reduction. Given the stress in the agriculture sector and the need to focus on rural incomes, it may be appropriate for the government to revise its minimum wage policy and allow minimum wages to start declining. At the very least, minimum wages 7 This conclusion is maiaid even if one simulat dh tas rfomns. incbdig the VAT increase of 1993. s To assess the effect of minimum wges on the income of the poor in Colombia. dte model simulaes how te economy would behave if fomal tCtor wages adjusted to te level requied to expand labor demad and to reduce unemployment of unskldled urban woers ftrm te 8.8 pec estinated for 1992 to 7 percet. Tis lvl of umpoyme has been identfied as fdcin in Colombia by, among others, dte Chmety Employmemn Mission (Mission de Empleo (1987)). It provides the necessary space for the nol entrance and uans of woes betwee occupatio and for sti be sly rd demand of labor due to oduaonal hactoWs. - 39- should be set on the basis of expected inflation rather than past inflation. This should help the government bring inflation under control. The payoff of this policy change will be gained by th, rural poor. 2.21 Changes in Agricultural International Prices. The simulation results indicate, as shown in Annex 2, Table 2.18, that the drop in agriculture prices observed between 1990 and 1992 has led to reduced wages and capital incomes in rural areas, at 3.6 percent and 6.6 percent, respectively. The difference in behavior of wages and capital income in rural areas has been related to the effects of reduced coffee prices and other agriculture commodities. Whereas the drop in international coffee prices has directly affected the rents of the sector-accounting for a decline of 6.3 percent in rural capital income-the decline of prices in other agriculture conmnodities has been largely absorbed by the reduction of real wages. Ultimately, the change in factor remuneration has led to a decline of real rural income in the lower two quintiles of 5.5 percent. In urban areas, since production has experienced only minor changes and real wages of the unskilled formal workers have been protected by the minimum wage policy, variations of labor and capital income have been insignificant. Total real urban income of the lower quintile has improved 0.4 percent as a result of the drop in international agriculture commodity prices. 2.22 Stock Management and Domestc Agricultre Pricidg Policies. Taken together, trade liberalization, tax reforms, minimum wages, and intermational price changes can only partially -explain changes in agriculure relative prices and rural incomes. Relative prices of agriculure food products (with respect to the GDP deflator) fell 13.3 percent during 1990-92, and modern agriculture goods fell 16.0 percent. The four factors mentioned above account for a reduction of 8.4 percent in the case of agriculure food products (leaving 4.9 percentage points to be explained) and for less than half a percentage point in the relative price of modern agicultral goods (leaving 15.6 percentge points unexplained). Similarly, these factors reduced rural incomes by 5 percent of a total calculated reduction of 14.7 percent. Clearly, other elemems have been at play. Although mnmerous other policy changes and exogenous shocks influenced prices and incomes,9 the most important remaining factor relates to the restructring of IDDEMA and the corresponding changes in stock management and agricultre pricing policies. 2.23 Indeed, as presented in Annex 2, Table 2.20, the relaxation of support prices, the reduced importance of purchasing agreements, and the release of agriculture stocks have caused a severe reducdon in producer prices of agrculture food products, 15.7 percent, and of modem agriculture goods, 18.1 percent. The simulation indicates, in turn, that these changes have had significant repercussios for prices of processed food products, and milder but still noticeable ones for the remaining urban sectrs. In all, domestic prices have fallen 6.5 percent. This has made domestic products more competitive, inducing a 1.8 percent expansion of exports and a 3.8 percent reulction of imports. Although the most significant 9 See Tabl 2.7 in Annex 2 for a comee Flasug of al doe exogaeus dwne ntoduced in de model to epcte die obmrved b:havior of di econoay beween 1990 an4 1992. - 40 - changes in exports and imports have taken place in the agriculture sector, the spillover effect has also changed exports and imports in the manufacturing sector. Improved competitiveness has provided room for a 1.6 percent acceleration in manufactring activity, 0.6 percent in the service sectors, and a 4 percent appreciation of the real exchange rate. Economic growth has expanded 0.6 percent as a result of restructuring agriculture. 2.24 Increased production levels have generated labor demand, leading unsilled workers from the informal to the formal sector. As a result, as shown by the simulation outcome presented in Annex 2, Table 2.20, formal employment of unskilled workers has increased 2.8 percent while informal employment has fallen 0.8 percent. Given the rigidity of formal real wages, incomes of unskilled workers in the formal segment of the market have grown in line with employment. The reduction of the informal working force has in turn been accompanied by an increase of 2.2 percent in real wages, with total real incomes of unskilled informal workers increasing 1.6 percent. Real wages of the skldled workers and urban capital incomes have increased 1.3 percent and 3.9 percent, respectively. This responds to the acceleration in activity levels and improvements in relative prices of industrial goods and services. Thus, all urban factors have benefited from the reduction of agriculture prices. Total real urban incomes have increased 3 percent, with very similar advances across all groups along the income spectrum. The counterpart of these improvements has been a substanial deterioration of rural incomes. Agriculure relative prices have fallen 4.5 percent (in reiation to the GDP deflator), resulting in a 4.3 percent decline in real income of the lower quintile in the rural sector. The slight deterioration in income distribution results from the more significant decline in wages than in capital income: 5.4 percent compared with 3.1 percent. 2.25 The relaxation of support prices in agriculture in conjunction with the release of agriculture stocks led to an important redistribution of income from the nrual to the urban areas, increasing the levels of poverty in the countryside and the country as a whole. Indeed, these agricultural supports had been operating as a system of income transfers to rural households. Their costs, as estimated by the simulation model, were reflected in a less compedtive economy, both in the rural and urban sectors, which constrained total exports in 1.8 percent, encouraged imports in 3.8 percent, and limited total private consumption in 1 percent. In terms of poverty alleviation, the support mechanisms were costly instments given their inability to target the poor. The top quintile of rural households got as much benefit from the support mechanisms as the lower quindles. Leakages to the non-poor were substantdal. Eliminatng agricultural support was justified on efficiency grounds. But the results suggest the government should have anticipated the negative effects and adopted pallel compensatory measures. 2.26 The relaxation of the agriculture support system and the drop in international agriculture conmodity prices explain the decline of 10 percent in real income of the two lower quintiles in the rural area, which is almost 70 percent of the observed deteriortion. The extent and magnitude of this deterioration clearly warrnted govenunent intervention. - 41 - 2.27 Policy Response to the 1991-92 Crisis in the Rural Sector. As described before, the rural poor have been negatively affected by developments between 1990 and 1992. The macroeconomic policy framework pursued by the government, however, did not include any corresponding compensatory measures. It is only in reaction to these events that authorities tried to define a policy response to the crisis in the rural sector. Initally, in May 1993, the Minister of Agriculture reacted by proposing an emergency policy package (Plan de Reacivacitn del Sector Agropecuario). This, in turn, was followed by Law 101 of 1993 (Ley General de Desarrollo Agropecuarno y Pesquero), which set the new policy framework for the nrual sector. As a whole, the main policy measures taken by the authorities are as follows: (a) Trade policy. The measures presented were to: (a) impose minimum import prices on forty-seven agricultural products based on prices quoted in international commodity exchanges (the original proposal included sixty-nine items); (b) modify the system of agricultural price bands by converting import tariffs under the variable duty scheme from a specific to an ad-valorem basis; (c) establish, for price band products, a limit of 50 percent for the sum of the basic and resulting additional tariff; (d) raise effective protection on some crops by temporarily eliminating or reducing import duties on certain agricultural raw materials and intermediate goods (for example, pesticides, ferdlizers); (e) establish prior import licenses for poultry and dairy products; (f) harmonize with its Andean Pact partners both the price band system as well as all other sectoral policies, like production subsidies, which are a source of distortion to trade flows; and (g) postpone the scheduled elimination by the end of 1993 of tax credit certificates (CERTs) in the case of farm goods. Among other proposals to be regulated the most important would establish "absorption agreenents" that ban certain imports of farm products as long as domestic supplies are deemed sufficient and introduce safeguard mechanisms against injury to domestic producers from sectoral surges of imports. (b) Domestic pricing policies. Policies pursued were to: (i) reinitiate market interventions to support the purchase of crops like rice, cotton, and beans (in the case of rice, for example, the intervention price was orginally set 6 percent above the established minimum price); (ii) establish purchase agreements between the industrial sector and agriculture producers for barley and sorghum; and (iii) establish a stabilization fuNd for cotton. (c) Credit policies. Approved measures were to: (i) extend by one year, until he end of 1994, the date at which interest rates to small farmers would be increased from DTF + 2 percent to DTP + 6 pement, but make the subsidy explicit and finance it through budget allocations -42 - (7,500 million pesos were budgeted for 1994); (ii) refinance on preferential terms outstanding loans of farmers producing cotton, rice, coffee, sorghum, corn, barley, tobacco, or cassava; (iii) recapitalize Caja Agraria (90,000 million pesos in 1993), purchase part of its non- performing debt, and transfer it to municipalities who would use any collections they receive for social investments; and (iv) create an investment subsidy scheme (Incentivo a la Capitalizaci6n.Rural) to support rural capital expenditures by the private sector. (d) Public employment scheme. An Emergency Rural Employment Progrm was put in motion. The program, being coordinated by the Ministry of Agriculture, will be managed by local governments, fmanced through cofinancing arrangements, and executed in connection with projects implemented by HIMAT, IDEMA, INDERENA, DRI, and PNR. The total cost was budgeted at 12,000 million pesos for 1994. (e) Government expediture programs. An important expansion of rural expenditure programs was included in the budget.10 If implemented, public investment in the rural sector would double between 1992 and 1994. Programs like DRI would expand 144 percent, HIMAT 208 percent, IDEMA 90 percent, and INCORA 16 percent. Although no specific measures were proposed, the policy package urged authorities at the different levels of govermment to better target expenditures. Programs specifically mentioned in ths regard include, among others: DRI, PNR, INCORA, and ICA. 2.28 Based on the results presented in this report, one would conclude that-in Ums of poverty reuction-the efforts to respond to the crisis in the agriculture sector have been pauly misguided. Because the government lacks the instruments to properly address amporay shocks in the nural sector, it has relied on policy options which, in retrospect, are inadequate. Most of the new measures represent a temporary regression on trade liberalization and domestic agriculture pricing policies. As has been shown, the trade liberalization program is estimated to have had a positive effect on mrual incomes, and domestic support mechaisms have been an inefficient insument to protect the rural poor. In tem of credit policies, it has been proven, time after time, that it is the access and not the price of credit that is the main constraint to lending in the rural areas. In the context of poverty reduction, therefore, the govenmnent is moving in the wrong direction. Unfornately, the auithorities have been unable to resist the pressure of interest groups which have used the recent deteioration of mrual incomes to pursue their own objectives. By doing so, the government is not only sending the wong signals but putting in question the success of its own stuctural reform efforts. Takig advantage of the recent improvement in 10 See C 5 fr an ovetvls of i Se nual expm pogma have paed in reducif a poey in Coomb - 43 - international commodity prices, and the expected good performance of the agriculture sector during 1994 (3-4 percent real growth), the government should move away from crisis management to define a medium-term strategy in the sector and take corrective measures to be better prepared for any future shocks. In this context, the government should be focusing instead on developing the last two components of its package as presented above: putting in place the rural employment program and restructuring government expenditures. Government efforts in this direction would probably have a better payoff, as both of these instruments are more suitable for targeting and should prove more cost effective in helping the poor. 2.29 In line with the above arguments, the government should consider reversing-as soon as possible-most of its recent policy changes related to trade, domestic pricing, and credit policies in the agriculture sector. Clearly, however, there are some measures in these areas which should be pursued. In terms of trade policy, probably the only area where the government's recent changes need to be maintained is in the area of harmonization of trade and sectoral policies with its Andean Pact partners. Of particular importance is the elimination of existing subsidies to production in other countries. As a result of these policies, Colombia experienced a huge inflow of imports of soybeans from Bolivia and rice from Venezuela early in 1993 that had an important effect on domestic prices. Given the strong possibility that these subsidies will have to be removed sometime in the future, doing it sooner rather than later is better to avoid distortions in relative prices which will influence resource allocation in agriculture. As for domestic pricing, instead of further promoting the development of commodity stabilization fumds, the government should encourage producers to use the international futures commodity market to manage fluctuations in international prices. The authorities should devote resources to defrne the legal and institutional framework for the development of transparent and well-functioning commodity markets. These should be linked to both spot and future international commodity markets and subject to clear rules and regulations. With respect to credit policies, the government should fmcus on stimulating new financial intermediaries to lend to the rural sector. If a subsidy is considered, it should subsidize, for example, the transction costs of lending to small farmers-as in Mexico or Chile. 2.30 The proposed public employmemt scheme should receive the government's ful attention. Until mid-1994, the budget for the program had not even been used. If proper attention is given, the program could set the basis for a rual safety net to deal with future shocks to agricultue. It would also reduce unemployment costs during the expected adjustment to the oil boom (see next section). As presented in the Bank's Poverty Report (1990), such programs have proven effective in countries such as Chile and Peru. In both cases the programs successfuzlly tuageted the poor and provided a safety net for the unemployed during recession. hnporant modifications to the recently established Colombian mual employment program, however, would be rquired. Chapter 7 discsses the program as part of a future social safety net in Colombia and provides specific recommendatons for its structure and implementation. - 44 - 2.31 The increased priority given to rural expenditure programs should prove effective. The government has already started redirecting public sector expenditures, which heretofore had been allocated to inefficient input and product pricing policies, to improving the rural financial system, and to interventions in factor markets, including technology, infrastructure, and education. In designing programs to enhance the skills of rural people, the potential for off-farm employment in the countryside also needs to be recognized. This would provide better support services to enhance agriculture's corapetitiveness. Public infrastructure programs in roads and small scale irrigation schemes offer the double advantage of providing employment in the short tenn, and an increase in the sector's productivity and employment prospects in the long term. Road infrastructure is especially important in Colombia, where transport costs are very high. Given the progressive nature of existing rural expenditures programs, their simple expansion would certainly benefit the rural poor. However, implementation of such expansions could be problematic. Chapter 5 makes specific recommendations aimed at restructuring existing ural government programs to deal both with implementation and targeting issues. These recommendations are most relevant giveren the recent steSS in the rural sector and the planned expansion of the programs. 2.32 Importantly, the government's reaction to the recent crisis should be consistent with a longer-term strategy for the rural sector. Expansion of government expenditure programs, for example, should be aimed at improving the sector's competidveness and enhancing its producdvity-through better infrastructure, training, technology, and diversification. Policy adjustments that are deemed appropriate to put Colombia back on tack for fiuther poverty reduction should be implemented in the context of an overall medium-term policy framework. The next section develops such a framework, focusing on Colombia's upcoming macroeconomic policy challenges. III. Macroeconomic Policy Challenges In the Medim Term: Adjusdng to Cusiana's OU Bonanz 2.33 The discovery of oil reserves in Cusiana and Cupiagua provides Colombia with new wealth to raise the population's standard of living." Attaining this objective is probably the country's biggest challenge during the coming years. This section presents an overall perspective of this challenge. It describes the complex problems of macroeconomic management entailed and proposes a policy framework that would allow (a) the potential benefits of the expected oil revenue to be realized; and (b) the poor segment of the population to benefit from the corresponding windfall gains. 11 Official estimaes indicate ht the pesen value of te net icome flows to be geeratd by fte Cusiana ad Cupiagtu projects between 1993 and 2005 amounts to US$15,300 million. This resents an o windfl Pin of approximately 28 pere of GDP. To put it in perspective, this windfall gain is lager dtn those experienced by Indons and Nigeria with te fourfold oil price euae in 1974 corresponding to 17 peent ad 23 perent of GDP, respectvely. For tie ntrsted reader. Section m in Anx 2 descrbes in mome detail the magnitade of the new oil discoveries and its reltion to the Colombian economy. - 45 - Colombia's Medium Tenn Macroeconomic Perspectves: Adjusting to the Oil Boom 2.34 There is extensive literature about adjustment to oil booms. The theoretical framework of the "Dutch Disease" clearly portrays the main elements of what would be efficient adjustment to an oil boom of the kind currently facing Colombia.12 There is also ample evidence of the experience of different countries in their effors to adjust to similar external shocks. Colombia itself is not unfamiliar with this phenomenon. As recently as 1975-80, Colombia encountered a coffee bonanza that increased international coffee prices more than fourfold over a period of two years. 2.35 Many lessons can be drawn from the existing literature and the intemational experience. First, the natural and efficient adjustment to a permauentt3 windfal gain, like the one now facing Colombia, is to move toward a higher growth path and to experience restucturing of the composition of output. Resources move away from the (non-booming) tradeable goods and services toward non-tradeable, domestically produced goods and services. This transformation is desirable and ultimately necessary for the improvement in the standard of living to endure. Second, the cases in which windfall gains did not prove to be welfare-improving relate to bad macroeconomic management in general, and relaxation of fiscal policy in particular. In the case of Mexico, for example, not only did the government use the oil revenues to embark on a very ambitious public expenditure program, but it bonrowed heavily against expected fitre income. The outcome was excessive appreciation of tle real exchange rate and a significant waste of resources, as the economy encountered allocative bottlenecks. And third, policy makers should be concerned, not with protecting the tradeable sector per se, but with enhancing resource mobility in the economy and removing distortions that could lead to misallocation of resources. This will enable the economy to take advantage of the v indfall gain and enhance its capacity to absorb the investment boom productively. 2.36 Based on these lessons and the particular circu nces facing Colombia, the long-run version of the general equilibrium model is now used to explore the country's medium-term macroeconomic prectives. Special attention is given to better understand the way in which the poor would share in the benefits of the oil windfall gain. The analysis is carried out in two parts. The first explores the adjustment process to the oil boom, while pursuing prudent fiscal policies and implementing the recendy approved reform of the social sewrity system. The genera fgs suggest that the oil discovery will enable Colombia to atain a higher growth path. But without specific policies to enhance rural incomes, these 12 See W.M. Corden, and J.P. Netay (1982). and W.M. Cordon (1984). for clear eosions of tbe nalydts of theo -Duc Disease. The amn of this phenomesno comes from the observed devlopmet in te Nethelads folowing lrge nal gps discovedes duTing the 1960s. 13 Athugh dwe is an ongoi dsdsinn wheh the oi windl in Coombia i panent or a, -y.there is sm agreement thi ie new oi discov ae pbably he ftiat a series. As new nesmut are mA on expaIon, te pmbabit of finfing new oi rsves i sgnifln To ta et one can thik of at least par of the wif haing a Pemanent naur - 46 - incomes could fall below long-run trends even in a well managed adjustment to the oil bonanza. The second part of the analysis is devoted to identifying policy options for the authorities to use the oil windfall gains to put Colombia not only on a higher growth path but one that would ensure a generalized improvement in the standard of living of its population. 2.37 Adjusting to the Oil Boom: Prudent Fiscal Policies and Social Security ReLorm. The oil boom simulated in the model represents the base scenario used by the govenmment for new oil field production. As shown in Annex 2 (Table 2.21), production from the Cusiana and Cupiagua Project will increase from its current level of 10,000 bpd to 600,000 bpd by 1997, when it will plateau. This will increase total oil production in Colombia 124.3 percent, from 479,800 bpd in 1993 to 1,076,100 bpd in 1997. The simulation imposes this oil production increase and assumes the economy will reach its new equilibrium by 1998. Consistent with existing official and private medium-term projections, the model assumes Colombia will attain new equilibrium with an overall fiscal deficit of 1.5 percent of GDP and a corresponding current account surplus of 1.5 percent of GDP. The simulation incorporates the expected implementation of the social security reform (Law 100 of 1993) in the fiscal position. Finally, it assumes total non-coffee agricultural productivity will increase 1 percen per year. The real exchange rate is endogenous in the model. 2.38 The results clearly portray the expected "Dutch Disease" adjustment process, as indicated in the sectoral breakdown of price, production, and trade changes in Annex 2, Table 2.25. As a result of the oil boom relative prices are expected to move strongly in favor of non-tradeable goods and services. Besides mining, commerce, housing, and other modem services would experience the largest relative price increases. Parallel changes would occur n the structure of production, with the construction sector (being both a non-traded and an investment good) expanding most significantly. Other non-durable manufactured goods, intermediate manufactured goods, and processed food products are all expected to grow below their long-term trends. The model captures the different degrees of substitutability between domestic and foreign goods, as well as the relative importance of trade in different sectors. It thereby distinguishes whether a sector will behave as a tradeable or non-tradeable sector. Interestingly, agriculture food products and durable manufaured goods would be expected to expand production as a result of the oil boom-behaving more like non- tradeables. Finally, the oil boom has the anticipated effect on foreign trade. Imports rise in all traded sectors and exports decline, both because productive resources are drawn into non- tradeable sectors and because domestic tradeable goods prices, not perfectly tied to international prices, are allowed to rise. The resulting appreciation of the real exchange rate would be 9.5 percent (or 1.6 percent per year). 2.39 In the final analysis, oil sector expansion in Colombia could increase the country's GDP by 5.3 percent (0.9 percent per year) above its long-term trend. The additional income would accrue to the urban areas, whose productive resources would be more itensively used and which would benfit from the 11.6 percent decline in the prices of agriculture goods relatve to urban goods and services. Although urban incomes would only increase 2.9 percent (see Annex 2, Table 2.26), urban welfare would improve significantly. - 47 - By being able to finance social security reform, expected urban income would increase as future social security benefits become viable. Rural incomes, on the other hand, would fall 7.6 percent. The oil boom would thus provide the necessary room to expand private investment in the economy (3.9 percent per year) at a time when major social security reform is implemented, and would allow Colombia to move to a faster growth path. However, the benefits of the windfaU gain would not be equally distributed. If no corrective action is undertaken, the rural sector, and thereby most of the poor in Colombia, would not share the benefits of the bonanza. Indeed, the gap between rral and urban areas that has been opened since 1990 would firther widen. 2.40 Adjusting to the Oil Boom: Proactive Policies to Enhance Rural Incomes. The base case scenario described above shows that good macroeconomic policies will not ensure rural incomes respond positively to the increased oil revenues. Complementary measures will be required. The model identifies the potential of alternative instuments to raise nurm incomes. Overall, the results indicate that the government has room to maneuver within a p;7_determined fiscal position (see Annex 2, Section IV). Three specific areas of policy action can be identified: (a) develop the network and financing of a rural employment program; (b) fmance rural -ifrastructure projects that can enhance nrual productivity and employment; and (c) reverse the agricultural trade protection of 1992-93. 2.41 The rural employment program can be an effective instrument, not only as a safety net, but to minimize the adjustment costs to the oil bonanza. Spending 0.15 percent of GDP in this program (only for labor costs, with wages set 30 percent below market levels) over the next six years would increase rural incomes an estimated 0.5 percent annually. 2.42 Rural infastrucue projects that enhance rural productivity-particularly in colnmercial agriculture-can also have significant effects on rural incomes. A 1 percent productivity increase in commercial agriculture would accelerate rral incomes by 0.2 percent a year. This is pardy because of the relatively large size of the subsector, as well as the high degree of tradeability-increased production can be channeled abroad without major impact on domestic prices. On the other band, productivity gains in agricultural foods (with existing concentration mostly in non-tradeables) have more limited effects. For each percentage point incra in productivity in this subsector, mral incomes only grow by 0.06 percent a year. Consequently, the effort to increase rral productivity, research, and extension should not only be focused on the commercial sector but provide the basis for diversification toward tradeable commodities. 2.43 Removing the variable tariff scbeme on all agricultural products (other than maize and sorghum) will have negligible effects on mual incomes.14 It would, however, reduce the effective protection recendy granted tc agriculture, make the system more urnsrent, avoid future distortions in effective protection in the economy, and provide 14 hum policy opon wasad th O Goveamet of Colomba dug the BDks Agkule Stegy Review mlm dbt visited Bogota in May 199. - 48 - incentives for agriculture to become more competitive. In doing so, it will reduce the pressures for exchange rate appreciation during the oil boom adjustment process. 2.44 Finally, labor migration will also likely reduce the widening income gap between rral and urban areas. Having 0.3 percent of the rural worling force migrating every year would increase per capita rural incomes by 0.14 percent amually. To ensure the effectiveness of this mechanism, the govermment should eliminate labor market distortions and provide high-quality basic rural education and health care. Migration decisions would then be made by a better equipped labor force. 2.45 As part of a medium-term strategy to increase rural incomes the government should focus on two areas: developing a rural safety et to deal with expected adjustments and any negative future shocks, and enhancing agricultural competitiveness. Part of the oil windfall gain can, in addition to developing a rural safety net, be used to improve rural infrastructure, enhance agricultrl extension, improve access to modem technology and market services, and consolidate agricultural land. This would not only improve competitiveness but integrate the agricultural sector into the rest of an economy that is expected to grow 'nore rapidly and adjust more dynamically to a changing external environment. Proposed Medium-Tenn Macroeconomic Policy ftamewo*k 2.46 Based on the analysis just presented and the general lessons described by the "Dutch Disease" literature, one can propose a policy framework that would enable Colombia to attain the full benefits of its oil bonanza. The main elements of this policy framework are: (a) Set targets for a continuous abatement of inflation. A key to a smooth adjustment to the oil boom is to avoid real exchange rate appreciation through a burst of inflation. It is important to build consensus around a targeted path for further reductions in inflation. Given Colombia's gradual approach to macroeconomic management, targeting a 1 to 2 percentage points reduction in inflation each year would be reasonable. This could put Colombia's inflation in the single digits by the year 2000. (b) Define a macroeconomic program each year consistent with inflatonary targets. Given the limited role of monetary policy in an open economy such as Colombia's, this will define the fiscal position which would allow the governm program to achieve its inflation target. (c) Given the appropriate fascal positon, excess oil revenues should be used to prepay debt or be saved abroad. There should be no - 49 - predefned targets for reduction of external debt. The prepayment of debt should be done only when it is financially more attractive than saving abroad. An analysis of existing debt and the costs involved in its prepayment should be carried out. Maintaining good standing for Colomb,a in international capital markets should be an objective. (d) Establish an oil fund"5 to institutonalize the option to save abroad. Given the political difficulties of maintaining fiscal discipline in the context of an oil bonanza, the establishment of the oil fund may provide an instrument to build consensus to save part of the oil revenues abroad. Given the existing law concerning the transfer of central government revenues to localities, the fund should be placed between ECOPEMTOL-Associates and the cental government. Oil revenues saved abroad will not then be considered part of central government revenues. Local government authorities, however, must be assured that their claim on those resources would still be effective and would be implemented as those resources are brought back into the economy. (e) Identify a public Investment program, focused on infrastucture and human capital development-for example, projects in education and health-with high economic and soclW rates of return. Given supply and implementaton constrains, public investment should be spread out over time to ensure that the execution of the projects does not hamper their rates of return. In no case should projects be undertaken when their rate of return is equal to or lower than what would be obtained by investing abroad. While the public investment program faces rapidly decreasing returns, saving abroad would attain constant returns (a larger transfer of funds abroad will bave no discemable impact on international interest rates). (f) Invest in rural Infstructure and services. Increasing rural productivity through economically attractive rual infrastructure projects is probably one of the most effective means to use oil revenues to reduce poverty in Colombia. Programs to improve rural infrastucure (roads, water supply, and irrigation), sevices (agriculture extension, market information, and access to modem technology), and land consolidation (through land reform) should be importa instruments during the coming adjustment process. The challenge is to be able to identify, design, and implement projects with high economic and social rates of retr Private sector participation should be encouraged. In is Altuough die dwiscusi. in Colombi Is rferin o do fnd a a stabti fbd, gin th penut nr of t dbocit of 6e oil discoDveies an e observd nta bdior of el prics I i difiu to hin abow it as a s umm. Ineed. dr oil find wud be peteednp to R-accum deposs for a probnged prid of time. - 50 - the short to medium term, probably the most important constraint in nrual project delivery will be implementation capacity due to decentralization. (g) Increase the institutional capacity of local governments for planning and budgeting, and for properly identifying, preparing, and implementing projects. These objectives, clearly set out as part of the decentralization effort, have now attained particular importance. Given the sharp increase in resources to be transferred to local governments as a result of the oil revenues, proper use of these resources will be essential to maintain fiscal discipline. (h) Derme the Royalties Fund (Fondo de Regalias) to ensure that additional resources received by local governments are used to execute regional infrastructure projects. Oil revenues can thereby allow the cental government to discharge some of its expenditure responsibilities, and thereby enable greater fiscal discipline. (i) Transfer any additional resources to the private sector. Given the appropriate fiscal position, and an attainable and attractive public investment program for the year, any additional resources should be transferred to the private sector to invest, save, or consume. By increasing public savings, such resources would be available to the pnvate sector through the fncial system. (j) Keep opening the economy. One way to transfer resources to the private sector is to frther reduce tariffs by eliminatng recent increases in agricultural protection. This would promote a competitive economy, encourage an efficient resource allocation, and reduce the pressures for exchange rate appreciation during the oil boom adjustment process. (k) Incres factor mobilit. With the subsantial shift of resources expected in the economy, factor mobility wfll be critical to ensure both mininum costs of tansition and efficient allocation of resources. In the formal labor market, real minimum wages should be allowed to fall slowly. At the very least the minimum wage should only be adjusted for expected inflation. Labor mobility should be considered in its broadest sense. Colombia needs to reduce its frictional unemployment and facilitate nual-urban migration. Education and health expenditures, particularly in the rural areas, are essenti insuments to attain these objectives. 2.47 Given the resuts presented in this report, 1992's rura stress may persist. Correcive measures should be consistent with the eqed policy adjustments in the longer - 51 - term. In that context, the emergency employment program should serve not only as a safety net, but as a way to enhance labor mobility during the transiton period. The expansion and restructuring of rural expenditure programs should not only improve targeting, but also directly address the urgent need to increase rural productivity. The government should understand that its stance on agriculture is not a response to a temporary phenomenon, but a strategy to address a medium-term problem. It is in this context that the measures so far undertaken by the government seepi even more counterproductive. On the trade front, for example, any increase in protection would result in a furither appreciation of the exchange rate during the adjustment process to the oil boom. The expected gains on one side would be lost on the other. The loss, however, would be accompanied by a misallocation of resources at a time when significant shifts are taking place. Trade liberalization has put the Colombian economy in a better position to smoothly adjust to the oil boom. The gevernment should be careful not to lose this advantage. 2.48 The challenge for the government is to look beyond existing problems in the rural sector, and implement a comprehensive medium-term policy strategy. Colombia will thus be able to regain its path for poverty reduction, and steer its economy toward higher sustainable growth with equity. The poverty reduction strategy should encompass a well- targeted public expenditure program, an efficient and equitable institutional framework for the provision of social services, and an effective social safety net. The following chapters provide specific recommendations in these areas. In terms of government expendit : Chapter 3 reviews the programs for human resource development; Chapter 4 focuses on the provision of energy, water, sewerage, and shelter to the poor; and Chapter S taies a closer look at the rural expenditure programs. Chapter 6 urns to the institutional framework, focusing on the issues of equity, efficiency, and accountability in the provision of social services in a decentralized setting. Finally, Chapter 7 explores the issue of a social safety net in Colombia. Chapter 3 Human Resource Development Human resource development is a key element of a poverty reduction strategy. Colombia needs to intensify its efforts to increase access by the poor to basic education and health services, and to improve the quality of those services. In both sectors, Colombia has not emphasized enough those programs that most benefit the poor. As a result, public expenditures are not sufficiently targeted toward the needy. The decentralization of responsibilities to the municipal level in education and health can be a positive step if it makes service delivery more responsive to the needs of local communities. However, considering the institutional weakness of the social sectors, it could further increase community disparties. I. Introduction 3.1 Human resource development-investments in education and health to incase the human capital of the poor-must be a central component of poverty reduction. Public support for these activities has the double benefit of improving living standards both in the short and long term. Younger generations benefit the most, since poor children can acquire the tools and oppornmities to break with patterns of low productivity and income. 3.2 Chapter 1 shows that Colombia faces important gaps in social conditions that must be bridged to make further progress in reducing poverty. Human resource development must become a priority for Colombia and be reflected in acdve and effective government intervention in education, health, and nutrition. 3.3 This chapter analyzes government programs in education and health.' The purpose is to assess how effectively existing programs serve the poor, to identify the forces that constrain more eff'vcive action, and to recommend change when necessary. 3.4 Public expenditures in education and health are not well-targeted to the poor. Thn is generally the r:sult of a history of public support for activities and programs such as higher education and higher-level health services at the expense of programs-such as primary education and basic health care-that most benefit the poor. The poor are also affected by the low quality of services they receive. 3.5 The weakness of Colombia's social institutions is the most serious constn to effectively serving the poor in education and health. This chapter considers how low capacity I Mo of f ijtonado i ft del,ra is ob_t die_mf bakgmwo pae eaborated for do rpoit (Mouns- ad ad (1993 a. b)). - 53 - has limited the efficacy of Colombia's social programs.2 Section II considers education, and section m considers health. Each section begins with a sector overview, including an analysis of public expenditure and programs serving the poor. Next, the question of who benefits from public spending is discussed, with results compared with those observed two decades ago. This is followed by an analysis of the main constraints on health and education programs for the poor. Each section closes with conclusions and recommendations. IL. Education 3.6 During the past several decades, Colombia has made important gains in extending educational opportunites and increasing the average schooling of the population. This was made possible by steady and substantial increases in national spending on education through the mid-1980s, which were not sustained in the latter part of the decade. There remain considerable inequities in the provision of education both among income groups and between urban and rural areas. The overall quality and efficiency of basic public education at both the primary and secondary levels are extremely poor. Students at the lowest income levels are most seriously affected. Selected national government programs, particularly the Escuela Nueva, have made significant contributions to extending education to the poor. But Colombia will need to reorient its policies and make major new investments in primary and secondary education and preschool programs if the poor are to gain access to quality public education. It will also have to increase recurrent expenditures that cover basic operating costs of school systems if performance is to improve. Sector Overview 3.7 The Colombian education system begins with one year of preschool for children under age six, followed by five years of primary schooling for children seven to eleven, and six years of secondary education for children from twelve to seventeen. The five years of primary and the first four years of secondary comprise the basic education cycle. These and preschool are constitutionally mandated for all Colombian children. Prvate schooling is available at both the primary and secondary levels. The system provides further general education or diversified technical-vocational education culminating in general or vocational bachiller degrees. Post-secondary technical education and vocational training are widely available through the National Apprenticeship Service (SENA) and private schools. Higher education is provided in numerous public and private universities offering undergraduate and graduate degrees. 3.8 Expenditures in Education.3 Public expenditures in education grew steadily from 1970 to the mid-1980s at an annual rate of 9.7 percent (Table 3',1). Public education 2 11K specific challenges decentizatin bea to the capacW bulldlg process ae anyzd in apter 6. 3 A complet pictue of social expenditues (education healt, social y, housing. IC. nuail pmgns, and so on) ad tbi role in total public expenditur is presented in Annex 3. Tables 3.1-3.4. - 54 - funding doubled as a share of GDP (from 1.8 percent in 1970 to 3.6 percent in 1984). enabling Colombia to expand access to public education at all levels. In contrast, the period from the mid-1980s to the present has been one of stagnation. In 1991 the share of expenditures in public education in GDP fell to 2.7 percent. This trend reversed during 1992, with education spending increasing to 2.9 percent of GDP.4 Table 3.1: Public Expenditure in Education, Selected Years, 1970-92 Annual Expenditure Year (billions of 1992 pesos) As percentage of GDP 1970 238.5 1.82 1974-80 464.6 2.38 1981-85 770.1 3.24 1986-89 791.3 2.80 1990 851.0 2.72 1991 869.0 2.72 | 1992a 946.3 2.86 a. Preliminary Source: Mclina and others (1993a). 3.9 Stable funding levels combined with growth in enrollment has reduced per student public education expendiure. The impact has differed for the three levels of the system. For primaiy education, per student spending increased until 1984 and has remained relatively stable, since enrollments have not increased significantly. For secondary education, the combination of lower real expenditures and enrollment growth has resulted in a reducton in per student spending since 1982. Per student spending for higher education was relatively stable throughout the 1980s. 3.10 Private expenditures on education are higher in Colombia than in other counties. Measured in terms of tuition and fees paid by famrilies, private expenditures have remained relatively constant over the years, ranging from 1.2 to 1.5 percent of GDP. Estimates based on infonnation from a 1992 household survey indicate that, on average, families with children in public primary schools spend 1.6 percent of their income on education-related expenses. The poorer the family, the higher the percentage spent on education. Families in the bottom 20 percent of the distribution spend 4.4 percent of their income, while those in the top 20 percent spend only 1 percent. 3.11 The share of primary education in total public education spending has declined to 38 percent from 50 percent in 1970, though Colombia still has not achieved universal 4 ThI vaue excd te paomems of debts to th teah pesion nds. When toe fids ae nuded, tot eduation expendturein 1992 auntd to 1.153 biion pesos. 3.5 pern of GDP. - 55 - primary education. At the secondary level the share of educational expenditures is 29 percent and for higher education, 21 percent (Table 3.2). The distribution of expenditures differs only marginally from Latin America as a whole, with Colombia allocating a lower share to primary education than the regional average. Table 3.2: Distribution of Public Educational Expenditure, Selected Years, 1975-90 (percent) Year Pniwy Seconday Higher Otheh 1975 50.5 25.0 16.6 7.8 1980 41.7 26.7 24.9 6.7 1985 40.0 32.0 21.5 6.6 1988 40.9 31.7 20.9 6.4 1989 40.1 30.3 20.5 9.1 1990 37.9 29.2 20.6 12.4 a. Includes cultural activities, sports, and nonfornal education. Source: Molina and others (1993a). 3.12 Colombia's skev,ed pattern of per student public spending at different education levels generally follows that of other Latin American countries. In 1990 per student spending at the higher education level (US$1086) was nearly six times the secondary level (US$183), which was twice the primary level (US$92). Compared with other countries in the region Colombia shows low spending per student at the primary and secondary level-only Bolivia, El Salvador, and Paraguay spend less (Table 3.3). Although Colombia's absolute spending levels for higher education seem average for the region, in relative tenns Colombia spends about twelve times more on higher education than on primary (the highest rado in the region with the exception of Brazil and Paraguay). Colombia spends about 30 percent less (in per student terms) on both primary and secondary educadon than the average. 3.13 As in most other countries in the region, public spending in education has only a small investment component. Per student investment in Colombia in 1990 was the equivalent of US$6.2 for primary education, US$7.9 for secondary education, and US$80.4 for higher education. As a share of total operational spending, investment represented only 6.7 percent for primary education, 4.3 percent for secondary education, and 7.4 percent for higher education. 3.14 Since education was nationalized anCfer Law 43 in the mid-1970s, public education has been primarily funded by the national govermnent. The national government's average annual share of funding for education has increased from 70 to 95 percent. Until recently, departnental and municipal fumding for educadon has been mainly used to recrunt - 56 - Table 3.3: Spending per Student by Level, Selected Countries (U.S. dollars) Country, year Pnmwy Secondry Higher Higher/Primay Arr-ntina, 1985 142 310 577 4.1 Bolivia, 1990 72 101 504 7.0 Brazil, 1989 243 434 5,911 24.3 Chile, 198711993 181 363 1,006 5.6 Colombia, 1990 92 183 1,086 11.9 Costa Rica, 1986 144 222 1,541 10.7 Ecuadot, 1987 97 218 652 6.7 El Salvador, 1990 63 66 227 3.6 Honduras, 1990 115 185 1,233 10.8 Mexico, 1984 135 589 1,035 7.7 Paraguay, 1990 50 144 710 14.2 Uruguay, 1989 256 306 614 2.4 Venezuela, 1989 198 347 1,625 8.2 Source: For Colombia, unit costs from Molina and others (1993a) were converted into U.S. dolis at an exchange rate of 502.2. For Chile, Secondary Education Project, World Bank. For other cowunries, Psachatopoulos and Ng (1992). teachers outside the official national roster.5 This explains, in part, the reduction in their share of official public expenditure. The departments and municipalities also cover some educational expenditures from their own resources. For example, in many municipalities a Fondo de Servicios Docentes derived from student fees funds supplies for and maintenance of secondary sriools. This, however, has been limited. 3.15 Under the decentralized framework established by the 1991 Constitution and the newly adopted Ley de Competencias y Recursos, local governments will be responsible for the provision of preschool, primary, secondary, and middle education. The main source of funding will be the growing transfer of resources (through the Situado Fiscal and the TIansferencias Municipales)' that the central govermnen will be making to departments and municipalides. S Sic dc rst was frozen in 1982 by dMe raoa govanmethe enly way depatme and municipa coul hite mome teacbr was to recwwl dthm as docrus tewpeua (pime tachen) and pay them dimecly fom fteir own esoure 6 See Chapter 6 for a discssion of the sysem of uergovemmeVa tdn in Coombia and the Ley de qmpnds y Renws. - 57 - 3.16 Programs Serving the Poor. Since the mid-1970s, various governments in Colombia have launched programs to expand and improve educational opporunities for the poor. The Escuela Nueva, an essentially rural program, is the best known of these programs in view of its poverty focus and innovative features (Box 3.1). Sustained evaluation of the Escuela Nueva shows sound results. Most notably, evaluation outcomes confum that it is a viable means to provide the complete primary cycle in rumral areas. The condnuing basic education program, another targeted effort, was designed to serve the needs of adults and children who were either illiterate or had only rudimentary schooling. 3.17 The Gaviria Administration has adopted new goals that give priority to expanding access among the poor and improving the quality of preschool and secondary education. An example is Aflo 0, a program to reduce high repetition rates in the first year of primary school and to prepare children for effective learning in later years. The program helps children move from home to school through activities that foster interest in learning, positive social attitudes, and independence. 3.18 Another innovative program is the use of vouchers for secondary education. The lack of public secondary schools was recognized as one of the reasons for high dropout levels between primary and secondary schooling. To expand access to secondary education, the national government launched a program offering vouchers to low-income students to attend private secondary schools that have excess classroom capacity. This program, cofmanced by the national government, offers municipalities a short-term, cost-effective, and efficient alternative to expand access to secondary schooling. From 1991-93 the government funded 72,000 vouchers in municipalities throughout the country. Box 3.1: Escuela Nueva - A Successful Program Serving the Rural Poor Launched in 1976, Escuela Nueva is one of the government's most successful programs for extending educadional opportunities to the rural poor. The program initially covered 50 schools, and by 1989, 17,948 schools were serving 800,000 students. The program seeks to improve student achievement, enhance educational efficiency and productivity, and integrate the school and the community. It employs a multigrade teaching technique that provides a full five-year primary course with no more than three classrooms and tbree teachers.' The curriculum is rural-oriented, recognizing differences in student aptitudes and the inevitability of student absences, especially during periods of peak agricultural acdvity. It advocates a flexible promotion policy that:.. encourages students to progress at their own speed and to reenter school after periods of absence. Students . .,<-. advance based on learning rather than grade level. Emphasis is given to student assessment and to remedialt' assistance. Student monitors and community members may assist in the provision of remedial suppon for. students. Some materials are drawn from the community (for example, a local artisan might prepare clay -fr small children to form letters or numbers). Other materials, though prepared centrally, relate to typical community situations and encourage student questioning and experimentation.. - 58 - Incidence Analysis: Who Benefits from Public Expenditures in Education? 3.19 Enrollment Trends. Since the 1950s coverage rates at all levels of schooling have grown steadily. Whereas only 50 percent of children between ages seven and eleven attended primary school in the early 1950s, primary school attendance was about 86 percent in 1992 (Table 3.4). Secondary education enrollment also grew steadily, doubling between the 1960s and 1980s. In higher education, participation remains modest, but nonetheless grew to about 10 percent in the early 1990s. According to the Ministry of Education, roughly 4.4 million students are now enrolled in primary education, 2.4 million in secondary schools, and half a million in higher education. An additional 300,000 students, twice the number in the mid-1980s, are now enrolled in technical-vocational training offered by the National Vocational Training Service (SENA).' Table 3.4: Enrollment by Level, 1992 . Level Public Private Tot4l Pimnuay 3,708,356 678,042 4,386,388 Secondary 1,549,632 880,155 2,429,787 Higher 202,798 323,171 525,969 Source: Ministy of Education and ICFES. 3.20 The private sector represents a sizable share of the education system. At the primary level. private schoc'. haie increased their share of enrollments to about 15 percent. Secondary plivate schools cover 40 percent of all enrollment, a drop from 60 percent in 1960. In higher education, private universities and technological institutes represent 60 percent of total enrollment. 3.21 Differences in Enrollment. Female enrollment rates are higher than male rates at all levels of schooling. This is true in rural and urban areas and across income groups. Colombia does not have a gender bias in access to formal education. Significant cross-country differences in coverage remain, however. Coverage for all three education levels is higher in urban than in nual areas. At the primary level, there is a 12 percent difference in enrollment rates; at the secondary level the difference is close to 16 percent. 3.22 Coverage also varies among regions. Departments like Antioquia, Caldas, and Quindio have primary school net enrcilment rates of over 90 percent. Others (mostly in the Adantic region) like Cesar, Magdalena, Bolivar, C6rdoba, and Sucre have rates below 70 percent. Similarly, at the secondary level, Bogota, Valle, and Quindfo have net coverage above 50 percent, while Magdalena, Choc6, and Bolivar approach 30 percent. Coverage in large mnnicipalities averages 58 percent, but is only 36 percent in smaller municipalities. 7 This uma is baed on 800.000 crses, wih an avena atndac of 2.5 curs per sumdn per year. - 59 - 3.23 Coverage by Income Groups. The significant growth in prmary and secondary sckool enrollment since the early 1970s reflects the increased participation of students from the lowest-income households. Children from the poorest 20 percent of households now constitute 40 percent of enrollment in public primary schools and 20 percent in public secondary schools. In 1974, these households constituted only 32 and 17 percent of public primary and secondary school enrollment, respectively The composition of private school enrollment, however, has not changed. 3.24 There remains a significant shortfall in coverage at all levels of the educational system, particularly among children from low-income households. Table 3.5 shows how school attendance varies by income, a phenomenon clearly observed in urban areas. For children six to twelve, the top quintile has almost universal attendance, while the bottom quintile has 20 percent non-attendance. The difference in attendance between the poorest and richest 20 percent increases with schooling level. For students between thirteen and nineteen, the difference is 25 percent. For those between twenty and twenty-five, the top quintile has an attendance rate four times that of the bottom quintile. Although a similar pattern is found in nrral areas, the picture is somewhat less clear since attendance rates do not always increase from one quintile to the next as they do in urban areas. Table 3.5: School Attendance by Income Level and Age (percent) 6-12 Years 13.19 Years 20-25 Years QuW e.k Urban Rwwl Urban Rurl Urban Rural 1 79.1 75.1 50.7 38.6 7.6 6.1 2 86.0 80.1 57.5 51.5 10.5 7.8 3 90.1 79.5 58.5 45.8 11.8 7.1 4 95.2 82.7 64.6 72.6 19.0 6.4 | S 96.0 89.2 75.41 51.2 30.0 9.2 Source: Molina and others (1993a). 3.25 Unit Subsidies. Table 3.6 presents the unit costs for public education for nural, national, and large, intermediate, and small urban areas.' 3.26 Primary Education. Enroilmert in public primary education is progressive. Households in the lowest income quintile represent 40 perce& of aL enrollment at this level. Thus primary education expenditures ought to benefit the poor more than the rich. But spending per student is higher in urban areas than in mral areas. Because average income is lowei in mral areas, where a large share of the pocr live, the difference in unit costs has a 8 These vaues exclude pe spendg. - 60 - Table 3.6: Subsidies in Education, 1992 (1992 pesos) Area .may Secondary Higher Large Cities 75,177 149,441 1,107,081 Iermedirte Cities 102,779 204,218 1,0?,1,835 Small Urban 111,639 186,294 860.277 Rural 78,784 168,259 1,010,954 National 86,649 | 170,916 1,010,954 Source: MWisin Social. regessive effect. Still, spending on primary education is highly progressive. The poorest 20 percent of households receive 39 percent of the total subsidy, an amount rine times that received by the richest 20 percent. And the distribution of subsidies today is more progressive than two decades ago, when the poorest quintile benefited from 32 percent of primary education spending (Table 3.7).9 Table 3.7: Distribution of Subsidies to Public Primary Education, 1992 D&stibutlon of Susddy Enrolment (percent) Subsidy per Household &Quk (stdents per household) 1992 1974 (pesos) 1 1.09 39.4 32.1 91,461 2 0.70 26.2 26.7 60,909 3 0.51 19A 20.5 45,026 4 0.28 10.8 14.5 25,137 S 0.11 4.3 6.2 9,912 Source: Misdn Social. 3.27 Secondary Education. Spending on secondary education is less progressive than spending on primary education because the poorest quintile has low enrollment and resides in mral areas (where spending per student is lower). At the secondary level quintiles two and tree benefit the most, together receiving 52 percent of the subsidy. The system has become significantly mote progressive since 1974. The first tbree quitiles benefit more from public spending in secondary education today than they did two decades ago (Table 3.8). 9 For zersaces W 1974 es Seowsky (1979). - 61 - Table 3.8: Distribution of Subsidies to Public Secondary Education, 1992 Disbibuion of Subsidy Subsidyper Enrollment (percent) Household Quindle (studenta per household) 1992 1974 (Pesos) 1 0.30 20.6 16.8 51.683 2 0.39 26.8 21.8 67,293 3 0.37 25.3 21.2 63,655 4 0.26 17.7 24.6 44,560 5 0.14 9.6 15.6 24,210 Source: Misidn Social. 3.28 Higher Education. Enrollment in highei education is concentrated among the two top quintiles-only one in four students at this level belong to nouseholds in the '-ottom 60 percent of the distribution. It is not surprising, then, that the distribution of benefits from public spending is distributed unequally (Table 3.9). Compared with the 1974 distribution, however, public expenditue on higher education benefits more low-income households. In 1974, individuals in the bottom quintiles simply did not attend universities. Though expansion of higher education has taken place, it has been at a relatively high cost-most public expenditure benefits the relatively rich. Table 3.9: Distribution of Subsidies to Public Higher Education, 1992 Disribudon of Subs*id Subsidy per Enrollment (percent) Household Qinte (students per household) 1992 1974 (pesos) I 0.01 5.1 0.8 8,932 2 0.02 9.4 4.6 16,434 3 0.03 18.6 10.7 32,593 4 0.06 33.2 23.5 58,186 S 0.06 33.8 60.4 59,428 Source: Misidn Social. 3.29 SENA. Students enrolled in technical-vocational training courses offered by SENA are primarily from middle-income backgrounds. Fifty-one percent of all SENA students belong to quintiles three and four; only 11 percent belong to the bottom quintile (Table 3.10). Because SENA's courses are geared toward in-service training rather than pre- service traiiing, most trainees already have positive earnings and are generally better - 62 - educated than the average worker. In SENA's case, public expenditure is targeted not to the poor but to middle-income groups. Table 3.10: Distribution of Subsidies by SENA, 1992 Quline Dbbuton of Enrlend Dstibuon of Subsidy (percent) (percent) 1 10.9 12.1 2 19.5 18.3 3 23.7 23.3 4 27.2 27.7 5 18.7 18.6 Source: Misidn Social. 3.30 Overall Distributive Imp,'ct. Public education expenditure benefits low- income households relatively more than other income groups. In 1992, the poorest quintile acounted for 23 percent of all subsidies and received, on average, a subsidy 63 percen igher than the rchest quntdle. However, considering that the poor have larger families, the distribution of pet capita subsidies is in fact almost neutral. Progress has been made since 1974. But the three education levels follow quite different patterns, and poor students benefit less as thy move from primary to higher education. This is reflected in the contribution the ree levels make to the total subsidy received by each quintile (Table 3.11). As one moves up the income distribution, primary education drops from 60 percent to 11 percent of the subsidies, while higher education increases from 6 percent to 64 percent. Secondary education explains a large share (47 and 45 percent, respectively) of the subsidies received by the second and third quintiles. Table 3.11: Total Education Subsidies, 1992 dubu Contributon Shari' Pesos per Pesos per Pdmy Seconday Higher Quie (percet) Household pita (ecent) (percent) (percent) 1 23.1 (19.8) 152,076 28,022 60 34 6 2 21.9 (20.2) 144,636 29,418 42 47 11 3 21.4 (18.6) 141,274 29,917 32 45 23 4 19.4 (20.1) 127,883 31,104 20 35 45 S 14.2 (21.3) 93,550 25,658 11 26 64 a. Inclues primary, secondary, and higher educadon. b. 1974 shares are in parenthesis. Source: Tables 3.7, 3.8. and 3.9. - 63 - Main Constraints and ChaUlenges 3.31 Coverage. Although Colombia has increased access to primary and secondary education, there are still significant shortfalls in coverage at both levels. With an overall coverage rate of about 86 percent at the primary level and near 50 percent at the secondary level, Colombia is still far from achieving its constitutional mandate for universal basic education to age 15. Coverage alsQ favors the rich and is substantially higher in urban areas. Some departments (mostly in the Atlantic region) experience exceptionally low coverage. 3.32 At the primar: school level, the problem is concentrated in certain departments.'° Evidence indicates that both supply and demand play 2 role. On the supply side, some areas lack sufficient facilities. On the demand side, high direct costs and the need to work limit the chances for low-income children to attend primary school. There is also evidence that in the poorest two quintiles the children of less-educated parents have lower attendance. In a highly centralized system local communities have little opportunity to voice needs and concerns. It is not surprising that the children of those with little schooling are left out, 3.33 Coverage problems at the secondary level are more widespread geographically. As with primary education, both supply and demand constraints explain current enrollment rates, though at the secondary level demand factors dominate. It is generally recognized, however, that past underinvestment in public secondary school facilites has created insufficient supply, particularly in low-income urban areas and in nrual areas. In both small municipalities and ural areas, secondary educaton typically does not extend beyond grade nine. 3.34 Insufficient demand is explained by two factors. Low promotion rates in primary education combined with high dropout rates after five years of schooling reduce the number of children that qualify for secondary schooling. The poor quality of primary school education discourages low-income children from pursuing secondary education. In addition, the direct and opporunity costs of secondary schooling remain too high for a large portion of the population, particularly male students. In medium and large municipalites, insufficient public facilities coexist with excess private sector capacity. Private capacity is not used because of the lack of effective demand. The voucher program (see below) addresses this concern. 3.35 The national policy of automatic promotion at the primary level adopted in 1988 and high retention rates resulting from school improvements introduced under Escuela Nueva are projected to increase both primary education completion rates and primary to secondary transiton. Secondary education enrollment grew at an average anmual rate of 2.2 percent over the past five years. It is now projected to increase at an annual rate of 4 percent 10 See Chapter I for a descpon of regional patterns in edwatonm and other social indiazon. - 64 - for the remainder of the decade. Thus, as one factor limiting secondary education demand is removed, new bottlenecks are likely to appear unless current investmnent levels are increased. 3.36 Quality. Public education in Colombia suffers serious quality problems. Tests indicate that the system is not educating students effectively, and the poor are most seriously affected. The results of Colombia's first national assessment of primary education," conducted in the 1990-91 school year, revealed serious deficiencies in reading comprehension and mathematics, as well as in socially accepted norms and values. The majority of pnmary school students in grades three and five are not working at grade level, nor can they apply the knowledge acquired in school to daily life. Reading comprehension and mathematics scores indicate that a majority of fifth-graders lack the ability to infer meaning from texts beyond the most basic comprehensior, and cannot apply math concepts to problem solving. 3.37 The quality of secondary education is also unsatisfactory and has been declining since the mid-1980s largely because of expanding enrollment and the three-shift system used in many public schools. Scores on the national college entrance examination (ICFES) reveal that the majority of public secondary school students have achievement levels substantially below their private school counterparts. A recent study of college entrance examinations found an increase, starting in 1985, in public and private school scores below the national median."2 The study also confirmed that the lowest achievement .was by students in the night shifts of public schools, where 76 percent had scores in the lowest achievement category. Not only is the quality of secondary education declining, the poor-most of whom attend low-budget private schools or public schools during night shifts-are receiving an even worse education than other segments of society. 3.38 Several factors contribute to this problem. Most spending at the primary and secondary levels is personnel-related, which cuts expenditres on essential insttuctional inputs, basic materials, and equipment known to impact education quality. The quality of teachers is also inadequate. More than 50 percent of teachers lack post-secondary degrees. The percentage rises among teachers hired locally. In addition, promotion policies rarely compensate performance. Finally, a highly centralized system has not encouraged comnunity involvement in schools and school management. 3.39 Ineffiendes. The average Colombian student requires seven years to complete the five grades of primary school. While 87 percent of students complete all five grades, 19.5 percent do so only after repeating three or more courses. After five to six years of schooling, students begin to leave the school system in increasing numbers, probably due to economic necessity. Frustration could also be a factor. it Colombia's National Evaluadon Systm (SABER) was inid in 1991 to appraise te quality of Colombb's educaoal system. SABER's goal is to inprove all dimensions of education and to establish an evaluation *culume from the school level to the levd of regional and national government'. 12 Alviar and Potania (1991). - 65 - 3.40 Colembia's public secondary education also suffers fro-' inefficiency. Estimaced 1991 promotion rates averaged 74.5 perent (oniy 62.2 perewn'. in grade .ax), dropout rates 9.6 percent (13.7 percent in grade six), vA repedsion rates 15.9 percenm. Actual figures may be even worse-many incomplete sxcondary schools (those widout Ministrv of Education certification or the full six grades) are not cwunted. Oniv 50 percent of sixth grade students reach grade eleven five yoars later, and those who co;nptl sw.cwuary school are most often from wealthier families. A combination of admintratve inefficiencies-bureaucratic procedures, inadequate hiring policies, and high rwes of staff turnover-have stymied attempts to improve course quality and rates of proinotion al.; retention. 3.41 Voucher Program for Secondary Education. An analysis of the vouchers awarded indicates the majority of vouchers reach low-income students.13 But the program raises several issues. Although the voucher program aims to increase enrollment, it is likely that the increases could well be substantially less than projected due to substitution (that is, if students receiving vouchers would have attended private school anyway). Another area of concern is the quality of education. Most voucher recipients-particularly the poorest- attend schools with low standards, many of which do not offer the full secondary cycle. Finally, the sustainability of the program is not assured after the rent cofinancing by the national government expires. 3.42 Voucheqcan be a powerful tool to generate effective demand for excess private sector capacity. Though no quantitative assessment of the magnitude of such capacity is available, it is unlikely to af.ect coverage rates significantly. For that, an increase in supply will be required, which the voucher program could encourage by offering incentives for private sector investment. But the program would fist need to achieve a level of credibility currently lacking, and the quality concerns discussed above would need to be addressed. In addition, the cost effectiveness of the voucher program as a tool to increase private sector investment in education will need reassessment. Finally, the voucher, by covering only the direct costs of schooling, is unlikely to generate demand by the poorest groups, for whom the opportnity cost of secondary education will probably remain high. 3.43 Decentralization. Decentralizing responsibility for primary and secondary education may be a potentially effective way to address many of the constraints described in previous paragraphs. Transferring responsibilities to municipalities is a necessary-though insufficient-condition to empower communities. Decisionmaking at the local level can more accurately reflect the needs of users. But it requires competitive political regimes in which authorities answer to their communities, and institutional channels through which the disadvantaged can voice their demands. It also requires effort, time, and resources to build the technical and administrative capacity to run school systems. When and if these conditions are met, the system should be positioned to address most of the factors constraining coverage, quality, and efficiency. 13 Sec Molina and others (1993a). - 66 - 3.44 Tne Ley de Corpetencias y Recursos, passed in 1993, changed education decentralization by giving departments more control over teachers and financial resources. Now, those teachers paid with resources from the Situado Fiscal (the majority of teachers) will he under the authority of departments, who will distribute them among municipalities according to need. 3.45 In the immediate future, the main challenge associated with decentralization will be the local capacity to make decisions and administer the system. It is likely that, in the short run, decentralization will lead to heterogeneous responses, with improvements in only a minority of municipalities. Financial resources, however, are unlikely to be a major constradnt. Under the new arrangements, most municipalities will receive a net increase in resources for prinmay and secondary education. Unfortunately, the initial conditions in municipalities are so diverse that analysis is limited to general statements regarding the limited capacity in smaller communities.14 Conclusions and Recommendatons 3.46 While Colombia has made solid gains in providing access to education to its population, the pace of progress has diminished since the mid-1980s. The education system is at a crossroads. The decentralization effort will pose opportunities and challenges and requires a genuine commitment to improve education, reflected in government priorities and expressed in a Minister of Education who embodies these principles. 3.47 Overall, public spending on education is distributed in a mildly progressive fashion. As Colombia moves toward universal primary schooling, low-income households are receiving a larger share of expenditures than richer ones. Likewise, the distribution of education spending is now more progressive than two decades ago. As coverage increases, public education has reached more disadvantaged groups-though not always the poorest. Thus, in 1992 the poorest 20 percent of households accounted for 40 percent of enrollment in public primary schools and received 39.4 percent of public expenditures at that level. This dynamic is particularly relevant to secondary education, which is bound to expand in the next decade. Today the second and third quintiles of the distribution benefit the most m'rom public expenditure at that level. If current expansion plans are successfully implemented, secondary education spending in ten years will be as progressive as spending in primary education is today. 3.48 In contrast, public spending on higher education is regressive. Households in the bottom 20 percent of income distribution account for only 5.5 percent of enrollment and S. 1 percent of subsidies. Though a much larger share than the group received in 1974, the poor still do not benefit from public expendiure on higher education. If public actions are to effectively reduce poverty, heavy subsidization of programs that reach only the wealthy households should be reconsidered. The high levels of private spending on higher education 14 A broader discussion of Ihe challenges posed by the decmizato prces s included in Chapter 6. -67 - indicate users' willingness to pay. Thus, an education strategy oriented toward poverty reduction will require increased cost recovery in higher education and expanded efforts on the primary and secondary levels. 3.49 As coverage grows and all income groups have equal access to public education, differences in per student spending and school quality become the main sources of inequity. Today, rural areas have spending levels below urban areas." Decentralization under the Ley de Competencias y Recursos will reduce these differences by transferring more resources to smaller towns. This will require local commitment to education. 3.50 Smoothing the sharp differences in quality is likely to prove more difficult. The devolution of power to tue local level is a positive step, bringing decisionmaking closer to users. But it can also contribute to disparities between municipalities. Addressing such tendencies is the key challenge to policymakers. While success will depend principally on the direction chosen by local authorities, the central and departmental governments play important roles in offering incentives and supplementing local efforts. The following are some of the priority areas for action by central and departmental governments. 3.51 Increasing Access to Basic Education. Local governments will be responsible for achieving the constit!itionally mandated goal of universal basic education. But the national govenment can provide incentives through technical assistance and cofinancing to assist those departments and municipalities having difficulty extending basic education coverage.1' Cofinancing through the FIS (Fondo de Cofinanciaci6n para la Inversion Social) shoOld be used to encourage local authorities to invest in additional classrooms where there are inadequate facilities. It should also encourage municipalities to allow students to attend public schools in their neighborhood. This would reduce the high incidence of dropouts between primary and secondary schooling. The FIS and the departmental government could also play a role in providing the technical assistance poorer municipalities will require. 3.52 The combination of mtergovernental transfers through the Siuado Fiscal and Transfferencias Municipales, the cofinancing resoumrces of the FIS (derived in part from external public financing), and the loans from FINDETER will help municipalities and departments cover the costs required to expand coverage and improve quality of secondary education. DNP has estimated the costs for expanding coverage from 46 percent to 70 percent during the 1990-95 period to be US$54 million, about 0.1 percent of GDP. Most of those costs will have to be paid from the increased resources local governments will receive through transfers, with the FIS concentrating its resources (estimated to represent only 7 percent of intergovernmental tsfers) on projects strictly targeted toward poverty reduction. 15 in uban areas, arge cides have he lowes kvels of per student spendit. Thais expled by the econoies of sale that opetate in laiger cities. 16 See Chapter 6 for a discussion of the PIS and otercofnaing nds. and teir ue as an inenive f_Imwk dft local govemment expenditures to-ar the poor. - 68 - 3.53 Improving Quality. The national government's priority must be improving educational qualiiy. This requires rewarding performance rather than tenure, improving the availability of instructional materials in schools (particularly textbooks), and providing adequate faciliti-s to limit overcrowding. Ultimately, the government should eliminate the three-shift system used in most public secondary schools. 3.54 Some of these improvements require new investments and entail new recurrent expenditures (purchating textbooks and other instructional materials, maintaining schools). Others, such as revitalizing the teaching force, do not depend solely on expenditures, but call for improvements in the quality of training, in the managerial capacity of school directors, and above all, in parental demand for quality education. The national teachers' union (FECODE) could play an important role in encouraging these inprovements. 3.5 In pursuit of educational quality and poverty reduction, the Ministry of Education is seeking to improve student achievement in math and reading during the first three grades of primary school. The program supports in-service training of teachers, instructional materials, and technical assistance to schools in low-income areas. It will initially be implemented in four departments (Cordoba, Bolivar, Boyaca, and Choc6) that have high poverty and low attendance. It will eventually be extended to as many as ten departments. 3.56 As mentioned before, the Escuela Nueva has an impressive record of achievements. Recent evaluations indicate that if Escuela Nueva is to continue to function effectively on a national scale, it will require renewed commitment from senior policymakers for its fimdamental principles. It will also require overcoming factors that have recently limited effective implementation, such as logistical problems involving training materials and dissatisfaction with the program's technical training. Also, more resources should be devoted to teaching materials and less to school infastructure. Finally, recent political decisions have tended to replace experienced teachers with untrained, inexperienced ones. 3.57 Building Local Commitment and Capacity. The projected growth in the Situado Fiscal and Transferencias Municipales should substantially increase the resources available to regional and local governments for public education. Theoretically these expanded resources should increase access and improve the quality of public education, but these depend on the commitment of local govemments. Building this commitment is as critical as building capacity. The national government, particularly the Ministry of Education and the FIS, should devise strategies to support these ends. School-based mangement should also support increased commitment by school officials and the community. 3.58 Strengtheing the National Government. The strategies for increasing coverage and improving educational quality will need national govenmnent support, a responsibility of the Ministy of Education and institutions like the FIS. But the Ministry itself needs substantial strengthening to play an effective role. A first step would be to recruit highly qualified professionals. A second would be to maintain contimnity in management, - 69 - avoiding frequent cicanges of Minister and vice-Minister. A third would be to support the decentralization of education by building management capacity at regional and local levels. Without these changes, it is unlikely that the national government can provide leadership, ensure continuity, and support the aforementioned measures. III. Health 3.59 Over the past three decades, Colombia achieved impressive gains in health. Between 1965-90 the population growth rate declined from 3.7 percent to 1.7 percent, and infant and -naternal mortality retes were cut in half. Continuous economic growth allowed for such broad gains, but the distribution of many social benefits related to growth remains highly skewed. Marked disparities persist in key health indicators and in the provision of related services across income groups and among rural, periurban, and urban areas. Colombia's health system has emphasized curative over preventive care, and high-level institutions over primary-level ones. Approximately one-third of the population lacks access to basic health services, most because of cost. Current health reforms address only some of these problems. Technical and administrative weaknesses must also be addressed for the new model to be successful. Sector Overiiew 3.60 Colombia's health system has three major parts: (a) the public health system (PHS); (b) social insurance services (SSS) provided mainly by the Social Security Institute (ISS) for private sector employees and by CAJANAL for public sector employees; and (c) private services. The public sector is the largest health services provider, officially responsible for about 60 percent of the population, including the poor. But effective coverage is esimated to reach only 35 percent, leaving a quarter of the population without access to health services. Social security and private providers are responsible for 22 and 17 percent of the population, respectively. The public health sucre is headed by the Ministry of Health (MOH). At the regional level, the MOH is represented by Direcciones de Salud. At the municipal level, the sector is represented by Direcc!4n Local de Salud and, increasingly, by an autonomous apparatus under each mayor. 3.61 In the public sector, health care is provided at three levels: a) local health posts, centers, and hospitals (level one, in terms of complexity); b) departmental hospitals (level two); and c) specialized and university hospitals (level tree).17 Infrastructr includes over 4,000 health posts and centers, 500 local hospitals, 124 level two hospitals, and 27 level three hospitals. The ISS maintains several health centers and 40 hospitals. In 1990, the total nmber of health personnel included about 30,000 physicians, a ratio of 9.4 per 10,000 population. But only 2.5 per 10,000 worked in the public health system. About 50,000 17 Lewl wo bospitals provide series in the basic specialties (hding inteal medicin, s ,rgety obics and gynecology, ad peditcs), peneally not avaible at farst lv. Level hee hospils an equipd whh advanced diagwic and ta thnologies and stafe wibt specwli mrses and nurse assistants and about 8,000 he;Jh promoters work at the local level in the public health system."8 While overall numbers are acceptable, the distribution is skewed toward higher-level institutions; nmrse assistants and promoters are often the only staff available at health posts and smaller rural centers. 3.62 Expenditures In Health. Colombia spends about 6 percent of GDP on health services. This share is comparable with the high end for middle-income countries. Yet the profile by type of expenditure fits that of the low-income countries (a high share of private spending on drugs, a modest share of public spending on direct services, and limited insurance coverage)."9 Private expenditures (about 4 percent of GDP) account for the largest share of health spending. Pu Ž1 e expenditures-excluding social security programs-constitute 1 percent of GDP.20 Finally, social security programs in health represent 1.4 percent of GDP. 3.63 Public health spending has grown at high rates since 1986. Spending increased at an annual rate of 7.$ percent between 1986 and 1991 (Table 3.12). During this period, however, the two public subsectors showed very different degrees of dynamism. While expenditures by social security programs grew by 10 percent annually, expenditures by the PHS simply recovered the growth pattem they had during the 1970s. Prelimnary informaon for 1992 indicates an acceleration in PHS spending. Table 3.12: Growth In Public Spending on Health (Annual rate of real growth) Period Cenrl GovenmenS" Socal Secitn TOW 1970-80 3.2 2.9 3.0 1980-86 1.3 1.4 1.4 1986-91 3.7 9.9 7.5 1991_92b 14.2 7.9 10.1 a. Funding by regional and local governments, which accounts for about 25 percent of spending by PHS, is not included. b. PreliHinay. Soure: Annex 3. Table 3.7. 3.64 In 1990, 40 percent of recurrent health expenditure (financed through national funds) was allocated for the PHS, and 60 percent for the SSS. This split underscores the ineuity in per capita expenditure and the inadequate funding for the public system. Social is The utis per 10.000 poulation in die public healh setice are 1.2 for nurs, 12 for muse assist ansl 3.8 for p _ons 19 See WoMd Bank (1993d). P. ItI. 20 This edetspending by ICBF. typic comdsred nder healt eepditu which is anlyzed septely Chiaptur 7. -71 - security expenditure per capita (recurrent) represented almost six times the expenditure per capita in the public health system, about US$69 compared with US$12.21 Moreover, per capita expenditure in the PHS, which services mainly low-income groups, is insufficient to provide adequate health care-per capita costs for a minimum, standard package of preventive and curative primary care are US$22 for a middle-income country. 3.65 Adjusted for regional variations,23 expenditure per capita in 1990 averaged US$35.24 With few exceptions, expenditure levels are higher in richer departments. Thus, per capita levels range from about US$16-20 in some of the poorest regions (Cesar, Choc6, Magdalena, Narifio, Sucre) to US$50-60 in Bogota or Quindio. Expenditure under the Situado Fiscal (the share of general government revenues allocated for health and education expenditures, mainly recurrent) shows a neutral distribution with regard to income, and somewhat favors regions with the most critical health problems. Thus, variations in per capita expenditure are largely attributable to differences in social security expenditure-and concentrated in large urban areas. 3.66 Government's Strategy , X le the Poor. A centerpiece of the government's stategy is the universal provision of a stanuard package of health services-the Plan Obligatorio de Salud-that includes health prome "-n and prevention, education, curative, and emergency services.25 Law 100 of 1993, whic: reforms the social security system in terms of pensions as well as health, was approved by Congress and has been in force since Aril 1994. The new system will integrate the current public health ard stcial security systems, as well as draw on private providers. A dual regime wi1l cover all Colombians regardless of income. Those with relatively high earnings (formal sector workers and certain independent workers) and their families will be part of the rEghnen conributivo. The rest of the population-mostly the poor-will be part of the rEgimen subsidiado, fianced tbrough a solidarity fund, the Fondo de Soldaidad y Garantta (PSG). Given resource limitations, the initial package available to those in the rEgimen subsidiado will be less comprehensive than that available to those in the rigimen contribuaivo. Over time, however, the former will be expanded with the goal of achieving a single package by the year 2000. 21 Sie World Bank (1993c). 22 See World Bank (1993d), p. 117. 23 An adjusunm factor is applied to a nmber of sevices, Iluding ouade. matenty, swy, and hostlizaton catm, baseud on institutonal data of dte natonal heat sytm and soci securty. for a descripon of the me_odology. see Moliu and otiers (1993b). 24 Basd on Iformation from Table 3.8 in Anoex 3 and usng an exge ate of 502.2 pesos per U.S. dollar. 25 Educational serices Inude hygiene, nuiti sex educo tam of water, and eady cddhod sthmaton. Preventve serAvices include Iniuo. bamly planning, pred cats. detn of cancer, occupadona sfty, assssing faiy dsk. epidemlologic contiol, monttorin mu a intae, ad monitoing childhood dielopmnt of audiry an visual capace Basi cuivec re incud pediics. intl edcine gyecolbgy and obsw , miotr srery,. palaioa and rehailittionfor problat of low complexity. ddldbit6 and emaorgency serr es. - 72 - 3.67 Under the new system, families enroll in a health organization of their choice. The Entidades Promotoras de Salud (a variety of public, private, and social organizations), are now responsible for the provision of the standard health package, either directly or through other institutions (hospitals, health centers). The Entidades Promotoras de Salud collect contributions from members belonging to the regimen contribuivo equal to 11 percent of their incomes and receive a subsidy from the FSG for those members belonging to the regimen subsidiado.6 3.68 The regimen subsidiado is financed through a combination of resources. National resources are channelled through the FSG and include a 1 percent contribution by the affiliates to the regimen contributivo, matched by a similar contribution from the national budget." Municipal and departnental resources are derived from the Transferencias Municipales, the Situado Fiscal, Ecosalud, and a share of the revenues that municipalities and departments receive from the Cusiana and Cupiagua oil fields. 3.69 To promete the package, the government plans to create 60-80 user groups, Empresas Solidarias de Salud, during 1993-94 as forerunners to the broad expansion of prepaid services. These would comprise 200 to 1,000 urban families and 2,000 to 5,000 rural families each. Community boards would secure technical assistance from independent or govermnent-affiliated experts to contract with local health providers to deliver the package. Initially, a national grant equivalent to US$33 per person (80 percent of the average cost of the package at US$41 per person, per year) would be provided. This grnt would be reduced to 20 percent of the per person cost by the program's third year, when the bulk of the cost would be covered by fees and regular government revenue transfers. Despite initial relucae of municipalities to buy into the program, early outcomes are promising. 3.70 The Bank supports health sector reforms througi, an operation that introduces the package to PHS users at the municipal level in nine of the poorest regions. The program includes support for municipal institutional and service development. It also introduces pricing and quality controls for primary health care, develops improved incentives for PHS pesomnel, and incorporates evaluation of project impact. 3.71 Essential Drugs Program. An essential drugs program (EDIP) inidated by the Ministry of Health has established 2,500 conununity phannacies. A list of 420 basic drugs-about 10 percent of those currently available in the country-has been prepared, and 60 of these are assigned to first-lev'l hospitals. The drugs are sold to community members at cost. Continued central support for the EDP and its expansion is pr ed. But first-level 26 The value of the heal packae wi vay accordg to age and gender an be refenred 1 as te Vdtad de Pago por q*mOn (UPC). 'he PG will credit dte E£idades Pmtoras de SaIkd one UPC per member. When member conution dffer from these aedits, dhe FS0 and Emidades will reoi dw dUirene. 27 Thus total contibuon by membs of the righme contii will be 12 per of their Incomes, up from dhe curmt 7 percent tax for SSS sevices paid by socI-scuty affliaes. Two4thds of tids conuton wM be pid by the employer and one4tird by dte enpioyee. - 73 - hospitals, where needs extend beyond the 60 assiged drugs, remain vulnerable. Most do not participate in cooperative purchasing and continue to buy at high cost from private pharmacies, thus discriminating against the poor. Incidence Analysis: Who Beneflts from Public Expenditures in Health? 3.72 Service Use. The public health system is marred by limited coverage. Depending on the type of service, only 40 to 50 percent of health services are delivered by the PHS and the SSS. The rest is provided by the private sector. A significant number of Colombians are left without access to basic health care. Almost 20 percent of individuals in need of medical assistance did not attend a health center because it was too expensive or too far from their home (Table 3.13). Among the poorest 10 percent of the population, this share is 36 percent. Differences exist between es of the country, with rural areas experiencing the highest incidence of non-attendance. Table 3.13: Non-Access to Health Serv1zs (percentage of individuals In need of medical assistance) Area AU Populton Poost 1 pereant Nadional 19.2 35.7 Large Urban 14.4 25.2 Intermediate Urban 12.9 23.0 Small Urban 13.5 25.2 Rural 28.5 40.3 Source: Molina and others (1993b). 3.73 The PHS and SSS serve very different populations. Most PHS beneficiaries belong to the lower half of the income distribution, with a high participation of the poorest 20 percent, while SSS beneficiaries tend to come from the upper half, particularly the top 40 percent. Thus the belief that the PHS serves mostly the poor and the SSS mostly the better- off, wbile oversimplifying reality, is not erroneous. 3.74 A breakdown of beneficiaries is shown !n Tables 3.14 and 3.15, which include similar results for 1974.21 For all services, tu o different popuWaons are served. More than 40 percent of ISS surgery was performed on individuals in the top income quintie. In contrast, one in three pregnancies covered by the PHS was for women in the bottom quintile. Within the PHS, however, hospit ion and surgery are more often used by the top quiniles, while the poor make relatively more uw of outpatien and matrity services. Use 28 Give at ht hMnfuma reardn CAJANAL was _Mn oly tha peran to 1t1S i presnd. M th rrces o 1974 Womon are fom Sdowsky (1979). m reulls for tSS e ot suctly copaable a ft 1974 data icudes CAJANAL - 74 - of health services is lower in rual areas. The ISS concentrates its facilities in cities, and the rural poor make less use of PHS services than the urban poor. Table 3.14: Distribution of PHS Services (percent) Oulpabent Matenuty Surgery Hospitaion Quind le 1974 1992 1974 1992 1974 1992 1974 1992 1 22.2 28.5 29.5 33.3 26.6 18.0 33.7 29.5 2 21.1 24.9 27.3 26.4 15.6 24.8 21.6 25.9 3 21.1 22.1 19.9 17.1 13.8 15.9 20.2 17.9 4 23.2 16.0 14.8 9.9 22.9 21.8 15.1 14.5 5 12.4 8.4 8.5 13.4 21.1 19.4 9.4 12.2 Total 100 100 100 100 100 100 100 100 Source: Molina and others (1993b) and Selowsky (1979). Table 3.15: Distribution of ISS Services (percent) Q'-tie Outent Materny Swrery H_OSPiaizo 1974 1992 1974 L992 1974 1992 1974 1992 1 5.7 7.9 9.1 9.7 13.6 8.8 9.0 5.6 2 16.7 13.5 12.7 22.6 13.6 23.6 14.8 19.4 3 22.3 22.7 27.3 23.4 36.5 10.0 31.1 20.3 4 28.0 27.8 29.1 30.9 13.6 19.8 28.1 19.0 5 27.3 28.1 21.8 13.4 22.7 37.7 17.0 35.7 Total 100 100 100 100 100 100 100 100 Source: Molina and othes (1993b) and Selowsky (1979). 3.75 A comparison with 1974 data on service use is not very positive. For the PHS, lower-income groups are using outpatient and maternity services more often. The reverse is true for surgery and hospitalization-the lowest quintide of the distribution represents a smaller share of total use today than twenty years ago. ISS use has become more regressive sinice 1974. The top 20 percent of the distribution shows a dmatic increase in the use of surgery and hospitalization services. - 75 - 3.76 Subsidies. The distribution of subsidies for the PHS and ISS in 1992 and 1974 is shown in Table 3.16. PHS and ISS allocate more than half of total subsidies for hospitalization and surgery, services least used by the poor. This pattern confirms the Colombian health system's lack of emphasis on basic and preventive care. Compared with 1974, the PHS has somewhat corrected this orientation, now allocating a smaller share to those two s:rvices. The opposite trend is found for the SSS-the share of subsidies allocated to surgery and hospitalization has increased since 1974. Table 3.16: PHS and SSS Subsidy Distribution (percent) PHS sss_____ Ser*e 1974 1992 1974 1992 Outpatient 12.2 22.1 46.1 25.3 Maternity 0.7 5.7 0.5 6.7 Surgeay 17.3 15.6 17.4 17.8 Hosphalization 69.8 48.7 36.0 41.0 Dentistry n.a. 7.9 n.a. - 9.2 Total 100 100 100 100 Source: Molina and others (1993b) and Selowsky (1979). 3.77 Distibutive Impact of the PHS. PHS spending is distributed progressively. The lowest income quintile receives 27.4 percent of the subsidies and the highest quintile 12.5 percent (Table 3.17). Thus the poorest quintile receives an amount twice as large as the richest quintile. Once differences in family size are considered, the distnbution of subsidies among individuals is remarkably less progressive, with the second quintile receiving larger subsidies than the first. Subsidy distribution today differs only marginally from two decades ago. In spite of a public policy aimed at reducing poverty, no notable progress has been made in targeting puiblic expenditure toward the poor. 3.78 Because the PHS is supposed to serve the 60 percent of Colombians without access to other health services, the distribution of subsidies is not successfully targeted. Almost 30 percent of PHS spending benefits the richest 40 percent of households. This pattern is observed for all services-the top 40 percent receive subsidies from the PHS for sugery, hospilization, outpatient, and othffer services.2' 29 TsM cmposion of the bWis reced by diffecentqles Is not dsgncy dyIfe ftquintdeives a bqar dr f*em hos_lralln and oupae service-. whie te higes qumie does so fom surgedie See Anx 3. Table 3.9. - 76 - Table 3.17: PHS Subsidies by Income Level, 1992 [ is&ibudlon (percent) Subsidy per Household Subsidy per capia Qaindle 1992 1974 (pesos) (pesos) 1 27.4 28.0 41,922 7,720 2 25.6 21.0 39,170 7,961 3 18.7 20.1 28,578 6,055 4 15.9 17.7 24,370 5 S,929 5 12.5 12.2 19,110 5,236 Tota 100.0 100.0 30,627 6,716 Source: Molina and others (1993b) and Selowsky (1979). 3.79 Some regional differences are worth mentioning. The distribution of subsidies in smaller urban areas is the most progressive, and distribution in rural areas the least progressive. Overall, the poor in smaller urban areas receive a higher level of support than their counterparts living in marginal conditions in large cities or in rural areas.30 3.80 Distributive Impact of ISS. The health services provided by ISS involve a subsidy equal to the difference between the cost of services delivered and the contribution toward that cost paid by affiliates. ISS subsidies, it is estimated, amount to approximately one-third of the institution's health budget, or 31,125 pesos per affiliate (US$46) and 19,521 pesos per beneficiary (US$29). When ISS subsidies are averaged among all households (including non-beneficiaries), average subsidy per household is 11,220 pesos-about one-third of the average subsidy implicit in services provided by the PHS, which reaches twice as many individuals as ISS. 3.81 The distribution of health spending by ISS is shown in Table 3.18. The distribution indicates a regressive pattern in which the top quintile receives almost one-third of spending and the first quintile ordy 7 percent. Given the pattern of use shown in Table 3.15, this is not surprising. Table 3.18 also shows the distribution of contributions by affiliates. One-third of ISS affiliates belong to the highest income quintile and, given the use of progressive scales to determine contributions, this group pays more than half of total contributions (the first quintile pays only 1.6 percent of the total). The result is a distributive pattem of subsidies ir which the top quintle subsidizes the second and third ones, which together receive almost 80 percent of the subsidies. On average, the poorest quintile benefits relatively little from ISS subsidies, receiving 1,831 pesos per capita (less than US$3), about one-fifth of what they receive through the PHS. 30 See Arex 3, Table 3.10. - 77 - Table 3.18: Distribution of ISS Subsidies, 1992 Distdbution (percent) Subsidy per Subsidy per Quinlile Spendfag Conhibudons Subsidies Household capita UHt*lO Spending Contributions Subsidies (pesos) (pesos) 1 7.1 1.6 17.7 9,939 1,831 2 18.7 6.8 41.6 23,365 4,749 3 19.7 12.0 34.6 19,425 4,116 4 22.9 23.9 20.8 11,695 2,84S 5 31.6 55.6 (14.7) (8,298) (2,273) Total 100 100 100 11,220 2,460 Source: Molina and ohers (1993b). Main Constraints and Challenges 3.82 Non-communicable diseases among adults and the elderly are increasing in Colombia. High levels of infectious diseases and malnutrition reflect and influence the extent of poverty. This bipolar health characteristic is reflected in mortality and morbidity patterns.31 For children, perinatal conditions and infectious diseases constte the major cause of deaths, while malnutriton continues to affect 13 percent of children under age five. Among teenagers and adults aged 15 to 44, homicides and accidents account for 55 percent of deaths, undersconng the public health implications of Colombia's unusually high level of violence. For those between 45 and 59, non-infectious cbronic diseases are the major contributors to mortality and morbidity. 3.83 Most of Colombia's poor rely on the public health system for services. Sufficient services of appropriate quality are not reaching this group. Effective care through the PHS reaches only 35 percent of dte population, compared with the target of about 60 percent. Overall, PHS expenditure is not sufficiently well-targeted toward the poor-the richest 40 percent of households receive almost one-third of the PHS's expenditures. In addition to insufficient expmnditure, the health services for the poor have low average quality and efficiency, high private costs, and management weakrtesses. 3.84 Quality and Efficiency. Systematic evaluation of health service quality is not available. But with the exception of advanced urban areas, poor average quality is indicated. Services provided directly through the PHS vary among institutions, regions, and municipalites, making it difficult for the average user to rely on primary services as the entry point into the system. The system has emphasized curative over preventive care and 31 For infomtio on regional dlflnes in sodrI n a s Chapt I of is rot - 78 - provision through higher-level institutions over primary care institutions. As a result, primaty care facilities are bypassed by those that can pay for secondary and tertiary level facilities.32 Primary services are - lso mainly clinic-based, with limited outreach to the community. 3.85 The PHS's quality suffers from an incomplete set of prinary services. A standard package of basic services that includes preventive measures, health education and promotion, and curative and emergency services, has not yet been provided at the local level. The ability to respond to the population's needs and emergencies is furher hindered by inadequate provision of supporting inputs. Most facilities cannot adhere to standard diagnostic and treatment protocols because they lack drugs, supplies, and equipment. 3.86 Shortfalls in centrally managed PHS programs for infectious and vector-borne disease also detract from effective preventive care. Vector-borne diseases are prevalent in urban slums and coastal and river valley areas. Malaria and dengue fever control programs have become less effective in recent years. Immunization campaigns are effective but not yet universal. 3.87 Staff weaknesses further contribute to low productivity and coverage, especially at the primary level. Professional training institutions (for medicine and nursing) emphasize curative care at higher levels. Training does not focus on prevention and primary health care or on meeting the needs of the poor. Rural areas especially suffer from inadequate numbers of health care professionals and high staff turnover in primary care facilities, where conditions are the least attractive. Moreover, 40 to 50 percent of staff at primary health facilites are recent graduates completing a mandatory year of social service. Use of recent graduates is associated with low efficiency, high levels of absenteeism, and inadequate supervision of lower-level health workers. 3.88 Private Costs. Quality weaknesses in primary health cae, combined with fees and other out-of-pocket costs at higher levels, effectively limit access of the poor. Charges are detennined at the regional level within Ministry of Health guidelines. Fees are prorated by income. While collection is not systematic, data suggests that cost recovery at first-level hospitals ranges from 0 to 50. Institutions in poor communities often lack funds for medical supplies and drugs, which must then be supplied by patients. It is estimated that households in the bottom quintile of income distribution spend 7.5 percent of their income on health.33 Medications represent about half of their expenmes. 3.89 The poor tend to use less costly services. The average private or own cost of hospitalization (for drugs, supplies, and fees, pro-rated by income level) is about 30 percent of annual income at the lowest income levels. Accordingly, non-use of services increases at 32 In 1990. prmay levl hospials rpoed an aveage occupay me of only 35 percent. compared Ith 59 pee rnd 7 peent at the secoy and teniMy heah leves, especive. See Ministy of Health (1991). 33 MThs Includes thos households that do not have access to medIcal services. Thus. those tat do have acces spend a higher toan average propordon of their incom on heath. -79 - lower income levels and in rural areas, reaching 40 percent in the first quintile. Fifty percent of sick persons do not seek health services because of cost. This percentage is higher among low-income families. Deficiencies in supply of health services account for only 15 percent of non-use. 3.90 Management. PHS management weaknesses contribute to quality and coverage problems. Insufficient technical and supervisory support from the Ministry of Health is notable. In the past the Ministry of Health focused on direct management of centrally-provided services. Other functions, including policy management, sectoral guidance, operating standards, and staff training and support, are not developed. Effectve planning, targeting, and delivery of PHS services is incomplete. Effective monitoring and evaluation of services is lacking. 3.91 Dk. 'tralization. Devolution of authority to municipalities is essential for Colombia's health ategy. A recent World Bank study34 analyzed several municipalities dealing with new health sector responsibilities and id_tndfied four constraints limiting decentralization. First are political constaints-such as patronage-based hiring and nvestment without regard to cost-that contribute to low efficiency. Second, financial constraints affect municipalities. Decentralization costs make many authorities unwilling to assume the financing for local services. Third, administrative constraints-weak systems and inadequate civil service-limit efficient reaction to the challenges of decentralization. Finally, there are informatonal constraints. Officials often do not understand decentralization procedures, and technical assistance from the Ministry of Health often focuses on legal requirements rather than on improving service capacities. Conclusions and Recommendations 3.92 Over the past 20 years the Colombian government has implemented public health activities designed to strengthen services for the poor. While particular acdvities-a food stamp program in the early 1970s, a program to integrate primary, secondary, and tertiary care in the PHS in the mid-1980s-have made progress, none have been institutionalized. Inadequate expenditure has been a constraint for the PHS system. But weaknesses in technical and administrative support and inadequate monitoring and evaluation are more enduring problems. 3.93 Recent efforts, beginning with the Barco administration and extending to hie Gaviria governent, address the sector's problems more comprehensively. Low overall health service expendimres for the poor are addressed by increased resources through the Sitado Fiscal and Transferencias Municipales programs. There are also plans to increase expenditure for the universal provision of basic services. While the poor will not have the same bealth benefits as the better-off (since private expenditure on health insurance or prepaid medical plans buy expanded services), the provision of a standard package of 34 Wodd Bank (1993a). - 80- preventive, clinical, and emergency services would be a major advance toward health care equity. The affordability of such a package has not yet been determined. 3.94 The universal care program is incomplete and promises to be difficult to institutionalize. The Ministry of Health has been unable to guide sector development, and the difficulties of decentralization are just emerging. Most local governments and communities lack planning and management experience. In addition, the issue of improving public health institutions in a decentralized context is not resolved. Continuity and flexibility in pursuing universal basic coverage will be required by future governments. Attention to monitoring and evaluation will also be needed. 3.95 Defining and Refining the Central Government's Role. Decentralization has reduced the central govenmment's role in health care. This should change. The central govermment must play an active role in guiding the system and in setting standards for quality and efficiency control. Certain responsibilities should be retained by the Ministry of Health, such as vertical programs for immunization-including expanded immunization-and endemic disease control. Information and education campaigns have been identified as important continuing responsibilities of the Ministry of Health. 3.96 Inefficiencies in the move toward integration and improved competitiveness among health providers requires the central government set standards and prices for isrn and services. This will require reorientation of the Ministry of Health. The current Minister has reorganized the Ministry, replacing about one-third of staff with a smaller cadre of qualified, well-remtunerted staff. But the reorganization has focused largely on planning and support for decentralization. Emphasis on additional responsibilities will be required. 3.97 Building Local Management Capacity. Regions and municipalities will be responsible for management of higher-level and primary health services, respectively. During the transition period, rsource transfers earmarked for health must be effectively utilized. Financing formulas are designed to promote equitable resource distribution among regions, but local incentives do not ensure a continued focus on equity. The current system of national fimncing has contributed to overinvestment in local health facilities, where recurrent costs may not be sustainable. Cofrancing incentives-through the new social investment fund, FIS-that encourage equitable service provision are promising. The effectiveness of this system remains to be seen. Local demand is critical for better quality, and officials should be accountable for services provided. The Ministry of Health and its regional offices should help build demand for basic services. Experience suggests that public and private interest groups promote accountability and help programs continue from one political generation to the other. 3.98 Developing a Rural Health Model. The rural poor are the least well-served of the PHS client groups. While the urban poor are also inadequately served, improvements in coverage and quality of clinic-based primary care services will have a positive impact. But the clinic-based care system is poorly suited to the rural target group. A model involving - 81 - outreach services should be considered. An attractive career package for rural-oriented public service also needs to be developed. 3.99 Strengthening Monitoring and Evaluation. Attempts to establish information and monitoring systems for public health services have not succeeded. Effective planning and delivery of priaary services in large urban centers should serve as a model. The Ministry of Health will need to build on this experience to support municipal, departmental, and national health information systems. Periodic surveys by the National Statistics Department (DANE) provide information on the impact of public health services, but sustained monitoring and evaluation has not been introduced. This information base is required to support the provision of appropriate services to the targeted population. Plans for extending health care coverage require the tracking of progress and problems. Chapter 4 Providing Energy, Water, Sewerage, and Shelter to the Poor The strategy of the Colombian government to provide energy, water, and sewerage to the poor has been anchored in the belief that by maintaining low corresponding prices, poor households would have access to these basic services. But the resulting massive subsidies have not been well- targeted. Therefore the poor population has not benefitedfrom them. The mainfactors explaining this outcome are: (a) regressive service coverage; (b) poorly designed tariff structures; and (c) an ineffective stratification scheme. The strategy has resulted in sector-wide financial crisis and inefficient resource allocation. To ensure future efforts to provide basic services to the poor are successful, sector-wide reforms should be implemented. For energy, water, and sewerage, stratification should be simplified or eliminated to allow better targeting of subsidies to the poor. Alternative modes of provision (such as isolated generating units for energy, or tanker trucks for water) should be used for residents in inaccessible or very poor areas, whether peripheral settlements near urban centers or dispersed rural zones. Housing policy should meet the needs of residents in inaccessible nonseruiced areas without creating incentives for others to settle in even more difficult terrain. This framework should provide shelter to the poor and, more important, improve their access to basic services. The direct subsidy approach the government has recently adopted for home purchase and improvement, land titling, and accessing sanitation services is a step in the right direction. The government should consider directing more of the program's resources to properly maintain and upgrade existing houses and neighborhoods rather than financing new units. l. Introduction 4.1 Basic services in Colombia have improved substanially in the last two decades. Coverage has been expanded in electricity, water, and sewerage-with major advances m electncity and slower ones in sewerage. The govermnent has devoted significant resources to provide these services. In 1992 public subsidies to urban households amounted to 615 billion pesos, which represents 1.8 percent of GDP, 7.5 percent of total public expendiures, and almost one-third of total subsidies in social government expenditures.1 Subsidies for basic services, however, have not been well-targeted. The richest 30 percent of the households have received 29 percent of the subsidies. The use of price controls hindered economic efficiency. For example, enrg use in Colombia is uneconomical. Electricity is I Abot So b;Ho pe5w in sublics, or 8 pescea of do tW amount nomloned in the w ta is exaied by ila of awi access to basic icms . - 83 - used intensively in households to generate heat (cooking and water heating). Lower-cost fuels are more appropriate and provide similar service. For water and sewerage, a vicious cycle of low cost recovery, low investment, poor service, and poor maintenance has led to a chronic crisis in most water entities. Ultimately, the strategies pursued have misallocated resources without any clear benefit to the poor. 4.2 This chapter evaluates the government's strategy to provide basic services to the poor in energy, water, and sewerage.2 A sector overview identifies the main problems in delivering services to low-income households. Then the chapter examines government expenditure in these sectors, considering the extent to which poor households benefit from public subsidies. The presentation concludes with recommendations to improve the delivery of basic services to the poor. 4.3 The final section reviews Colombia's new housing policy. In most countries, housing programs are supply-driven operations that rarely reach the poor. The program targets are based on subjective estimations of poor households' perceived needs, without consideration of willingness or ability to pay. The Government of Colombia has recently moved to a demand-driven approach in which direct subsidies are provided to the poor for new housing, housing improvements and expansions, land titling, and sanitary facilities. The new approach focuses resources on the poor's shelter problem and on integrating unattended marginal settlements into municipalities and avoiding further proliferation of unauthorized settlements-with corresponding savings in unplanned urban developments. i. Energy Sector Overview3 4.4 For low-income households, energy is a major unavoidable cost of daily life. It can take the form of direct energy purchases or indirect costs, such as fuelwood gathering or queuing for scarce petroleum products. In major cities, the lowest income decile spends up to 11 percent of its income on energy. The average direct energy expenditure in the general population is 2 to 3 percent. Certain forms of energy also represent significant health hazards to low-income households. Carbon dioxide inhalation is a risk with fuelwood stoves, as is fire when using cooking gasoline. Wider energy availability in terms of quantity and variety can significantly increase welfare in low-income households. Increased productivity results when less time is spent on chores such as woodcutting, or when leisure and work time are 2 Most of die infonmaton in this chapter is obtained from tie bawground paper aborat for the report Econometra Ltda. (1993). For die enegy sector, a study prepared by ESMAP (1993) is also used as a basic reference. 3 A comprehensive analysis of household energy is pardary complex due to dte varety of fuels, applances. and habits that inluence fue choice. 7he analysis in ts repot will therefore be limited to modem fuels for domestic use, namely electricity. rnAual gas, and oil products. Without igrng the importance of traditiona fuels such as biomass, carcoal, or miea coal realistic policy choices at die sector level are limied to the modem faus whose impact is greatest in terms of increased productivity and energy bil reduction. - 84- extended by the introduction of modem energy forms such as kerosene, liquid petroleum gas (LPG), or electricity. 4.5 Pridng of modern energy products is controlled by the government. Until 1993, electricity tariffs were set by a Tariff Board. They are now the responsibility of aka Energy Regulatory Commission. Hydrocarbon prices are set by the Ministry of Mines and Energy. The market sets prices for traditional fuels such as fuelwood, charcoal, or mineral coal; these products are mostly traded in rural areas. 4.6 Giving lower-income households access to energy products has long been a desired policy objective in Colombia. Regulated fuels, particularly electricity, have been viewed as a means to redistribute income. Reasoning that consumption increases with income, residential electricity has been sold based on increasing price blocks. In the early 1980s a typical tariff schedule covered ten blocks, providing users no infonnation about the actual cost of electricity. Around 1983, a discriminatory stratification scheme was added. Dwellings were assigned a stratum according to location, building characteristics, condition, and access to public services. Six strata were defined. One to two is the lowest dwelling quality, five to six the upper-level. The price was set according to stratum. Users had to pay a fixed charge and an energy charge; both varied by stratum. Within each stratum, the energy charge was structured in four blocks with increasing prices.4 For mosi residences prices are below costs. 4.7 To show the magnitude of subsidies in the electricity tariffs, Table 4.1 presents the tariffs for different subsectors as a percentage of long-run marginal costs (according to the new 1993-97 targets). Even though the new targets reduce the subsidy, they do not eliminate it. The information in the table identifies three important features of current pricing policies: (a) subsidies in the current schedule reach the relatively well-to-do (strata ree and four) until 1996; (b) before 1996, cross-subsidies are practically nil, most residential subsidies are generated either by other subsectors (such as commercial or industrial users) or require external (government) financing; and (c) industrial and commercial tariffs are consistely above cost. 4.8 Natural gas tariffs exhibit a similar patern, as seen in Table 4.2. As with electricity, natural gas tariffs increase with consumption and provide below-cost prices which, in some cases (Barranquilla), are enjoyed by the higher stata. 4.9 Prices for fuels such as LPG and kerosene are also set by the government. As administered by the Ministry of Mines and Energy, the price fixing system keeps energy prices well below opportunity costs. This policy inhibits production, particularly of fuels such as liquid petroleum gas (LPG) that displace gasoline production in the refinery. In turn, 4 This gaiftiou sdiew hs oow bn appied to aodir public serviens I W1 S swply, telephon serc s Mo IC*.y, namtu gs. - 85 - Table 4.1: Subsidies In Electricity Tariffs as a Percentage of Cost December December D Dcember mber ber December Stratum 1993 1994 j 1995 1996 1997 Residential One 38 42 45 48 50 Two 41 45 49 56 70 Three 48 54 68 100 100 Four 62 76 99 100 100 Five 113 115 115 115 115 Six 135 135 135 135 120 Non-Residential Industrial [ 123 [ 122 [ 121 [ 121 120 Commercial 128 125 122 120 120 Other 110 103 102 102 102 Souce: Junta Nacional de Tarifas. Table 4.2: Residential Natural Gas Tariffs as a Percentage of Cost, December 1991 Bogot Banmquifl Consumpton Range Consumption Range Stratum 1 2 3 1 2 3 26 36 97 44 55 70 2 56 63 98 64 66 73 3 76 79 100 103 85 79 4 104 99 104 128 98 82 5 157 130 112 153 112 86 6 252 1 182 1 129 178 123 89 Source: National Energy Commission, World Bank, and ESMAP (1992). disruptions in LPG supply encourage the use of less economical substitutes, such as electricity. Prices remain below cost despite recent government correction. 4.10 The ill-conceived policy of artificially depressing household energy prices has had unfortunate consequences. The first was a shortage of LPG and kerosene that lasted until at least the late 1980s, manifested by queues and extended waiting periods, during which - 86 - households used electricity as a substitute. With persistent shortages, consumers switched to electricity permanently. Electricity tariffs were also kept below cost, but producers could not restrict their output without major supply disruptions. This contributed to soaring residential electricity consumption, indiscriminate subsidization, and a massive, sector-wide financial crisis in the power industry. Without a change in the overall approach, electricity tariff policy tried to mitigate the problem through discriminatory pricing, with high rates for industrial and commercial customers, together with stratification, which sought to charge residential users according to income level. Electricity pricing has been unsuccessful. Massive subsidization persists and the stratification criteria have failed to target the poor (see following section on incidence analysis). The overall energy strategy paradoxically has made a country well-endowed with energy resources endure inefficient and costly domestic energy supplies. 4.11 Household Energy Use. LPG, natural gas, gasoline, and kerosene are mainly used for cooking. Electricity has multiple uses, including cooking. Table 4.3 shows how much energy is used by poor households to prepare food.5 The data show substantial variations among cities in the percentage of energy used to prepare food. These differences can be partly explained by end-use appliance efficiency (which would justify the high proportion of energy used in Bogota) or by different diets. In any case, at least half of all energy used by low-income households is spent in cooking. Given that other energy needs such as lighting and refrigeration depend exclusively on electricity, providing energy to the poor requires solution of two problems: (a) the supply of electricity for basic welfare needs; and (b) the availability of efficient alternatives for cooking.6 4.12 Programs for the Poor. In addition to sector policy actions, providing energy access for poor households has been pursued through more focused community programs. The best known program is the distribution of cooking gasoline-Cocinol-in BogotA. This program was initiated in the early 1970s. The product is delivered to private distributors who send a tanker fortnightly to local delivery points controlled by Community Councils. At these locations registered users can purchase up to five gallons of fuel. Long queues are common on delivery days. Because of Cocinol's extremely low price (curently about US15 per gallon), rationing is necessary to prevent its diversion to automotive uses. But even under optimistic conditions much of this gasoline is siphoned away by agents who intervene in distribution. 5 The infonrmtion is obtained from diffent surveys that wee integted as part of an energy efficiency study conducted with Wodd ank/ESMAP support. 6 Tabls 4.3 is based on suvey data from major cities-miller ton and dhe mad popaton are ignored. In dee cases, however, dte basic concluson probably rermans die same. Cooking needs stdll repst the mUor energy cost, pariularly smong the lowesa income segmen where fudwood gatherin time could be spent on more productive activities. - 87 - Table 4.3: Monthly Energy Used for Cooking per Lower-Income Household Electaiciy Hydrocarbons Totl Percentage of Region (kWh) (kBTU) (kBTU) (kTDM) Total Bogota 55 183 780 963 64 Medelifn 170 5, 70 3 573 61 Cali 120 400 43 443 _ 48 Barranquilta 25 83 420 503 56 Source: National Energy Commission, World Bank and ESMAP (1992). 4.13 Despite its low cost, Cocinol was considered dangerous by the public because of user-related fires involving children. This led the Ministry of Mines and Energy to launch a "gas massification" -ogram in 1986.1 ECOPETROL was ordered to increase LPG production and a natural gas pipeline was built for Bogota. The natural gas distribution network was initially built to serve lower-income sections of the city and is gradually being extended. At the same time, an LPG distri- bution program was initiated through COLGAS, the major LPG distributor. C. x 4.1: COLGAS traded registered Cocinol users.. Cocinol stoves for LPG stoves and cylin- Cecilia works as a cl+eani wom i a.i ders and has supplied LPG at the official residential building. She eah thei wage v price. This substitution program has been suppiemented by overtme income (about US$1251 month). She has electricity~ soerie cld nth well-accepted by users (see Box 4.1). t, but her landlord does not allow her. to coo with it. She used to own a.Codnol stove and a. 4.14 A neglected aspect of low- low-efficiency.:LPG stove asbackup. Her LPO income household energy provision relates cylinder used to last less than a month. She to smaller towns where there is no LPG exchanged her C1clnol stove for the new- LPG. distribution. In some areas, such as stove and cylinder provided as par of the LP - Cundinamarca, the electricity uility community CoIunci. Sh reprt that- a0 oun (CELGAC) runs an LPG distribution clinder (US$5.50) now lasts up to three mouths. business. Most communities outside major It also provides a good flame thiat doesntI cities, however, lack LPG supplies. In blacken kitchen uiensils. Her slightyi sed general, government energy policy has cooking expeses have been compensated by:; focused on the urban poor. Natural gas is a better quality and safer energy. good example. It should reach the major cities of Medellin and Cali in 1995, while programs for the rural poor are practically nonexistent. 7 In some areas, such as Boyac and Cw rnaas , Codno has been almost totaly replced. In 3ogoL, more dian 70 percet of those consming Coinol at the begwning of 1993 had swtdW to either raunal or propane gs by the end of the year. - 88 - Incidence Analysis: Who Benefits from Public Subsidies?8 4.15 Coverage. As seen in Table 4.4, Colombia has made impressive gains during the last two decades in electricity coverage. Between 1974 and 1992 the number of households receiving electricity increased from 63 to 92 percent, to a total of 6.6 million households. This achievement required new access for 4.3 million households, double the number of users in 1974. New users increased from I percent per year during 1970-74 to 6 percent per year for 1974-92. Added coverage during 1974-92 was distributed more or less evenly among income groups, with a Gini coefficient of -0.03. This partially corrected the regressive coverage in 1974, but could not reverse it. In 1992 coverage of the lowest quintle was 81 percent, compared with 98 percent of the highest quintile. 4.16 National figures on electricity coverage mask important regional differences, and the fact that some connections to the system are illegal. In urban areas, 98 to 99 percent of households are covered, compared with 81 percent in rural areas. Illegal connections account for about 5 percent of total coverage and vary substanially among regions and income groups. In Barranquilla and Pasto, for example, illegal connections account for 8.3 and 12.6 percent of total coverage, respectively. As expected, lower-income households more frequently attain their services illegally. In Barranquilla, for example, 23.8 percent of connections of the lowest decile are illegal. Whereas the Gini coefficient for legal national coverage is 0.045 (slightly regressive), for illegal coverage it is -0.15 (progressive). 4.17 Amount of ]Public Subsidies and Distribution. In 1992 the total household electricity subsidy amounted to 469 billion pesos (1.4 percent of GDP), 8 percent of which was due to illegal connections to the network. The total amount represents 90 percent of the subsidies for all basic services. Approxinately 20 percent of this subsidy is covered through higher prices imposed on industry and commerce. As shown in Table 4.5, however, the distibution of the subsidy has been slighly regressive, with a Gini coefficient of 0.07.9 The subsidy per decile increases steadily up to the eighth decile. Households in the lowest 20 percent of the distribution receive a smaller subsidy than the richest 20 percent of the households. It is not the poor who benefit the most from the electricity subsidy. 8 The analysis in this secton focuses mainly on the elcticity ector. Natur gas, for which nformation is available (Ecormeta (1993)), has only recenty started to be distdbuted lo the Colombian househod Covrage only rea 8 percet of houseds and the amount of the subsidy rempesens 1.5 perent of tie total amon of subsidies for basic services. 9 Given t poor households have a lager numbet of menm, die disrbuion of the subsidy is mote regressive if cad on a per capha basis - 89 - Table 4.4: Coverage of Electricity by Income Group, 1970, 1974, and 1992 Simulated 98 percent 1970 1974 1974-92 Change 1992 Homogeneous Coverage Distribution Households Disnibuton of New to be of Required Connections Coverage Connecdons Coverage Connectons Connections Connections Coverage Connected Connections uIncome Quntie (millions) (perent) (millions) (perent) (millons) (percent) (millions) (percent) (millions) (percent) 1 0.268 37.7 0.298 41.4 0.869 20.2 1.167 81.3 0.240 53.8 2 0.327 46.0 0.353 49.1 0.943 21.9 1.297 90.4 0.110 24.7 3 0.414 58.2 0.444 61.7 0.897 20.8 1.340 93.4 0.067 14.9 4 0.511 71.9 0.529 73.5 0.849 19.7 1.378 96.0 0.029 6.6 5 0.639 89.9 0.657 91.3 0.750 17.4 1.407 98.0 0.000 0.0 Total 2.158 60.5 2.280 63.2 4.308 100.0 6.588 91.8 0.447 100.0 Gini Coefficient 0.171 0.156 -0.030 0.034 -0.502 Annual Growth Rate 1970-74: 1.1 percent 1974-92: 6 percent 1992-Target: 6.3 percent Source: Velez (1994). - 90 - Table 4.5: Distribution of Electricity Subsidies Total Subsidies Distbutfion Subsidies per household (millions of t992 pesos) of Subsidies (percent) (thousands of 1992 pesos) Iech Lega llega Total Legal illg Total Legal lltegal Total 29,666 3,057 32,724 6.86 8.35 6 98 41.3 4.3 45.6 35,903 4,279 40,182 8.30 11.68 8.57 50. 1 6.0 56.0 38,346 3, , :3 42,059 8.87 10.14 8.97 53.5 5.2 58.6 40,381 4,299 44,680 9.34 11.74 9.53 56.3 6.0 62.3 43,769 3,967 47,736 10.12 10.83 10.18 61.0 5.5 66.5 47,063 3,903 50,966 10.88 10.66 10.87 65.5 5.4 71.0 48,711 3,992 52,703 11.27 10.90 11.24 67.9 5.6 73.4 51,332 4,177 55,509 11.87 11.40 11.84 71.6 5.8 77.4 52,855 2,542 55,397 12.22 6.94 11.81 73.6 3.5 77.2 0 44,353 2,698 47,050 10.26 7.37 10.03 61.8 3.8 65.6 'otal 432,378 36,627 469,005 100.00 100.00 100.00 60.3 S.1 65.4 Hni Coefficien, 0.080 -0.038 0.070 Dw: Velez (1994). - 91 - 4.18 The slightly regressive distribution of the subsidies results from several factors. The first is the previously mentioned regressive coverage of services. When poor households are not connected to the services, other mechanisms cannot compensate. Even if coverage was evenly spread among income groups, two other factors lend to regressive distribution of subsidies: stratified tariffs and stratified households. 4.19 The urban electricity tariff structure has directed substantial subsidies to the first four strata-between 6,100 and 9,400 pesos per month for legally connected households (Figure 4.1). The fifth strata has received a reduced subsidy. Only the sixth strata has paid a low tax. This implies that even under perfect stratification, middle-income households would have received substantial subsidies. Compare the design of the gas tariff structure (Figure 4.2). In this case only the first two strata receive subsidies. The four higher strata pay increasing taxes. With appropriate household stratification, the tariff structure could create the desired progressivity in the distribution of subsidies. 4.20 As Table 4.6 indicates, the household stratification used in Colombia in 1992 is inadequate. The correlation coefficient between distribution by income and by strata is less than half that under an "ideal" stratification scheme. For 58 percent of urban households, there is a Type II classification error-households are being classified in a lower strata than would correspond according to income. Only 20 percent of the households in the first strata have income levels that put them in the lowest two deciles of income distribution. Successful correction for Type II errors in the stratification would double the correlation coefficient. On the other hand, only 12 percent of households are subject to Type I classification errors-households classified above their corresponding strata. The analysis is quite Fgure 4.1: Monthly Electricity Charges, Costs, and Subsidies per Legal Urban Household, by Strata, 1992 (1992 pesos) 35000 30000 25000 20000 15000 10000 r ~~~~~~~SubsidieBs 5000 0 1 2 3 4 5 6 -a--- Charges - Cost Source: Velez (1994). - 92 - Figure 4.2: Monthly Gas Charges, Costs, and Subsidies per Legal Urban Household, by Strata, 992 (1992 pesos) 5000 4500 \/ Taxes 34000 3500C Subsidies 3000 2500 1 2 3 Strata 4 5 6 - Charges Costs Source: Velez (1994). Table 4.6: Household Stratificato and lIcome Distribution, 1992k Sbif cation EFwS (pwent) Comen fdoa Coe,ffts No_TypeI No 2peH lde4 No Areas and Ciies l)pe I T)pe H 7)pe I+H Reat Enr r err Totd Urban 12 58 70 0.316 0.578 0.867 0.965 Traditiona Urban 10 60 70 0.443 0.632 0.891 0.964 Four Major Cities 9 62 71 0.452 0.628 0.896 0.965 Other Cities 14 S0 64 0.388 0.653 0.863 0.959 Rest Urban 17 55 72 0.037 0.439 0.630 0.966 BogotA 9 64 73 0.432 0.624 0.901 0.966 Medellin 11 57 67 0.451 0.644 0.879 0.961 Cali 5 71 77 0.457 0.584 0.905 0.968 Barranquilla 12 47 59 0.538 0.681 0.898 0.964 Bucaramanga 9 57 66 0.413 0.618 0.868 0.948 Manizales 16 46 62 0.358 0.663 0.861 0.963 Pamtt 23 4 63 0.336 0.737 0.876 0.969 Caragena 15 48 62 0.417 0.660 0.862 0.966 a. See Anmex 4, Table 4.3 for reference. Soure: Velez (1994). - 93 - conclusive: the stratification scheme used in Colombia has diverted subsidies to middle- and high-income households. The government is currently revising the stratification system to address these problems. Conclusions and Recommendaions 4.21 Concerted sector-specific actions are required to address the inefficient use of energy resources and the financial disruption created by uneconomical pricing policies, and to ensure that public subsidies target the poor within financially manageable levels. A reform program should be guided by the following principles: * Encourage fuel supply by pricing different energy sources at a level that allows cost recovery in monopoly cases that require price regulation, and by allowing prices to respond to market forces. * Target subsidies to lower-income households through rebates on fixed nvestment charges (connection costs) in electricity and natural gas, or through rebates on LPG cylinders and basic appliances such as two-burner stoves. * Limit complementary subsidies on the product to lifeline rates for electricity and natral gas with an overall ceiling that does not disrupt suppliers' finances. * Promote end-use efficiency through awareness campaigns. * Address rural energy problems by encouraging the supply of household fuels-such as LPG and kerosene-through pricing policies. Subsidies can then be targeted by providing high-efficiency LPG stoves and pressunzed kerosene appliances to low-income households. 4.22 The main elements of the approach involve substituting inefficient electricity uses with natural gas, LPG, and kerosene use through appropriate pricing policies and targeting public subsidies to the poor. 4.23 ElecticIty Options. With 92 percent coverage, electricity coverage should be examined in light of the costs of reaching more of the population. Isolated rural households make up the majority of those lacking electricity. The cost of extending networks to serve them is high. Alternative solutions would be more cost-effective, such as isolated generators providing basic lighting and communications. Small communities can be served by installing generatig equipment such as diesel motors (see paragraph 4.29). These solutions are best fostered through rural energy promotion. 4.24 Special attention needs to be given to tariff reform. Two policy alternaives are available: - 94 - * Maintain stratification-while incorporating revisions to address problems (see paragraph 4.20)-by gradually readjusting all tariffs to economic costs in the kWh charge and allowing a certain degree of price discrimination through fixed charges. This would eliminate the major distortions of the stratification scheme. * Eliminate stratification altogether, establish a general residenial tariff, and complement it with targeted social tariffs for low-income users (the latter tariffs could be fixed charges with capacity limiters). 4.25 Revising the existing stratification while gradually eliminating kWh charge differences would encourage substitution for more efficient energy uses. The discri.:;natory fixed charge would provide some cross-subsidization. This gradual approach should be complemented by alternative substitutes to avoid unacceptable increases in consumers' total energy bill. It will be more successful in cities (such as Medellin) where no electricity alternatives have been developed. This approach unfortunately requires persistent tariff reform, which can be threatened by short-term political pressures. 4.26 The second option would entail simplification of the tariff structure by creating a general tariff for most consumers and complementing it with an additional rate targeting low- income users. The general residential tariff would have a low initial priced block, to allow low-income users who cannot take advantage of the complementary rate to reduce their energy bill by cuting back on their consumption. The complementay rate would target low-income, low-load dwellings where metenng is uneconomical. The rate would be a flat charge with a capacity limiter. 4.27 A lifeline rate with capacity limiter is a good option for low-income users in single family dwellings, particularly in smaller towns. The option is not satisfactory in large cities where many low-income users share a common meter in multiple household dwellings. This increases consumption far beyond the basic rate each household would be entitled to. This points out two pitfalls of block pricing. The basic rate of the first block may not reach the very poor, and utlities should avoid rising multiple block tariffs (sometimes with penalties for excess consumption), since boarding houses located in urban ghettos with numerous households end up paying the highest rates and being denied the use of appliances by the landlord. To address the problem, boarding houses could be allowed to register with the utility and a boarding house rate could be devised with a larger basic block. 4.28 The second option is more desirable for the functioning of the sector and for poverty reduction. But policy makers in Colombia have decided to follow the first one. In eitber case, the successful implementation of tariff reform is essential. If privatization of distribution enterprises is to be taken seriously, the only alternative is reducing subsidies to manageable values-a level where subsidization becomes a minor matter for electricity. To achieve this tariffs should be adjusted to economic levels for the vast majority of consumers. The poor could be helped tbrough reasonable cross-subsidies that contain the problem within the concerned enterprise. The current tariff policy, shown in Table 4.1, is moving in the right - 95 - direction.'0 It attempts to reach its goals by the end of 1997. The government should consider speeding up this timetable. 4.29 Hydrocarbon Options. Economical alternatives for household energy are the only feasible option for reforming electricity pricing without provoking major consumer opposition. Natural gas, LPG, and kerosene are the alternative fuels to be promoted. Hydrocarbons should also play a key role in providing energy to the rural poor. Given that transportation costs are the major obstacle in reaching this target population, LPG and kerosene are the realistic sources of modem energy. LPG is the more efficient alternative for small towns given the costs of transporting, filling, and exchanging cylinders. In more scattered rural populations kerosene is the best alternative, since it is less volatile than gasoline. Overall, the strategy calls for promoting: (a) supply of economical alternatives to avoid shortages; and (b) demand for economical substitutes by making end-use equipment available and protecting consumers in areas prone to market failure. 4.30 LPG and kerosene supply-side promotion. On the supply side, disruptions in the provision of LPG and kerosene can be avoided by allowing opportunity costs to prevail at the wholesale level. This would ensure that ECOPETROL has an incentive to produce these fuels and allow private entrepreneurs to import the products direcdy. Wholesale LPG and kerosene subsidization should be discouraged on two accounts: (a) if the subsidies are to be absorbed by the producer, the burden would automatically shift to ECOPETROL, since pnvate importers would not penetrate this market; and (b) there would be iiscriminate subsidization of all income classes, particularly in the case of LPG. 4.31 For large cities, the distribution of LPG and kerosene fuels is a competitive industry that can easily sustain multiple firms. Supplying smaller towns implies higher transportation costs and monopolistic behavior, particularly for LPG. The small communities subject to these limitations would most likely become predominantly kerosene consumers, for which a competitive market could be sustained. Consequently, consumer LPG and kerosene prices should be allowed to respond to market forces. One option to promote these fuels while keeping prices low is tax exemption. 4.32 Demand-side LPG and Kerosene Promotion. The market can accomplish the switch from electricity to alternative fuels. The rate of change could be improved by complementary measures to reduce market failures in appliance efficiency and LPG cylinder fill controls. 4.33 Enhanced appliance efficiency and the use of more practical equipment can be achieved. To increase stove efficiency, manufacturers could be provided with design guidelines by INEA (Insutiao Nacional de Energias Alternas). The goal would be to improve LPG bumer design and adapt existing Cocinol stove designs to pressurized kerosene. To assure consumers choose correct equipment, the program could incorporate appliance 10 See Junta Nacimal de Tarifas (1993). - 96 - labeling to ensure minimum efficiency standards. An awareness campaign could then be launched to inform consumers of the meaning of efficiency labels, the benefits of economic substitutes, and energy-saving habits. 4.34 LPG cylinder fill concerns many consumers. It is a consequence of market failure-there is no practical way for the purchaser to ascertain that the cylinder has been filled to capacity. The authorities should consider imposing controls and sanctions, through licensing LPG distributors and periodically inspecting for underfilling. 4.35 Demand-side management could incorporate significant subsidies for low- income users. LPG and kerosene stoves can be sold at subsidized prices and distributed through Community Councils (as in the Cocinol program in Bogota). In rural areas, subsidies can be targeted by providing high-efficiency LPG stoves and pressunzed kerosene applianes at a nominal fee. Quality control of cylinder fill and cylinder maintenance would also complement subsidies. 4.36 Natural Gas. As a network-delivered commodity, natural gas mimics electricity, ,with prices indiscriminately subsidized because of stratification. In Bogota, natural gas was introduced as a substitute for Cocinol with a very low tariff and was limited to very low-income neighborhoods. Monopolistic behavior by the Bogota electric utility-a shareholder of the gas company-pardy explains this strategy. Higher strata customers yield most of the company's residential income. To avoid losing these consumers, the company purposely delayed the extension of the natural gas urban pipeline to high-income sectors of the city. This collusion between monopolies should be addressed by forbidding cross- shareholding among energy concessionaires and promoting the use of LPG by liftng price controls and alowing free entry to this industry. 4.37 As a natural monopoly, natural gas requires regulation even in the presence of competing fuels such as LPG. In residential markets, natural gas distributors should be allowed a regulated price margin with incentives to extend the network. The pitfall to avoid is indiscriminate subsidization that would limit the distributor's incentive to serve profitable industrial and commercial markets, and perpetuate the inefficient use of energy. 4.38 To reach low-income households, network extensions could be subsidized in poor neighborhoods through low hook-up charges and the subsidization of efficient two- burner stoves. The Community Council infrastrucure could manage such a scheme. Such a program is unlikely to reach the poorest, such as those living in unauthorized settlements, where haphazard building imposes uneconomic costs on the supplier."1 It is more likely to be successfil in low-income but planned developments, where the typical family has two or more wage earners. ii SOppig the proliferation of unauthorized settlements, therfore, does not only solve a housing and environmen problem but prvides the basis for more efficient use of resoures in the provision of services. See Secdon IV on housig poltcy framewok - 97 - HI. Water and Sewerage Sector Overview 4.39 Main Sector Problems. Unlike power and natural gas, which have national interconnected networks with integrated systems and a centralized agency responsible for the industry, water and sewerage in Colombia is comprised of 1,028 separate companies ranging from 1,000 to 900,000 households served. These companies have diverse ownership and levels of administrative and operational efficiency. Many have additional funtions (services covering solid waste, markets, slaughterhouses, and so on). The main problems in the sector are well-known: institutional weakness and instability; low and inequitable coverage (with many of the poorest households and marginal areas left unattended); various administrative and operatonal inefficiencies (lack of qualified personnel); and poor service. These problems have restrained progress in water and sewerage delivery, with detrimental health and living consequences for the population, particularly the poor. 4.40 Tariffs. Since water is a basic necessity, many believe that users should have free access to it. This view has been bolstered by official ambivalence. Apart from collection deficiencies, most water companies have not systemically adjusted tariffs-even during penods of sustained double-digit inflation. Most tariff adjustments have only come with cash flow difficulties. A vicious cycle of low cost recovery, low investments, poor service, and poor maintenance has thus sent most water entities into a downward financial spiral. 4.41 As with electricity, the government has kept charges low to provide the poor with access to water and sewerage. To do this, Colombia has a rather complex tariff strucre composed of a fixed charge with three ranges of increasing per-unit charges for households, divided into six socioeconomic strata. A 50 percent sewerage surcharge is added to the water bill. Some water companies (Pasto) have less than six strata. Each has the flexibility to set its tariff structre, within certain guidelines. The purpose of such guidelines is to maintain some cross-subsidization. Table 4.7 presents the tariff strucure for Bogota and Medellin. 4.42 Unlike electricity tariffs, water charges are properly structured like natual gas prices to attain cross-subsidization across income groups. As seen in Figure 4.3, tariffs are set so that the last two strata pay a tax, whereas the first two are subsidized. Given the previously mentioned stratification problems (paragraph 4.20), however, the distribution of water and sewerage subsidies is still not progressive (see incidence analysis below). Efforts are underway between DNP and the water companies to steamline stratification and establish more objective criteria (housing-unit value, assets owned, and so on). It will take some time to obtain results from this exercise. - 98 - Table 4.7: Water and Sewerage Tawiff by Socioeconomic Strata (1992 Pesos) More than Average Cost. Bogot Fixed 0-30 cubic 31-60 cuc 60 cubic 242.11 per (Stiwiwn) Charge mete"s meters meters cubic meter 49.36 19.75 69.11 251.76 2 246.81 44.41 98.72 251.76 3 987.21 69.11 133.29 251.76 4 1,727.65 118.47 167.84 251.76 5 3,455.30 167.84 202.4 251.76 6 4,936.14 217.20 236.92 251.76 More than Average Cost. Medeli Fied 0-30 cubic 31-60 cubk 60 cubc 202.46 per (StsO4m) Charge meters meters meters cubc meter 1 334.83 7.77 122.72 253.55 2 622.04 19.27 137.93 253.55 3 1,274.51 55.44 164.64 253.55 4 2,275.18 107.17 187.63 253.55 5 3,826.90 114.94 202.83 253.55 6 6,893.15 122.72 214.33 253.55 Sowvze: Velez (1994). FIgue 4.3: Monthly Water and Sewerage Charges, Costs, and Subsidies per Legal Urban Household, by Strata, 1992 (199 Pesos) 16000 14000 I12000p Taxes 8000 " 000 2000 1 ~~~2 3 Strata 4 56 r~- Charges . C-s Sore: Velez (1994). - 99 - 4.43 Programs for the Poc!. From 1968 to 1987, the Division of Basic Rural Sanitation (DBRS) of INPES carried out the government's most effective program to assist communities below 2,500 people in constructing, operating, and maintaining their own low- cost water systems. It was called the National Program for Basic Rural Sanitation. Working out of regional offices, 100 DBRS facilitators provided technical guidance and cofinancing directly to community water committees, called Juntas Comunitarias. The juntas contributed 15 percent of the project costs up front. Of the remaining 85 percent, 40 percent was covered by the national government over 15 years at an interest rate of 6 percent. The juntas then administered their own local water utilities. Each juta hired a DBRS-trained community member to keep accounts, maintain water quality, and execute simple repairs. This work required only a part-time commitment, and was compensated by a nominal salary. Operating costs were low, financed by monthly users' fees. DBRS technicians provided follow-up support, including major repairs. By 1987, DBRS had facilitated the construction or expansion of 2,500 water systems, bringing or upgrading water supply to 2 million people (6 percent of Colombia's total population). This impressive result accounted for only 30 percent of potential coverage. By the end of 1988, 90 percent of the corresponding water systems were still operating. Many development agencies outside Colombia, including the highly regarded Guatemalan NGO Agua del Pueblo, still use this model with great success. 4.44 Through Decree 77 of 1987, the national government liquidated the DBRS as part of decentralization. The decree transferred authority for rural services to the municipalities in January 1989. The users quickly became little more than consultative bodies to the mayors. Departmental governments were supposed to provide technical assistance; very few have continued this practice. Monthly user fees were sometimes diverted to general municipal treasuries instead of the users' associations. Costs increased, service quality deteriorated, and user contributions dwindled. The strong institutional network, which supported all operations, was dismantled. It was replaced by a feeble structure that quickly proved inoperative. Some financing continues to be channeled to ural areas through the World Bank's 1988 Water Supply and Sewerage Sector Project, the national government's Programa de Desarrollo Rural Integrado (DRI),12 and a direct subsidy fund for rural housing, water, and sanitation. These funds, estimated at US$140 million in the past four years, often operate without the necessary technical assistance, and no follow-up is provided in the field. Mayors mostly invest in the municipal capitals, where influential local stakeholders reside, instead of in rural villages with the greatest need. In short, efficiency and equity collapsed when users were cut off from administrative and financial control. 4.45 Ironically, decentralization led to the program's disintegration, reversing the efficiency and equity of water and sanitation services for the poor. The case demonstrates the qualitative differences between community participation in the form of information sharing (consultation) and community participation in the form of direct control (see Chapter 6). 12 See Chapter 5 for a detiled ptesentation of the program. - 100- Indidence Analysis: Who Benefs from Pubic Subsies? 4.46 Coverage. National coverage statistics for water and sewerage, as presented in Tables 4.8 and 4.9, clearly indicate that provision of these services is biased against the poor." For both services, coverage of the lowest quindle is well below country averages. In 1992, water coverage of the lowest quitile was 65 percent versus the national average of 83 percent. For sewerage it was 37 versus 67 percent. The least progress has been observed in sewerage. Coverage has only increased from 51 percent in 1974 to 67 percent in 1992. New coverage during this period has been distributed regressively, and the resulting distribution of services among income groups for 1992 has been the most regressive of all basic services. The Gini coefficient is twice that for water and four times higher than electricity. 4.47 Lack of access to water and sewerage is mostly a problem of the poor. As seen from the coverage information presented before (see simulation results in the last columns of Tables 4.8 and 4.9), the nber of connections must increase by 910,000 (water) and 1.8 million (sewerage) to atain the level of coverage of the highest income quintile. At least 70 percent of these required connections belong to the poorest 40 percent of Colombian households. This lack of coverage does not reflect unwillingness to pay. In the main cities, households in the lowest income decile spend up to 9 percent of their income on water and sewerage, while the population at large spends only 1 to 2 percent. 4.48 An imnortant finding from the household survey is the number of illegal or unauthorized water and sewerage connections. Of all basic services, illegal water and sewerage connections are the most pervasive. ilegal connections average 7 percent nationwide. Among the lowest income decile they represent 10 percent. But these figures reflect wide variations among regions and income groups. In Barranquila, illegal connections average 18 percent of the total, and 43 percent of services rendered to the lowest income decile are illegally obtained. As with energy, legal national coverage follows a regressive pattern (Gini of 0.1); illegal coverage is progressive (Gini of -0.2). 4.49 Publc Subsidies and Ditibution. The total 1992 water and sewerage subsidy was about 125 billion pesos (0.4 percent of GDP), of which 22 percent is explained by illegal comections. Subsidies are better targeted than in the case of electricity. Nevertheless, targetng is sdll far from adequate as indicated in Table 4. 10. 13 The _nfrmati preseted in te te is based on tme household survey conducted by DANE in Septmber 1992. It is intesdng to note at the covemage figures differ siniay from those presented in the 1990-94 Developmen Plan. Table 4.4 in Annex 4 presents a compaison of hese twc sets of figures. While the Plan cites esimad nationa water and sewerage covemge of 66 and 51 percent for 1990, and seas tgas of 76 ad 57 percent for 1994, for example, th survey indicates the nadonal coveage is algeady above those agets. It should be noted, however, that the Plan's greus correspond t eftive coverae (ta is. coverae adjusted for conunuty and frequecy of the service) and thus are not sticty comparable with the figures obtained fom dIe househl survcy which correspond to service coverage. - 101 - Table 4.8: Coverage of Water by Income Group, 1970, 1974, and 1992 Simnuled 98 percent 1970 1974 1974-92 Change 1992 Homogeneous Coverage Distibuton Households Distibution of New to be of Required Connections Coverage Connections Coverage Connections Connections Connections Coverage Connected Connections Income Quintile (milons) (percent) (millions) (percent) (millions) (percent) (millons) (percent) (millions) (percent) 1 0.261 36.7 0.316 44.0 0.611 16.6 0.927 64.6 0.447 49.1 2 0.320 45.1 0.356 49.5 0.771 20.9 1.127 78.5 0.248 27.2 3 0.401 56.4 0.447 62.2 0.780 21.2 1.228 85.5 0.147 16.2 4 0.495 69.7 0.516 71.8 0.790 21.4 1.306 91.0 0.068 7.5 5 0.627 88.2 0.640 89.0 0.735 19.9 1.375 95.8 0.000 0.0 Total 2.104 59.1 2.276 63.2 3.686 100.0 5..963 83.1 0.910 100.0 Crini Coefficient 0.171 _ 0.141 0.023 0.072 _ -0.472 Annual Growth Rate 1970-74: 1.9 percent 1974-92: 5.5 percent 1992-Target: 13.2 perce Source: Velez (1994). - 102- Table 4.9: Coverage of Sewerage by Income Group, 1970, 1974, and 1992 Simulated 98 percent l______________ 1970 1974 1974-92 Change 1992 Homogeneous Coverae Distribu1ion Households Distibution of New to be of Required Connectons Coverage Connections Coverage Connectons Connections Connections Coverage Connected Connecaons income Qunke (milions) (perent) (millions) (percent) (mMons) (percent) (millions) (percent) (millions) (percent) 1 0.178 25.1 0.198 27.6 0.334 11.4 0.533 37.1 0.783 43.8 2 0.249 35.1 0.263 36.6 0.552 18.8 0.816 56.8 0.500 28.0 3 0.320 45.0 0.347 48.2 0.638 21.7 0.985 68.6 0.331 18.5 4 0.434 61.1 0.442 61.4 0.700 23.8 1.142 79.6 0.173 9.7 5 0.581 81.7 0.599 83.3 0.716 24.4 1.315 91.7 0.000 0.0 Total 1.762 49.4 1.849 51.3 2.941 100.0 4.790 66.8 1.800 100.0 Gini coefficient 0.225 0.212 0.124 0.158 -0.424 Annua Growth Rate 1970-74: 1.2 percent 1974-92: 5.4 percent 1992-Target: 27.1 percent Source: Velez (1994). - 103 - Table 4.10: Distribution of Water and Sewerage Subsidies Total Subsidis Dsbibudm Subsdi per household (mons of 1992 pesos) Of Subsis (perce t_) (thousands of 1992 pesos) Dedk Lga lga TOl Legal egal ToW Legal illegal Total 1 12,316 3,072 15,388 12.61 11.35 12.33 17.2 4.3 21.4 2 13,163 3,665 16,828 13.47 13.54 13.49 18.4 5.1 23.5 3 14,069 3,052 17,121 14.40 11.28 13.72 19.6 4.3 23.9 4 12,617 3,890 16,507 12.91 14. 37 13.23 17.6 5.4 23.0 5 12,691 2,714 15,405 12.99 10.03 12.35 17.7 3.8 21.5 6 - 13,826 2,5S0 16,376 14.15 9.42 13.12 19.3 3.6 22.8 7 11,681 2,329 14,010 11.96 8.60 11.23 16.3 3.2 19.5 8 9,087 2,703 11,791 9.30 9.99 9.45 12.7 3.8 16.4 9 3,850 1,644 5,494 3.94 6.07 4.40 5.4 2.3 7.7 10 (5,593) 1,447 (4,146) -5.72 5.35 -3.32 -7.8 2.0 -5.8 Totad 97,706 27,066 124,772 100.00 100.00 100.00 13.6 3.8 17.4 fo;du cwl -0.2(9 -0.131 -0.231 5kuow. Velez (1994). - 104- The distrioution of subsidies is close to neutral up to the sixth decile. Households in the fifth decile received the same subsidy as those in the first. Only the tenth decile paid a small tax. 4.50 The distribution of subsidies relates to two policies: coverage and pricing. Water and sewerage coverage is worse than energy, both lower and more regressive. This is more than compensated by the better design of the water and sewerage tariffs. As a result, even with the stratification errors the ultimate distribution of subsidies is more progressive than in energy. Conclusions and Recommendations 4.51 Information Systems. The large coverage discrepancies identified in the 1992 household survey, and those used for the Development Plan, are only one example of the unreliable data on water and sewerage. The information gap is too wide to conduct in-depth diagnoses, monitor sector activities, and use the results as feedback for reform. Both the Ministry of Development and DNP are giving high priority to this issue. A bottom-up inventory of assets and liabilities, conditions of infrastructure, needs for new investment, and operations and maintenance is required. 4.52 A Geographical Information System (GIS) could reduce compartmentlization among the various entities, develop a comprehensive, reliable data base, identify key bottleneck areas, and define follow-up implementation actions for this complex task. A GIS can also lend a spatial dimension to sector problems. The needs of inner-city groups, those in distant and inaccessible terrain, or those in isolated rural areas call for distinct approaches that may require different technological, institutional, and other solutions. 4.53 Service Expansion and Quality. Given the discrepancies in coverage between urban and rural populations, and within the urban sector (see Table 4.4 in Annex 4), the government must set coverage goals for each of these areas. The goals would consider the relatdve costs and benefits of higher coverage and improved service.' For larger cities nearing universal coverage, extending services to marginal settlements could be too costly, with technological and environmental implications. In these cases, government should conctrate on reducing UFW rates and improving quality. Any acton plan aimed at attaining 100 percent coverage should avoid providing seivice to residents in precarious areas. Providing such service would create an incentive for others to establish new settlements in worse terrain and demand access. Legal prohibitions and threatened sanctions have not worked so far.15 14 The GIS referred to in the previous patagraph is essential for this purpose. 15 For example, in the Bank's Cartagena Housing Project, massive land reclamation works were execoted along a lake and sever thousand fimilies beneftd from roads. ectri*. water and sewerage and other amenities. However, shordy afterwards, in spite of efforts keep out squatters, unauthorized setdlemes mushroomed at the ew periphery with demands for more reclamaion works. - 105 - 4.54 Meeting the needs of residents in inaccessible, unserviced areas without creating incentives for others to settle in even more difficult terrain is not easy. Apart from technological constraints, the per household cost is generally prohibitive. This brings into focus two issues: (a) intermediate technologies for residents in inaccessible or very poor areas, whether in peripheral settlements near urban centers or in dispersed rural zones; and (b) housing policy that prevents proliferation of unauthorized settlements.16 4.55 The government should promote such alternatives as community standpipes, cisterns, tanker trucks, and water purification tablets as solutions for unserviced areas. It should explore ways of upgrading the quality of service provided under such alteratives by encouraging private sector participation through a bidding process that would assign both lucrative and nonlucrative neighborhoods to be serviced by the bidders. A subsidy of the same magnitude as that provided to low-income households through conventional services could be awarded to ensure service in the poorest neighborhoods. A regulatory framework would ensure such operators comply with minimum sanitary norms in storing, transporting, and distributing water, and in their tariffs.'7 The effort could be linked to the current cholera program so that unattended areas receive improved water and sewerage services. 4.56 Tariff Structure. For those already connected to the system, an urgent priority is improved service. Most water companies currently lack the resources to deal with their technical, financial, and admiistrative constraints. The DNP, FINDETER, and the reorganized Ministry of Development institutional efforts to tackle these problems should be complemented by appropriate tariffs. As shown above, the sectoral subsidy is not currently received by the poor. 4.57 As with electricity, the government should consider eliminating stratification, establihing a general residential tariff, and complementing it with targeted social tariffs for low-income users. If this is politically unfeasible, stratification needs to be revised to ensure it more closely reflects income distrbution. Consideration should be given to incorporating an explicit accounting mechanism to separate the subsidy in the water bills. At a minimum, this would make users more aware of the true cost of water. 4.58 Sewerage Treatment. Current neglect of wastewater treatment raises serious environmental concerns. The curent situation negatively impacts the health and economic welfare of all residents, especially the poor, who live in marginal settlements along polluted waterways. The 1990-94 Development Plan recognizes the importance of this situaton. It includes provisions for an unteated wastewater program. Apart from the logistical, technical, and environmental issues, however, this is a very costly endeavor, especially in Colombia, where the current treatment system is almost non-existent and has been long neglected. A 16 The government's curent housing pohcy frmework and its role in promotng the more ordedy devdopment of human settlemems is prsented in Secdon IV of this chapter. 17 Refer to Cbapter 6. pangraph 6.23, for a more detailed disussion of the issues enaile in contcing municipal sewrv - 106 - well-formulated, long-term strategy should be developed to cover both industrial and other wastes. There should be clear priority areas defined in stages to assist in a systematic and coordinated approach. 4.59 Community/Private Sector Participation.18 To ensure the new decentralization effort is successful, the role of non-governmental participants (private sector, NGOs, and users) must be clearly defined. Several water companies have already begun to actively encourage private sector participation through equity contributions. This approach should be amplified. Companies should look at subcontracting billing, collection, and maintenance. This could be pursued through clearly defined bidding processes with clear terms of reference and performance criteria, and appropriate rewards and penalties. 4.60 Consumers want services that meet their needs in a safe, reliable, and affordable way. Given the importance of water and sewerage to the community, they also want to help make major decisions about the provision of such services. Upper-income groups have greater access through established distribution networks and their own investments in reserve tanks, pumps, and so on. Urban and rural households in the lowest three deciles of income distribution generally receive marginal attention and suffer relatively more inconvenience and losses. 4.61 The stuccess of the DSBR shows the benefits of active community groups. In- kind participation reduces costs. But efficiency can be achieved in many other ways, aided by a sense of ownership and active participation in the system. These groups can help keep drainage systems clear, and ensure heavy commercial vehicles do not damage streets and underground distribution lines. They can detect maintenance problems early. They should be given an active early role in prioritizing works. They should help educate the community on the negative effects of illegal water connections. The Development Plan envisions resumption of models like DBRS through local municipal agencies. Most towns lack the technical capacity to assume this role and will need initial assistance. IV. Housing Policg Framework: A Programn to Benefi the Poor Sector Overiew 4.63 Main Sector Problems. In 1991, 1.3 million households in Colombia-one in every five-lived in unsatisfactory conditions in terms of space, public services, and quality of housing (as defimed by the corresponding NBI indicators). Housing needs differ between urban and rural areas. In urban centers, as Table 4.11 indicates, the most significant problem is overcrowding, where rooms must be shared by more than three persons, or by siblings of different sex. In rural areas, inadequate services is the main constraint. Inadequate housing 18 Although the topic of stitutional developmem in a decentraon framework is the main subject of Cbapter 6, it is wordh menfioning at this point a few seCtor Specific issues. - 107 - Table 4.11: The Housing Problem in Colombia, 1991 Percenftge Number of Incidence Of NBI Indicator Households (percent) Total Inadequate Housing 152,831 2.2 11.7 Urban 43,780 1.0 13.4 Rural 109,051 3.9 11.1 inadequate Services 377,196 5.3 28.8 Urban 18,818 0.4 5.8 Rural 358,378 12.7 36.4 Crowded Households 391,387 5.5 29.9 Urban 216,256 5.1 66.4 Rural 175,131 6.2 17.8 inadequate Housing and Services 154,440 2.2 11.8 Urban 7,820 0.2 2.4 Rural 146,620 5.2 14.9 Inadequate Housing and Crowded Households 67,061 0.9 5.1 Urban 26,033 0.6 8.0 Rural 41,028 1.5 4.2 Inadequate Services and Crowded Households 86,025 1.2 6.6 Urban 7,150 0.2 2.2 Rural 78,875 2.8 8.0 inadequate Housing and Services, Crowded Households 80,285 1.1 6.1 Urban 5,823 0.1 1.8 Rural 74,462 2.6 7.6 Households with Housing Problems 1,309,225 18.5 100.0 Urban 325,680 7.7 100.0 Rural 983,545 34.9 100.0 Source: Estimates based on household surveys (ENH74). (in terms of structure and constuction materials) is the least of Colombia's problems. Unlike some developing countries, homelessness is not a problem in Colombia. 4.64 Chronic deficiencies in inner cities, peripheral, and rural areas have lowered standards of living. The poor bear most of the direct costs of crowded living, inadequate water, sewerage, and power, and environmental degradation (such as loss in productivity because of chronic health problems). Uncontrolled and peripheral settlements are also costly for the general population. They contribute to pollution, erosion, and other environmental degradation, loss of life through landslides, floods, and other disasters, and the high cost of providing on-site services to distant settlements in unsuitable terrain. - 108 - 4.65 Like many other countries in the region, Colombia has long intervened in housing to help the poor. Despite decades of government intervention, spending US$1 billion over forty years in direct funds, and directed credit schemes (Annex 4, Table 4.6) filling the housing gap and meeting the housing needs of the poor remains an elusive goal. As presented in the 1990-94 Development Plan, such failure reflects past housing policies: excessive centralization, high government participation crowding out private sector initiatives, inefficient public systems used to construct and fnance low-income housing, and inability to target subsidies to the poor. 4.66 The New Housing Policy Framework. The govenmment is currently implementing a new program to avoid the problems of the past. The program moves from the direct provision of housing units to directly subsidizing some housing needs of the poor. Certain key characteristics set this framework apart from previous efforts. In particular: (a) it aims at integratng unattended marginal settlements into their respective municipalities, making control of settlement practices an important objective; (b) it emphasizes decentralization and greater participation of local entities, NGOs, the private sector, and community groups; and (c) it focuses on the poorest strata, specifying target groups and type of assistance (explicit subsidy with an indexed mortgage instrument), type of solutions to be fnanced, beneficiary selection conditions, and procedures. 4.67 Subsidies: Dermitions and Targets. The government's social housing program (as presented in the Development Plan) is aimed at two categories of low-income families: those witlh mo2thly income between two and four minimum salaries (SM) (between US$213 and US$426), and those with monthly income less than two SM. For rural areas and those with less than 15,000 inhabitants, subsidies are targeted at families with assets less than 70 SM (US$7,448). Potential beneficiaries can apply individually or collectively. The subsidies are for the purchase of housing among eligible options: urbanizable lots, urbanized lots, new units, used units, home improvements, water and sanitation works, and titling. Table 4.12 shows the program of subsidies envisaged in the Development Plan by eligible beneficiaries and types of housing solutions. 4.68 For applicants in the 2-4 SM range, the cost of eligible new housing units may vary between 50 and 135 SM (US$5,320 and US$14,364). Applicants are expected to have cash savings of at least 5 percent of the value of the unit. In these cases the beneficiaries receive an initial subsidy of 12 SM, to be complemented by UPAC-denominated mortgage loans from private lenders."9 Among families in similar economic cicumstances, those with higher personal savings would receive priority under the program. To avoid misuse, once the subsidies are approved by INURBE they are placed with a private tmst agency (Fiduciaria) 19 Ihe subsidy is actually denominated in Unidadta de Poder Adqudsivw Constame (UPAC): 210 UPACs. The UPAC is an indexed monlgne unit which is continwusly adjstwed for inflaion. It was insdtuted in 1972 with the purpose of attracing funds to financial intermedaries which in turn on-lend tdem to die constuton sector. In Febraty 15, 1993, 1 UPAC = 4,575.91 pesos (equivalent to US$6.11). - 109 - Table 4.12: Program Targets and Observed Performance (Number of Housing Solutions) 1991 1 IM 9 1993a 1994 1991 1992 | 1993a 1994 Housing Ent Planned Soluons Cwnulaive Figures Inurbe-Cajas-Cavs 47,884 41,979 56,709 62,195 47,684 89,663 146,372 208,567 FNA 4,865 4,959 5,012 5,358 4,865 9,824 14,836 20,194 Caja Vivienda Militar 2,122 2,656 2,838 3,089 2,122 4,776 7,616 10,705 Rural housing 22,723 23,996 22,997 26,302 22,723 46,719 69,716 96,018 Total 77,394 73,590 87,556 96,944 77,394 150,982 238,540 335,484 Generated Soluions Cnulati've Fyguns Inurbe-Cajas-Cavs 39,684 88,279 39,684 127,963 127,963 FNA 5,530 5,928 3,782 5,530 11,458 15,240 Caja Vivienda Militar 1,949 2,976 2,712 1,949 4,925 7,637 Rural housing 600 20,241 7 600 20,841 20,841 Total 47,763 17,424 6,494 47,763 165,187 171,681 Approved Subsies Cwndative Figures Iuurbe-Cajas-Cavs 35,000 |45,860 35,000 80,860 Rural housing 600 20,241 600 20,841 Total 35,600 66,101 35,600 101,701 a. Actuals as of April 1993. Note. Does not incluxt dding or housing units priced over 135 SM. Sowme: DANE. which then verifies the documentation and gives the funds directly to the seller (not to the beneficiary) upon compliance with contractual conditions. 4.69 To promote community participation, a 15 SM subsidy (as opposed to 12 SM for individuals), is targeted to families who form organized groups to obtain new housing through self-help programs.20 These families are expected to have monthly incomes below 2 SM. They are the ones who would try to substitute their own labor instead of paying others to do the work. Consequently, their solutions are expected to cost no more than 50 SM. For both individuals or groups, subsidies for lots, home improvements, or titling are set at 12 20 Set at 262 UPACs. - 110 - SM. Once approved, individuals have twelve months to find appropriate solutions and claim their subsidies. Those in community groups have eighteen months. 4.70 In the rural subsidy program, given the high ownership rate and poor access to water and sanitary services, 50 percent of the subsidies are reserved for sanitation services. Individival subsidies of 12 SM are available to leverage assistance from other sources through muniicipalities . 4.71 Program Targets and Initial Results. To ensure that housing meets required standards and costs, subsidies apply only to housing solutions declared eligible by INURBE or Caja Agraria. Table 4.12 details the expected participation of various entities in the supply of housing and their role in meeting the 1990-94 Development Plan targets-in all, 335,484 housing solutions.2" The performance of the program, taking the end-1992 as the benchmark, has been mixed. On the one hand, total housing (165,187) has aleady surpassed the Plan's original target (150,982) by 9 percent. But the total number of approved subsidies (101,701) represents only 62 percent of all solutions. Of those, only 13,373 have been fully paid out (as of December 1992) by the tmst entities-which account for 8 percent of the solutions generated and 13 percent of the approved subsidies.' 4.72 Results indicate the program is successfully providing adequate incentives to supply low-income housing by enhancing available credit on the expectation of increased capacity to pay among low-income groups. This is corroborated by direct discussions with private developers, which indicate that low-income housing provides equivalent rates of return to alterative uses of capital. The only complaint is the lag in subsidy payments. This seems to be compensated by the lrge scale of projects and the rapid turnover of capital. But subsidy payments delays deserve attention from the authorities. As Figure 4.4 shows, 56 percent of eligible housing solutions have been for individual applicants. Seventy percent of the supply is provided by private developers. Half of the solutions are urbanized lots priced below 30 SM. In contrast, the available solutions for collective applications-with prices below 50 SM-seem evenly spread between new units, urbanized lots, and housing improvements.23 4.73 On the demand side, the program has yet to generate the anticipated response from the targeted population. Of particular concern is the large gap between approved and disbursed subsidies. CENAC's survey (see next section) explored this issue and provided 21 'This excludes tiig-the responslbiliq of municipalties-and housing units pticed over 135 SM-as they are not considered pat of low-Income housig. 22 Part of die observed disbursement lag, of course, can be aruted to dte fact dtat beneficiares have between twelve to eipteen monhs to claim their ubsdes. 23 Annex 4. Table 4.7 presents information on the regional disrbution of eligibe housing solutions. - 111 - Fgure 4.4: Supply of Housing Solutions Collective Available Housing Solutions Type of Supplies 44.1% 61.69% bddus Solutions Mv 0.03% 5S9% 38.28% Type of Housing Solutions Price Range (SM) Now urnt 3 34.7% 87.0% Urbarized lot 29.8% -~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~16 Individual Avalable Housing Soduions Type of Suppiles 44.1% 17.0 . 55.% ?j 70.0% Type of Housing Solutions Pric Range (SM) 30 0.7% 31.50 V40% 6.0% 502% 101-135 12.1% 25.7% So~:CENAC - 112 - some indicative explanations:24 (a) communication problems (12 percent of interviewed beneficiaries did not know their application had been approved); (b) beneficiaries could not fimd the housing solution they wanted (location is probably the typical constraint); (c) lack of access to the necessary complementary credit; (d) lack of knowledge on how to obtain the approved subsidy; (e) problems with required documentation; and (f) lag of up to eighteen months between approval and disbursement. Incidence Analysis 4.74 Figure 4.5 summarizes the results of the survey conducted by CENAC to identify the beneficiaries of the housing subsidy program.35 One of the most striking findings is that subsidies are not well-targeted. Households successful in completing the process to obtain subsidies are, on average, better off then those for whom subsidies have been approved, and disproportionately higher in the income scale than the targeted population. At the national level, only 25 percent of the subsidy is received by the poorest 30 percent of the population. The constraints deterrng households from claiming their subsidy correlate closely with income. The constraints are relatively more binding for poorer households. Unless these constraimts are relaxed, further refmement of the target population may not yield substantial gains. The data point to the need to re-examine providing the poorest deciles with completed units, rather than trying to increase their access to basic services (urbanized and urbanizable lots, water, and sanitation works) and improve their existing title, housing, and neighborhoods. 4.75 The survey highlights two other important points. INURBE's records show all of its beneficiaries falling within the 4 SM range of household income distribution (that is, 73 percent of households). But the sample survey indicates there may have been some leakages, with 10-20 percent of the subsidy going to households with incomes above 4 SM. Second, the survey identifies a few cases where the original beneficiaries have sold or rented the units, contrary to regulations. These results warrant closer monitoring and follow-up to verify the situation and take corrective measures. Conclusions and Recommendadons 4.76 Colombia's housing policy has moved away from the supply-driven, product- oriented programs it pursued in the past. As in many other countries, such programs shackle public housing agencies (and in some cases, private banks) with the responsibility of meeting the needs of people who cannot consistently pay. Policymakers have distinguished between housing demand (for which households are willing and able to acquire solutions), and housing needs (essentially a subjective estimation of what households are perceived to need, without consideration of willingness or ability to pay). They have removed the public sector 24 Note that the program is sll in its early phases, and the reasons for the discrepancy are probably very dficult to determine. 25 More detailed information is provided in Annex 4, Tables 4.8 to 4.11. - 113 - f F%gure 4.5: Subsidy Program - Distribution of Target Population, Approved and Disbursed Subsidies 15 4 16 2 14 12 10 4 2 1 2 3 4 5 6 7 8 9 10 inome pr Capita - Decles + Target Population * Approved Subsidies -Disbursed Subsidies from constructiom and opted for a direct demand subsidy targeted to the poor. They have focused on brinng order to urbanization and providing a framework to absorb the ongoing 40 percent of settemert activity now taking place in disutat, marginal areas and unsuitable terrain. The framework involves government, private sector, and community participation. Although the changes are significant, and focused in the right direction, evidence indicates the program could more effectively reach the poor by de-emphasizing attention to new housing, direcig more resources to basic services for the poor (through urbanized and urbanizable lots, water, and sanitation works), and improving existng titles, housing, and neighborhoods. 4.77 There are reasons to be encouraged by the performance of the program so far: (a) the private sector has responded to the new incentives; (b) increased availability and demand for urbanized lots provides a solid basis to tackle peripheral settlements; and (c) many housing solutions (three out of every four) offered are in the 0-50 SM price range, providing options to the poorest strata. One can still identify changes required if the progrm is to attain its desired objectives. Four areas deserve special atention in the poverty reduction context.: subsidy stucue and delivery, information and communication, land titling and legalization, and construction norms and standards. 4.78 Subsidy Structu and Delivery. The housing program seems to miss the lowest income groups. To some extent this relates to the current level and strcture of the stbsidy. Table 4413 presents some calculations for a family that receives a subsidy for a low- cost solution costtng 50 SM, having a monthly family income of 2 SM. - 114 - Table 4.13: Subsidy for a 2 SM Household and the Esthmated Required Payments for a Housing Solution US$ SM Maximum monthly family income 213 2.00 Cost of unit or solution 5,320 50.00 Prior savings (5 percent) 266 2.50 Subsidy provided 1,596 15.00 Loan 3,458 32.50 Monthly payment at 28 percent interest for 15 years 82 0.77 Monthly payment as percentage of monthly income: 39 Source: Bank staff calculations. 4.79 As can be seen from the above figures, for a family with a 2 SM income, a 50 SM solution would call for 39 percent of its income (assuming an interest rate of 28 percent and a 15 year repayment period, with a 5 percent downpayment and 15 SM subsidy). For the 1 SM family (the lowest 17 percent of income levels), the mortgage payment would be 77 percent of monthly income. Sector officials are aware of this problem and made some adjustments to deal with it. Emphasis has been given to providing the option to acquire urbanizable or urbanized lots for 30 SM (an option more suitable for the lowest deciles in household income distribution). Additionally, the CAV are authorized to use variable payments for low-income groups, under which early mortgage payments are lower than the amount required to amortize the loan, and there is negative amorization. But the CAVs would still not provide loans for more than 70 percent of the property value, and they have other income stability criteria. This implies that the assumptions underlying the subsidies may be too optimistic (in terms of the types of solutions those in the lowest few deciles can afford) overestimating beneficiary capacity to repay mortgages. Even with a 30 SM plot of land, the beneficiaries would still have to fnance some type of on-site superstructure or incur shelter costs at another site, either of which would require additional funding.36 4.80 The above results partally explain the fact that only one-third of paid subsidies have been awarded to collective applicants. This is disappointing given that this opton was included precisely to reach the very poor. To overcome this problem the government could increase the subsidy for this group so that it: (a) represents a larger share of the cost; and (b) provides more incentives for the CAVs to lend. This does not necessarily require increasing the required resources. The adjustment can be attained by shifting resources away from individual applicants. This would have the added advantage of encouraging new supply of low cost solutions (below 50 SM), which are not appealing to the middle-income groups. 26 Indeed, te wReocIn Pac(flc acknowledges this constt and assumes the sbsidy would be complemented by resoures provided by local govermments. lhose resources, however, have not been forthcomig. - 115 - Another option is to de-emphasize new housing and direct more resources to improve the condition of existing housing and poor neighborhoods. In this case the subsidies for new housing could be reduced, and the resources used to increase the 12 SM subsidy for individuals and groups to buy lots, home improvements, or titling. 4.81 The govermnent needs to eliminate delays in the delivery of the subsidy to the developer. This would reduce housing costs and prices. The government should decentralize this part of the process by giving CAVs responsibility for direct payment of subsidies. This would be both more expcdient and less costly. 4.82 Information and Communication. There is a severe lack of information about the housing program among the target population. CENAC's survey indicates 40 percent of households do not know ICT, 55 percent have not beard about INURBE, and half know nothing about the subsidy. The population seems unaware of the radical changes in housing policy. A publicity campaign should explain the basic objectives and features of the subsidy. NGOs should play an important role in 3,ch an outreach program. The campaign could follow the basic design of Mexico's Solidanidad Program promotion. 4.83 About 12 percent of current beneficiaries are unaware their applications were approved. Twenty-two percent did not receive the communication letter from the official entity required to process final payment. Forty-three percent of beneficiaries did not know how to find their housing solutions. Thirty-eight percent did not know how to claim the subsidy. To tackle this information gap, authorities should more effectively communicate with beneficiaries throughout the process. During application, each part of the process should be explained. Information about available housing solutions should be presented in a user- friendly manner. Finally, beneficiaries should be able to complain about process irregularities to responsive authorities. 4.84 Availability of information will speed implementation and ensure that supply responds to demand. Currently, 17 percent of beneficiaries do not find the housing solution which, given their resources and needs, they wouid like to acquire. Part of the problem may be that developers have not had the time to respond to these needs. Information on the composition of current applications, interest in given housing solutions, price ranges, and locations should be made public. In all, better communication and information should help deal with existing constraints (referred to in paragraph 4.73) precluding better targeting. In general, access to information decreases with income. 4.85 Land Titling and Leaizaton. Like other countries m the region, Colombia has many land titling problems, ranging from multiple ownership to absentee owners and lack of legal documentation. Current figures on land title and registration are not available, but institutional bottlenecks apparently exist in several areas: outdated or non-existent data bases and technical facilities, poorly trained staff, low compensation levels, high turnover rates, and cumbersome legal requirements. As in other countries, the government is among the largest landowners, but as the experience with ICT shows, there is no record of its - 116 - holdings in terms of location, size, use, and other such characteristics. A well-designed housing program must improve land management, registration, and related activities in legalization. Appropriate technical assistance, better-staffed and equipped land-registration institutions, and simplified systems and procedures to update exising records would provide incentives to attain legal titling and improve the access to legalized plots. 4.86 Widespread land title problems also have other far-reaching impacts. A city's municipal revenue base is limited, since property tax could be a key source of revenue with decentralization. Outdated tax valuations and inefficient collection help explain the limited role of property taxes. This contributes to the inability of local governments 0o meet sector needs. Furthermore, tax and title problems hinder direct levying of betterment taxes, and encourage land speculation by reducing the costs of owning vacant lots. 4.87 Construction Norms and Standards. An important complementary action to promote better housing for the poor in Colombia is a review and simplification of current minimum norms and standards and land-use regulations, which are required by municipalities, water companies, etc. Although important improvements have been made in modifying norms and standards in most municipalities, licensing requirements and land-use regulations have been made more cumbersome. The reality of illegal, pirate, and other settlement patterns arising from urbanization and population pressure contrsts with the norms and standards and land-use criteria of municipal authorities through-out the country. As municipal codes are ignored and titling and other services are denied, unauthorized developers and residents in the informal sector continue to operate below prudent stanidards in terms of location, type, size, and so on, of the solutions provided. Families settle in precarious areas without regard for environmental or safety factors and they, along with the government, pay a high price for this. To achieve improvements in housing, the authorities should reconcile unrealistic municipal requirements with the unrstined excesses of unauthorized settlements. 27 In 192. the central govemmet imposed homogeneous procedwues for licensing acoss the country. bis bas increased the breaucatic process in an imporaMt miber of cities. Chapter 5 Rural Programs and the Poor Government programs in rural areas play an important role in poverty reduction. Overall, rural spending has been progressive. But since 1988, spending on the principal rural programs has decreased. In most cases, weak institutions and inadequacies in policy design have been the main factors limiting the effectiveness of rural programs. Many of these constraints are currently being addressed by reforms in Fondo DRI, the proposed new Land Reform Law, and the overhaul of Caja Agraria's operations. But institutional weaknesses remain a liability. Decentralization will require consistent supportfor capacity building at the departmental and local levels. With 70 percent of Columbia's extremely poor living in rural areas, afully coherent rural development strategy is urgently needed. To ensure the effectiveness of rural expenditures for infrastructure and social services and to avoid misallocation of resources, the government needs to adequately coordinate all executing and financing agencies in the rural sector. The government needs an institutional framework compatible with the broad objective of rural development, not strict adherence to subsectoral lines, agency responsibilities, or levels of government. i. Overnew of Main Rural Programs 5.1 Several rural government programs address poverty reduction, with varying success. This chapter examines the most important of these rural programs: Fondo DRI (Desarrollo Rural Integrado), PNR (Plan Nacional de Rehabilitaci6n), INCORA (Instituto Colombiano de la Reforma Agraria), and the agricultural credit bank (Caja de Cridito Agrario).) Jointly, these four programs aim to improve small producer access to infrastructure (roads, water supply, irrigation), services (agricultural extension, market information, social services), land (through land reform programs), rural credit, and markets. 5.2 Investment by DRI, PNR, and INCORA have fallen dramatically since 1988. Combined spending in 1991 represented 45 percent of the 1988 level. The reduction in those programs was accompanied by a sharp reduction in Caja Agraria's lending to small farmers. While the shifts in expenditures have been accompanied by an increase in the block transfer of resources to local governmens, between 1988 and 1991 overall public rural expenditures declined. This trend reversed in 1993, and (as discssed in Chapter 2) current plans call for expanded rural expenditure programs in 1994. I Ibe analysis does not include HMAT (opeatg in the field of hirgaton) and II)MA (in cbarge of markeing agHcuI prod&cts). - 118- Table 5.1: Investment Spending by Selected Rural Program RPwd Progrms DRf DPNRb LNCORA Year (miwons of 12 psos) (sham) (she) (sha) 1988 166,561 23.3 45.1 31.6 1989 146,503 22.5 35.9 41.6 1990 125,560 17.5 32.9 49.7 1991 74,261 35.5 31.1 33.4 a. Corresponas exclsively to executed investment. b. Corresponds to budgeted investment by DPNR. Source: Tables 5.4, 5.6 and 5.8. 5.3 The four programs distribute expenditures progressively among income groups. As shown in Table 5.2, households in the bottom 20 percent of the nrual income distribution receive 29 percent of the public subsidies while households in the top 20 percent receive less than 11 percent. In monetary terms, these expenditures add 10.5 percent to the incme of the poorest 10 percent of the households, and less than 0.2 percent to that of the richest 10 percent. In per capita terms, the first quintile of the distribution receives almost twice as much subsidy from these programs as the fifth quintile. Table 5.2: Distribution of Subsidies by Selected Rural Program, 1991 (peos) Subsid, per SubsWd per DWbudon Quindle Hohseold Capa (percent) 1 39,020 6,846 29.1 2 33,593 6,279 26.4 3 28,017 5,548 22.8 4 13,826 2,973 10.8 5 13,548 3,518 10.9 Source: Tables 5.5, 5.7, 5.9, and 5.11. 5.4 Composition of subsidies to households varies significantly among income groups. For households in the first three quiles, INCORA is the source of approximately half the subsidies, while for households in the upper two quintiles, PNR and DRI provide approximately 80 percent of the subsidies. - 119 - Table 5.3: Contribution of Selected Rural Programs to Total Subsidies (percent) Cqja Qduinle DRI PNR INCORA Agiare Total 1 23 25 42 11 100 2 21 22 51 6 100 3 22 23 52 3 100 4 41 41 12 6 100 5 45 38 13 4 100 a. Corresponds to interest rate subsidies in small farmers' lending. Source: Tables 5.5, 5.7, 5.9 and 5.11. 5.5 This chapter will consider each of the four programs, describing their services for the rural poor, analyzing their distributive impact, and discussing the institutional and policy challenges they face in reducing rural poverty. Section II considers DRI, followed by PNR in Section III, INCORA in Section IV, and Caja Agraria in Section V. Conclusions and recommendations close the chapter. IL Fondo de Desaollo Rural Integrado (DRI) 5.6 The rural development agency (DRI) has operated sinmc the mid-1970s, first under the Planning Ministry, and under the Ministry of Agriculture since 1983. While limited in resources, it has comprehensively addressed the social and economic problems of small farmers. In 1987, Fondo DRI was given responsibility for the Plan de Desarrollo integral Campesino (PDIC) to improve production, market conditions, and incomes in areas with large campesino populations. The Plan emphasized production, market access, and basic social services. While DRI ensured overall coordination of the Plan, its cofinancing was limited to investments in agricultural extension, marketing, irrigation, natural resource management, aquaculture, water supply, sanitation, and rural roads. In the early 1990s, as decentralization took hold, the agency's mandate was fundamentally changed to cofinancing municipal investment projects. 5.7 Recent Reform. Decree 2132 of December 1992 restructned DRI into the Rural Investment Cofinancing Fund-DRI. Under the new regime, DRI continues to cofinance projects with local governments, but can no longer do so jointly with other centrally funded entities. The Decree expands DRI's coverage to all 1,033 municipalities and allows it to cofinance land reform, flood control, electrification, and rural housing. The proposed Land Reform Law being considered by the Senate (Draft Law 114) allows DRI to cofinance projects with community groups, a practice previously accepted only for PNR. The Decree obliges DRI to close down its departmental offices and transfer their functions to departmental governments, carrying the decentralization of public institutions one step further. - 120 - DRM Spending 5.8 Total spending by DRI has fallen since 1988, when it totalled 38.7 billion 1992 pesos. In 1991 DRI's investment levels increased, but they fell to 24.5 billion pesos in 1992. External resources contribute, on average, 40 percent of DRI's funding. The reduction in DRI's spending between 1988 and 1990 was the result of lower budget appropriations (from own and external sources) and lower execution levels for the own-source components.2 Table 5.4: DRI Program Spending, 1988-92 (millions of 1992 pesos) Year ToWl Own Funds (percent) 1988 38,749 66 1989 32,919 58 1990 21,913 53 1991 26,349 62 1992 24,450 n.a Source: Perfetti and Guefra de Mesa (1993) and Mdaterio de Agricultura (1993). 5.9 The two main components of DRI's investments are infrastructure and production support. Together these represent 80 percent of DRI's spending over the period 1980-91. While the production subprogram has been falling in size, the share of the infrstructure subprogram has increased. This pattern could affect who benefits most from DRI's operations since production activities may be better targeted toward low-income households than infrastructure investments. 5.10 Resource Allocation. At the planning stage, DRI allocates a share of total resources to each department based on an estimate of needs. Each deparment receives an indicative allocation drawn from 70 percent of DRI's funds. Departments can compete for the remaining 30 percent once their annual quotas are exhausted. At the municipal level, allocaton responds to demand for specific projects appraised by DRI for soundness and eligibility. To influence the pattern of demand, DRI has constructed a cofinancing matrix that allows it to fund more projects in less developed municipalities.3 The range of cofnancing rates within each class of municipality corresponds to different project types and is used to lower the cost to the municipality of priority projects with less than optimal demand (for example, natural resource management projects). Other projects, such as in ucture-that 2 The execution ratio (with respect to appropnations) for ordinary resources fell from 100 percent in 1988 to 87 percent in 1989 and 85 percent in 1990. 3 Municipalities areclassified into seven groups. according to tdeirrclativeeconomicdevelopment. DRIfaimsto cntr its activiies in the poorest four categories. - 121 - reWire no special inducements because of their appeal to local govermments-are assigned lower cofinancing rates. Box 5.1: Rural Development hnwetmnt Project A Decentralized, Demand Drven Program of Rural Investnents Fondo DRI is implementing a number of exteraly funded operatons. The largest is te Rural Developmet Investment Project, cofinanced by the World Bank and the IDB. Me project provides fimds to municipalities to finance small investment subprojects, on a matchig gnmt basis. Local communities prvide labor and local materials, and municipalities contribut a portion of the cash cost. The Project fosts community participation, with mandatory community involvement in the ideatification of needs and the election of subprojects, as well as direct participation in execution, operation, and mantenae of the works. Subproject preparation is the responsibility of the municipalities. DRI ai subproject based on te:ical and eaviromental criteria, and the reimenat that beneiciaries be low-inme rur families. Projet execution is open to a broad range of pottal contt, incg private conractors, stae agncies, NQOs, ad universities, which compete to supply the works and services. By fnlog a wide variety of local initiatives, the Project responds to very rel needs, while mimizing te rol of govMMnt and creang a market for technical and other seraves. Fondo DRI includes an important capacity bildig component, desigd to X stente t waker .mmicipalities in the area of community participation, need diagnosis, inest£me planning, and projet idetification. It also provides fuding and teical supportfor the desigad f lty stdies required t bring subprojects to the requsite state of readines. It thus incra effectie project demad by the Poorwt .unicipalities. Di uiblve Impact' 5.11 Overall, the distribution of DRI's expenditures is modestly progressive. The lowest income distribution segments receive a larger than proportional share of public subsidies. Households in the lowest quintile of the rural income distribution received 26.1 percent of DRI's total subsidies. 5.12 The degree of progressiveness, however, is not enough to guarantee that the poor benefit from DRI's expenditures more than the rich. Subsidies received by the poorest households are much larger in proportion to their incomes than those received by the highest segments of the distibution. For example, subsidies average 1.5 percent of income for the poorest quintile of the households, and only 0.1 percent for the richest quintle. But since poor households are larger than rich ones, one finds that the richest quintle benefits as much as the poorest in terms of per capita subsidy (Table 5.5). 4 The detiled prseadon of the resdts of he incidence analysis for the rual progams hre considered can be fund in Annex 5. - 122 - Table 5.5: DRI Distribution of Subsidies, 1991 Subs*d per Subsi per Disibu*ion 0ntie Househoid Capita (percent) 1 8,875 1,557 26.1 2 7,142 1,335 21.0 3 6,191 1,226 18.3 4 5,650 1,215 16.7 5 6,079 1,579 17.9 Source: Perfetfi and Guerra de Mesa (1993). 5.13 Among DRI programs, support to production activities shows the most progressive distribution, followed by infrastructure investment. Marketing is the least progressive program, with the lowest and highest quimles receiving about the same share of total expenditures (23.6 percent and 22.3 percent, respectively).5 Main Constdats and Chalenges. 5.14 Activities fnanced by ural development programs such as DRI are directed not toward the poor, but toward the rural population with good productive potential in the short to medium term. Programs focus on investment, enhancing the earnings of small farmers and other small, rural entrepreneurs through the provision of infrastructure (road links to markets, irrigation water, erosion and flood control) and services (adaptive research, agricultural extension, organization of farmers). DRI's success lies in its decentralized, participatory, and demand driven approach. Despite relatively modest funding, DRI is a potent instrument for aiding productivity and diversity in the rural economy. 5.15 Weak instituional capacity is the biggest constaint on Colombia's rural development program, particularly among poor municipalities. Many municipalites experience difficulties identifying and preparing sound projects. Inexperience in executing and monitoring projects and financial reporting coupled with insufficient capacity to support implementation create a gap between project targets and results. In addition to unequal capacities among municipalities, the poor are disadvantaged in voicing their demands within municipalities. 5.16 The challenge of decentalition dominates analysis of DRI's ability to reach the poorest communities. Decentralization necessitates the relinquishment of authority, limiting the national govermnent's capacity to regulate the program and target its beneficiaries. Under a decentralized system, the national government decides the program's s See Table S. mAmex 5. - 123 - departmental budgets and influences allocations through municipal ranking, project selection criteria, and cofmancing matrices. Beyond the municipal level, the national government's influence is weak. No special mechanisms ensure that an equitable share of resources reach the municipalities' poorer communities.6 5.17 Weak institutions are the main challenge rural development programs face in a decentralized environment.7 Extended coverage and an expanded set of project types could thin DRI's operational capacity. And more responsibilities will be invested in municipalities that already face difficulties under the centralized system. The transfer of project approval and monitoring responsibilities to departmental governments will be a challenge for most. 5.18 DRI's success is influenced by other programs. Access to credit by small producers, for example, constrains their ability to fully participate in DRI's productive programs, including technology transfer, minor irrigation, processing, and marketing projects.8 Inefficiencies and distortions in other programs impose additional constraints on the performance of the rural development program. Conclusions and Recommendations 5.19 Resources allocated to rural development programs fluctuated widely in the past. They were low relative to the needs of rural areas and the agricultural sector and the size of the poverty group. In the long run, more public resources should be channeled into projects that improve rural productivity and competitiveness. In the short run, institutional capacity is the main limiting factor. An acceleration of these programs will first require resolving the institutional issues constraining implementation of the public investment program in rural areas. 5.20 Absorpdve capacity is weak in rural municipalities, and worst in municipalities most needing investments. Given the pace of decentralization, building up local capacity to plan and monitor public investment is imperative. Departmental capacity building to appraise projects and monitor execution is equally urgent. Better systems for monitoring the physical execution of investment projects are also required, together with improved financial controls. Resolving these institutional constraints is critical to increase the flow of public funds for widely scattered, municipality managed, rural subprojects. 5.21 External support, as well as support from the central and departmental govermments, should be directed toward capacity building. The main central government 6 Loy de Competencias y Recursos, however, does ensure that the allocation of resources within municipalities responds to the share of mral population in each municipality. 7 See Chapter 6 for an indepth analysis of the challenges involved in the process of decentralization. 8 Jr. those programs DRI cofinances the supportive infrastructure and technical services and, in some cases, a share of star-up costs. However, the balance on farm or private investment cost is expected to be financed through the banking system. - 124 - agency for rural development, Fondo DRI, has an increasingly large role to play in strengthening the investment planning and monitoring capacity of regional and local entities in charge of ural development investment, and it needs to devote more efforts and resources to this function.9 5.22 Recent reforms could significantly affect DRI's operations. The decision to limit DRI's cofinancing to joint projects with local entities-excluding those involving other centrally funded institutions-should help foster decentralization. The decision to increase coverage to all rural municipalities will doubly challenge DRI. First, its human and financial resources will be stretched. Second, without geographical targeting, other forms of targeting will have to be improved. The incidence analysis shows that a significant share of DRI's expenditures benefit the richest rural households. Third, the new decree extends DRI's authorized activities. With few exceptions, DRI currently possesses little experience in the new activities. In the case of flood control (one of the added activities and typically a high- cost operation), for example, cofinancing rates will have to be high to attract municipalities. 5.23 One major element missing from Colombia's rural development program has been the provision of timely reliable market information to nrual producers. Improving the operations of rural markets is especially important with Apeflura. The govermnent has recently started work on the establishment of such a system. Another weakness of the pacl,age, probably the most important one, is inadequate research and extension services for smail farmers. The government recognizes this, and a new approach, known as SINTAP (National System of Agricultural Technology Transfer), is currently being established. iIL Plan Nacional de Rehlabfilacln (PNR) 5.24 Since its inception in 1982, PNR has been a tool for successive adnistations to pursue rural political and social objectives. In its earliest years, it primarily provided economic assistance to zones troubled by guerrilla violence. Today, PNR gives social and economic assistance to municipalities with high incidences of poverty or violence, emphasizig community development and insttutional coordinaon. As PNR's role has expanded, so has its geographical coverage. In 1992, the program operated in 404 municipalities covering some 7.7 million people, or 25 percent of Colombia's population. Most of those municipalities are small and mrual, with 323 below 25,000 inhabitants. In many, campesino and indigenous communities predominate. 5.25 PNR establishes dialogue between the state and communities that, for reasons of poverty or violence, have been alienated. Consequntly, PNR places great importance on 9 Mhe Rur Developentm Inves=men project (financed by the Wodd Bak) is now spportng fte developmen of tainig modoles, plamng tools, and monitoribg systems adapted to the needs of municpal governmens with low maname capaciy. DRY is also peparig sh componen to increase departml govermeMn' apaity to manage centraly fnded Invmen pm,icts. - 125 - responding to communities. It does not apply rigorous criteria to influence rural investment patterns. 5.26 A municipality graduation system was instituted in 1991. Municipalities were classified into four categories according to relative economic development. The poorest were eligible for full PNR assistance for up to six years. Wealthier municipalities received assistance for no more than two years, and only for institution building. PNR Spending 5.27 Direccid6n PNR's (DPNR) 1991 and 1992 investment program was smaller than DRI's. Nonetheless, these resources constitute one-sixth of total resources spent in PNR municipalities. The rest originate in cofinancing by other institutions, such as INCORA, DRI, HIMAT, and FNCV. DPNR's spending has been significantly reduced since 1988 (Table 5.6). Table 5.6: PNR Program Spending, 1988-92a (millions of 1992 pesos) Year Total 1988 75,162.7 1989 52,620.1 1990 41,287.9 1991 23,125.6 1992 17,104.3 a. Corresponds to budget appropriations for Direccidn PNR. Source: Perfetti and Guerra de Mesa (1993). 5.28 Because of nrral needs, PNR primarily finances infrastructure projects such as roads, bridges, and sanitation. Such projects absorb about half its budget. The importance of production investments in the PNR portfolio declined from 17.2 percent in 1983-86 to 7.5 percent in 1991-92. In contrast, the relative significance of social and community development projects increased from 27.8 percent to 41.9 percent during the same period. 5.29 Nine departments absorb half of PNR's resources: Antioquia, Santander, CaquetA, Meta, Cauca, Choc6, Norte de Santander, Sucre and Huila. This is consistent with the fact that these departments have the larger number of qualifying municipalities. In certain cases, the distribution of resources also reflects the higher capacity of local authorities. 5.30 Special Projects. PNR's largest program is Special Projects, representing nearly half of PNR's 1991 budget. Its purpose is to finance local rural development and to solicit the active participation of communities in selection, design, and execution of projects. - 126 - Ten percent of the Special Projects budget is set aside for indigenous communities. Each department is assigned a budgetary quota, which is then distributed among competing municipal demands. Communities identify and elaborate project proposals, which are submitted to PNR-Bogota for financing. One of the attractions of this fund is the relative ease of its approval and disbursement procedures. Once approved by PNR, project supervision and follow-up are the responsibility of the community. 5.31 Special Projects require that the projects be in the rural sector, last no longer than a year, involve at least three million pesos of PNR support, and not be under consideration by another national executing entity. Eligible projects are wide-ranging: roads, water and sanitation, construction (for education, health, and storage), reforestation, and electrification. 5.32 Increasingly, Special Projects act as liaisons between communities and national executing agencies. Finance for these projects from sources other than PNR's central budget (such as HIMAT and FNCV) has become much more significant in recent years. Direct PNR finance to Special Projects has declined from 63 percent in 1988 to 8 percent in 1992. Distributve Impact 5.33 The distribution of PNR's spending among income groups is more progressive dtan DRI's. The first quintile of the income distribution receives 28 percent of total subsidies. In absolute terms, this group receives higher subsidies even when larger family size is considered. Nevertheless, as with DRI, households in the highest quintile receive per capita subsidies as high as those in the second quintile, because of smaller family size (Table 5.7). Table 5.7: DPNR Distribution of Program Subsidies, 1991 (pesos) Subsid per SUbd per DIstron | Quinte Household Capita (percent) t 1 9,565 1,678 27.8 2 7,313 1,367 21.2 3 6,585 1,304 19.2 4 5,710 1,228 16.6 5 5,221 1,356 15.2 Source: Perfeti and Guerra de Mesa (1993). - 127 - 5.34 Among the different components, social development (including education and other social programs) is the most progressive, followed by production activities.10 The lowest quintile of the rural income distribution receives 37.4 percent and 30.7 percent of total subsidies in these two subprograms respectively. Special Programs are the least progressive component. The poorest 20 percent of households receive only 24.6 percent of total subsidies. The fact that this component is growing in importance implies PNR's overall targeting could diminish. Conclusions and Recommendations 5.35 Under the approach initiated by Decree 2132, the focus of PNR activities will be on organizing municipalities, informing communities of their rights under the new constitution, and of their entitlements under public programs. PNR thus should continue to operate under its mandate to secure social and political links with communities alienated from the state because of violence and extreme poverty. 5.36 Future PNR activities will enable communities to participate more fully in municipal decisionmaking. PNR's Consejos Municipales de Rehabilitaci6n (CMR) will perform the functions of the Comites de Desarrollo Rural Municipal (CDRs) as a forum of discussion between municipal authorities, local communities, and national development agencies. The Mayor will chair CDRs, formulate the Plan de Desarrollo Municipal, and establish priorities among the projects to be submitted to the National Cofinancing System. 5.37 It is important that PNR transfer the expertise it acquired in operating with local community groups to departments and municipalities. This will require strengthening its regional presence, particularly in areas in the process of pacification. Their needs are less related to specific investment requirements than to the positive effects that PNR's presence could have consolidating local institutions. 5.38 Cuts in PNR's budget are expected in 1994. PNR's financing role will likely be reduced to the Special Projects program, and it is unclear whether this program will be retained. In this new context, however, Special Projects could be a useful vehicle to reinforce institutional development, especially in the most disadvantaged communities. In its more specialized role, PNR could become an essential complement to the local interventions of other rural programs and to the outreach of social investment programs in the marginal communities. IV. Insdtuto Nacional Colombiano de la Refonna Agraria (INCORA) 5.39 Established in 1961, INCORA administers agrarian reform: land acquisition, titling, distribution, and the delivery of aupporiug services. From its inception to 1990, it acquired 1.3 million hecares through purchase or expropriation, and another 4.8 million 10 See Table 5.2 in Aniex S. - 128 - hectares through the annulment of ownership of underutilized or abandoned land.1" INCORA distributes land under two programs. One, operating in agrarian reform zones, is targeted almost exclusively to campesinos with inadequate land. 12 The other program, which distributes relatively larger tracts of public lands, is not targeted toward the poor. Between 1961-90, INCORA distributed 1.07 million hectares to more than 60,000 families in the agrarian reform zones and issued more than 300,000 titles covering 9.2 million hectares of public land. In addition, more than 256 reserves, covering 25 million hectares and benefitting some 37,000 indigenous families, have been established by INCORA. 5.40 INCOR4 has been the principal agent in the Colombian land market. Up to two-thirds of the value of its land purchases have been financed with five-year government bonds, the balance with cash. Sales to farners have been heavily subsidized. The value of the land is financed with a 15-year INCORA loan, with a two-year grace period, at an annual interest rate of 4 percent. The total estimated effective subsidy of the land sale has been as high as 80 percent. 5.41 During 1984-87, INCORA concei1trated its operations in colonization zones of Arauca, Meta, CaquetA, and Magdalena Medio. In 1988-91, emphasis was shifted to the Atlantic coast and areas in Cauca and Tolima with indigenous population. In 1991, INCORA had a presence in 237 municipalities with a total population of 6.7 million. INCORA Spending 5.42 Since 1984, INCORA's investment program has been significantly larger than DRI's. Since 1990, however, INCORA's most important funding source (revenue from agricultural import duties) has decreased sharply because of trade liberalization. As a result, its investment budget fell by 60 percent in 1991. Under the proposed Land Reform Law (Draft Law 114) submitted to Congress, 1 percent of the national budget would be earmarked for INCORA. 5.43 The agency's effectiveness has been low. It has suffered from a shrinking budget and increasing administrative costs which, as a share of total expenditure, have inacreased from 16 percent in 1987 to nearly 40 percent in 1991. Its ratio of staff to annual issued land parcel is 1:31.13 Within the investment component, acquisition, distribution, and titling-the core of INCORA's program-constitute approximately 80 percent of the spending. It Under current Colombian law, the State may annul the ownership of land that been detemined to have been underutized or abandoned for three years. 12 Ihe determination of the agrian reform zones is made annually by INCORA's Junta Directlva, headed by the Minister of Agriulture. Crteria for selectng participating municipalities within these zones include the Gini coefficient of the conation of land ownership, quality of ifie indicators, and the demand for and availability of land. In 1991, there wete 79 agrian reform zones, encompassing 237 municipalities. 13 In 1992 staff was reduced by about 1,200, pursaing the specialization of INCORA's functions. - 129 - Table 5.8: INCORA Program Spending, 1987-91 (milions of 1992 pesos) Administmton Debt Investment Year (percent of total) (percent of total) (percent of total) TOtal 1987 10,818 (16) 12,816 (19) 42,150 (64) 66,079 1988 10,818 (15) 9,680 (13) 52,649 (72) 73,136 1989 18,550 (2)) 8,935 (10) 60,964 (69) 88,444 1990 18,041 (21) 6,836 (8) 62,359 (71) 87,236 l 1991 17,966 (37) 5,693 (12) 24,786 (51) 48,445 ource: Perett and uerra de Mesa (199). Disiutdive Impact 5.44 INCORA's subsidies are concentrated among the bottom 60 percent of the income distribution. The top 40 percent receives less than 7 percent of the program's subsidies. This pattem is largely the result of the targeting character of INCORA's main component-land acquisition and titling-which establishes strict eligibility criteria.'4 Within the lowest three quiniles, the distribution of subsidies is close to neutral. The second quintile receives a somewhat higher subsidy than the first and third. Table 5.9: INCORA Distribution of Program Subsidies, 1991 (pesos) Subsd per Subsidy per DiAuibuin Qindke Hou8ehold Ct (percet) 1 16,416 2,880 31.9 2 17,222 3,219 33.4 3 14,493 2,870 28.2 4 1,646 354 3.2 5 1,767 459 3.4 Source: Perfetti and Guerra de Mesa (1993). 14 Thse rests ate based on inforabon that does not alow clear ideniaion of possible ermr of Inusion m dte land acquiston and iding component As a result, the rsults shown in the text may overstate the level of progr_ssivns in the distuton of INCORA's subsidies. - 130 - Main Conshraints and ChalUenges 5.45 Despite INCORA'S efforts, the agrarian stucture remains very skewed. landlessness is a severe problem. An estimated 200,000 farm families are without land. Another 750,0000 families are mininfwistas, with too little land to make an adequate living.15 Although the share of mrual population has declined rapidly in the last 30 years, absolute numbers of rural people remain the same. Because of rapid population growth, the pressure on land still exists, despite rapid migration to cities and a 20-year old land reform program. Between 1960-88, the concentration of land ownership decreased only slightly; the Gini coefficient declined from 0.87 to 0.84.16 Nevertheless, in areas where the concentration of land ownership traditionally has been high Bolivar, Atlantico, C6rdoba, Magdalena, and Boyaca, INCORA's strong presence has been effective. The departmental Gini coefficients ranged between 0.86 and 0.90 in 1960 and fell to between 0.69 and 0.78 in 1988. 5.46 In recent years reform efforts have been hindered by the agrarian counter- reform resulting from increased land holdings by narcotraffickers who, according to recent estmates, have bought more land than INCORA did in the last 25 years. In certain regions (Uraba, C6rdoba, Magdalena Medio, Cesar, and Norte de Santander) the agrarian conflict also involves guerilla and paramilitary forces, limiting reform's implementation. 5.47 The 1993 Market Based Land Reform Prog . Under draft Law 114, the administrative approach to land reform would be replaced with a strategy that relies on the market. INCORA would facilitate land transactions, with direct negotiations between the buyer and the seller. INCORA's administrative budget has been severely cut, since its role has been strictly limited to land matter: allocation of public lands, land acquisition, titling, and distrbution; the establishment of reserves for indigenous communities; and the annulment of ownership of underilized land. Complementary rural development assistance to land reform beneficiares, including ifrstucture, credit, social, and techmcal assistance programs, will now be provided by specialized agencies (such as Fondo DRI for rural development and PIS for social investments). 5.48 Land subsidies are also being reduced. In the agrarian reform zones, beneficiaries will be entitled to a subsidy for the purchase of one unidad agricola familiar (UAF).' The land reform beneficiary (landless worker or minif ) will receive a IS CONPES, DNP, Ministeo de Agruta (1993). 16 Mlnterio de Agf_utr (1990). 17 A UAF is the bnd area required to yield the equivalent of iee mbinmu salaes-about 18 bectes on average, though acal UAF iu vay widely between areas. - 131 - subsidy, equivalent to a maximum of 70 percent of the value of one UAF.X' This amount will be paid to the seller in government bonds. The balance of the sale price will be paid in cash, from credit obtained by the beneficiary from the banking system. To this end, a special line of credit of US$50 million will be established by FINAGRO. Larger farmers entitled to buy public land would not qualify for the land acquisition subsidy and would finance the entire purchase of the land. They would, however, have access to credit from the Corporacion de Ahorro y Vivienda. 5.49 The main constraint on the market-based land reform of draft Law 114 will likely be funding, not land. INCORA reports an adequate supply of land for sale, partly because of rural violence. To resolve the land shortage among the land-poor would require 5.8 million hectares. The land reform program, however, would aim only to reduce that shortage. Its purpose is to enhance the functioning of the land market and to provide options to producers' access to land through other means (rental, sharing arrangements, and so on). Of late, INCORA has had resources to purchase only 90,000 hectares a year, less than half the land currently for sale in the agrarian reform zones. With the resource increase allowed under draft Law 114, INCORA could purchase approximately 200,000 hectares. The government aims to adjudicate nearly 700,000 hectares during the next four years. Conclusions and Recommendaions 5.50 INCORA's specialization in land acquisition, tidtling, and other land matters is a step in the right direction. The proposed new land reform policy relies heavily on market mechanisms. It would resolve most supply-side constraints by reducing cumbersome administrative procedures in land transactions, and providing high de facto subsidies on distributed land. 5.51 Once land transaction processes accelerate under a market-based system, funding will quicldy become the main supply constraint (both for the subsidy and credit components). Land availability is not likely to be a major overall constraint, though it could be in high population pressure zones. In such cases, expropriation with full compensation is an option that could be used as an exception. 5.52 Under the new regime, demand will be relatively constrained by the more rigorous criteria applied to beneficiary selection. These criteria emphasize both farm management ability and the potential for running a small commercial farm. Not all of the landless or land-poor are potential land reform beneficiaries. Some will benefit indirecdy from expanding employment opportunities on the larger number of small commercial farms created by the reform, but land reform is not directed to the most destitute. s FParmers quallfying for the subsidy to buy land must come from a poor ral hosehold whose primry activity is agriculotue, derve at leat 70 percent of their income from agriclture, and use family labor in agricutural acdvities. In addition, the head of the household must be younger than 55 years old, the famfly monthly income must be under two minimum salaries, and toa household assets need be valued at less than one UAP. - 132 - V. Agricwiwuul Credit Bank (Caja de CreMto Agrrio) 5.53 Caja de Cr&Jito Agrario (also called Caja Agraria) was established in 1931, primarily as a development bank for small and medium farmers. It is the chief source of formal credit for small farmers.'9 In 1991, Caja was the nation's largest financial intermediary in terms of active accounts (836,000 million pesos), offices (886), staff (15,000), and clients (over three million). Over the years, its operations have diversified to urban banking, insurance, agricultural input marketing, and the provision of technical assistance. In 1990 Caja's already fragile financial situation seriously deteriorated. Real lending levels fell by 19 percent in 1990 and 37 percent in 1991. The budgeted 1992 lending program represented less than 54 percent of the 1990 program. 5.54 In 1991, because of its financial predicament and its perceived social and economic importance, the government entered into an agreement with Caja that set the basis for recapitaliaton if it would cut its staff by 30 percent and operate under competitive performance criteria.20 As a result, many of Caja's non-banldng development activities are being eliminated or transferred to other agencies. The agreement also specified that interest rates should be linked to commercial rates and that, eventually, these would equal commercial funding rates plus a spread sufficient to cover costs and a normal profit. Lending to Sma Farmers 5.55 Small farmer loans represent approximately 60 percent of Caja's lending portfolio. The sources for such loans are Caja's own resources and rediscounts from FINAGRO.21 Due to its fiancial crisis, Caja's own resources for small farmer credit have fallen. They practically disappeared in 1991, when they represented only 2 percent of total lending. As a result, Gaja's total lending to small farmers in 1991 was only 26 percent of the 1988 volume (Table 5.10). 5.56 Since 1983 small farmer loans have been primarily (65 percent) short or medim-term production loans for a limited number of short-cycle crops. These loans have concentrated heavily in a few departments. More than half the small farmer loans extended in 1991 were in Boyaca (9.40 percent), Cundinamma (16 percent), Huila (11.03 percent), Tolima (11.20 percent), and Antioquia (5.31 percent). 19 A smal farmer is defied (h 1992) as an agricltural producer wih wealth not exceeding Col$7.6 milion (about US$11,000) who has either 75 perent of his assets invested in the agricultural sector or at least two-thids of his family annaul income derved from agricutral actives. Credit to a small faner cannot exceed 50 percent of the value of his asets. 20 Wodd Bank (1992a). 21 ForPmely Pedo Fbadro Agropcaro (FFAD). - 133 - Table 5.10: C4ja Agrana Lending to Small Farmers, 1988-91 (millions of 1992 pesos) Year Totl 1988 214,940 1989 201,540 1990 114,323 1991 55,185 Source: Perfetfi and Guerra de Mesa (1993). Distribuive Impact 5.57 Caja Agraria provides small farmer credit at subsidized rates. In 1991 the average interest rate charged was an estimated 10 percent below market rates. As a result, the small farmer credit program implied an estimated subsidy of 4.5 billion pesos. 5.58 The distribution of subsidies in Caja's lending to small farmers is extremely progressive. More than half the total subsidy accmes to the lowest quintile of the rural distribution, three quarters to the poorest 40 percent of households. It is not possible to determine the distributional impact of Caja's overall lending program. Only 60 percent of it is related in the results shown in Table 5.11. The other 40 percent corresponds to lending to medium and large farmers. Such lending will be directed mostly to the highest quintiles of the rural income distribution, and thus will probably be regressive. But, the interest rate subsidy for those loans is lower than in the case of small farmers. Table 5.11: Caja Agrana Distribution of Program Subsidies, 199l (pesos) Qake Subsidy per Subsidy per Dis_bu _on Household Capita (percent) 1 4,168 731 51.3 2 1,917 358 23.5 3 743 147 9.2 4 825 177 10.2 5 477 124 5.9 Source: Perfetti and Guerra de Mesa (1993). - 134 - Main Constrs and Challenges 5.59 There is general agreement that rural credit markets are currendy a constraint in Colombia. Insuffi'cient access to production and investment credit hinders the intensification and diversification of agriculture. Similarly, lack of capital for new ventures constrains small, rural entrepreneurs. Finally, credit availability is required for complementary government programs to yield their full benefits. For instance, credit for investment and input purchases are a necessary adjunct to agricultural research and extension, and to agro-industries and marketing subprojects promoted by Fondo DRI. The proposed land reform policy calls for purchase of at least 30 percent of the land value on credit by beneficiaries, who will also require working capital for initial input and equipment purchases. 5.60 The provision of subsidized credit through Caja Agraria has constituted the core of policy in this area. The result has been a financial system that lacks competiton, in which lending to small farmers by private institutions has been an unattractive proposition. As a result Caja Agrana has a virtual monopoly in rural credit and is not subject to competitive market pressures that improve performance. This is reflected in poor management and aggravated by political interference, especially at the branch level. 5.61 The incidence analysis indicates that Caja Agraria's subsidies have been progressive. But the effectiveness of these subsidies for poverty reduction is seriously questioned when their overall effects on rural finances are considered. In addition to the crowding out of private sector lenders, both the direct cost to the public sector (represented by Caja Agraria's periodic recapitalization with public resources) and the opporunity cost involved in the forced investments by commercial banks are very high. Alternative uses of those public resources providing cofinancing for infrastructure and social investments are likely to have a stronger impact on the productivity of the rural poor without the distortionary effects of credit policies. 5.62 The approach to rural credit markets is being changed to a more competitive market environment. The new approach, contained in the Caja Agrania reforms, proposes increased interest rates and an increase in the FINAGRO rediscount margin to 90 percent for small farmer credit to attract other financial institutions to lend to small farmers.22 Loan processes will be simplified for small farmers. To lessen credit applications turned down for lack of acceptable collateral, a Guarantee Fund (Fondo Agropecuanio de Garantia, or FAG) has been established with a capital of 10,000 million pesos (US$14 million).23 22 DNP (1992b). 23 PAG guamats up to 80 percent of the vaue of the loan, with die borrower paying an annual commission of I percent of the outtnding guarateed loan vahe. Unlike banks, which eque formal land tides as collaeal, FAG would accept evilence of possession as guaatee. - 135 - Conclusions and Recommendatons 5.63 Subsidized credit is not the way to revitalize rral credit markets. There are much more desirable uses of public funds. Interest rate subsidies aggravate small farmers' credit access problems. Under a market based system, risks are factored into interest rates, which tend to be high for small farmers relative to commercial rates for larger and more secure borrowers. As a result, small farmers with inadequate collateral are not automatically excluded from borrowing from commercial sources. 5.64 Despite intentions to change past policies of relying on credit subsidies to lower the cost of credit to small farmers, the Government of Colombia has recently decided to postpone the adjustment of small farmer interest rates-which were to increase by four points to equate commercial rates-because of the poor economic performance of agriculture during 1991-92. This measure is unlikely to significantly impact on short-term levels of agricultural production, but it could delay the development of a more efficient and broader mrual credit market. 5.65 Colombia has yet to move to a fully market-based credit system. That would entail eliminating interest rate ceilings and involving a plurality of financial intermediaries in the countryside. Rural credit liberalization is overdue. It would result in greater competition, lower informal interest rates, more viable financial intermediaries (because of reduced forced investments), and improved credit access for investment by asset-poor farmers and rural people. It would also release public resources for financing physical and social infrastructure expansion. Any subsidy should be directed toward reducing the transaction costs of lending to small farmers, as is done in Mexico and Chile.24 VI. OvemU1 Conclusions and Recommendatons 5.66 Overall, spending in rural programs has been progressive. The poor receive a larger share than the rich. But a significant share of the subsidies accrue to high-income groups. While in some cases this results from imperfect targeting, in many cases it is the natural result of programs that by design are not targeted. Investments in infrastructure from which the whole community (both rich and poor) benefit are examples. 5.67 Much of the effectiveness of existing rural expenditure programs has been impaired by the lack of coordination of activities. In the past, there have been substantial overlaps among programs run by DRI, PNR, and INCORA. Of 620 municipalities covered by DRI, some 200 were also covered by PNR, and some 150 by INCORA. About 40 municipalities had all three programs in operation. In many cases, these programs brought the same services to the communities, which meant that the efforts of each agency were too widely dispersed. 24 See Chapter 2, paragraph 2.29. - 136 - 5.68 The need to streamline institutions emphasizing comparative advantage and removing duplications and to better target public sector expenditure to disadvantaged areas are considered in the ongoing insttutional reforms that accompany fiscal and administrative decentralization. Under these reforms, considerable order has been put into the institutional framework of rural development in the last two years. Fondo DRI, PNR, INCORA, and Caja Agrana, which all provided support services, infrastructure, or both in support of rural development in the past, have had their functions streamlined. They have become specialized, complementary institutions. Fondo DRI has become the main executing agency for rural development programs, providing funding for a wide range of rural development projects. PNR's role in funding similar projects is winding down. Its main focus is now community mobilization, organization, and training, where it has a comparative advantage. INCORA has stopped providing support to rural development. It now has the more limited role of implementing land distribution. Caja Agraria has eliminated extension services and is concentrating on its main business, rural credit. 5.69 But reform seems to fall short of what is required. Education, health, and nutrition expenditures are not articulated with productive investments in the rural areas. With 70 percent of extreme poverty in Colombia occurring in rural areas, a fully coherent rural development strategy is urgently needed. An essential feature of this strategy is an institutional framework conducive to the broad objective of rural development. 5.70 Institutional mework for Rural Development. The government must adequately coordinate all executng and fmancial agencies in the rural sector to ensure the effectiveness of rural expenditure programs for infrastructure and social services and to avoid misallocation of resources. The integrating principle should be rural development, not strict adherence to subsectoral lines (agriculture versus health, for example), agency responsibilities (such as DRI versus HIMAT), or even levels of government. 5.71 As a first step in this direction, the govermnent has recently established the Municipal Councils for Rural Development to serve as local coordinating units for rural development. The Municipal Councils will bring together municipal authorities, rural community representatves, and public officials involved in productve and social projects and services. If properly implemented, this can aid more rural development, with appropriate consideration of the linkages between natonal, regional, and local rural projects and expenditures. The expected creation of a Vice-Ministry of Rural Development within the Ministry of Agriculture will give increased attention to the small farm sector and improve coordinaton among government programs. Within the central government probably in DNP a unit should be responsible for overseeing the allocation of expenditure programs for productive and social purposes in the rural areas, and for assessing their mutual interaction and performance. 5.72 The Colombian productive rural development programs are well conceived and have been relatively successful. But from a poverty reducton perspective they suffer certain weaknesses. These programs have not effectively reached the economically, socially, or - 137 - geographically marginal segments of society. With decentralization, these groups risk further marginalization. The decetalization of resource allocation will require safeguards to ensure rational poverty reduction goals are reflected in departmental and municipal planiing and budgeting. Minimum percentages for rural expenditures are not sufficient by themselves. Additional incentives, including ligher cofinancing percentages for poor municipalities and poverty reduction oriented projects, need to be included. Finally, whether the program benefits the poor depends on adequate participation in local planning. This will depend on giving adequate priority to special programs implemented by the PNR to involve these groups in the development process. Chapter 6 Poverty Reduction in a Decentralized Setting: Equity, Efficiency, and Accountability The future effectiveness of Colombia's social expenditures and poveny reduction efforts depends on a risky new balance of power between central and local governments. With the decentralization of its public sector, Colombia is shifting the responsibility for delivering social programs for the poor to local governments. Local governments have eagerly embraced the new powers and autonomy. However, effectiveness and efficiency of social spending and management are seriously in question in all but a handful of cities and departments. To induce effective spending on social priorities, the government is developing a framework that combinesfinancial incentives and administrative controls, consisting mainly of: (a) earmarked revenue transfers; (b) matching grants; and (c) regulatory provisions. Even if these instruments work well, there is no assurance that the poorest will benefit from social spending, that resources will be efficiently spent, or even that correction and control mechanisms will right errors and reflect shifting priorites at local and central levels. The biggest risk in Colombia's strategy is the limited capacity of local governments to deliver assistance. This chapter offers guidance-both normative ground rules for policy and operational guidelines for social funds-to take advantage of the potential efficiency benefits of decentralized administrations, and to avoid potential inefficiencies and inequities of local control over spending. Thie next steps for the Government of Colombia and the Bank are instituional development for local governments, departments mediating where municipalities cannot handle things themselves, linking the social funds to municipal development, and mobilizing local governments in more intensive monitoring and evaluation. L Introduction 6.1 As indicated in the three previous chapters, the success of government programs to reduce poverty in Colombia depends on the institutional framework that delivers social services to the poor. This chapter focuses on the challenge to policymakers in merging decentralization and poverty reduction. Section II reviews the recent decentralization process in Colombia and its implications for poverty reduction. Section m proposes a framework to guide and evaluate the provision of social services to the poor within a decentralized system in coming years. The analysis identifies Colombia's weak institutional capacity as the major stumbling block for the ongoing transition between a centralized public delivery of services and a decentralized system. Section IV develops a corresponding strategy for institutional development. - 139 - I. Recent Evolution of Insttuonal Arrangements for Pove y Reducdon' 6.2 Since 1986 Colombia has undergone ambitious decentralization to local and intermediate levels of government. This effort has achieved many milestones, including a new constitution, electoral reform, local elections, functional and financial reform of governments, and most recently, the Ley de Competencias y Recursos (law on local government functions and financing) and the creation of three social funds to operate on a matching-grant basis. By no means completed, the process is slowing as reforms reach lower levels of government, and the pace of decentralization is uneven among regions and municipalities. Matters are likely to remain unsettled for some time. For instance, the financial and institutional strategy to coordinate intergovernmental public spending is still being defined after the early goal of full spending autonomy for each level of government (1986-91) was revised. Implementation is slow. After seven years of decentralization, no municipality is fully responsible for the functions it shoulu fulfill. Once the impact of decentralization on public sector efficiency and the cohesion of the whole process are fialy assessed, there might even be reversals of earlier decisions. As local governments gain autonomy over social spending, many new questions arise about the effectiveness, efficiency, and equity of decentralization. 6.3 The Barco Government (1987-90) made poverty reduction the most important item on its policy agenda. Its national "War on Poverty" was the centerpiece of its development plan. To a large extent, however, poverty reduction programs were planned, executed, and controlled by the central government. More recently, with the Ley de Competencias y Recursos and the matching-grant programs, decentralization and poverty reduction have been brought closer together. Decentralization means transferring some poverty reduction responsibilities to the lower levels of government.2 But with power sharing, buttressed by electoral reforms, the scope of policy options may be significantly reduced since retrenchment in spending is difficult and recission of power sharing is extremely unlikely, at least in the short run. The challenge is to make poverty reduction work under a decentralized system. The fate of the poor depends on the outcome. 6.4 Under the present fiscal system in Colombia, and in most Latin American countries, the most substial and dynamic sources of income remain with the central government. No local or regional government with significant responsibilities can ever be I This chapter is based on findings of the main nission of May 1993, with side visits to the departments of AtIintico, BoyacA. and Valle and to officials of the Distrito Especial. Both depatment officials and municipal authorities were interviewed. Poverty reduction in Colombia is conceived as a nmltisectoral effort touching bealth, education, water, sanitation, utilities, family welfaTe, rural development, and housing. For purposes of clarity and simplicity, this chapter focuses on basic servces-water, seweage, and primary health and education-that all departnents and municipalities tmst, by law, deliver. For the most patt, the analysis and conclusions included here petain to the broader set of services as weil. 2 Among others heath, education, water, sanitation, and low-income housing. - 140 - financially independent.3 Furthermore, revenue sharing schemes severely enervate local tax efforts. Figure 6.1 shows that insufficient-and technically ill-designed-instruments to stimulate local fiscal effort have been of no help in overcoming municipal tax apathy.4 Combined with increasing transfers from central government revenues, the transfer of spending responsibilities to local governments for social and urban development needs will increase the ratio of revenue transfers to own-revenue sources in Colombia. This trend is apparent in Figure 6.2, which shows projections of total municipal income by source.5 These projections are based on optimistic assumptions regarding the growth of local tax revenues and underestimate royalty transfers.6 As a result, the fiscal dependency ratio is likely to be even higher than the one in these figures. 6.5 A more productive goal is to focus on mechanisms that confront users with the costs of services local citizens prefer and with what central governments require and local governments can deliver toward poverty reduction. Local governments should have freedom to set local welfare levels above minimum standards. They should be responsible for covering the cost of services that exceed minimum levels, but expect help when own-source income, levels of wealth, or costs of living depart from national or regional averages. Under the present system, most mayors and municipal councils have no way to detect and close gaps when local welfare departs from national standards of service. 6.6 Moving Toward Expenditure Coordination. Because fiscal interdependence is inevitable, presumed fiscal autonomy has gradually been replaced by expenditure coordination and revenue hannonization, including proposals for several tax and revenue sharing mechanisms. This new goal has meant a redefinition of expenditure responsibilites among different levels of government, widespread imposition of conditionality on all sorts of transfers-including those that were previously conceived as block grants-and the creation of new credit cofinancing mechanisms, most recently consolidated and transformed into tbree social funds. Two main instruments in this transformaton-the Ley de Competencias y Recursos and the social cofinancing fluids-seek to resolve intrinsic tensions between opposing tendencies in social spending, one at the local level to follow local priorities, the other at the national level to meet basic needs for all. They are the key fiscal instruments for 3 The same is tnre of G-7 countries, where local governments finance no more than 40 percent of local functions with own-source revemnes [see Dillinger (1993)]. 4 The Ley de Competendas y Recursos passed by Congress in mid-June 1993 includes some refinem in iricentives for local tax effort. However, since incentives in the new law are relatively insignificant and dfficult to esimate, deparmns and municipalites are not likely to attain significant increases in mncipal current income. 5 The prjections are based on laws, decrees, and bils submitted to Congress by the Executive, as of May 1993. Congress has modified some of these bills. Laws passed in June 1993, and others that are pening will change the ealier projections. But there is no reason to assume the general dependency rati will change substantally. 6 The central govemnent expects a substantial expansion of the property tax base as a result of the self-reporting system being authorized by the bitl before Congress. Judging from the experience of provinces and municipalities that have introduced self-reporting in other countries, the Colombian governmme may be overly optmistc with respect to the medium- to long-term effect of dte self-reporting tectnique. - 141 - Figure 6.1: Munidpal Property Tax Income (mlions of 1980 pesos) 40,000- 35,000n i0.000 25,000X 20.000-E Municipal Transers NProperty Tax 15,000- 10,000 5.000 ,- X _7 X EX_ 1988 1989 1990 1991 1992 Source: DNP. Figure 6.2: Local Government Revenues, 1992-2000 $201 - _ jss $0 1992 2000 Year *own-source *Royalies *Credit *Tranfers and Situado tmCofinance Somrce: Vargas and others (1994). - 142 - managing expenditure and revenues under Colombia's new decentralized system for the provision of social services. 6.7 The Ley de Competencias y Recursos sets the framework for the provision of social services, defiing overall financing and expenditure responsibilities. The law redistributes responsibilities for the delivery of social s-,rvices.' Local governments are assigned the principal executing responsibility, and depirtments mainly have an oversight role. But departments will play an active part in investment, because many local govermnents will not be institutionally or financially strong enough to assume their new roles. The law gives departments responsibility for complementing municipal action, cofinancing, advising, training, or temporarily acting for those local governments unable to assume these duties. This wil be the case for more than lf the country's municipalities (Table 6.1). More than 900 municipalities will likely need help. Additionally, departments are responsible for secondary and tertiary health care, and for specialized education. Departments act as liaisons between central and local governments. They design their regions' development plans-including the expansion of health and education coverage-based on input submitted by each municipality. They must also prepare a decentralizaton plan for their regions before December 1994. This must include a timetable for transferring to departments responsibility of those educational and health facilities previously under cental government jurisdiction. Table 6.1: Local Goverments and Population 1985 Populaion Percentage of Number of Totd Tota Populaton Range Clitesa (thoua) Populatdon Growth Rate Tss than 15,000 655 4,977 18 0 15,000-30,000 216 4,400 16 -1.5 31,000-100,000 97 4,652 17 -2.3 100,000-O00,000 25 4,954 18 4.1 More than 500,000 5 8,169 30 3.2 a. The munber of cities in 1985 is not Identical to the number of municipalities in 1994. Source: Cardenas (1992). 6.8 The prospects for continuous progress in poverty reduction are seriously threaned by the weak institutional capacity of most departments and municipalites. Most are not equipped to take on complicated new responsibilides as intermediate managers in the layering of Colombia's government structre. Aware of this, the government initiated the Municipal Development Program (Programa de Desarrollo Institucional Mwuicipal-PDI). The program is aimed at municipalities with less than 100,000 inhabitants. Its purpose is to provide them with technical assistance in five different areas: planing, financial 7 Annex 6, Table 6.1 detials msponiblfles among differ levels of govememnt. - 143 - administration, development and implementation of projects, management, and community participation. The basic strategy is to establish stable technical groups in the departments, which could in turn provide permanent assistance to the municipalities that require their services. Although the PDI has a record of experience and training, issues like managing intergovernmental spending to achieve equivalent basic services have remained outside its scope. A comprehensive insdtutional development strategy should revise the mandate, expand the scope of work, and broaden coverage to departments.8 6.9 In 1993 the government also created the Misi6n Social to provide technical assistance on decentralization and social spending targeting (especially of health and education services). The Misi6n Social is a three-year project supported by the United Nations Development Programme (UNDP). Its main objectives are: helping selected departments and municipalities build institutional, financial, and planning models to administer their Situado Fiscal resources autonomously; designing and helping to implement effective instruments to target social spending to the poorest and most vulnerable, as required by law; and developing and promoting information systems for the administration of FIS resources by departments and municipalities. The Misi6n Social works closely with the Ministries of Health and Education, DNP, and the FIS in the design and implementation of technical assistance. Currently, its goal is to assist six pilot departments to meet the certification requirements of the Ministries of Health and Education that would allow them to autonomously administer their Situado Fiscal. In doing so, the Misi6n Social is trying to develop a model of technical assistance that could be replicated in other departments. The Misi6n Social has already completed a preliminary diagnosis of institutional development needs in four departments in coordination with sectoral ministries and other central government agencies. Given the experience so far, the major challenges the Mision Social faces are: im1 roving coordination with other technical assistance projects in the Ministries of Health and Education, the Social Funds, the PDI, and PNR; developing effective models that can be easily translated and adapted to other departments and municipalities; defining a strategy for dissemination of the experiences it develops; promoting and financing horizontal cooperation am