Report No. 39251-HO Honduras Public Expenditure Review (In Two Volumes) Volume II: Background Chapters and Annexes November 19, 2007 Poverty Reduction and Economic Management Central America Department Latin America and the Caribbean Region Document of the World Bank ENESF Encuesta Nacional de Epidemiologia IHSS Instituto Hondurefio de Seguridad y Salud Familiar (National Social (Honduras Social Security EpidemiologicalandFamily Survey) Institute) ENIGH Encuesta Nacional de Ingresos y IMF InternationalMonetary Fund Gastos de 10sHogares INB indice de NecesidadesBBsicas EPHPM Encuesta Permanente de Hogares de INE Instituto Nacional de Estadisticas Prop6sitos Mfiltiples (Household (National Statistics Institute) Survey) INFOP Instituto Nacional de Formaci6n ERP Estrategia para la Reducci6n de la Profesional (National Vocational Pobreza Training Institute) ERSAPS Regulating Entity for Water & INJUPEMP Instituto Nacional de Jubilaciones y Sanitation Services Pensiones de Empleados y FDI ForeignDirect Investment Funcionarios del Poder Ejecutivo FDI Fondode DesarrolloInstitucional (Public EmployeesPensionFund) FHIS Fondo Hondurefiode InversibnSocial INPREMA Instituto Nacional de Previsi6n del (Honduran SocialInvestmentFund) Magisterio(Teacher PensionFund) FIDAS Fondo Innovadorpara el Desarrollo y IRA InsuficienciaRespiratoriaAguda la Asistencia Social (Innovation Fund I S 0 InternationalStandardOrganization for Development and Social JSAN Nota del Grupo Consultivo (Joint Assistance) StaffAdvisory Note) FIDE Investment& Export Development LSMS Living Standards MeasurementStudy Fund MCA Millennium ChallengeAccount FONAPROVI Fondo Nacional para la Production y MCC MercadoComb Centroamericano la Vivienda (National Fund for MDG Metas de Desarrollo del Milenio Productionand Housing) (Millenium DevelopmentGoals) FSDP Financial Sector Development Policy MDG Millennium DevelopmentGoals Credit MDRI Multilateral Debt ReliefInitiative GdH Gobierno de Honduras MECOVI Programa de Mejoramiento de GDLN Global Development Learning Encuestas de Condiciones de Vida Network MIGA Multilateral Investment Guarantee GDP Gross DomesticProduct Agency GEF GlobalEnvironmentFund MOE Ministry ofEducation GNI GrossNational Income MOF Ministry ofFinance GoH Governmento f Honduras MOH Ministry ofHealth GPOBA Global Partnership for Output-Based MSME Micro-, Small- and Medium- Aid Enterprises GPS Gasto PublicoSocial NBI Necesidades Bhsicas Insatisfechas HDI HumanDevelopmentIndex (UnsatisfiedBasicNeeds) HIPC Heavily IndebtedPoor Countries NGO Non-GovernmentalOrganizations HIV/AIDS Human immunodeficiency NPV Net PresentValue virudAcquired immune deficiency OBA Output-BasedAid syndrome OECD Organization for Economic HOGASA Hogares Gestores de la Salud Cooperation andDevelopment (HouseholdManagerso fHealth) PAPIN Programa de Apoyo a 10s Pueblos HONDUTEL Empresa Hondurefia de Indigenas y Negras (Program to Telecomunicaciones (Honduran Support to Indigenous and Black TelecommunicationEnterprise) Peoples) IBRD InternationalBank o f Reconstruction PER Public ExpenditureReview and Development PES Programa de Escuela Saludable IDA International Development (Healthy SchoolProgram) Association PHRD Japan Policy and Human Resources IDB Inter-AmericanDevelopment Bank DevelopmentTrustFund IDF InstitutionalDevelopment Fund PIB Product0InternoBruto IFC InternationalFinance Corporation PNUD Programa de Naciones Unidas para el IFPFU International Food Policy Research Desarrollo Institute PPIAF Public-PrivateInfrastructureAdvisory IHNFA Instituto Hondurefio de la Nifiez y Facility Familia (National Institute for PPP Public-PrivatePartnerships Childrenand Family) PRAEMHO Programa de Apoyo a la EnseAanza Media de Honduras (Support to Secondary Education Teaching in SANAA Servicio Nacional de Acueductos y Honduras) Alcantarillados (Water and Sanitation PRAF Programa de Asignaci6n Familiar Company) PRAF Programa de Asignaci6n Familiar SAP Suplidoras de Abastecimiento Popular (Family Allowances Program) (Popular Shops) PRALEBAH Programa de Alfabetizacibn y SEDLAC Base de Datos Socio-Econ6micos Educaci6n de Adultos (Adult Literacy paraAmerica Latinay el Caribe Program) SEFIN Secretaria de Finanzas PRGF Poverty Reduction and Growth SETCO Technical Cooperation Ministry Facility (IMF) SIAFI Financial Management Information PRIESS Programa de Reorganizaci6n System Institucional y Extensi6n de 10s SIERP Sistema de Informaci6n para la Servicios de Salud (Institutional Estrategia para la Reducci6n a la Reorganization and Extension o f Pobreza (Poverty Reduction Strategy Health Services Program) Information System) PRIMHUR Programa de Mejoramiento SMEs Small & MediumEnterprises Habitacional Urban0 (Urban Housing SOPTRAVI Secretaria de Obras Publicas, Improving Program) Transporte y Vivienda (Ministry o f PROEIMCA Programa de Educaci6n Intercultural Public Works and Housing) MultilingUe de Centro America ss Secretaria de Salud (Centro America Intercultural SWAP Programas Sectoriales con Enfoque MultilingualEducation Program) Ampliado (Sector Wide Approach) PROHECO Programa Hondureflo de Educaci6n TSC Supreme Audit Authority (Tribunal Comunitaria (Honduras Community Superior de Cuentas) EducationProgram) UN UnitedNations PRONEEAAH National Education Program for UNAT Unidad de Apoyo Tecnico de la Indigenous and Afro-Antillean Ethnic Secretaria del Despacho Presidencial Groups inHonduras (Technical Unit o f the Ministry o f the PRS Poverty Reduction Strategy Presidency) PRSC Poverty Reduction Support Credit UNDP UnitedNations Development Program PRS-TAC Poverty Reduction Strategy Technical USAID United States Agency for Assistance Credit InternationalDevelopment PVMR Programa de Vivienda Minima Rural WB World Bank (Rural MinimumHousing Program) WBI World Bank Institute ROSC Report on the Observance o f WTO World Trade Organization Standards and Codes s Vice President LCR PamelaCox Director, LCC2C Jane Armitage Director, LCSPR Ernest0 May TaskMana er Ulrich Lachler ACKNOWLEDGMENTS This report was preparedby ateam ledby UlrichLachler (LCSPE) and comprising Mario Cristian Aedo (LCSHE), Katherine Bain (AFCWl), Ana Bellver (LCSPS), Stephen Brushett (LCSTR), Maurizio Bussolo (DECPG), Florencia Castro-Leal (LCSPP), John H. Edwards (consultant), Ricardo Figueroa (consultant), Adrian Fozzard (LCCHN), Hugo Godoy (consultant), Olympia Icochea (consultant), Jorgie Irias (LCSPS), Albert0 Leyton (LCSPS), Humberto Lopez (LCSCE), Jose S. Marques (consultant), Denis Medvedev (consultant), Eduardo Melinsky (consultant), Lars Moller (LCSPE), Miriam Montenegro (LCSHS), Dante Mossi (LCCHN), Christine Lao Peiia (LCSHD), Rajeev Swami (LCSFM), Claudia Varela (consultant) and Fernando Yitzack (JPA). The peer reviewers for his report are Claudia Sepfilveda(DECVP) and Yasuhiko Matsuda (EASPR). Theprincipalauthorsofthe separate chaptersinthe PER are as follows: Chapter 1: UlrichLachler Chapter 2: UlrichLachler, Eduardo Melinsky, Chapter 3: Humberto Lopez, EdwinGoni Pacchioni, Guillermo Raul Beylis; Chapter 4: Maurizio Bussolo, DenisMedvedev, U.Lachler, Lars Moller Chapter 5: John Edwards, Cristian Aedo Chapter 6: Christine Lao Peiia, Hugo Godoy, Olympia Icochea Chapter 7: Ricardo Figueroa, StephenBrushett Chapter 8: Jose Marques, Claudia Varela, ManuelSalazar Chapter 9: Adrian Fozzard, Rajeev Swami. Annexes: Lachler (A, ByC), Lopez (E), Bussolo (F), Edwards (G, H), Lao Pena (I,Figueroa,Bmshett(K,L),Moller(D,StatisticalAnnex). J) The main mission to collect the informationused inthe report visited Honduras inmid- October 2006. The PER team would like to thank the Honduran authorities for their cooperation infacilitating accessto allthe dataandinformationrequestedbythe mission. Financial support provided by the Department for International Development (DFID) of the UnitedKingdom for the preparation of this report is gratefully acknowledged. NOTE On November 26, 2007, the Central Bank of Honduras published a revised series o f national income accounts, based on the 1993 System of National Income Accounting and shifting the base year from 1978 to 2000. An important outcome of this revisionof the national income accounts is to raise the value of Honduras's Gross Domestic Product by approximately 19.3 percent. This Public Expenditure Review was completed before these revisions became public and, therefore, is based on the unrevised macroeconomic figures. Readers of this report are therefore advised to make the appropriate adjustments in any analyses involving ratios o f GDP . HONDURAS PUBLICEXPENDITURE REVIEW VOLUME I1 CHAPTER COUNTRY I. 1 B Key Development Issues andDocument Outline ofthe PER............................................ A Country Context .. CONTEXT AND MAIN FISCALCHALLENGES ........................................... .................................................................................................................. 1 - 6 CHAPTER OVERVIEWOFTHE PUBLICSECTOR AND OF PUBLICSPENDING I1. ............................ 9 B TheEconomic Composition of Central Government Expenditures.................................. A The Composition of Central Government Revenues 11 C The Sector Composition of Central Government Expenditures........................................ ... . ......................................................... 13 17 E. The EfficiencyofPublic SpendingInHonduras.............................................................. D The Institutional Composition ofCentral Government Expenditures............................... 19 -20 F. The HonduranPension System .......................................................................................... 21 CHAPTER 111 FISCAL . GROWTH ......................................................... 27 A The LinkbetweenFiscalPolicies andEconomic Growth................................................. B. The EvolutionofFiscal Policy inHonduras ..................................................................... . REFORMAND ECONOMIC 29 ............................................. 32 D. MainConclusions andRecommendations........................................................................ C. The Effects of Budget Composition on Economic Growth., 38 39 CHAPTER Iv.THECHALLENGEOF ACHIEVINGTHEMILLENIUM DEVELOPMENT GOALS 41 ....... A Modelingthe Attainment ofthe MillenniumDevelopment Goals.................................... .. 43 C. The Available Fiscal Space to FinancePriority Expenditures B Alternative Scenariosfor Attaining the MDGs................................................................. .......................................... 46 D. MainConclusions andRecommendations........................................................................ 49 55 CHAPTER v.EDUCATION DEMAND TEACHER AND ............................ 57 A Sector Overview ................................................................................................................ . SALARIES INHONDURAS B The Demand for Schooling................................................................................................ 59 .. 57 ..................................................................................... D. ThePASCEAgreement andits EstimatedFiscalImpact.................................................. C The Structure of Teacher Salaries 66 E. Will PASCE Buy IncreasedEducation Quality?............................................................... 74 79 F What To Do Instead?......................................................................................................... . G. Main Conclusions andRecommendations ........................................................................ 82 85 CHAPTER VI REVIEWOF HEALTH . ............................................................. A Summary ofMainFindings on Healthinthe 2001PER................................................... . SECTOR SPENDING 91 91 B. Health Sector Overview.................................................................................................... -91 D Access andEquality......................................................................................................... C. HealthExpenditure Analysis ............................................................................................. 96 E Efficiency of Public Health Spending............................................................................. .. 104 108 F. Quality of Services .......................................................................................................... 120 G. MainConclusions and Strategic Implications................................................................. 123 CHAPTER VI1 REVIEW OFPUBLIC SECTOR SPENDINGONTRANSPORT . A The Roads Sub-sector ..................................................................................................... . .................................. 129 132 B. The Ports Sub-sector....................................................................................................... .................................................................................................. 138 D The Railways Sub-sector................................................................................................ C The Airports Sub-sector 140 142 E Transport Subsidies 144 F MainConclusions and Recommendations .... ......................................................................................................... ...................................................................... 145 CHAPTER VI11 REVIEW OF SOCIAL PROTECTIONPROGRAMSAND SUBSIDIES . ...................... 151 A ConceptualFramework.................................................................................................... 151 B Social Risksand Groups-At-Risk .................................................................................... 152 C. The Principal Social Protection Programs .. ....................................................................... D The Main Subsidization ProgramsinHonduras.............................................................. 165 E Assessment ofthe Social Protection System................................................................... 171 F Main Conclusions andRecommendations...................................................................... ... 182 194 CHAPTER ..................199 A Honduras' Experience inManaging Poverty Reduction Spending ............................... B Improvingthe Alignment and Impact of PRS Spending................................................. .. Ix.PUBLICEXPENDITURE MANAGEMENT FOR POVERTY REDUCTION -201 210 D Decentralizing PRS Spending......................................................................................... C. Managing for Results....................................................................................................... 227 E Aligning ExternalAssistance withthe PRS.................................................................... .. 232 239 F. Promoting Participation 244 G MainConclusions and Recommendations . ................................................................................................... ...................................................................... 252 BIBLIOGRAPHY ............................................................................................................................ 259 ANNEXES Annex A Technical Note on the Efficiency of Public SpendinginHonduras ..................... 261 Annex B Budgeted Entitiesofthe National Public Sector ................................................... 271 Annex C A SimpleAnalysis of Debt Sustainability for Honduras ...................................... 273 Annex D Measuring the Gross Domestic Product of Honduras ........................................... 279 Annex E A Framework for Analyzingthe Impact ofFiscalReformon Economic Growth ................................................................................................. 281 Annex F The Maquette for MDGSimulations (MAMS) ..................................................... 293 Annex G ModelingContinuation inthe School System...................................................... 297 Annex H The Fiscal Impact of Early RetirementofTeachers ............................................. 299 Annex I Institutional andGovernance Issues inthe HealthSector Annex J Preparing the Medium-termExpenditureFramework for the Health Sector........309 ..................................... 313 Annex K Recent and Projected Port Traffic inHonduras; 2000-2006................................. 319 Annex L andthe 2003 PPIAF Rep0rt................................................................................... Summary Recommendationsfrom the 2001 and 2007 PERs 321 Table 1. Honduras: Key Macroeconomic Indicators.................................................................. STATISTICAL ANNEXTABLES 325 Table 2 Honduras: Summary Central Government . 326 Table 3. Honduras: Summary Central Government Operations (inmillions of Lempiras) ........................................................................... Operations (as percentageof GDP) ............................................................................. 327 Table 4. Honduras: Summary Central Government 328 Table 5. Honduras: Summary Central Government Operations (as percentageoftotal revenuesor expenditures)..................................... Operations (inmillions of 1995 Lempiras) ................................................................. 329 Table 6. Honduras: Central Government Expenditure by Function (inmillions ofLempiras).......................................................................... 330 Table 7 Honduras: Central Government Expenditure . 331 Table 8. Honduras: Central Government Expenditure by Function (as percentageof GDP) ........................................................................... 332 Table 9 Honduras: Central Government Expenditure .by Function (as percentageoftotal) ............................................................................. Table 10 Honduras: .byFunctionCentral (inmillions of 1995Lempiras) ................................................................ 333 Table 11. Honduras: Central Government Expenditure by Function (inmillions of Lempiras) ........................................................................ 334 Government Expenditure by Institution(as percentageof GDP) ........................................................................ 335 Table 12. Honduras: Central Government Expenditure by Institution(as percentageof total) ......................................................................... 336 Table 13. Honduras: Central Government Expenditure by Institution(inmillions of 1995 Lempiras)............................................................. 337 TABLES Table 1.1 Honduras: Key Macroeconomic Indicators. 2007-07 .................................................... 3 Table 2.1 Honduras: Financial Structure of the NationalPublic Sector Budget, 2006..................49 Table 1.2 Policy-Dependent Debtand Debt Service Thresholds................................................... Table 2.2 Honduras: AdjustedBudget ofthe Non-Consolidated Public Sector, 2006 ................10 Table 2.3 Honduras: Fiscal Balances of the Combined Public Sector......................................... 10 Table 2.4 Honduras: The Composition of Central Government Revenues ................................. -12 Table 2.5 Honduras: Economic Composition of Central Government Revenues........................ 14 Table 2.6 The Evolutionof the Public Sector Wage BillinHonduras ........................................ 15 Table 2.8 The Evolution of the Teacher Wage Bill...................................................................... Table 2.7 The Evolutionof Central Government Employment inHonduras............................... 15 16 Table 2.9 The Proportion of Public to Private Sector Wages....................................................... 17 Table 2.10 Honduras: Sector Composition of Central Government Expenditures...................... 18 Table 2.11 Honduras: Institutional Composition of Central Government Expenditures ............................................................................. 19 Table 2.12 Ranking of Efficiency Scores for Public Spendingin Education andHealth ................................................................................................ -21 Tabel 2.13 Key Parametersof the Social Security Institutes inHonduras, 2006........................ -24 Table 2.14 Estimated Evolutionof Income and Capital of the Social Security Institutes inHonduras .................................................................. 25 Table 4.1 MDGAttainment inHonduras and Latin America ...................................................... Table 3.1 Honduras: Contribution to Growthof SelectedPolicy Areas (percent) ....................... 29 Table 4.2 Determinants ofMDGAttainment inMAMS............................................................. 41 -44 Table 4.3 Baseline Macroeconomic Projections .......................................................................... 45 Table 4.4 Baseline Scenario for MDGIndicators ........................................................................ 46 Table 4.5 The Volume of Public Services Requiredto Achieve the MDGs intheAbsenceofEfficiency Improvements ............................................................... 47 Table 4.6 Structureof Public Financesinthe Baseline andExpenditure Scaling Up Scenarios (percent of GDP) ...................................................................... 48 Table 4.8 Average Efficiency Scores for Public SpendinginEducation andHealth ..................53 Table 4.7 Central Government Tax RevenuesinLatin America & the Caribeean...................... 52 Table 4.9 Sources of Potential Fiscal Savings for the Central Government ............................... -54 Table 5.1 A Half Century of Schooling EvolutioninHonduras: Inequality and Attainment ............................................................................................ 57 Table 5.2 Education Outcomes inCentral and LatinAmerica..................................................... 58 Table 5.3 A Half Century of Schooling EvolutioninHonduras: Inequality and Attainment ........................................................................................... Table 5.4 ProjectedEnrollment Growth2007 to 2017, by Grade (MECOVI 2004) ...................60 63 Table 5.5 Honduras: Projected Increase inPublic Sector .2006 ............................................................ Table 5.6 (1) All Jobs by Level (2006 MINED) ......................................................................... Teaching Expenditures (Aug Lempiras) 65 66 Table 5.7 All Public Schools ...................................................................................................... (2) Teachersby Level Taught inPrincipal Job (2006 MINED)................................. -66 Table 5.8 Public Primary Schools MultipleJobs (MINED2006)................................................ -67 69 Table 5.10 Public Primary Schools Compensation andNumberof Jobs..................................... Table 5.9 All Public Secondary Schools MultipleJobs (MINED 2006)...................................... 70 71 Table 5.12 Average Earnings by Post and Sub-sector................................................................. Table 5.11 All Public Secondary Schools MultipleJobs Held..................................................... 71 Table 5.13 Public School Teacher Professional Training (YO)..................................................... 72 Table 5.14 FullTime Teachers Earning Lessthan 40 LempirasElour ........................................ 72 Table 5.15 OLS Log-Wage RegressionEstimates(t-statistic) ..................................................... 73 Table 5.16 Statutory and Actual Seniority Bonus ....................................................................... 74 . 12, -75 Table 5.18 Ratesof Increase ........................................................................................................ Table 5.17 PASCE Salary Adjustment Program of Aug 2006.............................................. 76 77 Table 5.19 Poverty Reduction Strategy Goals inEducation Table 5.20 EFA/FTIGoals........................................................................................................... ....................................................... 79 Table 5.21 Measuresof Teachers Efforts..................................................................................... 80 Table 5.22 SecondaryEnrollment by Age and Level, 2004......................................................... 83 Table 6.1 Health Outcomes and Outputs, 2001and 2005/06....................................................... 84 Table 6.2 Number of HealthFacilities, 2005 ............................................................................... 92 Table 6.3 TrendsinM O HFacility Expansion, 1990-2005.......................................................... 94 95 95 Table 6.5 Population, GDP, andHealthExpenditures inCentral America.................................. Table 6.4 IHSS Facilities by Type o f Provider in2005 ............................................................... 97 Table 6.6 Total HealthExpenditures and Private Health Expenditures by Financing Sources.................................................................................................. 97 Table 6.7 Public HealthExpenditures as Shares o f Total Public Expenditures andGDP, andper capita Public Health Spending....................................................... Table 6.8 Health Expenditures and Shares by Government Institutions, 1999-2005 99 Table 6.9 Public HealthRecurrent and Capital Expenditure Shares, 1999-2005 ....................... ...................98 100 Table 6.10 Changes inMOH Total andper capita Expenditures, Recurrent 101 Table 6.11 MOH Programs: Trends inExpenditure Shares, 1999-2005.................................... and Capital Expenditure Shares, 1999to 2005........................................................ .......................... 101 Table 6.13 MOH Expenditure Group Shares, 1999-2005.......................................................... Table 6.12 Change inM O HExpenditures by Program Category: 1999-2005 102 Table 6.14 MOH Budget Execution Rates, 1999-2005.............................................................. 102 102 Table 6.15 IHSS Expenditures as a % o f Government Spendingand GDP, Table 6.16 IHSS: Expenditure Shares of Expenditure Categories, 1999to 2005...................... Spendingper Beneficiary, 1999to 2005 .................................................................. 103 103 Table 6.17 Household Health Seeking Behavior during Previous Month ................................. 104 Table 6.18 Reasons for Not Seeking Care Among Those Who SaidThey Were Ill,2004...................................................................... 105 Table 6.19 Problems Cited by Women InAccessing Health Services ....................................... 105 Table 6.20 Average Waiting Time per Consultationby Facility, Table 6.21 Average Household per capita Yearly Expenditures................................................ Income Quintile, andUrban/Rural ........................................................................... 106 106 Table 6.22 Percentage of Households Financedby the MOH Only and By a Last Consultation by Income Quintile ..................................................................... Combination of Own Resourcesand MOHFunds during Their 107 Table 6.23 Regional Health Spending, Population, and Poverty Index, 2005 ........................... Table 6.24 M O H Staff Salaries: Trends from 1996and 2006.................................................... 108 Table 6.25 Ambulatory Consultations per 1000Persons by type of MOH Facility ..................111 114 Table 6.26 Average Daily Medical Care AttentionProvided Table 6.27 MOH HospitalProduction Indicators, 1999to 2005................................................ by Type of HealthCenter ......................................................................................... 114 115 Table 6.28 CMI Productivity Indicators, 1999to 2005.............................................................. 115 Table 6.29 IHSS ProductionIndictors, 1999to 2005 ................................................................. Table 6.30 SelectedCharacteristics and Quality Indicators by Provider Type (2006) 122 Table 7.1 Summary of Transport Sector Expenditures .............................................................. ..............117 129 Table 7.2 ProjectedPublic Sector Expenditures inthe Transport Sector................................... 131 Table 7.3 Desiredversus RequiredInvestment Spending inRoads ........................................... 134 Table 7.4 Tax Revenuefrom Petroleum Derivatives, 2002-2005 .............................................. 136 Table 7.5 Mandated versus Actual RevenuesAllocated to the RoadFund 136 Table 7.6 Operating Expensesof the Road Fund ....................................................................... ............................... 137 Table 7.7 Total Expenditures ofthe NationalPorts Authority (ENP)........................................ 138 Table 7.8 Total Cost of Ports InvestmentsContemplated inMaster Plan.................................. 140 Table 7.9 Summary of Gross RevenueandRoyalties Paidby InterAirports S.A. ..................... 141 Table 7.10 IASA Investments inConcessionedAirports ........................................................... 142 Table 7.12 Budget Execution of Expenditure for Granting Subsidies....................................... Table 7.11 Passenger Volume and Air Cargo Traffic by Airport .............................................. 142 Table 7.13 Summary of Transport Sector Expenditures, 2002-06 ............................................. 144 146 Table 8.2 Malnutritionand DiarrheaAmong Children Under 5 Years ofAge, 2004................153 Table 8.1 PrenatalCare inPrevious 5 Years, 2004 (Women 12-49 years)................................ Table 7.14 Projected Public Sector Expendituresinthe Transport Sector................................. 150 154 155 Table 8.4 Net Primary and Secondary Enrollment Rates, 2004................................................. Table 8.3 Kinder andPreprimary Enrollment Rates, 2004 ........................................................ Table 8.5 Education Indicators................................................................................................... 155 156 156 Table 8.8 Incidence of TeenPregnancies, 2004......................................................................... Table 8.7 Children and TeenagersOccupations, 2006............................................................... Table 8.6 Reasons for Not Attending School, 2004 (7-11 year olds) ........................................ 157 158 Table 8.9 Unemployment andUnderemployment, 2006 ........................................................... ...................159 Table 8.11 Membersof Labor Force with HealthInsurance, 2004............................................ Table 8.10 Population of 65 Years and Older with Pensionand HealthInsurance 160 161 Table 8.12 Inadequateand Overcrowded Housing, 2004 .......................................................... Table 8.13 Households Without Basic Services, 2004............................................................... 162 162 Table 8.14 Households With Access to Water Services, 2004 .................................................. 163 Table 8.16 Indigenous and Black Population............................................................................. Table 8.15 Households With Access to Electricity, 2004 .......................................................... 163 164 Table 8.17 Principal Diseasesthat Affect the Indigenous andBlack Population...................... Table 8.18 Placeo f Deliveryand BirthAttendance, 2004......................................................... 165 Table 8.19 Vaccination inChildren Under 5 Years Old, 2004 .................................................. 167 167 Table 8.20 Honduras: PensionPlans, 2005 ................................................................................ 170 Table 8.21 Beneficiaries ofthe 300 KwhSubsidy, Dec .2005, Oct 2006................................. . 172 Table 8.22 ENEE's Income Statement, 2005-2006 173 Table 8.23 Bono 80 Subsidy, 2006 174 Table 8.24 Fiscal Cost of FuelTax Exonerations for GeneratingElectricity ............................ ............................................................................................. .................................................................... 174 Table 8.26 FuelSubsidies, 2005, 2006....................................................................................... Table 8.25 Subsidiesto Hypothetical Electricity Consumers, 2005/2006 ................................. 175 Table 8.27 Market Price of Fuel,November 2006 ..................................................................... 175 176 Table 8.29 Average Residential Water Tariffs inSelectedCities.............................................. Table 8.28 Estimate Fiscal Costs of Reduction inFuel Taxes. 2006......................................... 176 178 Table 8.30 Water Tariffs. by Type of Consumer. 2006 ............................................................. 178 Table 8.3 1 Price Structure of Telecommunication Services. 2006 ............................................ Table 8.32 HONDUTEL Transfersto the Government. 2004-2006.......................................... 179 180 Table 8.33 Summary of Principal Subsidies. 2005-2006 ........................................................... 181 Table 8.36 Distributionof Social Assistance Resources. by Age Group. 2005 .........................183-184 Table 8.35 Populationat Risk. Program Coverage and Program Cost. 2005...................... Table 8.34 International Comparison of Public Social Spending .............................................. 182 185 Table 8.37 EstimatedIncidence of PRAF Programs .................................................................. 186 Table 8.38 EstimatedIncidence of FHIS Program., ................................................................... 187 Table 8.39 Distributionof Beneficiaries of Education Programs .............................................. 188 Table 8.40 Scholarships .............................................................................................................. 188 Table 8.41 State Subsidized Housing Programs. 2004 ............................................................... 188 Table 8.42 Incidence of the 300 KwhSubsidy 189 Table 8.44 Mass Transportation Subsidies................................................................................. ............................................................................................. ........................................................................... Table 8.43 Fuelfor Cooking. 2004 190 Table 8.45 StudentTransportation Subsidies............................................................................. 190 191 Table 8.46 Households with andWithout Access to PipedWater. 2004 191 Table 8.47 Progressivenessofthe Principal Assistance Programs ............................................ ................................... Table 8.48 Potential Savings From Subsidies............................................................................ 192 Table 9.1 Total PRSP SpendinginHonduras............................................................................. 196 206 FIGURES Figure 2.1 The Evolutionof Real Monthly Wages ...................................................................... Figure 3.1 Twenty-Five Years ofper capita Growth inHonduras (percent) ............................... inthe HealthandCivil Service Sectors 16 27 Figure 3.2 Growth inHonduras versus Growth inLAC (percent) ............................................... 28 Figure 3.4 Public Spending and Social Outcomes inEducation.................................................. Figure 3.3 Public Spending and Social OutcomesinHealth ....................................................... 31 Figure 3.5 Total Public Debtand Central Government Deficit: 1980-2006................................ 31 33 Figure 3.6 Central Government Expenditures, Revenues, and Deficits, 1990-2006 .................... 34 Figure 3.7 Measures of Government Size: ExpendituresandRevenuesof Figure 3.8 Components of Total ExpenditureinHonduras, 1996-2006...................................... theNon-Financial Public Sector (percent of GDP) .................................................... 36 Figure 3.9 Public Expenditure on Wages and Salaries................................................................. 37 37 Figure 4.1 Trade-off betweenEfficiency Improvement and SpendingExpansion Figure 5.1 Secondary Education Gap........................................................................................... to Achieve the MDGsinHonduras ............................................................................ 50 Figure 5.2 Survivor FunctionPlot by Years of Education ........................................................... 58 58 Figure 5.3 Quality: National Student Assessment Results ........................................................... 59 Figure 5.4 Honduras (2004) School Enrollment by Age .............................................................. ............................................... -62 61 Figure 5.6a Salary Distribution, Non-Teaching Administrators .................................................. Figure 5.5 Honduras-Survival inthe School System, byGrade 68 Figure 5.6b Salary Distribution, Non-Administrative Teachers., ................................................ Figure 5 . 6 ~Teaching andNon-Teaching Salary Distribution...................................................... -68 68 Figure 5.7 Primary School RepetitionRates. by Grade in2004 ................................................. -82 Figure 6.1 ProgressTowards Meetingthe MDGsandGovernment Healthand Figure 6.2 Trends inPublic HealthExpenditures..........................................................................93-94 NutritionTargets. Honduras. 1990-2005 .............................................................. 98 Figure 6.3 Trends inPublic HealthFinancingby Funding Source 99 Figure 6.4 MOH Sources of Financing. 1999-2005 ................................................................... .............................................. 100 Figure 6.5 Relationship betweenChild Stunting Levels and Per Capita Health Expenditures inPPP Terms ........................................................................... 109 Figure 6.6 Average GeneralPractitioner Salaries: Figure 6.7 Cost per Bedfor M O Hand IHSS Hospitals ............................................................. Comparison Among 6 CA Countries........................................................................ 112 118 Figure 6.8 Cost per Discharge for MOH and IHSS Hospitals. 2005 .......................................... Figure 6.9 Revenueand Expenditure Trends of IHSS Health InsuranceRegime...................... 118 119 Figure 7.1 Budget Execution Ration inSOPTRAVI.................................................................. 130 Figure 8.1 Illiteracy Rates for the Population Older Than 15 Years. 2006 ................................ 159 Figure 8.2 Average Years of Schooling for the Population Older Than 15 Years..................... 160 Figure 8.3 DifferentialBetweenMarket and FrozenFuelPrice. Figure 9.1 Poverty Reduction Spendingas Percentof GDP inHIPCs ...................................... April 9-November 18. 2006..................................................................................... 176 Figure 9.2 PRS andNon-PRS Spending.................................................................................... 204 204 Figure 9.3 PRS Spendingby Source of Funds........................................................................... 205 Figure 9.4 PRS Spendingby PRS Pillar..................................................................................... 206 Figure 9.5 PRS Structure of Financing by Pilar.2005 to 2007 ................................................... 207 Figure 9.6 PRS Spendingby Economic Classification .............................................................. 207 234 Figure 9.9 MunicipalTransfers and Municipal PRS Projects.................................................... Figure 9.8 FHIS Projects. 2006 .................................................................................................. 234 Figure 9.7 PRS Financing Allocated to Municipalities.............................................................. 234 Figure 9.10 PRS andNon-PRS External Financed Expenditures .............................................. 240 Figure 9.11External Assistance Trends...................................................................................... 240 BOXES Box 1.1 The Looming Energy Crisis inHonduras....................................................................... 5 Box 7.1 Classification of Health Care Interventions According to their Delivery Mode ............................................................................... 176 Box 9.1 The Poverty Reduction Strategy................................................................................... Box 9.2 Do Real Poverty Reduction FundsBringReal Benefits............................................... 200 Box 9.3 .............................................................................................................................................203 Box 9.4 AssessingProgram's PRS Alignment andPerformance Box 9.6 SIAFI............................................................................................................................. 220 Box 9.5 Expenditure Tracking Surveys...................................................................................... .............................................. 213 219 Box 9.7 Legislative Oversight of Public Spending .................................................................... 224 Box 9.8 A Cautionary Note on Result-Based and PerformanceBudgeting 230 Box 9.9 ProgressinHarmonization and Alignment under Education for All............................ ............................ Box 9.10 Congressandthe PRS................................................................................................. 243 246 Box 9.11 PEM Recommendations .............................................................................................. 254 CHAPTER I COUNTRY CONTEXT AND MAINFISCAL CHALLENGES 1.1 Public Expenditure Reviews (PERs) are core diagnostic studies periodically preparedby the World Bank to help countries establisheffective andtransparent mechanismsto allocate and use available public resources in a way that promotes economic growth and helps to reduce poverty. The proposed PER is being undertakenat the request of the Government of Honduras, with the objective of providing useful advice on the allocation of public resources, as well as to develop necessary analytical foundations for the Bank-supported program in Honduras, as contemplated in the Honduras Country Assistance Strategy (CAS). The intended audience of this report comprises government officials, Honduras' civil society and the donor community. The PER aims to contribute to a better understanding of public expenditures and the process of expenditure determination, especially in areas related to the PRSP, and should serve as an analytic inputto assist the processo f aid coordination. A. CountryContext 1.2 Honduras is a lower middle-income country, with a per capita income of US$1,200 (Atlas methodology) and a population of 7.4 million inhabitants in2006. It stands out as one of the most open economies in Central America, with exports of goods and services accounting for around 40 percent of GDP. Furthermore, it has a fairly diversified export base, with the three main traditional export commodities (coffee, bananas and shrimp) currently accounting for less than 20 percent of exports. In spite o f having a very open economy, Honduras also stands out historically as the slowest-growing country inthe region. This largely explains the slow progress made inpoverty reduction, as Honduras i s one the poorest countries inLatin America. Honduras also has among the lowest social indicators in the region, lagging behind in the areas of child malnutritionand education quality. 1.3 The 2004 Living Standards Measurement Survey showed that 51 percent of the total Honduran population was living inpoverty and 24 percent in extreme poverty. These indicators had hardly changed since 1997 in spite o f a steady increase in anti-poverty spending and above average growth since 2002; see Chapter IX. Preliminary household survey results for 2006, however, indicate that poverty rates finally have begun to respond to the higher growth rates in recentyears and fallen significantly. Recent Political Developments 1.4 President Manuel Zelaya assumed power on February 27, 2006, after the Liberal Party defeated the incumbent National Party in hotly contested general elections. The Zelaya Administration has affirmed its commitment to the PRSP, but is facing a very challenging political environment in seeking to implement its government program. Since the winning Liberal Party did not obtain a clear Congressional majority inthe November 2005 elections, the Executive cannot automatically count on the support of Congress, even for ordinary legislation that does not require a qualified majority. Furthermore, the Executive is likely to face greater difficulties thanprevious administrations inensuring the support of even its own party members 1 inCongress. This is a by-product ofthe election law reform in2004, which replacedthe earlier party-based indirect method of electing Congressmen with a process of direct elections. While the process of direct elections is arguably more democratic, it also tends to weaken party discipline. This means that all aspects of PRSP implementation involving legislative action will requirelarge amounts of prior preparation, consultation andpolitical compromise. 1.5 The implementation of the PRSP also faces important operational challenges, as many of the members of the Zelaya administration still need to become familiar with the PRSP. Even though the PRSP was originally preparedby the Liberal Party government that was inpower in 2001, most of the implementation to date took place under the National Party administration of previous President Maduro. Also, the replacement o f government staff with each change in administration, which has become customary in Honduras, exacerbates the loss of policy continuity that inevitably surrounds any government transition. This PER should help to engage the new authorities inapolicy dialogue that re-establishescontinuity with previous reforms. Recent Macroeconomic Developments 1.6 While the Zelaya administration faces some daunting obstacles on the political front, it has been facing a much more benign environment on the economic front.' Macroeconomic developments between2004 and 2005 have beengenerally favorable: real GDP growth averaged around 4.8 percent inthat period, or almost 2 percentagepoints higher than in2001-02. Annual inflation appeared to have stabilized below 10 percent, in spite of a temporary surge in 2004 induced by rising international oil prices. The external sector also exhibited a significant improvement in spite of a significantly higher oil import bill, with the current account deficit (before grants) declining from 7.4 percent of GDP in2003 to 3.1 percent in2005. The country's gross international reserves have increased rapidly, keeping up with the acceleration of import growth in 2004-05 to maintain a ratio of around 4.4 months' worth of imports. These improvements are attributable to a sharp increase in current remittance inflows, as well as to an improved export performance, aided in part by favorable coffee prices. Meanwhile, the real exchangerate has remained fairly stable 2004. 1.7 These positive developments also carried over to the fiscal balances, which had improved significantly in response to the fiscal reforms introduced in 2002 and 2003: the overall public sector deficit declined from 5.2 percent of GDP in 2003 to under 2 percent in 2005. This improvement is due inpart to higher public revenue collections, which rose by almost 2 percent of GDP over the last 3 years, and greater spending discipline. Particularly important inthis last regardhadbeenthe gradual decline inthe size of the public sector wage bill during 2003-05. 1.8 Although it continues to be a source of vulnerability, the financial sector has been undergoing a gradual recovery since 2002. The portfolio share o f non-performing loans has been steadily decreasing over the last three years, while the capital-asset ratio, the provisions against non-performing loans andthe return on equityhave been steadily increasing. 1. This presents a stark contrast to the situation facing the previous administration when it came to power in early 2002. On the political front, the Maduro administration could generally count on a simple majority in Congress for most ordinary legislation. On the economic front, however, the country was still reeling from the impact o f hurricane Mitch and inthe midst o f a post-9-1 1 economic slump, the fiscal balances were deteriorating, the financial systemwas indanger o f collapsing, anda previous programwith the IMFhad fallen apart. 2 Table 1.1 Honduras:KeyMacroeconomicIndicators;2000-07 2000 2001 2002 2003 2004 2005 2006 2007p Annual growth rates (?A) Real GDP 5.7 2.6 2.1 3.5 4.1 4.1 6.0 5.5 Inflation (eop) 10.1 8.8 8.1 6.8 9.2 7.7 5.3 6.0 Real Effective Exchange Rate (eop; + = appreciation) 4.5 3.8 -0.7 -3.0 -1.4 -1.2 __ _- Shares of GDP PA), except when noted otherwise Public Sector Revenues and Grants 29.6 29.1 26.8 27.4 28.4 28.2 28.3 28.0 Public Sector Expenditures 30.3 32.0 30.8 32.3 31.4 29.9 30.3 31.3 o/w wages and salaries 11.6 12.8 13.5 13.2 12.9 12.3 12.5 13.5 Overall Public Sector Balance -0.8 -3.2 -3.5 -5.2 -2.9 -1.6 -2.0 -3.3 Export f.0.b. (% change) 11.5 -2.7 3.5 1.2 15.8 11.4 7.8 8.0 Imports f.0.b. (% change) 9.3 4.4 1.0 8.2 21.1 13.9 20.1 14.4 BOPC A Balance before grants -8.8 -10.4 -7.3 -7.4 -8.9 -3.1 -3.6 -4.6 BOPC A Balance after grants -3.8 -4.1 -3.1 -4.0 -5.9 -0.4 -1.2 -2.5 Gross International Reserves (months o f imports) 4.4 4.9 4.7 3.7 4.5 4.5 4.4 4.4 Source:Central Bank o fHonduras and JMF. Figures for 2007 are World Bankprojections. 1.9 Adding to these generally positive macroeconomic developments is the decline in the country's external public debt, which fell fiom around 70 percent o f GDP at the beginning o f the decade to 50 o f GDP in2005 and an estimated 29 percent at the end o f 2006. At the same time, the country's public sector external debt service fell from 8.3 percent o f exports in 2003 to an estimated 4.3 percent in 2006. These declines reflect the HIPC and beyond-HIPC debt relief provided to Honduras, which reached the HIPC completion point in April 2005, as well as subsequent assistance agreed b the IMF and World Bank under the Multilateral Debt Relief Initiative inlate 2005 and2006. Y 1.10 As a consequence o fthese debt relief initiatives, Honduras's current public debt and debt service ratios are well below the thresholds that signal potential debt distress; Table 1.2. The only indicator in which Honduras comes close to the threshold level i s in terms o f the Net Present Value o f public debt as a share o f GDP. Even here, however, Honduras is far from facing any immediate risk o f debt distress, considering that the quality o f its policies and institutions has earned a "strong" rating inrecent years and that its debt indicators are poised to fall even more in 2007 on account o f additional debt relief to be provided by the IDB under the MDRI. 1.11 In2006, economic growth accelerated to 6.0 percent and the CPI inflation rate declined further, to 5.3 percent. However, there have been a slight deterioration inthe fiscal and external accounts, and the authorities were confronted with several important challenges, discussed below. These challenges eventually resultedinthe expiration o f the PRGF arrangement with the IMF.3 Similarly, the pace o f implementation o f many o f the programs supported by the World 2. The total amount o f debt relief expectedfrom IDA i s SDR808 million(or roughly US$1,200million) innominal terms, which represents about 20 percent of Honduras's' total medium- and long-term public and publicly guaranteedexternaldebt remainingafter HIPC andbeyond-HIPCdebtrelief. 3. A three-yearPRGF arrangementwas approvedon February27,2004 for SDR 71.2 million, ofwhich SDR 30.5 1 million had been disbursedby end-2005. The third review was completed December 16, 2005. An IMF mission visitedHondurasinearly May 2006 to initiatethe fourthPRGFprogramreview, but could not complete the review. Theprogramwas allowedto expire inFebruary2007 without completingthe final reviews. 3 Bank and other multilateral institutions slowed down in2006, as the new authorities have been reconsideringvarious aspects o f the program and o fthe PRSP. Table 1.2 Policy-DependentDebtandDebt ServiceThresholds (percent) Quality of Policies and Institutions Honduras Debt Indicators Poor Medium Strong 2006 NFV Debt as shareof. GDP 30 40 50 29 Exports 100 150 200 73 GovernmentRevenues 200 250 300 152 Debt Serviceas share of Exports 15 20 25 4 GovernmentRevenues 25 30 35 12 Note: Debt indicators above the indicated threshold levels signal a high risk of debt distress. The quality of policies and institutions is measured on the basis of Country Performanceand Institutional Assessments (CPIAs) carried out by the World Bank. The figures for Honduras are fiom the Joint World Bank-IMF Honduras Debt Sustainability Analysis 2006. Source: Debt threshold values are from IDA and IMF (2005), "Operational Framework for Debt Sustainability Assessments inLow-Income Countries -FurtherConsiderations", Table 2, March28. Macroeconomic Outlook 1.12 Honduras continues to benefit from several positive external developments, but its medium-term macroeconomic outlook i s clouded by several major challenges. With respect to the external environment, Honduras concluded all the actions required to fully implement the CAFTA-DR agreement, resulting in the agreement's entry into force as o f April 1, 2006 (together with El Salvador and Nicaragua). Meanwhile, Honduras will also benefit from lower debt service payments to the Inter-American Development Bank (IDB), which also agreed in December 2005 to participate in the MDRI initiative. The additional reductions o f the external debt and the entry into force o f the CAFTA-DR agreement should raise Honduras' attractiveness to private investors andhelp create opportunities for achieving faster growth. 1.13 Less favorable external developments include the deceleration o f construction activity in the USA, which has dampened the growth o f wage remittances inHonduras. Also, the terms at which Honduras can borrow abroad are likely to become more expensive as its per-capita income grows beyond the threshold levels for concessional lending. For example, with a GNI per capita o f US$1,200 in 2006, Honduras already exceeds the IDA lending threshold (the operational cut- off is US$1,025 inFY 2007) by a substantial margin. 1-14 Most importantly, there are several major fiscal challenges that the new authorities must tackle successfully to avoid future macroeconomic destabilization and maintain a sustainable growth environment. These challenges include the rising operating losses o f the state-owned public enterprises and, to a lesser extent, the likely (albeit temporary) decline in tariff revenues as the CAFTA-DR agreement takes effect. The financial losses o f the state-owned electric utility, ENEE, inparticular, are estimated at around 2.0 percent o f GDP in2006, and on a rising trend; Box 1.1. The operating surplus o f state telecom company, HONDUTEL, inturn, has been declining rapidly with the adjustment o f international tariffs in response to the competition generated when the sector was liberalized inDecember 2005. 4 Box 1.1 The LoomingEnergyCrisisinHonduras Inthe early 1990s, Honduras' electricity sector experienced a severe financial crisis, attributable to ENEE's poor performance and a failure to adjust electricity tariffs in spite o f sharply higher debt service obligations (El Cajdn). The crisis was triggered by a severe drought, coinciding with a lack of generation reserve capacity. Inresponse, sector reforms were undertaken in 1994 under a new Electricity Law that (i) established a competitive power market, (ii)separated the roles o f policy, regulation and service provision, (iii) mandated cost-recovery tariffs and (iv) permitted private participation. The reform proved too ambitious for Honduras, however: it succeeded in attracting new private investment, but ENEE's performance deteriorated again and the system became increasingly dependent on thermal generation. Total electricity losses increased over 2001-06, from about 20 to 25 percent; largely on account of commercial losses (15 percent) due to fraud, billing errors and illegal connections. (All other countries in the region, except Nicaragua, exhibit total electricity losses below 15 percent.). Meanwhile, delays inthe procurement o f new generating capacity meant that ENEE has relied regularly on emergency generation capacity to meet demand. As a consequence, Honduras' largely hydro-based generation system o f the 1990s was transformed into a thermal-dominated system. Honduras now depends on imported fuels for about 70 percent o f its power generation, with the result that the cost o f energy doubled between 2001 and 2006. Electricity tariffs, on the other hand, were only adjusted to cover about 81 percent o f the economic cost o f supply. The electricity tariff structure is distorted, with large cross subsidies provided to residential consumers well beyond the lifeline estimated at 130 kwldmonth. Households with a monthly consumption o f 500kwh pay a tariff equivalent to only 65 percent o f the economic cost o f supply. As a result, Honduras is poised once again to suffer an energy crisis similar to the one in 1993. ENEE's operating losses averaged Lps. 2.6 billion per year since 2001 (equivalent to 2 percent o f GDP in 2006) and are projected to reach Lps. 4.4 billion by 2010 if no corrective measures are taken. Meanwhile, the supply/demand balance has tightened up, with a capacity reserve o f only about 5 percent, which substantially raises the risk of unscheduled energy cuts. SuppIy/demand baIance 2007-2010 2,000 peak dimand [Smallrenewables [ [Shortfall [ The main challenge inthe short term is to improve ENEE's financial situation and avoid power cuts. This will require the reduction o f commercial losses, while raising electricity tariffs to cost-recovery levels over a 3-year period and targeting subsidies better. Toward this end, implementation o f a bold program o f energy efficiency i s recommended, together with preparation o f environmental and feasibility studies for new hydro and thermal power plants to reduce the supply-demandgap. The main challenge inthe mediumand long term is to ensure the sector's financial sustainability, mobilize private finance to ensure reliable energy supplies, diversify energy sources away from hydrocarbon fuels and increase access to electricity services by the poor.) Source: World Bank (2007), "Honduras: Power Sector Issues and Options". 1-15 Another critical challenge concerns the evolution of the public wage bill inthe aftermath of the agreement reached with the teachers union inAugust 2006 on the determination of teacher 5 salaries. The government had succeeded in containing wage pressures with the enactment o f public sector pay framework legislationinDecember 2003. That legislation gave greater control over the process o f public sector salary determination to the Executive branch o f government, in part by temporarily suspending several previously negotiated fringe benefits and bonuses. This permitted the central government wage bill to decline from 10.8 percent o f GDP in 2003 to 9.8 percent in2005. Parts o f this legislation were revoked by Congress inmid-2006, and inAugust 2006, the government negotiated a new agreement with the teachers union. Preliminary estimates o f the fiscal impact o f this agreement indicate that teacher salaries will double in nominal terms during 2007-09, which would raise the public wage bill between 1.0 and 1.5 percent o f GDP over the medium term. 1.16 Finally, the government also has beenfacing increased popular pressures inthe aftermath o f recent oil price hikes to freeze gasoline and energy prices, and to expand public subsidies more generally intransportation, fuel and electricity, as well as to expand public credit programs, especially in agriculture. These pressures have resulted in an increase in the total amount o f public sector subsidies granted - from around Lps. 4.0 billion in2005 to Lps. 6.0 billion in2006 -which reduces the fiscal space available for priority spending. B. KeyDevelopmentIssuesand Outlineof the PER 1.17 The preceding overview draws attention to two overarching problems that afflict the Honduran economy and have guided the research agenda of this PER. One problem is the country's historically low growth rate, which largely explains why Honduras ranks among the poorest countries inthe region and exhibits among the worst social indicators. Another problem i s the persistence o f poverty, which has not declined in spite o f significant increases in anti- poverty spending over the last five years. 1.18 Chapter I1 provides an overview o f the public sector in Honduras. It reviews the evolution o f total public revenues and expenditures, as well as the composition o f public expenditures, along economic, institutional and functional dimensions. A section is also devoted to the structure and actuarial situation o f the public pension systems in Honduras, with a view toward identifying potential fiscal risks that could emerge within the next decade. Finally, Chapter I1also summarizes the findings from cross-country research on the efficiency o f public expenditures in Honduras vis-$vis other countries, with the technical details relegated to Annex A. 1.19 Chapter I11focuses on the nexus between public expenditures and economic growth. It begins by reviewing Honduras's growth and fiscal performance over the last decade. This i s followed by benchmarking and econometric analyses that link fiscal reform to economic growth, and seek to identify potential areas o f fiscal reformthat could boost economic growth. 1.20 Chapter IV focuses on the link between public expenditures and poverty reduction, with specific reference to achieving the Millennium Development Goals (MDGs). Using a dynamic general equilibrium model, the first part o f this chapter simulates how close Honduras i s likely to come to achieving the MDGs if it continues to follow the current growth and public expenditure trends. Findingthat Honduras in unlikely to achieve the MDGs by 2015 under this scenario, it also simulates the additional volume priority public spending or improvements in the productivity o f public spending that are required in order that the MDGs can be met by the 6 indicated time. The second part of this chapter estimates the fiscal space available to the Honduran authorities for the purposeof expanding priority public spending without destabilizing consequences. This i s done by examining (i) the scope for expanding the fiscal deficit (by way of a debt sustainability analysis elaborated on in Annex C), the scope for raising tax revenues, the scope for switching expenditures away from non-priority items, andthe scope for improving the efficiency of public spending (which draws on the technical findings summarized inAnnex A). 1.21 Chapter V focuses on the education sector, with special attention to the evolution of teacher wages. The findings inprevious chapterspoint to the large public wage bill inHonduras as a major source of fiscal concern. Teachers represent, by far, the largest and most visible group of public servants (accounting for more than halfof the total public labor force), and so the level of wages paid to the teachers has a direct and very important fiscal impact. Also, the labor arrangements negotiated with the teachers generally set a standard for the rest of the public sector. This has significant implications from a fiscal, equity and sector efficiency viewpoint. As indicated earlier (para. 1.19, the expiration ofthe 2003 public wage framework law raisesthe possibility of an uncontrolled wage expansion and threatens to weaken further the government's capacity to enforce quality standards in education. To address that threat, this chapter estimates the impact of the August 2006 agreement on the future evolution of teacher wages and human resource management inthe sector and explores some options for complementary measures that could mitigate any adverse effects of the August agreement. One possible option that has been proposed is to retire the older, more expensive teachers from the central government payroll, which has drawn attention to the financial capacity of the teacher pension system, INPREMA, and to the rest of the social security system, A discussion on the contingent liabilities presented bythe teacherspension systemis provided inAnnex H. 1.22 Chapters VI reviews public spending on Health sector and Chapter VI1 reviews public spending on Transportation. These sectors were selected on the basis of their expenditure size and because they represent important sources of fiscal concerns. Following Education, Health and Transport are the next two largest sectors interms o f total public expenditures, and all three account for the lion's share of the government's overall spending. In addition, public infrastructure development was found to be one of the two leading growth determinants in Honduras (para. 3.9, the other being education attainment. Inthis context, Transport deserves attention as the most important public infrastructure sector in Honduras in terms of total public investment4 1.23 Chapter VI11 reviews the system of social protection and subsidization in Honduras. It identifies the groups in society that are most vulnerable to risks and that do not have the instruments and means to manage those risks adequately. Then it reviews and assesses the main public programs that are in place to address specific risks. Inthis context, special attention is 4. To prevent the PER from becoming too broad and unfocused, several important sectors and topics were not addressed. These include agriculture & forestry, disaster risk management and the civil service. While rural areas harbor a disproportionately large proportion o f the extremely poor, the agriculture & forestry sectors absorb a very small proportion o f total fiscal resources and, therefore, are not analyzed here. (This sector is being reviewed in a separate sector study.) Disaster management has been omitted for similar reasons, even though Honduras ranks as one o f the countries that is most vulnerable to natural disasters (as became very evident inthe aftermath o f hurricane Mitch). Finally, the PER only touches on civil service issues in the context o f teacher management and remuneration, but leaves a more general civil service assessment to be dealt with in a forthcoming Institutional and Governance Assessment. 7 drawn to the different public subsidies that are provided inthe electricity, fuel, transport, water, agriculture and telecommunications sectors, which are reviewed in terms of targeting accuracy and fiscal impact. 1.24 Finally, Chapter IX focuses on public expenditure managementwith special attention on the link betweenpublic spending and poverty reduction. It starts off by tracing the evolution of Honduras' Poverty Reduction Strategy, including a discussion of the `real' poverty fund that was created to track anti-poverty spending at the time that Honduras reached the Decision Point of the HIPC debtreliefinitiative, andhowthis was transformed into a `virtual' fundin2004. Italso reviews the progress made in developing the integrated financial management system for the central government, together with other reforms to strengthen public expenditure management. Another focus i s the decision process whereby programs and projects are selected for inclusion inthe portfolio of programs classified as part of the Poverty Reduction Strategy Paper (PRSP), and the monitoring and evaluation mechanisms whereby the effectiveness of anti-poverty programs is periodically re-assessed with a view toward rearranging the PRSP portfolio. While the beginning of this chapter is largely concernedwith the quantity of anti-poverty spending, the latter part deals more with issues o f quality, and reviews the role of civil society and donors in the determination of anti-poverty spending, while referring to the findings of the recently concluded Poverty Assessment (World Bank, 2006) on the poverty orientation of PRSP spending. Finally, this chapter also examines briefly the scope for improving the quality of public expenditures, and anti-poverty spending in particular, through greater fiscal decentralization by drawing on recent work from CEPAL (2005).' 5. G6mez Sabaini, Juan Carlos (2005), "Honduras: El Papel de 10s Municipios en el Combate a la Pobreza", (CEPAL, Santiago de Chile), May. 8 CHAPTERI1 OVERVIEW OF THE PUBLIC SECTOR AND OFPUBLIC SPENDING 2.1 The public sector in Honduras at the national level consists o f the central government (comprising 15 ministries, 3 powers o f the State and 27 subordinated agencies), 14 decentralized institutions, 5 social security institutes, 3 public universities, 9 public enterprises and 4 public financial institutions; see Annex B. At the sub-national level, Honduras is divided into 18 departments, and these are divided, inturn, into 298 municipalities. Departmental governors are appointed by the President o f the Republic and serve in an administrative capacity, without legislative or executive functions, financed entirely through the central government budget. In contrast, the municipalities have their own budgets and (limited) taxation authority, and are governed by mayors, deputy mayors and municipal council members that are elected by popular vote every four years. Table 2.1 Honduras: FinancialStructureof the NationalPublic Sector Budget,2006 - (National, non-ccmsolidatedadiusted mblic sector budget figures) No. of Adjusted YOof Domestic Sources External Financing Entities Budget Total Treasury Self Fin. Grants Credits Debt Relief I n million Lempiras; except whet c! noted otherwise ' CentralGovernment 45 47,828 54% 36,529 281 2,188 5,984 2,846 DecentralizedInstitutions 14 1,312 1Yo 399 759 32 122 _. Social SecurityInstitutes 5 11,590 13% 8 11,582 Public Universities 3 2,255 3yo 2,064 191 ____ __-- -. Public Enterprises 9 17,222 20% 2,025 14,655 177 366 -. Public FinancialInstitutions 4 7,824 9% 199 7,036 21 568 _. Total 88,031 100% 41,224 34,504 2,418 7,040 2,846 (% ofTotal) 100% 47% 39% 3% 8% 3% Note: Treasury resources include Lps 1,890million of internal credit received by the public enterprises; external grants includes Lps. 96 million provided from the Millenium Challenge Account to the Central Government. Source: Secretariade Finanzas,PortaldelNuevo SIAFI. 2.2 The national level accounts for around 95 percent o f total public spending. Within that level, the central government accounts for about one-half o f total spending, followed by the public enterprises with 20 percent and the social security institutes with 13 percent; Table 2.1. The central government finances most of its expenditures with domestically raised treasury resources (76 percent), while the remaining expenditures are mainly financed externally, through credits and grants (or debt relief) in an approximately equal proportion. The public enterprises, social security institutes and public financial institutions mainly rely on self financing through the sale o f public services, with little recourse to treasury resources or external financing. The public universities stand out by relying almost entirely on treasury resources, with less than 10 percent o f their total expenditures in 2006 financed through cost recovery, and by having no access to external resources. 2.3 Total non-consolidated spending by the national public sector represents around one-half o f GDP (50.4 percent); Table 2.2. Ofthis amount, 8 percent o f GDP consists o f transfers, mostly to other national public sector institutions, and 5 percent o f GDP consists o f the accumulation o f financial assets. Netting out both o f these items, we arrive at a total adjusted budget for the 9 consolidated non-financial public sector of approximately 37 percent of GDP. By this measure, the size of government in Honduras is somewhat larger than in the other Latin American countries. Another aspect that stands out inTable 2.2 is the size of public spending on personal services, which accounts for 13 percent of GDP and appears to be significantly larger than in other countries inthe region; see Chapter 111. Table 2.2 Honduras:AdjustedBudgetof the Non-Consolidated Public Sector; 2006 Services Goods Fin. Debt Serv.i % of Personal Other Materials Canital Transfers Assets &other ! GDP Central Government 17,072 4,263 2,501 4,060 8,997 85 10,851i 47,829 27.4% Decentralized Institutions 602 468 59 94 80 6 4; 1,313 0.8% Social Security Institutes 1,090 247 359 456 3,126 6,263 51; 11,592 6.6% Public Universities 1,651 153 69 177 45 __ 159; 2,254 1.3% Public Enterprises 2,353 1,234 7,846 1,397 1,623 29 2,740; 17,222 9.9% Public FinancialInstit. 742 515 159 128 168 2,685 3,426: 7,823 4.5% Total 23,510 6,880 10,993 6,312 14,039 9,068 17,231; 88,034 50.4% as % ofTotal Public Exp. 27% 8% 12% 7% 16% 10% 20%; Yoof GDP 13% 4% 6% 4% 8% 5% 10%; Note: All figures refer to the Adjusted 2006 Budget. Source: Secretariade Finanzas,Portal delNuevo SIAFI. 2.4 Table 2.3 describesthe financial balances of the combined public sector, consisting of the non-financial public sector and quasi-fiscal deficit of the Central Bank. This shows a rapid deterioration of the overall balance during 2000-03, accompaniedby an equal deterioration of the primary balance. Fiscal measures taken in 2002 and 2003 succeeded in reversing this negative fiscal trendin2004, contributing to a muchimproved fiscal outlook by 2005. Table 2.3 Honduras:FiscalBalancesof the Combined PublicSector 2000 2001 2002 2003 2004 2005 2006 inpercent of GDP TotalRevenues 29.6 29.1 26.8 27.4 28.4 28.2 28.3 Current Revenues 28.1 27.0 25.5 26.2 27.2 26.6 26.6 Tax Revenues 17.6 17.1 16.8 17.1 18.0 17.7 18.2 Non-Tax Revenues* 3.8 4.1 4.7 5.0 4.9 4.7 5.2 Oper. Bal. ofPub. Ent. 5.1 4.1 2.5 2.5 1.5 2.3 1.4 Interest earnings 1.7 1.7 1.5 1.5 1.7 1.8 1.8 Capital Revenues 0.4 0.3 0.1 0.1 0.3 0.3 0.4 Grants (incl. HIPC) 1.1 1.8 1.1 1.1 1.o 1.3 1.3 Total Expenditures 30.3 32.0 30.8 32.3 31.4 29.9 30.3 Current expenditures 21.4 23.4 24.1 25.6 24.2 23.7 23.3 olw Wages & salaries 11.6 12.8 13.5 13.2 12.9 12.3 12.5 Interest 3.0 2.3 2.2 1.9 2.0 1.7 1.4 Operating lossesBCH 0.0 0.4 0.9 1.1 1.1 0.8 0.5 Capital Expenditures 7.3 7.2 5.5 6.4 6.5 6.0 5.6 NetLending 1.6 1.3 1.3 0.7 0.7 0.2 1.4 Overall Balance -0.8 -3.2 -3.5 -5.2 -2.9 -1.6 -2.0 Primary Balance 0.6 -2.3 -3.3 -4.5 -2.7 -1.8 -2.4 Source: IMF. Non-tax revenuesinclude social security contributions. * 10 2.5 The fiscal deterioration between 2000 and 2003 was attributable in roughly equal measure to a decline in current revenues and an expansion o f current expenditures. The decline inrevenues mainly came about through a fall inthe operating surplus o f the public enterprises, particularly o f the electric utility, ENEE. ENEE's operating costs rose steeply during 2000-03, in part reflecting a growing dependence on thermal generation and rising oil prices, while electricity tariffs were adjusted ina more gradual manner. The operating surplus o f the telecom utility, HONDUTEL,also declined inthis period, though ina lesspronouncedmanner. This was mainly the consequence of declining operating revenues that reflected greater private competition inthe sector, which lowered the price o f international calls. The increase inpublic expenditures, inturn, was mainly attributable to the increase inthe public sector wage bill, which grew from 11.6 percent o f GDP in2000 to 13.5 percent in2002. 2.6 To reverse these trends, the government passed a series o f fiscal reforms during 2002- 2003. These reforms did little to improve the financial balances o fthe public enterprises, but had a significant impact on tax revenues, which increased by about 1percent o f GDP between 2002 and 2005. On the expenditure side, the greatest reform success was achieved by containing, at least temporarily, the growth o f the public wage bill. With the passage o f the public wage framework law in December 2003,6 the public wage bill declined as a proportion o f GDP, from 13.5 percent in 2003 to 12.3 percent in 2005. Meanwhile, the continued decline in debt service payments after reaching the HIPC decision point in2000 helpedto strengthen public finances. 2.7 The fiscal outcome for 2006 shows that overall revenues remained steady as a share o f GDP. While tax revenues have continued to perform well, buoyed by robust economic growth over the last year, the operating surplus o f the public enterprises experienced a sharp downturn related to the spike in oil prices facing ENEE and the liberalization o f the telecom market in December 2005. An important development on the expenditure side has been the abrogation o f the public wage framework law in August 2006, which is likely to result inrising public wages over several years. The overall public wage bill already increased by a modest amount in2006, and is projected to rise by at least another 1 percent o f GDP in 2007. In spite of these developments, the overall fiscal deficit only increased to 2 percent o f GDP in 2006. The main reason that the deficit did not increase more is because o f a decline in public investment spending during 2005 and 2006. Although this investment decline has helped to produce a favorable fiscal balance outcome, it raises concerns about the emergence o f bottlenecks in the government's capacity to provide public infrastructure and other services. A. The Compositionof CentralGovernmentRevenues 2.8 Honduras currently has one o f the highest tax ratios in the Latin American region.' Its tax system relies mainly on four types o f taxes: the income tax, the sale tax, selective consumption taxes and import taxes. The tax structure in Honduras has been subject to detailed 6. Ley de Reordenamiento del Sistema Retributivo del Gobierno Central. 7. The observed tax ratio may be biased upward by a possible under-valuation o f GDP in the national income accounts. Based on the estimates presented inAnnex D, GDP may be around 15 percent higher than is suggested by the official figures, which would reduce the true tax ratio by a correspondingamount. 11 analysis in several IDB reports (e.g., G6mez Sabaini, 2005)8, while the system o f tax administration was analyzed in a recent IMF report. A series o f recommendations emanated from G6mez Sabaini (2005) that focused on measures to (i)raise the overall level o f tax revenues, (ii) reduce distortions induced by the tax system, (iii) the regressiveness o f the reduce tax system, and (iv) strengthen tax administration through improved transparency and simpler norms. With the implementation o f these measures, the author estimated that the total tax yield inHonduras could be increased by around 2.1 percent of GDP inthe short run and by up to 4 percent o f GDP inthe longrun. 2.9 In2002 and 2003, the Honduranauthorities implementeda series o f tax reforms inline with these recommendations. This reforms includedmeasures to broadenthe sales and selective consumption/production tax bases, unify tax rates, close tax loopholes and expand the application o f taxes to the public enterprises. At the same time, a series o f ongoing measures have been implemented to strengthen tax administration. Noteworthy initiatives in this area are the government programs to restructure and professionalize the government tax office (DEI), institutionalize audits, improve arrears collection, apply computerized cross-checking, update the tax payer data base and improve the customs office. Table. 2.4 The Compositionof CentralGovernment Revenues 2000 2001 2002 2003 2004 2005 2006 1Difference percentages of GDP 2002 - 2000 Total revenueand grants 18.7 19.9 19.4 19.6 20.3 20.5 21.5 2.1 Current revenue 17.7 18.2 18.3 18.5 19.2 19.1 19.8 1.5 Tax revenue 16.4 16.2 15.9 16.3 17.1 17.0 18.1 2.2 Directtaxes 3.5 3.8 3.7 3.9 4.6 4.9 5.6 1.9 o/w: incometax 3.5 3.7 3.7 3.7 4.2 0.1 5.4 1.7 Indirecttaxes 12.8 12.4 12.2 12.4 12.5 12.1 12.6 0.3 Goodsand services 10.5 10.0 10.2 11.1 11.1 10.7 11.2 1.o o/w sales tax 5.8 5.4 5.5 6.0 6.1 6.1 7.6 2.2 International trade 2.3 2.5 2.0 1.4 1.4 1.4 1.4 -0.6 Import 2.3 2.5 2.0 1.4 1.4 1.4 1.4 -0.6 Export 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Nontax revenue(incl. interestrevenue) 0.8 0.8 0.9 0.9 0.9 1.2 1.1 0.1 Transfers 0.6 1.1 1.4 1.3 1.3 0.9 0.6 -0.8 Capital revenue 0.0 0.0 0.1 0.0 0.1 0.1 0.0 -0.1 Grants 1.0 1.8 1.1 1.1 0.9 1.3 1.7 1 0.7 Source: SEFINand IMF, 2005 Article lV ConsultationReport. 2.10 Revenue outcomes have improved significantly over the last four years; Table 2.4. Though part o f that improvement can be traced to the relatively favorable growth environment that prevailed over this period, government efforts inreforming the tax system also have played an important role in this outcome. Current revenues increased by 1.5 percent o f GDP between 2002 and 2006, while tax revenues increased by 2.2 percent. The increase intax revenues was brought about mostly through increases indirect taxes. This switch from indirect to direct taxes 8. G6mez Sabain, Juan Carlos (2005), "Honduras: Hacia un Sistema Tributario mls Eficiente y Diversificado", in M. Agosin, A. Barreix and R. Machado, eds. "Recaudar Para Crecer" (Banco Interamericano de Desarrollo, Washington, D.C.). 12 should improve, albeit modestly, the distributional impact o f the tax ~ y s t e m . ~Within the different categories o f indirect tax revenues, it is noteworthy that the amount collected through the value-added tax has increased, while reliance on import taxes has declined. This switch from trade-inhibiting import tariffs to a VAT is generally deemed a positive development interms o f reducing the amount o f tax-induced distortions. 2.11 A matter o f concern, on the other hand, is the observed decline intransfers to the central government, which partly reflects the weakening financial situation o f the main public enterprises, ENEEand HONDUTEL. It is also a reflection o f recent changes inthe tax treatment o f HONDUTEL's net income, which i s now mostly transferred to the central government as income tax payments, instead o f dividendtransfers. B. The EconomicCompositionof CentralGovernmentExpenditures 2.12 The recent evolution o f central government finances broadly follows the same pattern observed in the combined public sector. The deterioration o f the overall fiscal balances in the aftermath o f hurricane Mitch (October 1998) continued until 2003; Table 2.5. Initially, the expanded deficit was accompanied by ample external financing on concessional terms. As this source o f funding eventually became scarcer, however, the government had to rely increasingly on non-concessional domestic financing, which rendered the fiscal deficit unsustainable. The primary fiscal balance had turned negative in 1999 and peaked at -4.4 percent o f GDP in2003. With the adoption o f the stabilizing measures described earlier (para. 2.6), the primary deficit declined again in2004 and 2005, and came out close to zero in2006. 2.13 The increase in the central government deficit following hurricane Mitch i s mainly attributable to the increase intotal expenditures, rather than to declining revenues. This overall expenditure increase was led by increases in current expenditures, which rose by 3 percent o f GDP from 1998 to 2001. Capital expenditures also increased during this period by a similar amount, considering that much o f the reconstruction spending was classified as net lending. By 2003, however, the GDP-share o f total capital spending had returned to pre-hurricane Mitch levels, while the share o f current spending continued to grow, placing an increasing burden on the fiscal balances. The post-hurricane Mitch increase incurrent central government spending i s mainly accounted for by the rapid rise inthe central government wage bill, from a little over 6.1 percent o f GDP in 1997 to almost 11 percent in 2002. This spending increase appears to have had little to do with hurricane reconstruction activities and may have reduced the fiscal space available for other public expenditures. As noted earlier, the continued decline in total capital spending after 2004 raises concerns about the economy's capacity to sustain growth. 2.14 Another significant trend that attracts attention in the economic composition o f central government spending is the decline in interest payments, which exceeded 3 percent o f GDP in 1997 and have since then declines to around 1 percent. This decline reflects the debt relief successively granted by the Paris Club and multilateral creditors under the Heavily Indebted Poor Countries (HIPC) initiative and, most recently, under the Multilateral Debt Relief Initiative 9. Gdmez Sabaini (2005) reports that the Gini index o f the income distribution in Honduras i s 0.543 before taxes and 0.571 after taxes. This regressiveimpact of the tax structureon incomedistribution was mainly attributableto the highdependenceonconsumptionandproductiontaxes inthe Hondurantax system. 13 (MDRI). The interest billonthe domestic public debt also hasdeclined, inpartreflectingfalling interest rates and more disciplined fiscal policies that succeeded in reducing the domestic public debt. Table 2.5 Honduras:EconomicCompositionof CentralGovernmentExpenditures (inpercentofGDP) 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Total Revenue 17.4 19.3 20.0 18.7 19.9 19.6 19.6 20.3 20.5 21.4 20.7 Total Expenditures 20.0 20.4 23.6 24.4 25.2 24.3 25.2 23.3 23.1 22.5 22.7 Current Expenditures 15.1 14.6 15.8 16.4 17.6 18.1 19.2 17.4 17.9 18.6 18.5 Wages and Salaries 6.1 6.9 8.0 8.9 10.0 10.7 10.5 10.0 9.6 9.8 10.2 Goods and Services 2.5 2.0 2.1 2.7 2.8 2.4 2.6 2.2 2.7 2.8 3.1 Interest 3.4 2.8 2.2 1.6 1.2 1.2 1.2 1.3 1.2 1.2 1.0 Domestic Debt 1.6 0.8 0.8 0.4 0.3 0.3 0.3 0.4 0.5 0.4 0.4 External Debt 1.8 2.0 1.5 1.2 0.9 0.9 0.9 0.8 0.7 0.8 0.6 Current Transfers 3.0 2.9 3.5 3.2 3.6 3.8 4.9 3.9 4.3 4.8 4.2 Capital Expenditures 4.5 5.8 5.9 6.6 6.8 5.0 5.4 5.9 5.3 4.0 4.3 Fixed Capital Formation 2.7 3.2 2.7 2.7 2.9 2.3 2.3 2.6 2.0 1.7 2.5 Capital Transfers 1.8 2.6 3.2 3.9 3.9 2.8 3.2 3.3 3.3 2.3 1.8 Net Lending 0.4 0.0 1.9 1.4 0.8 1.2 0.6 0.1 0.0 -0.1 0.0 Primary Balance 0.8 1.7 -1.4 -4.0 -4.0 -3.6 -4.4 -1.8 -1.3 0.3 -1.0 Overall Balance -2.6 -1.1 -3.6 -5.6 -5.2 -4.8 -5.6 -3.1 -2.6 -1.0 -2.0 Net External Financing 1.7 2.1 5.4 2.0 4.0 1.8 2.4 4.2 2.4 1.6 2.0 Net Domestic Financing 1.0 -1.0 -1.8 3.6 1.2 2.9 3.2 -1.1 0.2 -0.6 0.0 Notes:Figuresare expressedon acashbasis. Total Revenues includegrants. Total ExpendituresincludeNet Lending. Source:Ministry ofFinance. TheEvolution of Public Sector Employment and Salaries 2.15 An important development highlighted inthe 2001Honduras PERwas the increase inthe public sector wage bill duringthe secondhalf ofthe 1990s. This increase was traced primarily to the significant pay raises legislated under separate employment regimes (Estatutos) that applied mainly to teachers and health sector workers. Although the Estatutos arguably corrected for earlier real wage losses, the nominal salary increases granted over the period 1998-2001 did not take into account the country's limited fiscal possibilities and the subsequent declining inflation rates. Another finding inthe previous PER was that inthe effort to stem the rising fiscal burden generated by the Estatutos, the salaries of other civil servants - who are less unionized and politically weaker thanthe teachers and doctors -were being squeezed. This ended up distorting the public sectorwages structure and discouragingthe most able staff. 2.16 Table 2.6 describes the evolution of the central government wage bill since 1999. The Central Government wage bill, which had increasedto almost 9 percent of GDP in2000, peaked at 10.7 percent of GDP in 2002. In response to the ensuing fiscal pressures, a public wage fiamework law was passed in December 2003, temporarily suspending the Estatutos. Implementationof this law was largelyresponsible for haltingthe growing trend inthe wage bill ratio. The public wage framework law was partially revoked in July 2006, and the public wage billis once againshowing a risingtendency as a share of GDP. 14 Table2.6: Evolutionof the Central GovernmentWage BillinHonduras 1999 2000 2001 2002 2003 2004 2005 2006 Aspercent of GDP Public wage bill 8.0 8.9 10.0 10.7 10.5 10.0 9.6 9.7 civil servants 1.3 1.5 1.5 1.4 1.5 1.3 1.1 1.1 healthworkers 0.8 1.o 1.2 1.1 1.1 1.o 0.9 1.o other (incl. teachers) 5.8 6.4 7.3 8.2 7.9 7.7 7.5 7.7 As percent of Total Public wage bill 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 civil servants 16.8 16.9 15.4 13.4 14.7 13.3 11.2 healthworkers 10.0 10.8 11.5 10.4 10.1 9.8 9.7 9.9 other (incl. teachers) 73.3 72.2 73.1 76.2 75.2 76.9 79.1 78.91 Source:UPEGBEFIN 2.17 The main sectors driving the evolution o f the central government wage bill continue to be the teachers and health sector workers. The teachers comprise the single largest source o f public employment, accounting for around two-thirds o f the total; Table 2.7. The growth o f teachers and health sector workers has slowed down considerably during 2000-06, compared to the rapid increase in the previous decade. However, they represent an increasing share of total central government employment because the rest of the civil service exhibits a declining employment trend. lo Table 2.7: Evolutionof Central GovernmentEmploymentinHonduras 1990 1992 1994 1996 1998 2000 2002 2003 2004 2005 2006 Civil Servants 26,399 26,314 25,045 21,482 21,324 22,535 21,173 20,998 21,572 18,551 18,514 Health Workers 5,260 5,466 6,268 6,847 7,366 7,858 8,195 8,265 8,693 8,161 8,770 Doctors - 1,784 1,759 1,757 1,751 1,842 2,024 Nurses 869 888 909 912 884 900 Specialists -- 246 259 278 301 309 370 Auxiliaries 3,435 3,609 3,985 4,349 4,540 4,959 5,289 5,321 5,729 5,126 5,476 Teachers (primary & sec.) 30,176 32,829 35,298 40,923 45,025 48,086 - - 49,727 Source: Table 5.4 andWorldBank,HondurasPER, 2001, Table 2.2 2.18 While the dominance o f teachers and health sector workers in determining the evolution o f the public sector wage bill is partly attributable to the relative increase in the numbers employed inthese sectors, even more important has been the relative increase inreal wage rates. Figure2.1 describes the evolution o freal wages amonghealthsector workers vis-&vis other civil servants. Both sectors exhibited an increase in real wages during 1998-2002, only that it was much more pronounced in the case o f the health workers. Real wages in these to sectors remained steady after 2002, andbeganto rise again in2006. 10. These figures on total public employment andremunerationrequire caution since humanresourcemanagement records are kept indifferent government offices. Some of the data is maintainedinthe finance ministry, but other employment data(teachers) is maintainedinthe sector ministries. 15 Figure2.1: EvolutionofRealMonthly Wages inthe Healthand Civil Service Sectors e 10000 = ~ d2 Q 4 8000 a m 3$ 6000 3 - H g ; 4000 - 2000 BIV 0 I I I , I I I I 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: UPEGBEFINandWorld Bank (2001), HondurasPER. 2.19 The change in the teachers wage bill since 2000 obeys a similar pattern as the health workers. Between 2000 and 2006, the real wage bill of teachers increased by 22 percent, or slightly less than the 26 percent growth of real GDP over that period; Table 2.8. (Over the period 1998 to 2000, in contrast, the growth of the real wage bill exceeded the growth of real 1 GDP by a factor of 10.) O fthat total increase inthe teacher wage bill, only about 3 percent was due to an increaseinthe number of teachers, while 18percentwas attributable to increases inthe real wage ratesreceived by teachers. Table 2.8 The Evolutionof the Teacher Wage Bill I 1998 2000 2006 %Change 1998-00 2000-06 Teacher Wage Bill (inconstant 1990Lps.; millions) 412.7 614.0 747.6 48.8% 21.8% I NumberofTeachers 45,025 48,086 49,727i 6.8% 3.4% IAverageteachers monthly wage (inconstant 1990Lps) 654.7 912.1 1073.8 i! 39.3% 17.7%1 Source:Table 5.4 andWorld Bank, HondurasPER, 2001, Table 2.1 and2.2 2.20 The abrogation of the public wage framework law, combined with the August 2006 agreement on teacher wages, has once again raised the prospect of a rapidly growing public sector wage bill. As before, when the Estututos were first passed, such a development would seriously constrain the government's future ability to maintain fiscal discipline, rationalize the public sector salary structure and to improve the quality and poverty impact of public services. Such a developmentwould also take public sector wages further out of line with the wages being paid in the private sector. As shown in Table 2.9, public sector wages in Honduras are considerably higher than those paid on average in the private sector. In the early 1990s, Honduras exhibitedthe highest ratio of public to private sector wages of all the Latin American countries. This ratio diminishedinthe second half of the last decade (although it still remained 16 above the regional median), triggering the passage of the Estututos. By 2005, however, Honduras once again exhibitedamong the highest public to private wage ratios inthe region.l1 Table 2.9 The ProportionofPublicto Private Sector Wages 1990 1999 2005 Argentina na 1.29 1.21 Bolivia 1.17 1.27 1.61 Brasil 1.78 1.89 1.91 Chile na 1.35 1.45 Colombia 1.56 1.97 1.94 CostaRica 1.66 1.73 1.73 Ecuador 1.46 1.46 1.81 ElSalvador 1.77 1.73 1.91 Guatemala 1.92 1.55 1.87 Honduras 1.96 1.53 1.88 Mexico 1.39 1.44 na Nicaragua 1.06 1.67 na Panama 1.68 1.58 1.87 Paraguay (Asuncion) 1.55 1.44 1.76 Peru na 1.21 1.37 Dominican Republic na 1.33 1.21 Uruguay 1.11 1.37 1.64 Venezuela 1.05 1.28 1.66 Average 1.51 1.50 1.68 Source: CEPAL, "Panomara Social de America Latina, 2006", Table 24. Note: These ratioswere derived by dividing the average public wage by the average private sector wage, where both are expressedas multiples of the per-capitapoverty line. C. The Sector Compositionof CentralGovernmentExpenditures 2.21 The most noticeable changes in the sector composition of central government spending over the last decade have beenthe increase insocial sector spending andthe decline inthe public debt service, both as shares of GDP and as shares of total central government spending. While public spending inthe other sectors has fluctuated from year to year, their shares have remained more or less the same. 2.22 The sector that has benefitedthe most from the increase in social spending i s Education. Central government spending on education has more than doubled over the course of the last decade as a share of GDP, from around 3.2 percent inthe late 1990s to 7.3 percent in 2002-05. Total public sector spending on education has increased even more if we take into account the increased spending by the public universities, which rose from 0.7 percent of GDP to 1.1percent over this period. Another sector that was favored by substantial increases in central government 11. When each column in Table 2.9 is sorted from highest to lowest value, Honduras ranks 1st our of 14 LAC countries with data in 1990, 8th out 18 LAC countries in 1999 and 4th out of 16 LAC countries in 2005. In all cases, Honduras scored above the sample meansand medians. 17 spending has been Health, whose share o f GDP increased from around 2 percent in the late 1990sto an average o f 3.3 inthe 2002-05. The increase inpublic spending inthe social sectors was strongly encouraged in the aftermath o f hurricane Mitch, with donors pledgin to help reduce the social and ecological vulnerability o f the region as an overriding goal." Social spending received a further boost under the HIPC initiative, which explicitly links the provision o f debt relief to increased public spending on poverty reducing programs. In Honduras, as in other Central American countries, most programs identified as 'poverty reducing' have tended to be oriented toward the social sectors. 2.23 Much, if not most, o f the growth in education and health sector spending since 1998, however, consisted o f increased spending on wages and salaries. Teachers and health sector workers are by far the two largest groups o f public sector workers, and their labor union representation is correspondingly strong. This has raised real wages in these sectors, without evident improvements inproductivity. It is not clear, therefore, that the increased social sector spending has resulted ina corresponding increase insocial sector outcomes. Table 2.10 Honduras:Sector Comnositionof CentralGovernmentExnenditures 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 as percent of G D P Total Expenditures 21.5 20.2 20.0 20.4 23.6 24.4 25.2 24.3 25.2 23.3 23.1 General Public Services 2.7 3.0 2.8 4.1 4.0 4.8 4.7 3.3 2.4 3.1 3.8 Defense and Public Order 1.8 1.9 1.9 1.7 1.8 2.4 2.5 2.7 2.6 2.5 1.4 Social Services 6.3 5.8 5.5 6.0 7.5 8.5 9.3 9.8 11.6 10.9 10.7 Education 3.3 3.2 3.1 3.6 4.5 5.2 6.0 6.2 7.3 7.3 7.2 Health 2.4 2.0 1.9 1.8 2.4 3.1 3.1 3.3 4.0 3.4 3.2 Social Protection 0.2 0.3 0.2 0.3 0.3 0.2 0.2 0.3 0.2 0.2 0.2 Economic Services 4.6 4.4 4.6 4.7 4.6 4.6 4.7 5.2 6.1 4.5 5.0 Other 5.3 3.8 3.8 2.8 4.2 3.0 2.0 2.4 1.8 1.3 1.2 o/w Interest on Public Debt 3.8 3.5 3.4 2.8 2.2 1.6 1.2 1.2 1.2 1.3 1.2 Special Funds and Programs 0.8 1.3 1.4 1.1 1.6 1.1 1.8 0.9 0.8 1.1 1.0 FHIS 0.5 1.0 1.1 0.8 1.3 0.7 1.6 0.7 0.6 0.7 0.8 ._______________________________________------------------------------------------------ PRAF 0.3 0.4 0.3 0.2 0.2 0.3 0.2 0.2 0.2 0.4 0.2 Memo Item: Pub. Universities 0.7 0.7 0.7 0.9 1.0 1.1 1.0 1.1 1.1 1.1 1.1 as percent of total Central Government Expenditures Total Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 General Public Services 12.4 14.9 13.8 20.2 17.1 19.6 18.7 13.8 9.6 13.2 16.3 Defense and Public Order 8.6 9.3 9.5 8.3 7.5 10.0 10.1 11.1 10.2 10.5 6.3 Social Services 29.5 28.8 27.5 29.6 31.8 34.9 37.0 40.3 45.8 46.7 46.2 Education 15.5 16.0 15.4 17.5 19.1 21.5 24.0 25.6 29.0 31.4 31.1 Health 11.2 9.7 9.4 9.0 10.0 12.6 12.2 13.6 15.9 14.4 14.1 Social Protection 1.1 1.5 1.2 1.5 1.2 0.7 0.8 1.1 0.8 0.9 1.1 Economic Services 21.3 21.8 23.0 23.2 19.3 19.0 18.8 21.5 24.1 19.2 21.6 Other 24.6 18.6 19.1 13.5 17.7 12.2 8.1 9.7 7.0 5.7 5.2 o/w InterestonPublic Debt 17.5 17.1 17.0 13.7 9.4 6.6 4.8 4.9 4.8 5.4 5.3 Special Funds and Programs 3.6 6.6 6.9 5.2 6.6 4.4 7.3 3.7 3.3 4.7 4.3 FHIS 2.1 4.8 5.3 4.0 5.7 3.0 6.4 2.8 2.5 3.1 3.4 PRAF 1.5 1.9 1.6 1.2 0.9 1.4 1.0 1.0 0.8 1.6 0.9 Notes: Figures representactual expenditures on a cash basis. Spending by the public universities i s not included in sector totals. Source: Secretaria de Finanzas. 12. The Stockholm Declaration, Consultative Group for the Reconstruction and Transformation of Central America, May 28, 1999. 18 D.TheInstitutionalCompositionof CentralGovernmentExpenditures 2.24 The two largest budget entities within the public sector, by far, are the ministries of Education and Health. Education currently accounts for around 35 percent o f total central government spending and Health accounts for around 15 percent; Table 2.10. O f the remaining central government entities, only the Ministry of Public Works has consistently recorded more than5 percentoftotal central government spending. 2.25 With regard to expenditure trends, several important changes have occurred since 2001. Spending by the ministry of Education has continued to grow after 2001, in parallel with the overall growth in social sector spending observed earlier. By way of contrast, the share o f central government spending by the Health ministry appears to have stabilized in 2001, after having risen in the immediate aftermath of hurricane Mitch. Two other entities that have significantly increased their share of the total budget since the early part of the decade are the ministryof Governance & Justice andthe ministry of Security. This appears to be reflecting the rising concerns in Honduras with crime and public security. A more recent development is the notable increase in the share of central government spending associated with the Presidency of the Republic. The combined spending share of the Office of the President and the Ministry of the Presidency, which had been less than 2 percent of the total in 2001-02, rose to more than 6 percent in2006 and is budgeted at over 9 percent in2007. Table 2.11 Honduras:InstitutionalCompositionof CentralGovernmentExpenditures 2001 2002 2003 2004 2005 2006 2007 Actual Preliminary Budget Total Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Congress 2.0 1.6 1.8 2.2 2.3 2.5 1.3 Judiciary 1.9 2.4 2.3 2.2 1.8 0.3 3.2 ElectoralTribunal 1.6 1.2 0.8 0.9 1.3 0.2 0.0 Office ofthe President 1.1 1.0 1.5 1.7 2.1 4.3 3.7 Min. ofGovernance andJustice 1.4 1.3 2.8 2.8 4.8 4.3 4.2 Min.ofEducation 24.8 26.9 29.7 31.5 31.0 32.3 36.5 Min. ofHealth 12.5 14.3 16.3 14.5 14.0 13.2 15.6 Min.of Security 4.1 4.4 4.0 4.3 4.0 5.0 5.1 Min. of Presidency 0.5 0.6 1.1 1.1 1.2 1.7 5.0 Min. of ForeignAffairs 1.1 1.1 1.1 1.1 1.o 1.2 1.1 Min. ofDefence 3.5 3.6 3.1 3.0 2.9 3.2 3.3 Min.ofFinance 1.7 1.9 5.0 2.9 4.0 3.5 2.8 Min.ofIndustrv.CommerceandTourism 0.6 0.5 0.7 0.7 0.4 0.3 0.3 Min. ofPub. Works, Transport andHousing 7.1 9.1 8.2 7.9 8.4 6.5 5.2 Min.ofLabourandSocial Security 0.8 0.8 0.8 0.9 1.0 0.9 0.8 Min. ofAgriculture andLivestock 5.4 5.3 4.5 4.3 4.2 3.0 2.3 Min. ofNaturalResourcesandEnvironment 1.5 1.8 1.8 2.1 2.6 1.9 1.6 Min. ofCulture, Art andSports 0.5 0.4 0.4 0.4 0.3 0.3 0.3 Min. of Tourism 0.6 0.8 0.5 0.6 0.8 0.6 0.6 PublicMinistry(public prosecutor) 0.9 1.2 1.0 1.o 1.1 1.2 0.9 Comptroller General 0.2 0.2 0.1 0.1 0.1 0.3 0.1 Contingency/ Other* 21.4 14.6 7.3 8.4 5.2 7.6 1.7 Interest on PublicDebt 4.9 5.2 4.9 5.4 5.3 5.4 4.4 Note: *Includes transfers to FHIS, PRAF, Financial Services of Central Administration & National Registry of Persons. Source: Ministry of Finance 19 2.26 The most pronounced fall in the spending shares presented in Table 2.11 pertains to the ministries belonging to the `Economic Service' sector, namely the ministries o f Economy, Public Works and Agriculture. Total spending by these three ministries, which had been held steady at around 13 percent o f total central government spending during 2001-05, declined to a projected 10 percent in 2006 and 8 percent in the 2007 budget. This decline may be associated with the decline in total central government investment observed earlier; para. 2.9. Another expenditure category that exhibits a declining trend since 2001 is the one denominated `Contingency/Other' . This decline reflects lower spending by the social funds (particularly FHIS spending, which peaked in2001), as well as falls innet lending by the central government. E. The EfficiencyofPublicSpendinginHonduras 2.27 A general assessment o f the efficiency o f public expenditures in Honduras can be obtained from recent research at the World Bank on the measurement o f efficiency in public spending across ~ountries.'~This work involves the construction o f empirically-derived world production possibilities frontiers, relating the amounts o f public spending inparticular sectors to various outcome indicators ineach sector. Inthis context, efficiency is measured as the distance between a country's actual combination o f public spending and sector outcomes in each sector and an efficiency frontier. This distance to the frontier can be measured, alternatively, interms o fthe amount o f public spending made ingenerating a particular output, or interms o f the output generated for a particular level o f public spending. Inboth cases, the index measuring efficiency is constructed in a way that countries on the frontier exhibit an index o f 1.0 and less efficient countries exhibit an index that is between 0 and 1. See Annex A for details. 2.28 Table 2.12 compares the average efficiency scores for public spending on education and health in the Latin American region, based on data from 1996-2002. It suggests that the efficiency o f public spending on education generally i s lower in Honduras than elsewhere inthe region. This finding applies to both the input-oriented as well as output-oriented measures. With respect to healthspending, on the other hand, Honduras's efficiency scores do not appear to be appreciably different from the average efficiency scores observed inthe other L A C countries. As showninAnnex A, these findingsalso apply incomparisons with other regions. 2.29 Recent research at the Inter-American Development Bank14has extended this analysis to measure the operational efficiency o f public spending in terms o f a country's overall development performance in a sample o f Latin American countries. The efficiency scores derived from that research indicate that Honduras ranks significantly below the average and medianefficiency scores inthe region; see Annex A. 13. Herrera, Santiago and Gaobo Pang (2005), "Efficiency o f Public Spending in Developing Countries: An Efficiency Frontier Approach", World Bank working paper, May. 14. Machado, Roberto (2006), "El gasto en 10s paises centroamericanos y Republica Dominicana: eficiencia agregada, eficiencia-insumos y eficiencia-resultados", IDB Working Paper, September. 20 Table 2.12 Rankingof EfficiencyScores f: *PublicSpendinginEducationand Health Average EfficiencyScores in Education Average EfficiencyScores in Health Inputioriented Output-oriented Input-oriented Output-oriented Uruguay 0.97 Uruguay 0.97 Trin. & Tob. 0.89 Chile 0.98 Dominican Rep. 0.80 Argentina 0.75 Chile 0.88 Mexico 0.96 Guatemala 0.70 Barbados 0.72 Mexico 0.79 Trin. & Tob. 0.95 ElSalvador 0.66 Panama 0.70 Dominican Rep. 0.78 Panama 0.95 Chile 0.63 Chile 0.67 Guatemala 0.71 Uruguay 0.94 Peru 0.62 Mexico 0.67 Peru 0.71 Barbados 0.94 Trin. & Tob. 0.62 Trin. & Tob. 0.67 Jamaica 0.70 Jamaica 0.94 Argentina 0.61 Peru 0.66 Ecuador 0.69 CostaRica 0.93 Brasil 0.60 Brasil 0.66 Brasil 0.69 ElSalvador 0.93 Mexico 0.57 Jamaica 0.63 Honduras 0.68 Honduras 0.92 Sample median Paraguay 0.68 Argentina 0.91 Paraguay 0.52 Paraguay 0.61 Costa Rica 0.67 Peru 0.91 Colombia 0.52 Dominican Rep. 0.61 Sample median Panama 0.52 Guyana 0.61 Venezuela 0.66 Nicaragua 0.90 Nicaragua 0.52 Bolivia 0.60 ElSalvador 0.66 Brasil 0.90 Guyana 0.51 Colombia 0.58 Nicaragua 0.65 Ecuador 0.88 Bolivia 0.50 Costa Rica 0.56 Haiti 0.65 Guyana 0.87 Honduras 0.50 ElSalvador 0.55 Bolivia 0.63 Colombia 0.86 Costa Rica 0.49 Honduras 0.53 Colombia 0.62 Dominican Rep. 0.85 Jamaica 0.44 Nicaragua 0.47 Guyana 0.60 Paraguay 0.84 Barbados 0.36 Guatemala 0.47 Barbados 0.58 Venezuela 0.84 Ecuador na Ecuador na Panama 0.55 Guatemala 0.81 Venezuela na Haiti na 0.55 Bolivia 0.75 Haiti na Venezuela na IUruguay Argentina 0.50 Haiti 0.57 Source: Annex A. 2.30 These findings suggest that there is significant room for improving the efficiency o f public spending in Honduras, particularly in education. This is best seen by comparing the efficiency scores inHonduras to the scores achieved by countries ranked at the 75' percentile o f the regional and world-wide distribution of efficiency scores.l5Such a comparison again reveals marked differences in the efficiency o f education spending, but less so in health. From an overall public spending perspective, however, this represents an important opportunity cost for Honduras, considering that total spending on education i s more than twice as large as public spending inhealth. F. The HonduranPension System 2.31 This section briefly reviews the Honduran pension system from a fiscal perspective.16 One motivation for this review is to identify the potential size o f contingent fiscal liabilities represented by the pension system. While the 2001 PER did not identify the pension system as 15. See Table A3 inAnnex A. The comparisonwith countries at the 75th percentileofthe distribution of efficiency scores appears more reasonable as a benchmark for most developing countries. While it looks beyond a simple average, it is not as ambitious as seekingaposition on the efficiency frontier. 16. The coverage and quality of social protection provided by the system of social security inHonduras are treated separately inChapter VI. 21 anobvious source o f fiscal risk, various developments andreforms since that time, as well as the mixed experiences with the pension systems inneighboring countries, suggest that this area may require closer attention. Also, as the public wage bill is once again generating increased fiscal pressure, policymakers have begun to explore the option o f early retirement to mitigate these fiscal pressures. The feasibility o f implementing that option will obviously depend on the financial condition o fthe pension system. 2.32 Honduras has five public social security institutes: e IHSS (Instituto HondureAo de Seguridad Social) mainly services workers in the private sector and public enterprises, e INPREMA (Instituto Nacional de Jubilaciones y Pensiones del Magisterio) is the pension fund for teachers, e IMPREUNAH (Instituto Nacional de 10s Empleados de la Universidad Nacional Autonoma de Honduras) is the pension fund for employees o f the nacional university, e I P M (Instituto de Previsidn Militar) i s the pensionfund for the military, and e INJUPE (Instituto Nacional de Jubilaciones y Pensiones de 10s Empleados y Funcionarios del Poder Ejecutivo) provides pensions for the rest o f the employees inthe central government. The social security institutes are supervised by the National Bankingand Insurance Commission (CNBS). All of them provide pensions under various Pay-As-You-Go systems, while two institutes (IHSS and IPM) also offer health insurance. None o f the public institutes offers a defined-contribution or individual capitalization scheme. 2.33 The five social security institutes exhibit considerable diversity in terms o f size, retirement age, contribution and replacement rates and other pension parameters; Table 2.13. For example, the largest institute (IHSS) has a hundred times more affiliates than the smallest institute (INPREUNAH), while the minimum contribution period to become eligible for a pension varies from 10 years (INPREMA and INJUPE) to 24 years (INPREUNAH), and the total contribution rate varies for 3.5 percent o f base salary (IHSS) to 24 percent (IPM). 2.34 The five institutes also exhibit significant differences in actuarial soundness and, accordingly, pose different contingent liability risks for the central government. In some cases, reforms were recently undertaken to improve their actuarial situation, or at least postpone the emergence o f financial deficits. The fifth row in Table 2.13 presents the estimated "break-even contribution rates" that would allow the different pension systems to break even under the various replacement rate and retirement parameters. Inthe case o f INPREUNAH and INJUPE only minor adjustments would be needed (from an actual contribution rate o f 19 percent to a break-even rate o f 20 percent). Furthermore, IPM appears to be ina very solid actuarial position after a recent reform that raises its contribution rate to 24 percent by 2009, or well above its estimated break-even rate o f 20 percent. 2.35 In the other two institutes, more significant parameter adjustments will eventually be needed. The pensionbenefits offered by IHSS have historically been very modest. Even though the pension benefits have been raised inthe wake o f recent adjustments, they still only require an estimated break-evencontribution rate o f 15 percent. Nevertheless, the current contribution rate o f 3.5 percent i s very far removed from even this modest rate, so that either a significant hike in contribution rates and/or a further reduction in benefits will be needed to maintain the system solvent. At 18 and 19 percent, INPREMA's current contribution rates are significantly higher 22 than those of IHSS. However, its benefits are so much more generous than those offered by IHSS that the contribution rate would have to be around 30 percent to maintain actuarial balance. 2.36 Shortfalls in the actual contribution rates from the break-even rate signal an actuarial imbalance that should eventually lead to financial and fiscal distress. It does not necessarily translate into immediate cash flow problems, however, since most institutes have accumulated reserves that can be drawn on in the event of an operating deficit. To obtain a sense of the degree of urgency with which the pension parameters need to be adjusted inorder to prevent the bankruptcy of the social security institutes, Table 2.14 lists the years inwhich operating deficits are expected to emerge under different projection scenarios, as well as the years in which the capital reserves are expectedto runout for each ofthe institutions inthe absence of any changes. 2.37 As noted in Table 2.14, INPREMA already is exhibiting a financial deficit in 2007 independentof the recent teacher wage agreement (PASCE)." This means that INPREMA is already posing a fiscal burden- instead o f helping to finance the overall public sector deficit, it has to sell assets (including government bonds) to cover its own deficits. Nevertheless, INPREMA as an institution does not appear to be in jeopardy for another 8 to 15 years (depending on whether the average affiliate decides to retire at 53 or at 60),18 when its capital positionwould turnnegative.lg 2.38 IMPREUNAH also is shown to exhibit an operating deficit under Simulation I,even though Table 2.14 had shown earlier that its legal contribution rate is fairly close to its equilibrium rate. The reason for this divergence is that the university system (UNAH) has historically not beenpaying the 12percent employer contribution required by law. Only over the last year, has UNAH contributed about one-third of the mandated employer contribution to prevent an operating deficit. If UNAH were to fulfill its obligations by paying in full the mandated employer contribution (Alternative II),just continue to contribute at least one-third or of the mandated amount (Alternative 111), INPREUNAH's operating balance would remain positive for at least another15 years. 2.39 IPM was poised to generate an operating deficit in 2007 until it was reformed in 2006. That reform consists of several parameter changes that increased the contribution rates, as well as the incorporation of policemen and firemen, who had heretofore been covered under the general civil service pension institute (INJUPE). With this new injection of relatively young contributors and the rise in contribution rates, IPM is not expected to face financial difficulties for at least another 20 years. 17. Infact, the teacher wage agreement should tend to improve the financial situation o f INPREMA inthe short and medium term to the extent that the increase in the wages o f the current teachers boosts the contributions to the pension system. Later, when these teachers retire and demand pension benefits commensurate with the higher contributions, the financial situation o fthe pension system deteriorates rapidly. 18. The retirement age o f 53 was chosen arbitrarilywith a view to the age profile o f the teachers. Ifteachers could be encouraged to retire at that age, it would immediately remove 7,000 teachers f?om the education ministry payroll, which would generate just enough budgetary savings for the ministry to finance the wage increase mandated by PASCE. 19. A separate problem facing INPREMA concerns the investment of its capital reserves. It appears that a large portion o f these reserves have beenplaced as loans to INPREMA's affiliates, with uncertain rates of return. 23 R W 4 mI I 0m 0 ;; m 2 I Institution Estimatedyear whenfirst operating Estimatedyear when net capital deficit occurs position turn negative IHSS Base case 2012 2024 Alternative I 2008 2017 Alternative I1 2010 2020 Alternative IA 2007 2023 Alternative IB 2007 2025 Alternative IIA 2007 2015 Alternative IIB 2007 2016 INPREUNAH Alternative I 2007 >2022 Alternative I1 >2022 >2022 Alternative I11 >2022 >2022 IPM Pre-reform 2007 2014 Simulation I >2026 >2026 Simulation I1 2021 >2026 INJUPE NIA NIA Notes: All simulations assume a stationary population and constant salary mass, and that the rate o f return on capital reserves is 4 percent inreal terms. Institute-specific assumptions follow below: 0 IHSS o Base Case assumes that the current affiliates have contributed to IHSS during 50% o f their working life since age 18 untiltoday, and will also contribute 50% during their remaining working life untilretirement o Alternative Iassumes that the benefits uponretirement are 100%morethan inthe basecase o Alternative I1assumes that the benefits uponretirement are 50% more than inthe basecase. 0 INPREMA o Alternatives I(A and B) assumethat the affiliates wait untilthey are 60 years old to retire, while Alternatives I1(A and B) assumethat the affiliates retire uponreaching 53 years o f age. o Alternatives A (I and 11) assume that the salary base for computing pensions is not affected by the PASCE agreement, while Alternatives B (I 11)assume that it rises by the fillamount. and 0 INPREUNAH o Alternative Iassumes that the UNAH does not make any employer contributions o Alternative I1assumes that the UNAHpays up 100% o fthe requiredemployer contributions o Alternative I11assumes that the UNAHpays only 36% o fthe requiredcontributions 0 IPM o The pre-reform simulation assumes that the affiliates need 18years of service to retire. o Simulation Iassumes that the affiliates need to be 54 years old and have 25 years of service to retire, while Simulation I1assumethat they need only 25 years o f service to retire. o The pre-reform simulation assumes 12 contributions per year and an 11% government contribution, while Simulations Iand I1 assume 13 contributions per year and a government contribution o f 13% in2007, 14% in2008 and 15% as o f2009. 25 2.40 The transfer of contributors from INJUPE to IPM is bound to weaken the financial situation of INJUPE. However, INJUPE startedoff with a fairly solid financial situation, so that no imminent fiscal problems are foreseen. Unlike the military and the teaching professions, which tend to attract life-long career employees, a large portion o f the other civil servants rarely remain in their jobs long enough to satisfy the minimum contribution period o f 10 years to qualify for a pension. This i s inpart linked to the frequent turnover of government staff during political transitions. So, while pension contributions o f these political appointees finance the benefits o f the relatively few pensioners in the system during their brief tenure in government, many will have migrated to other jobs before they become eligible themselves for INJUPE pensions.20 2.41 To summarize, even though all five pension institutions present an actuarial deficit inthe long run, most o f them seem to have enough capital reserves to avoid bankruptcy over the next 10 to 15 years. A possible exception is the teacher pension institute, INPREMA, which is likely to face problems much earlier (within 8 years) in the event that teachers decide to retire earlier thanthey have inpast years. 2.42 The three smaller pension institutes (INPREUNAH, INJUPE and IPM) also are not expected to generate significant operating deficits over the next 15 years. However, the two largest institutes (IHSS and INPREMA) are projected to generate deficits within that time horizon. This raises concerns from an overall fiscal viewpoint. Once these institutes begin to generate deficits, they will be forced to sell assets, rather than contributing to the financing o f government deficits generated elsewhere. INPREMA already appears to be operating with a deficit, while IHSS i s projected to begin generating a deficit on its pension accounts within 5 years. Although these deficits initially remain small, they need to be addressed in a timely mannerbefore developing into a major fiscal burden.21 20. The phenomenon o f high staff rotation, where staff leave before they become eligible for pension benefits also characterizes the IPM. 21. By way o f an illustration, the annual deficit simulated for INPREMA increases fiom US$5 million in the 1st year to US$S million inthe 5th year o f deficitary operations, ifteachers continue to retire at the age o f 60. However, ifteachers are encouragedto retire earlier, the annual deficit wouldjump to US$45 million inthe 1styear to US$59 million in the 5th year. In the case o f IHSS, the simulated annual deficit in the base case increases from US$2.5 million inthe 1styear to US$28 million inthe 5th year. 26 CHAPTERI11 FISCALREFORMAND ECONOMIC GROWTH 3.1 Honduras's economic performance has improved significantly in recent years, with per capita GDP growing at an average annual rate of 1.7 percent over 2002-06; Figure 3.1. This is a remarkable achievement, considering that its per capita GDP growth had rarely averaged more than 1percentover any 5 year periodduringthe last 25 years. Figure 3.1 Twenty-FiveYears of per capita Growth in Honduras (percent)" 3.0 2.0 1.o 0.0 -1.0 -2.0 Note: '' Thefigure reports per capita GDP growthrates smoothed using a 5-year moving average. Source: Own calculations using WDI data. 3.2 While certainly encouraging, this performance should not give rise to complacency. To begin, even if the country manages to sustain this average growth rate over the long run, per capita GDP levels would barely double after 40 years. At this pace, the country's per capita income (in 2006 US$) would only reach a relatively modest level of US$2,500 by 2050, or roughly the per capita income currently exhibited by El Salvador. 3.3 Itis also important to recognize that the recent growth spurt has not beenunusually high compared to the growth performance observed inthe rest ofthe Latin American region.22 Figure 3.2 plots the deviations of Honduras annual growth rates with respect to the median growth rate o f the region. Positive deviations imply that Honduras is growing faster than the typical country in the region (as captured by the median), while negative deviations imply that the region is growing faster on average. Figure 3.2 indicates that Honduras's performance duringthe first half o f this decade has been broadly in line with that of the other Latin American countries. This means that the recent take-off appears to be mostly due to fortuitous conditions that have pervaded the entire Latin American region and not to any country-specific factors. In other words, luck may have beenlargely responsible for the growth acceleration ofthe 2000s. 22. Clearly Honduras' growth rates are also much lower that those observed in some of the successful experiences inEast Asia. Forexample, inrecent years bothChina andVietnam havebeen growingat sustained percapita growth rates above 6 percent. 21 Figure 3.2. Growth inHondurasversus Growth inLAC (percent)" 4.0 I -6.0 - -8.0 Note: */ The figure reports per capita GDP growth rates o f Honduras in deviations with respect to the mediangrowth rate ofthe region each year. Source: Own calculationsusing WDI data. 3.4 The preceding finding suggests that in spite of its recent achievements, Honduras still needs to focus on implementing actions aimed at achieving a faster growth rate. The problem o f low growth in Honduras was recently analyzed in World Bank (2004).23 Drawing on the empirical findings of Loayza, Fajnzylber and Calder6n (2002),24 it argued that the country's comparatively slow growthperformance is not due to insufficient investmentlevels, but to a lack of factor productivity growth. That is, poor growth is due more to inadequate factor allocation, thanto insufficient factor accumulation. From a public finance perspective, this finding draws attention away from possible inadequacies in the total level of public investment spending (which contribute to capital accumulation) and toward inadequacies inthe composition of public investment, or in the quality of public spending in general, as explanations of the poor growth performance. In that context, the analysis identified four broad areas (human capital, public infrastructure, financial market development, and government burden) where Honduras is lagging significantly behind inits development vis-his other Latin American countries and the rest of the world, and where policy advances could have a potentially great payoff in terms of raising growth. 3.5 This chapter mainly focuses on the last of these areas, namely the government burden, represented by the level of government consumption as a share o f GDP, but serving as a proxy for a broader concept of fiscal misconduct. Although improvements inthis fiscal variable are not expectedto have the highest impact on growth inHonduras, they should nevertheless result inan appreciable boost, considering that the historical trend in this growth determinant has been pointinginthe opposite direction. (By correcting this negative trend, economic growth would be enhanced relative to the level predicted in the absence of any corrections.) The simulation findings from Loayza, Fajnzylber and Calderon (2005) indicate that Honduras would grow more slowly (by as much as 0.3 percent per year) in the absence o f any fiscal corrections, and by almost 0.2 percent more quickly inresponse to certain fiscal corrections, for a net improvement 23. HondurasDevelopmentPolicy Review: Accelerating Broad BasedGrowth. World BankReportNo. 28222-HO. 24. Loayza, N.P., P. Fajnzylber and C. Calder6n (2002), "Economic Growth in Latin America and the Caribbean: Stylized Facts, Explanationsand Forecasts," The World Bank. 28 inits growth rate of almost 0.5 percent per annum; Table 3.1.25 The rest of this chapter briefly reviews the main channels through which fiscal policy may affect growth, reviews Honduras's recent fiscal experience and presentsa framework for assessingthe potential impact on growth o f a number of fiscal policy interventions. Table 3.1 Honduras: Contributionto Growth of Selected PolicyAreas (percent) ScenarioPolicy Education Financial Trade Government Public Area Sector Openness Burden Infrastructure Continuing Trend 0.12 0.08 0.05 -0.33 0.73 Sharp Progress " 1.37 0.14 0.00 0.14 1.07 Note: " Sharp progress assumes that the relevant policy indicator is taken to the top 25% percent of the regional distribution. Source: World Bank,NicaraguaDPR(2004), basedon figures fiom Loayza, Fajnzylber and Calder6n(2002) A. The Linkbetween FiscalPolicies and Economic Growth 3.6 In the traditional neoclassical growth models (e.g., Solow, 1956) long-run economic growth i s entirely determinedby the rate of growth of the labor force and an exogenously given rate of technological growth. Accordingly, the level and composition of public expendituresand revenues do not have a significant role to play in determiningthe rate of growth. A subsequent generation of growth models (e.g., Barro (1990), Lucas (1988) and Romer (1986)), focusing on the development ofnewtechnologies andhumancapital, has attractedrenewedattention to fiscal policy as an instrument for influencingthe rate of growth inboththe short and long run. Inthese models, fiscal policy can foster growth and human development through a number of channels, as discussedbelow. 3.7 At the macroeconomic level, fiscal policy can affect growth through the effects that the budget deficit and the existing stock of debt have on economy activity. As argued by Easterly, Rodriguez and Schmidt-Hebbel (2004), fiscal deficits appear to be unambiguously bad for growth, even though the correlation of fiscal deficits with any one indicator of macroeconomic imbalance may be close to zero. Whenever these deficits have been financed with monetary expansions over a prolonged period, inflation has tended to follow, constituting a well known brake to economic growth. When deficits have been financed with debt, other problems often appear, such as the crowding out of private investmentor fiscal insolvency. 3.8 It is sometimes argued that fiscal deficits are endogenous in nature, and that they represent policy outcomes more than the product of deliberate policy decisions, especially in developing countries. Easterly, Rodriguez and Schmidt-Hebbel (2004) have found, however, that more often thannot large fiscal deficits have beenthe result of conscious policy choices and rarelythe product o finternal or external shocks. 25. A similar point is made inthe 2005 Article IV Consultation report of the International Monetary Fund, which noted that "Failure to achieve fiscal sustainability (e.g., due to a reversion to the pre-program system for teachers' salaries and failure to compensate for revenue losses fiom CAFTA and the opening of the telecommunications market) would result in slower growth, rising domestic public debt, and a weakening external position. Slower growth would result fiom lower public and private investment, reflecting lower external financing and investor confidence." 29 3.9 The composition of public expenditures also can exert an impact on growth. In this regard, attention has focused on both the economic and sector composition of public expenditures. With respect to the economic composition of public expenditures, traditional debates focused mainly on the distinction between current and capital spending. Since only capital spending contributes to the accumulation o f productive capital, the conventional wisdom held that public investment was more favorable for growth than public consumption spending. While this result has been questioned in earlier empirical work (e.g., Devarajan et al, 1996), more recent research (e.g., Haque, 2004 and Gupta et al, 2005), has tended to reconfirm the conventional wisdom that switching resources from consumption towards investment spending i s growthenhancing. 3.10 One reason why the impact of public capital spending on growth can yield contrary results is that public investment can either complement or displace (`crowd out') private investment. Inthis regards, the analysis undertaken for the 2001 Honduras PER suggested that public investment generally tended to `crowd in' private investment, but that this positive effect rapidly declined when public investment rises above 9 percent of GDP (an event that occurred several times in the 1980s and early 1990s). This finding is plausible if the government prioritizes those investment projects with the highest social rates o f return. Inthat case, as the volume of public investment increases, the social rate of return at the margin would decline and so would the overall impact on growth. Another factor that could lead to this result, even if all the public investmentprojects maintain the same social rate of return, is possible diseconomies of scale associatedwith managing an increasingly larger public investment program. 3.11 At the sector level, the composition and efficiency of public spending can also have a dramatic effect on the impact of fiscal policy on growth. There is now ample evidence (e.g., Barro and Sala-i-Martin, 1995) that progress in the outcome indicators relating to education, health, public infrastructure, and research & development contributes to faster growth. To the extent that countries invest inthose areas, they would be expected to grow faster. This suggests that fiscal policy can affect growth through the composition of public spending. As argued in World Bank (2004),26 however, simply increasing public spending without considering the efficiency of such spending often fails to yield significant gains. As a result, the relationship betweenpublic expendituresand sector outcomes is less thanstraightforward. 3.12 For example, PanelA of Figure 3.3 presents a scatter plot of the under-5 mortality rate in 2000 and of averageper capita spending inhealth over the 1998-2000period. It suggests a clear negative relationship: countries that spend more appear to also have lower child mortality rates, with the regression line indicating that on average a 1 percentage point increase in per capita spending would be associated with a 0.6 percentage point decline inthe mortality rate. Panel B inFigure 3.3 presents a partial correlogram that controls for the general observation that richer countries tend to both spend more and have better social indicators. It presents a dramatically different story from Panel A. Once the effects of per capita income are filtered out, there does not appear to be a statistically significant relationship between public spending on health and health outcomes. The same lack o f cross-country correlation between social outcomes and public spending, once we control for income levels, also extends to education; Figure 3.4. As in the case of health, the unconditional correlation indicates that countries with larger public spending on education tend to have better education outcomes interms o f net enrollment. Once 26. WorldDevelopmentReport2004. "Making ServicesWork for PoorPeople". 30 we control for the countries' per capita income levels, however, the positive relationship not only disappears but inthis case, it even turns negative. 6 - E b 3 - 34 2 - 1 - 0 - I 0 2 4 6 8 10 -1.5 -1 0 -0.5 0.0 0.5 1.0 1.5 Per capita spending Per capita spending Note: *` Panel A presents the scatter plot o f child mortality (in 2000) and per capita spending on health (average of 1998-2000). Bothvariables are inlogs. Panel B controls for the level o fper capita GDP (in2000). Source: Own calculations usingWDI data. Figure3.4. PublicSpendingand Social Outcomesin Education *' PanelA PanelB .- 4.8 1 0.3 i -p c)4.64.4 - e64.2 - * * *** 4 3.8 3.6 4 -0.6 A 0 2 4 6 8 10 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Per capita spending Per capita spending Note: *' Panel A presents the scatter plot o f net enrollment rates in primary education (in 2000) and spending on primary education per student (average o f 1998-2000). Both variables are in logs. Panel B controls for the level o f per capita GDP (in2000). Source: Own calculations using WDI data. 3.13 This does not mean that the sector composition o f public spending is unimportant. Rather, it suggests that there are many other aspects beyond the specific level o f public spending in a particular sector that have to be taken into account. For example, if public spending increases, but i s badly targeted because o f mismanagement or inefficiencies, then we should not expect the extra spending to have a significant impact on any social outcome. Even ifthe public sector i s highly efficient and well managed, there may be other problems. Public spending i s generally perceived in terms o f affecting the quantity o f a public service, so that increasing the quantity o f services would move closer towards a desired outcome. However, public spending also generally affects the price o f the public service, which sometimes outweighs the impact on 31 quantity. Inthat event, higher public spending mostly means paying more for the same service and should not expect any gain in outcome. A case in point has been the increase in public spending on education inHonduras. According to the 2001 Honduras PER, this appears to have largely financed increases in teacher salaries, instead o f yielding more teaching hours and improved education outcomes. 3.14 Inother words, the level of sector spendingby itself may not be a good indicator of the overall resource allocation in that sector and, thus, may not reveal much about the potential impact o f increases in sector spending on growth. The sector composition o f public spending and its impact on sector outcomes are treated in Chapter IV and the respective sector-specific chapters o f this PER, while the remainder o f this Chapter will focus on the economic composition o fpublic spending and its impact on growth. B. The Evolutionof FiscalPolicyinHonduras 3.15 The optimal size o f government has been the subject o f many intense policy discussions. One view holds that governments are inherently inefficient and introduce distortions, so that larger public sectors yield slower growth.27 Another vied8sees the state as having an critical role in reconciling conflicts between private and social interests, thereby helping to overcome market incompleteness and externalities, which contributes to accelerating economic growth. While there may be disagreement on the link between the size o f government and economic growth, most agree on maintaining a sustainable fiscal policy as a necessary condition for growth. That is, whatever the total level o f public spending may be, it should match the level o f revenues. Fiscal sustainability 3.16 During the 1960s, the Non-Financial Public Sector (NFPS) deficit fluctuated at about 1 percent o f GDP, and public debt reached 15 percent o f GDP in 1970. The 1970s and 1980s witnessed a dramatic increase in the state participation in the economy and public investment grew sharply, peaking at 12 percent o f GDP inthe early 1980s. These high levels o f investment were financed with debt, as the fiscal deficit inthe mid 1980sreached 12 percent o f GDP. This had the consequence o f doubling the stock o f public debt during the 1970s (to reach 30 percent o f GDP in 1980) and once again during the 1980s (to reach 66 percent o f GDP in 1989). These fiscal disequilibria also were reflected inthe current account deficit o f the balance o f payments, which tripled betweenthe mid-1970s and the mid-1980s, reaching 11percent o f GDP. 3.17 Access to foreign assistance, provided generously while the civil conflicts inneighboring countries were at their peak intensities, allowed Honduras to stave off a major balance o f payments crisis. In the late 1980s, however, regional political circumstances changed and external finance became scarcer, forcing the government to cut back spending. Despite fiscal tightening, Honduras continued to face high external debt service obligations (at about 40 percent of revenue in 1988), andthe government resorted to external arrears to finance its deficit. In1989, Honduraswent into arrears withthe multilateral institutions. 27. Gemmel(l983); Landau(1983,1986); Barro (1989,1990,1997); Romer (1989); Easterly & Rebelo (1993) 28. Rubinson(1977); Ram(1986); Grossman(1990); Levine and Renelt(1992); Karras(1993,1996,1997) 32 110 b . 10 .' '- ,*', 90 * * \ . 8 b 6 ;.- 3 4 2 I # O 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: Own calculations using GFS data 3.18 Various government administrations since 1990 have tried to bring the situation under control, with mixed results. The Callejas administration (1990-94) introduced a stabilization package that included a sharp devaluation of the Lempira. As a result, the debt-to-GDP ratio jumped to more than 100 percent in 1990, while the external debt service burden increased to about 80 percent of central government revenues. The administration reversed course during the run-upto the 1993 elections, however, once againplacingthe fiscal accountson anunsustainable path. 3.19 The Reina administration (1994-98) cut central government expendituresby 4 percentage points of GDP during its first year inpower and introduced some tax adjustments to reduce the fiscal deficit to 3 percent of GDP in 1997. Combined with positive growth (averaging 4 percent a year) and a significant real appreciation of the Lempira, these fiscal adjustments resulted in a steady decline inthe external debt ratio, from 95 percent o f GDP at the end of 1994to 66 percent at the end of 1998, andbrought downthe total debt ratio to slightly above 70 percent o f GDP. 3.20 The fiscal situation remained largely under control during the initial years of the Flores administration (1998-2002), even after Hurricane Mitch struck Honduras in October 1998 causing major physical and social damage. The government revisedits economic program inthe aftermath of the hurricane and shifted spending priorities toward emergency relief, but simultaneously contained growth innon-emergency expenditures. Nevertheless, the fiscal deficit increased to almost 4 percent of GDP by 2001, eventually derailing the economic program agreed with the IMF. The Maduro administration (2002-06) reestablished greater fiscal discipline and reached agreement with the IMF in late 2003 on a new PRGF program. Under that program, the central government deficit was gradually reduced to around 2 percent of GDP in2006, while the stock ofpublic sector debt fell to about 40 percentofGDP. 3.21 These developments suggest that the improvements on the policy front over the past 5 years, coupled with external support through the HIPC and MDRI initiative^,^^ have brought the fiscal situation largely under control from a sustainability perspective. There are some concerns, 29. Honduras' debt burden had been declared unsustainable under the framework o f the Enhanced HIPC Initiative, such that the country qualified to receive debt relief. 33 however, about the nature o f the fiscal adjustment that was carried out so far. As described in Figure 3.6, sustainability has been achieved mainly through increases on the revenue side. Central government revenues as a percent o f GDP have increased over time (from 17 percent of GDP in 1990 to 21percent in 2006), while central government spending has remained basically constant (abstracting from cyclical fluctuations). The permanence o fthis type o f adjustments has been questioned by previous research on the link between the composition o f fiscal adjustments andthe likelihood o f successful fiscal reform, defined as a relatively permanent consolidation o f the budget3' Using a sample of OECD countries, Alesina and Perotti (1995) find that successful adjustments3' have relied mostly on expenditure cuts (especially reductions in transfers and the public wage bill), while unsuccessful ones have mostly relied on tax increases. Using a somewhat different methodology, Von Hagen and Strauch (2001) arrive at a similar conclusion. They found that `unsuccessful consolidations have relied almost exclusively on increasing revenues, while successful ones put heavy emphasis on cutting government spending'. Finally, applying a survival analysis to a sample o f 29 developing countries, Gupta et al. (2004) find that a reallocation o f current expenditures to capital outlays tends to render fiscal adjustments more permanent, while large levels o f wages and salaries, transfers, and subsidies tend to have the opposite effect. Figure3.6 CentralGovernmentExpenditures,Revenuesand Deficits; 1990-2006 *` I 30 8 25 7 ge! 6 20 F$ 5 - W 15 U 4.-s tu El 3 d E lo 2 5 1 0 0 Note: */ All variables are expressedas apercent of GDP. Source: Own calculations usingGFS data Is the Honduran Public Sector Oversized? 3.22 Going beyond considerations o f fiscal sustainability, there is little to go by in seeking to assess whether a country's public sector is too large, too small or about right. National preferences for public versus private goods differ across countries, and this i s reflected in significant variations in the size o f public sectors. One way of forming a judgment about the 30. Alesina and Perotti (1995) "Fiscal Expansions and Fiscal Adjustments inOECD Countries" 31. "A successful adjustment in year t is defined as a "very tight" fiscal stance in year t such that the gross debt/GDP ratio inyear t+3 is at least 5 percentage points of GDP lower than inyear t". 34 appropriate size of government is through the use of international comparisons. The top panelon the left of Figure 3.7 comparesthe leveloftotal NFPS expendituresacrossvarious countries and country groupings32as shares of GDP, while right-hand panel presents the same comparison in terms of total NFPS revenues. Honduras stands out as having the largest public sector among its Central American neighbors under both measures of government size. It also stands out in comparison to the Latin America average. Only the European countries appear to have larger public sectors on average. As noted earlier, however, there exists a positive and statistically significant relationship between the size of the government and income levels across countries; see middle panels in Figure 3.7. Accordingly, the bottom panels in Figure 3.7 present the measuresof government size controlling for differences inper capita income levels. That is, they measure the size of government as the difference between the actual size o f the public sector revenues (or expenditures) and the size predicted by the income level of that country. Based on these adjusted figures, Honduras does not stand out as before, though it continues to exhibit the largestpublic sector inCentral America. TheEconomic Composition of Public Expenditures 3.23 The most striking characteristic of the economic composition of public expenditures in Honduras is the enormous amount o f resources allocated to public wages and salaries. Figure 3.8 shows the evolution over time of the principal components of public outlays. Interest payments constituted a large burdenon the government, peaking at 18 percent of total expenditure (or 5.9 percent of GDP) in 1996-98. They have since then declined to 5.5 percent o f total expenditure (or 1.7 percent of GDP) in2006, aided by various debt relief agreements with the Paris Club and underthe HIPC initiative. 3.24 What stands out the most, however, i s that the fiscal savings in interest payments have beentotally offset by a radical increase inwages and salaries, which reacheda peaklevel of 44.3 percent of total expenditure (or 13.5 percent of GDP) in 2002. The growth inthe wage bill was stopped after 2002 with the passage o f the Public Wage Law inlate 2003. This law was largely repealed in 2006, however, setting the stage for increased pressures on public wages in future years. 3.25 Another remarkable aspect of the composition of expenditures is the large share of resources assignedto capital outlays. Capital expendituresrepresented 24 percent of total public sector expenditures (or 8 percent of GDP) in 1996, and have since then fluctuated between 19.5 percent of total expenditure 1998 and 26.8 percent in 1999 (in the wake o f post-Mitch the reconstruction spending). Inrecent years, capital outlays have maintained their share at around 24 percent of total expenditures. 32. The Latin American average is based on Argentina, Mexico, Chile, Peru, and Colombia. The average for Europeis basedon Belgium, Denmark,Germany, Ireland, Italy, andNetherlands. 35 Figure3.7 Measuresof Government Size: Expendituresand Revenuesof the Non FinancialPublic Sector (percent of GDP)" Public ExpendituredGDP Public RevenuedGDP 50 50 11 I 45 43.9 45 43.3 40 40 35 31.7 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 Cross-Country Relationship ExpendituredGDP Revenues/GDP BO , ' OI BO 50 L : * 40 :: 30 20 10 o moo0 zoooo 30000 40000 50000 BOOOO GDP par ~ a p l t a Actual - Predicted Note:*'Lefthandside alwaysrefers to Expendituresandrighthandsideto Revenues. All numbersare for (2004). Source: Owncalculationsbases on I M F ' s Article IV Staff Reports 36 Figure3.8 Componentsof TotalExpenditureinHonduras; 1996-2006 I 1 100% 80% 60% 40% 20% 0% 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Wages and Salaries (% of Exp)) 0 GoodsandServices(%of Exp) Capital Expenditure(% of Exp) 0Transfers (% ofET) Interest Payments (% of Exp) Source: Own Calculations basedon IMFArticle I V Consultations StaffReports. Note: Net lending is includedunder capital expenditures, Is thepublic wage bill in Honduras unusually high? 3.26 Compared to other countries in Central American, Latin America, and even Europe, Honduras exhibits a conspicuously large public wage bill as a share o f total expenditures; Panel A o f Figure 3.9. Once again it is important to take into account the generally positive and significant correlation between public expenditures and income levels in these countries. To correct for this correlation, Panel B of Figure 3.9 presents the shares of public wages and salaries as deviations from the linear prediction based on per capita income levels. It shows that Honduras exhibits the largest positive deviation by far o f any country in the sample. This suggests that the public wage bill inHonduras is indeedunusually large. Figure3.9 PublicExpenditureon Wages and Salaries(as % of GDP) PanelA PanelB 6 I GTM NIC LAT.AM C A M SLV PAN CRI Ell HND O N PER CHL NIC OERIRL MFX ARO COL NLD CRI PAN SLV ITA BEL HND Source: Own Calculations basedon IMFArticle I V Consultations Staff Reports 37 C. The Effectsof BudgetCompositionon EconomicGrowth 3.27 There are two major strands of thought inthe literature about the effects of fiscal policy on economic growth. One stresses the effects in the long run and the other the responses in the short term. The theoretical framework developed inthese strands concludes that fiscal policy can affect both, the steady-state and the short-run growth rate. On one hand, current studies addressingthe long runeffects suggestthat fiscal policy can either foster or inhibit growthvia its effects on the accumulation of physical and human capital (Barro, 1990; Barro and Sala-i- Martin, 1995; Mendozaet al., 1997). Onthe other, studies addressingthe short term effects have been devoted to understandingunder which conditions fiscal multiplierscan be small or negative (Alesina and Perotti, 1996; Alesina and Ardagna, 1998; Perotti, 1999). These studies conclude that budget composition also matters, besides the size and persistence of fiscal impulses. 3.28 In a recent study, Gupta et al. (2005) have empirically explored the impact of fiscal policy on growth using data for 39 low-income countries. The results of this study indicate that there is a strong link betweenpublic expenditure reform and growth in both the short and long terms, as fiscal consolidations achieved through curtailing current expenditures are, in general, more conducive to growth. Annex E presents a framework for analyzing the impact o f fiscal reforms on economic growth that builds upon the findings from Gupta et a1 (2005). This framework is developed by estimating the relationship between the composition of public expendituresand revenues, and economic growth. The elasticities derived from this estimation are then used to simulate the impact of certain changes in the explanatory fiscal variables on economic growth. 3.29 The following regressionfindings emerge from estimating the relationship between fiscal adjustment and growth across countries: 0 In general, fiscal adjustments do not appear to have a harmful effect on growth; i.e., improvements in the fiscal balance are associated with improvements in economic growth. 0 Domestic financing tends to be more harmfulfor growththan external financing. 0 The economic composition of expenditures affects economic growth in various ways: The largest long-run impact on growth is given by public spending on wages and salaries, which exerts a strong negative effect on growth. In contrast, capital spending exerts a significantly positive effect in the short and long run. Finally, expenditures on other goods & services also have a positive impact on growth, but only inthe short-run, while interestpayments do not yield a statistically significant impact. 3.30 As shown earlier, Honduras stands out in comparisons with other Central and Latin American countries with respectto several fiscal policy variables. Annex E simulates the impact of adjusting the values of various fiscal variables in Honduras toward the average values exhibitedby other countries inthe region, using the elasticities estimated by Gupta et a1 (2005) under various modeling specifications. This exercise leads to the following principal findings: simply aligning public expenditures on salaries and wages (as a share of GDP) to the average level observed in other Central American countries, and not considering any other adjustment, would raise Honduras's growth rate by roughly 2.4 percentagepoints. Although this estimate is probably on the high side, it is noteworthy that the realignment o f public expenditureson wages and salaries gives rise to the largest potential gain. When all the expenditure variables are 38 adjusted to the Central American benchmark levels simultaneously, Honduras's growth rate registers a more modest improvement of about 0.5 percentage points. Finally, ifthe whole fiscal bundle is adjusted simultaneously (Le., expenditures and revenues are both adjusted to their benchmark levels), growth declines by a modest amount. 3.31 The potential gains in growth are higher when Latin America is taken as the benchmark. The simulation results from adjusting all expenditures to the benchmark levels simultaneously indicate a potential gain o f 0.8 percentage points, where the largest contribution to growth continues to emanate from reducing expenditures on wages and salaries to the benchmark level. Changing the whole fiscal bundle (expenditures and revenues), inturn, yields additional growth o f 0.4 percentage points. When Europe i s taken as the reference point, the gain in growth from simulating the entire fiscal bundle is greater than 1 percentage point. Inthis case, however, most o f the gain comes from changes to the revenue side (and more specifically from changes innon- tax revenues). 3.32 The preceding results do not take into consideration possible measurement errors in the data, particularly with respect to GDP. (Guatemala and Nicaragua have recently updated and rebased their national income accounts, yielding significant changes: Guatemala's GDP was revised downward, which Nicaragua's revised GDP turned out significantly higher.) Clearly, a significant over or under-estimation o f the GDP figures could alter the simulation findings reported inthe preceding paragraphs, as that would exaggerate or reduce the differences between the fiscal variables exhibitedby Honduras and the benchmark averages. It turns out, however, that even if Honduras's GDP were underestimated by 30 percent, the same simulation exercise still yields important gains from bringing the level o f wa es and salaries in Honduras down to the average levels observed inCentralandLatinAmerica.3K D. MainConclusionsandRecommendations 3.33 While the performance o f the Honduran economy has improved significantly over the last five years, it does not provide grounds for complacency. Honduras's growth rates have been comparable to those observed in the rest o f the Latin American region and, therefore, are likely to be reflecting an overall favorable economic environment, rather than a significant reduction in barriers to growth that may previously have existed inHonduras. Furthermore, much faster and sustained growth will be required if Honduras is to make greater inroads than heretofore into poverty reduction. This issue was analyzed in the 2004 Honduras DPR, which found that that Honduras's comparatively slow growth in the past was due more to a lack o f factor productivity growth than to insufficient capital investment. Furthermore, it identified four growth determinants where Honduras was found to be lagging behind inits development vis-a-vis other developing countries, namely human capital development, public infrastructure development, financial market development (all o f which are positively correlated with growth), and the overall size o fthe government consumption (which i s negatively correlated with growth). 33. O f course, if Honduras's GDP were currently over-estimated, the potential growth gains from lowering the public sector wage bill to the benchmark levels would be even higher than was estimated under the earlier baseline scenario. 39 3.34 While the progress made over the last five years in bringing the fiscal situation under control has been good from a growth perspective, the manner in which the fiscal deficit was reduced raises questions, both from a growth perspective and a sustainability perspective. As noted earlier, most o f the reduction in the fiscal deficit since 2002 was achieved by increasing revenues, rather than curtailing expenditures. To the extent that government consumption expenditures are negatively correlated with growth, as indicated by previous cross-country research, the authorities may have missed an opportunity to promote faster growth by focusing on revenue mobilization to reduce the fiscal deficit. On the other hand, to the extent that expenditures were held steady in compliance with the HIPC initiative (which calls for replacing debt relief with higher anti-poverty spending), the negative growth effect may have been compensated by positive distributional effects. Another strand o f empirical research has investigated the link between the composition o f fiscal adjustments and the sustainability o f reform. That literature concludes that a successful fiscal reform is more likely to be sustained if it relies mainly on expenditure reductions rather than on revenue increases (para. 3.21), which raises concerns about the permanence o fthe previous fiscal reforms inHonduras. 3.35 To explore the potential impact o f possible fiscal reforms on economic growth in Honduras, an econometric analysis was carried out with the help o f an empirical model developed by Gupta et a1 (2005) that relates GDP growth to various fiscal indicators. Two important findings emerge from this analysis: first, it appears that fiscal reforms can result in significantly faster growth for Honduras over the long run. Secondly, o f all the different components o f fiscal spending, the main barrier to faster growth appears to be the high level o f wages and salaries. These findings are fairly robust, as they continue to apply under a number o f departures from the baseline specification that include (i) changes inthe underlyingassumptions used in the econometric models, (ii) changes in the reference groups used as benchmarks, and (iii)adjustments to account for a potential underestimationo f GDPby as much as 30 percent. 3.36 To explore the potential impact o f possible fiscal reforms on economic growth in Honduras, an econometric analysis was carried out with the help o f an empirical model developed by Gupta et a1 (2005) that relates GDP growth to various fiscal indicators. The two most important findings from this analysis are the following: first, it appears that fiscal reforms could result in significantly faster growth for Honduras over the long run. Secondly, o f all the different components o f fiscal spending, the mainbarrier to faster growth appears to be the high level o f wages and salaries. These findings are fairly robust, as they continue to apply under a number o f departures from the baseline specification that include (i) changes in the underlying assumptions used in the econometric models, (ii)changes in the reference groups used as benchmarks, and (iii) adjustments to account for a potential underestimation o f GDP by as much as 30 percent. 3.37 The following key recommendations emerge from these findings: 0 Honduras should aim for a more balanced adjustment approach in future fiscal reforms by also including measures to curtail expenditures, rather than concentrating mainly on revenue increases, while seeking to improve the efficiency o f public spending. 0 When considering expenditure reductions, Honduran policymakers should focus on reducing the size of the public sector wage bill inthe interests of accelerating economic growth and reducing poverty. 40 CHAPTER IV THE CHALLENGE OFACHIEVING THEMILLENNIUMDEVELOPMENT GOALS 4.1 Honduras is among the poorest countries inLatin America and the Caribbean, with more than 64 percent of the country's population living below the national poverty line, while 43 percent live in extreme poverty. Slow per capita economic growth, an unequal distribution of income and resources, and low labor productivity are some of the most relevant factors responsible for the continued high poverty incidence. Although the recent performance in poverty reduction has beendisappointing, the progressmade toward reaching the other MDGs i s similarto that achieved inthe rest ofthe region; Table 4.1. Table 4.1 MDGattainment inHonduras and LatinAmerica Latin America and Honduras Caribbean 1990 2004 2015 1990 2003 2015 (Target) (Target) MDG 1:People living on less than $1(PPP) aday (%of pop.)) 38 21 19 11 10 6 MDG 1:People living belowthe nationalpovertyline (% o fpop.) 84 64 42 MDG2: Primarycompletionrate (%ofrelevantage group) 65 76 100 57 97 100 MDG4: Under-fivemortality rate (per 1,000 births) 59 31 20 53 33 18 MDG 5: Maternalmortality rate (per 100,000 live births) 180 108 70 193 MDG7a: Access to an improvedwater source (%of population) 73 82 87 82 89 91 MDG7b: Access to improvedsanitationfacilities (%ofpop.) 66 17 83 68 75 84 Note: MDG2 is shown as net completion rate for Honduras, and gross completionrate for the LAC aggregate. For the LAC aggregate in2003, MDG 1andMDG 5 are 2000 values. While the table lists the MDG levels and targets as defined inthe UN Millennium declaration, the Government of Hondurashas set more ambitious targets for severalindicators. For example, the water and sanitation coverageis to be expanded to 95 percent of the population, significantly exceedingthe MDG of improving access by one-half. Similarly, using the national poverty line as the MDGI indicator, poverty is to be reducedfrom 64 percent in2004 to 42 percent in2015. Finally, the childmortalityrate target is set slightly higher at 24 instances per 1000livebirths. Source: www.developmentgoals.org, www.sierp.hn 4.2 Ineducation, the rate of alphabetization of the young has increasedfiom 79.7 percent in 1990 to 85.5 percent in 2001 and the coverage rates for primary education have reached 89.3 percent in 2004. Moreover, there is no apparent gender gap, as the data for primary education shows boys and girls having almost identical access and completion rates. These developments inthe education indicators exhibit a clearly positivetrend, but may be insufficient for achieving the target of universal primaryeducation completion by 2015. Furthermore, a number of studies have raised serious questions about the quality of education received by Honduran pupils andthe efficiency public education spending.34 34. For example, the World Bank (2001) HondurasPublic ExpenditureReview cites a study assessinglanguageand math skills inthe third and fourth grades of education, where, out of twelve Latin American countries participating inthat study, Hondurasranked last in languageand next-to-last inmath. Also, the World Bank (2004) Honduras Development Policy Review found that the recent expansion in public spending on primary education was accompanied by declining efficiency: even though spending on primary education per student increasedover the same period. 41 4.3 Inhealth, the under-five mortality rate decreased from 59 to 31 per thousand births and the infant mortality rate was reduced from 47 to 23 per thousand between 1990 and 2005/6. While data inadequacies do not permit a precise assessment of the evolution of maternal mortality, the available survey results3' suggest that considerableprogresshas been achieved: the maternal mortality rate was reducedfrom about 180 (per 100thousand) in 1990to around 108 in 2000. The percentage of childbirths taken care of by specialized personnel in health institutions has fluctuated during the 1990s, but follows an upward trend, rising from 45.6 percent in 1990/91 to 61.7 percent in the 2001. Although not large, there still exists a gap between the shares of urban (85.5 percent) and rural (80.7 percent) mothers receiving prenatal control. This gap is much larger for medically assisted childbirths: in2001, 82.4 percent of the childbirths in the urbanareas were taken care of ininstitutions with specialized sanitarypersonnel, as opposed to only 37.5 percent inthe rural areas. 4.4 The observed improvements in the 2015 Millenniumhealth goals36can be attributed to various measures adopted to promote the delivery and efficiency of health services - such as programs aimed at expanding ambulatory and hospital care and at strengthening the country's epidemiologic capacity to respondto emerging and other infectious diseases. Honduras has also recently started an initiative to improve maternal-infantile health, which includes preparing medical protocols to regulate the sector delivering mothers and infants-specific health services, the expansion of the services of the Atencih Integral a la NiAez (Integral Care to Childhood) network, and the implementation of a program for the monitoring and analysis of maternal and infantile deaths. However, additional progress is likely to require the expansion of coverage to under-served rural areas and the provision of more complicated (and more expensive) medical services administered by highly skilled healthprofessionals. 4.5 National coverage for potable water increased from 73 percent to 82 percent over 1990- 2004, while sanitation coverage increased from 66 percent to 77 percent. However, large disparitiesincoveragerates are observedacross rural and urbanareas, and even across large and smaller cities. Inaddition, Honduras faces severe challenges inreaching its ambitious coverage rates on account of the high growth rate of its population and the low efficiency of sector institutions. According to official forecasts, reaching a 95 percent coverage rate for water and sanitation in 2015 (a goal which is above that set by the Millennium declaration) means providing access to water for an additional population of 2.6 million intotal - 1.2 million inrural areas and 1.4 million in urban areas - and supplying sanitation services to an additional population of 3.5 million intotal, distributedbetween 1.3 and 2.2 million inrural andurbanareas respectively. Although the size of the investments requiredi s large, the government realizes the importance of starting these as soon as possible given the key externalities generated by universal access to these services. 4.6 Notwithstanding the positive developments discussed above, Honduras i s unlikely to achieve the full set of the MillenniumDevelopment Goals in the absence of additional efforts. This assessment holds true even with a significantly improved macroeconomic environment and 35. Surveys aimed at measuring maternal mortality rates were administered in 1990 and 1997, and the national statistical institute (WE) estimated the rate for the year 2000. 36. Inthis study we do not consider some o f the other important health goals, such as Malaria and HIV/AIDS. In Honduras, several efforts to fight these and other diseases are being currently financed through the Global Fund, which supports the consolidation o f the National Program against Tuberculosis and Malaria, executed by the Secretary ofHealth executes. 42 stronger aggregate growth performance. The rapid pace of reductions in infant and child mortality rates between 2001 and 2005/6 (see the discussion in Chapter 7 for additional details) bodes well for the achievement of MDG 4, but continued progress i s conditional on maintaining the recent growth of public health expenditures, which expanded at an annual average rate of 11.7 percent between 1999 and 2005 (measured in constant Lempiras). If this growth i s not sustained, additional inroads inimproving health outcomes are likely to be minor. Inother social sectors, past expenditure trends are likely to be insufficientto ensure the achievement of MDGs. While the country will come closer to achieving some targets (e.g., education), it i s likely to exhibit significant gaps inothers (e.g., water and sanitation). A. Modelingthe Attainmentof the MillenniumDevelopmentGoals 4.7 The extrapolation ofcurrent trendshasbeen carried out with the use of a dynamic general equilibriummodel -Maquette for MDGSimulations (MAMS) -which links public expenditures on social services and infrastructure to social outcomes in terms of MDG attainments and aggregate growths3' A key objective of MAMS is to capture the key interactions between the pursuit ofthe MDGs andthe evolutionofthe economy. For this reason, it focuses onthe sub-set of goals that is likely to be the most costly to achieve and that is expected to have the largest impact on the economy. These are: universal primary school completion (MDG 2), reduced under-fiveand maternal mortality rates (MDGs4 and 5), andincreasedaccess to improvedwater sources and sanitation (part of MDG7). To the extent that a package o f interventions that curtails child and maternal mortality helps to reduce the incidence of major diseases other than HIV/AIDS, the model also implicitly tracks MDG 6. In addition, achievements in terms of poverty reduction (MDG 1) are addressed, although the model does not contain mechanisms for specific MDG 1-related intervention^.^^ 4.8 Production of a typical MDG39i s modeled as a system of two functions: the first models the production of an aggregatemeasure of MDG service delivery which dependson direct MDG spending and on other determinants. The second function relates MDG outcomes with this aggregate measure. Inthe first function, the production of MDG-related services requires three broad categories of inputs: labor (which is disaggregated according to skill/education levels), capital goods and intermediaryproducts. Inaddition to these inputs, which basically account for spending on specific MDG interventions, the aggregatemeasureof MDG service delivery is also determinedby some other variables as listed in Table 4.2. For example, reaching the education MDG requires additional schooling services, but is also facilitated by improvements in health conditions (proxied by MDG 4 and 5), by better infrastructure (e.g., better roads to schools), by higher income levels (better-off parents may not need their children to work), and by good 37. A more detailed description o fthe model i s available inAnnex G and inLofgren and Diaz-Bonilla (2006). 38. Of the 8 MDGs of the 2000 Millennium Declaration, only three are left out o f the current version o f the model: M D G 3 -promote gender equality and empower women, MDG-7 - ensure environmental sustainability, and MDG 8 -develop a global partnership for development. 39. The modeling o f the education MDG is more complex given that M A M S considers the full educational system disaggregated into different cycles: primary, secondary, and tertiary. The primary cycle is needed to estimate attainments o f the MDG 2. The higher cycles are needed to link education to the labor market. They are also required to provide a complete picture of the dynamic fiscal consequences o f achieving MDG 2 and expanding the educational system. 43 returns to education (proxied by the wage premium paid to skilled workers). The aggregate measure is strictly increasing inall o f its components, and does not capture potential bottlenecks and/or the decreasing returns to scale as the target approaches (due to the difficulty o f reaching the most remote parts o f the population or, for example, necessity o f high-level medical care to diminishmaternalmortality beyond a certain threshold). Inorder to account for these effects, the second function i s introduced and this requires greater and greater improvements inthe level o f aggregate service delivery for the same rate o f improvement inthe MDGindicator. Other Determinants Per-capita Per-capita Wage Public Other P D G ServiceDelivery consumption incentives infrastructure MDGs 1 Poverty X 2 Education X X X X 4 4 InfantMortality X X X 7a,7b 5 Maternal Mortality X X X 7a,7b 7a Water X X X 4.9 Although improvements inpublic infrastructure are not part o f the MDGs, they serve as a key input in the MDG production function, and contribute to overall growth by adding to the productivity o f other production activities. This means that real GDP growth (and poverty reduction) is partly influenced by government policies and in particular by its investment in public infrastructure (roads, ports, energy, et^).^' TheBaseline Scenario 4.10 The baseline scenario is based onthe growth profile andexternal environment projections in the latest Debt Sustainability Analysis (DSA) and on sector expert asses~ments~l.The evolution o f key macroeconomic and public expenditure indicators projected under this scenario is shown in Table 4.3. In the baseline scenario, real GDP per capita grows at 1.8 percent per year-which coincides with earlier growth projections prepared by the IMF and Government o f Honduras, but i s much faster thanthe 0.5 percent average annual growth recorded over the 1990- 2004 period.42 The key assumption underlying MDG performance in the baseline is that the level o f government service provision in public infrastructure, water and sanitation, health, and education sectors i s assumed to grow at the same rate as real GDP (3.9 percent per year). 40. The empirical link between growth and public infrastructure has been analyzed widely, but there is no universal consensus on the strength o f this link. Given the lack o f consensus, this chapter does not consider any scenarios where poverty (MDG 1) is actively targeted by boosting growth via increased public investment ininfrastructure. 41. The M D G production functions are "calibrated" according to a set o f scenarios indicating the 2015 values for the variables in Table 4.2 under which the MDGs are likely to be achieved. The levels of the per-capita service delivery are provided by sectoral experts; inthis case, the required increases inthe level o f public service provision were derived from estimates published in Government o f Honduras - Grupo Consultativo, "Avanzando en la planificacion sectorial de mediano plazo. Plan Pluriennal de Ejecuci6n de la Em", Tegucigalpa, Honduras, May 2005. For MDGs 4 and 5, the costing estimates inthis document were revised to take into account the recent rapid decline inunder-five and infant mortality rates. 42. These macro projections are obtained from the Joint IDNIMF Debt Sustainability Analysis completed in2006. 44 Spendinginthe general government sector is also set to grow exogenously at this rate, so that total public consumption grows at the rate of real GDP expansion inthe projected period.43 Table 4.3 Baseline Macroeconomic Projections 2004 2010 2015 Annualgrowth Real GDP at marketprices (bnlcu) 136 174.3 206.1 3.9 Private consumption (bn lcu) 116 150.9 180.6 4.1 Government consumption (bnlcu) 17 21.0 25.3 3.9 Investment (bn lcu) 36 46.1 55.9 4.2 Private (bn lcu) 28 37.0 44.8 4.3 Public (bn lcu) 7 9.2 11.1 3.9 Exports (bn lcu) 57 72.7 83.8 3.5 Imports (bn lcu) 90 116.4 139.5 4.1 Real GDP per capita (thou lcu) 18,972 21,472 23,181 1.8 lcuper Exchange rate USD 1.oo 0.99 0.97 -0.2 Trade-to-GDP (W 108.5 108.5 108.4 Investment-to-GDP (%) 26.2 26.5 27.1 Debt-to-GDP ("/.I 68.2 62.0 58.2 Debtservice-to-exports (%I 2.3 2.1 2.0 4.11 Honduras would be unable to reach any of the MDGsunder the baseline scenario. Table 4.4 compares the evolution of the MDG indicators in education, health, and water-sanitation goals under the baseline scenario with the MDG targets. This comparison reveals significant variation across different goals: the biggest improvements (48 percent distance to target) are observed for the education goal, while progress towards health, water-sanitation, and poverty goals is much slower. Only 16 percent of the distance to target i s covered for the child and maternal mortality goals, while water and sanitation fare slightly worse at 14 and 15 percent o f total distance covered. Finally, the baseline improvement inthe poverty headcount is 24 percent of the total distance to target. That is, the 1.8 percent per capita income growth generated inthe baseline scenario is not sufficient to makemajorprogress inpoverty reduction.44 4.12 The baseline scenario is not to be confused with an extrapolation of recent trends inthe evolution of MDG indicators inHonduras. As indicated inTable 4.1,several of these indicators (e.g., people living less than US$1 per day, and the under-5 mortality rate) exhibit a rate of 43. A technical feature o f this model is that although public consumption grows at the same rate as GDP in real terms, the nominal share o f the public consumption to GDP, rises from 12.3 percent to 14.8 percent, i.e. by more than 2 percentage points between 2004 and 2015. This is explained by the fact that the services provided by the government-ducation, health, other public services-are more skill-intensive than the economy-wide average. As a result, the cost o f providing these services (purchasing intermediate inputs, paying worker's salaries) increases over time, requiring additional fiscal revenues to finance service provision in line with real GDP growth. To accommodate these fiscal space requirements, the baseline scenario assumes that fiscal measures are taken to increasesthe revenue from direct taxes by 2.4 percent o f GDP over the projection period. 44. In the baseline MAMS micro-simulation, poverty reduction is modest due to increasing inequality between 2004 and 2015. The Gini coefficient rises by 1.5 percentage points, while the Theil index increases by 4.8 percentage points. The trend towards rising inequality i s explained by high demand for skilled and tertiary-skilled workers, which drives up their labor earnings and increases wage inequality. Additionally, the wages of unskilled workers grow slower than the economy-wide average because o f the demographic structure o f Honduras, where large cohorts o f young people enter the labor marketat low skill levels. 45 improvement between 1990 and 2004 that, if sustained, would be likely to yield the respective MDGby 2015. What this simulation model suggests, however, is that sustaining these trends in the indicators will require an increase in MDG-related public spending that exceeds the baseline GDP growthrate. Table 4.4 Baseline Scenario for MDG Indicators 2004 2010 2015 Annual growth 2015 Target Poverty headcount (%I 65 60 -0.8 42 Primary completion rate (%I 76 79 87 1.3 100 Under-5 mortality (per 1,000) 31 30 29 -0.3 24 Maternal mortality (per 100,000) 108 105 102 -0.5 70 Access to safe water (%I 82 83 84 0.2 95 Access to sanitation (%> 77 78 79 0.3 95 4.13 Considering the shortfall in MDG attainment under the baseline scenario, what options does Honduras have for reaching the MDGs by 2015, or at least for coming closer than would be the case under the baseline scenario? One possible option is for the government to step up its public spending inthe most critical sectors by enough to ensure that the MDGs are met. Another possible option i s for the government to improve the efficiency o f its public expenditures by enough to meet the MDGs, without the need for significant expenditure increases beyond the baseline projection. Each o f these options is discussed next with the help o f additional M A M S simulations. B. Alternative Scenariosfor Attaining the MDGs Scaling UpPublic Expenditures 4.14 Drawing on government estimates o f the required scaling-up o f resources to reach the MDGS?~ section evaluates the economy-wide consequences o f a focused effort to achieve this the MDGs for the human development indicators. This scenario explicitly targets the attainment o f MDGs 2,4, 5, 7a and 7b through the expansion o f services inprimary education, health, and water and ~anitation.~~assumes that the provision o f social services in each sector grows at a It constant pace to meet each o f the MDGs by 2015. The simulated increase in each social service i s then adjusted to take into account the evolution o f the other MDG determinants shown in Table 4.2 and other general equilibrium effects. Finally, the simulated increase in service provision is used to calculate the volume o f public expenditures needed to attain these targets over the model horizon (2005-2015), assuming a constant level o f efficiency inpublic spending. Inthe education sector, the growth rate of primary school expenditure is such that graduation rates reach 100 percent by 2OlOY4' while expenditures on secondary and tertiary schooling are 45. See footnote 41. 46. As mentioned earlier, MAMS does not model government interventions directly linkedto the achievement o f MDG 1(such as cash transfers or other interventions targeted to the poor), but instead tracks its progress by means ofmicro-simulation techniques (described inAnnex G.) 47. This requirement is due to the lengtho fthe primary education cycle and the definition o f MDG2. Ifthe target is defined as reaching (close to) 100 percent primary school completion in 2015 and the length o f the primary education cycle is 6 years, achievement o f MDG 2 implies that 100 percent o f children o f primary school age must enter the first grade in2010 and complete grades 1through 6 at 100percent rates. 46 maintained at baseline levels. All other public expenditures not associated with the MDGs also are assumedto evolve inline with the baseline scenario. 4.15 The MDG results and the expansion in the level of service delivery required to reach them are shown in Table 4.5. The associated financing requirementsare shown in Table 4.6. The results of this simulation confirm that a large, sustained increase in government spending over the baseline level is required in order to reach the targets by 2015. In all instances, the required frowth in current and capital spending accelerates significantly relative to the ba~eline.~ Table 4.5 The Volume ofPublic ServicesRequiredto Achieve the MDGs in the Absence of EfficiencyImprovements Annual 2004 2010 2015 growth Poverty headcount (%I 65 59 -0.8 Primary completion rate (%) 76 91 99 2.5 Under-5 mortality (per 1,000) 31 28 24 -2.2 Maternal mortality (per 100,000) 108 92 70 -3.9 Access to safe water (%I 82 88 95 1.3 Access to sanitation ("4 77 86 95 2.0 Government consumption Primary education (bnlcu) 4.8 12.2 12.1 8.7 Secondary education (bn lcu) 1.9 2.4 2.9 3.9 Tertiary education (bnlcu) 1.5 1.9 2.3 3.9 Health (bnlcu) 3.6 7.0 11.9 11.5 Water and sanitation (bnlcu) 0.3 0.6 1.1 14.2 Public infiastructure (bn Icu) 0.8 1.o 1.2 3.9 Government investment Primary education (bnlcu) 0.2 1.2 0.0 40.4 Secondary education (bnlcu) 0.1 0.1 0.1 3.9 Tertiary education (bnlcu) 0.0 0.1 0.1 3.9 Health (bnlcu) 0.1 0.6 1.1 20.1 Water and sanitation (bnlcu) 0.4 2.6 5.0 25.0 Public infiastructure (bn Icu) 1.4 1.8 2.2 3.9 4.16 For the sake of analytical simplicity, this scenario assumes that all of the additional public expenditures required to achieve the MDGs are financed by foreign donors. That is, all domestic fiscal policy variables are fixed at their baseline levels and any budgetary shortfall is madeup by foreign grants. Table 4.6 shows that the total public spendingwould have to increase by 10.4 percent of GDP (9 percent in current spending and 1.4 percent in capital spending) between 2004 and 2015 to achieve the MDGs. It also shows that if all domestic fiscal policy variables are kept at their baseline levels, the annual flow of foreign grants would have to reach 48. Note that investment growth in education is reported for the 2004-2010 period, since the primary education system is assumed to reach its fillpotential by then and any future expenditure is directed towards maintaining the 2010 schooling levels. Also note that even though per capita GDP growth accelerates slightly to 2.1 percent per year, this is not sufficient to reach the poverty MDGwhere the target headcount is 42 percent o fthe population. 47 almost 10 percent o f GDP by 2015 under the expenditure scaling-up scenario.49 This level o f foreign grants is considerably higher than the 1.1 percent o f GDP considered in the Debt Sustainability Analysis (see footnote 42). Table4.6 StructureofPublicFinanceinthe BaselineandExpenditure ScalingUp Scenarios(percent of GDP) I I ! Baseline with foreigngrants MDGachievement 2004 I 2015 Total current revenues 19.5 'I 21.4 20.1 Directtaxes 4.6 I 7.0 6.8 Indirecttaxes 9.9 I I 9.6 8.8 Importtaxes 1.3 I 1.2 1.1 Exporttaxes 0.0 I I 0.0 0.0 Currenttransfers 3.7 I I 3.6 3.4 I Total Current Expenditures 17.0 I 19.4 26.0 Consumption:MDGRelated 8.9 I I 10.6 17.8 Consumption: other 3.4 I I 3.7 3.6 Transfers andinterest payments 4.7 I 5.1 4.6 I I Public Savings 2.4 I I 2.0 -5.9 I PublicInvestment 5.4 I I 5.2 6.8 I Overall Balance -3.0 I I -3.2 -12.7 Domesticborrowing 0.9 I 1.9 1.8 Foreignborrowing I - - 2.1 I 1.3 1.1 Foreigngrants 0.0 .I 0.0 _._ 9.9 ... Improving the EfJiciency of Public Spending 4.17 Rather than assuming that the increase in services required to achieve the MDGs is entirely brought about through an increase inthe volume o f public spending, it is also possible to contemplate increases in the efficiency o f public spending, thereby reducing the need for additional spending. This possibility suggests itself both in view o f the huge size o f the additional public spending that would otherwise be required to achieve the MDGs and inlight o f earlier observations on the apparently low efficiency performance inHonduras (footnote 34). 4.18 Figure4.1 describes a conceptual trade-off betweenthe efficiency o fpublic spending and the amount of (grant-funded) pend ding,^' as alternativepolicy variables for achieving the MDGs. At one extreme o fthis tradeoff, we have the previous scenario, where the MDGs (excluding the poverty target) are attained exclusively by scaling up MDG-related expenditures, while keeping efficiency constant at the current levels. As calculated earlier, this would require an increase in 49. Inthis scenario, foreign grants flow into the country throughout the 2005-2015 period even though Table 4.6 only showsthe inflow levelinthe final year. 50. Note that Figure 4.1 over-simplifies the problem by suggesting that the policy mix is two-dimensional. As discussed earlier, MDG attainment also depends on spending on infrastructure as well as the growth in household per capita income. 48 total public spending by around 10 percent o f GDP by 2015. At the other extreme, the non- poverty MDGs are attained exclusively by improving efficiency, while keeping expenditures at the same levels as in the baseline scenario. In the MAMS model, the efficiency of public spending i s entirely determined by labor productivity, and improvements in efficiency can be measured by the ratio o f the productivity level in 2015 versus their level inthe base year (2004). When simulating the model to achieve the MDGs while constraining additional (grant-funded) spending to zero, it turns out that the overall productivity level would have to increase by 89 percent. Finally, one can envision different combinations o f efficiency and additional public spending amounts between the two extreme cases, each o f which i s sufficient to attain the MDGs. Figure4.1 Trade-off betweenEfficiencyImprovementand SpendingExpansion to Achieve the MDGsinHonduras Efficiencyvs. spending expansion 100 ............................................................................................ c) 90 0 80 .-pd B3 70 60 L Y 50 .-8 40 e 30 ; .- E 20 I O 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 Additional MDG-related public spending (%of GDP in 2015) Note: The simulations to develop this figure assumed that all additional MDG-related spendingis entirelyfinancedthroughforeign grants, while all other fiscal parametersare maintainedat their baselinelevels. 4.19 The policy mix shown inFigure 4.1 represents a convex, rather than a linear relationship. That suggests that there are decreasing returns to scale in seeking to achieve the MDGs entirely along one dimension - be it by just raising efficiency levels or by just increasing public spending. Inother words, a combination o f efforts is more effective inattaining the MDGs. C. The Available Fiscal Spaceto Finance Priority Expenditures 4.20 The preceding sections indicated that Honduras is unlikely to meet the MDGs by 2015 by following the historical trend and argued that the only way to ensure that those goals could still be met is by stepping up public spending in priority sectors or by improving the efficiency o f public spending by sufficient amounts. This section estimates the potential fiscal space available 49 to the Honduran authorities for increasing public spending inpriority areas, as well as the scope for improvingthe efficiency o fpublic spending. 4.21 Fiscal space can be defined as the total budgetary room available to the public sector to finance the totality of public expenditures without prejudice to the sustainability of its financial position. By creating more fiscal space, additional resources can be made available for financing additional expenditures in priority areas. To estimate the maximal amount o f fiscal space available to the authorities, it i s useful to beginwiththe basic government budget identify: where G is total expenditures, R is total revenues and D is the fiscal deficit. Furthermore, total ex enditures can be separatedinto public spending inprioritized areas, GP,and, other spending, G . That is, B Finally, it i s also useful to separate other spendin (Go) into1aInserting this component, GOD, and discretionary a non-discretionary component, GON, or Go=(G0% + GON).5 identity into equation (2) and rearranging terms, we arrive at: Expression (3) simply states that the government's capacity to increase public spending in prioritized areas is constrained by (i) ability to raise total revenues, (ii) ability to increase its its the fiscal deficit (Le., borrow more from either the central bank or the private sector), and (iii) its ability to reduce other, discretionary spending. Each of these options will be examined in turn, beginning with the capacity to expandthe fiscal deficit. TheScopefor Deficit Financing 4.22 Honduras's total public debt has declined precipitously in recent years, from 77 percent of GDP in 2001 to an estimated 39 percent in 2006. Also, Honduras's current public debt and debt service ratios are well below the thresholds that signal potential debt distress. These developments might lead some observers to conclude that there i s ample room for conducting a more expansive fiscal policy, which keeps the public debt ratio constant, instead of letting it decline further. Inresponse to these perceptions, however, it i s important to note that the sharp decline in Honduras's debt indicators in recent years was due almost entirely to the implementation of several debt relief initiatives and is not attributable to particularly tight fiscal policies. 4.23 Annex C develops a simple model of debt sustainability to calculate how much room there may be for expanding the fiscal deficit in Honduras without jeopardizing macroeconomic stability. That analysis concluded that ifnet concessionalfinancing inflows from abroad as well 51. The non-discretionarycomponentof `Other' public spendingincludessuch items as debt servicepaymentsand constitutionally mandated expenditures that may be necessary, but do not have a direct impact on the MDG indicators. 50 as GDP growth remain at recent historical levels, then Honduras could afford a primary fiscal deficit as high as 2.3 percent of GDP without placing its non-concessionalpublic debt ratio on an unstable trajectory. IfHonduras were also prepared to contemplate a significant increase in domestic inflation rates for the sake o f maximizing seigniorage revenues, it could raise the sustainable primary deficit a little further, to perhaps 3 percent o f GDP. It turns out that this maximal sustainable deficit level is not far removed from the actual primary deficit exhibitedby the combinedpublic sector during 2001-05, whenitaveraged2.9 percent of GDP. 4.24 Basedon this analysis, there does not appear to be any significant room for increasing the primary fiscal deficit in Honduras over a prolonged period to support an expansion of priority public expenditures without jeopardizing macroeconomic stability or violating over public policy objectives, such as the maintenanceof single-digit inflation. The Scopefor Raising Tax Revenues 4.25 To examine the scope for raising tax revenues, it is useful to begin by comparing Honduras's performance with that of other countries in the region. As shown in Table 4.7, Honduras's central government collects a significantly higher share of GDP intax revenues than other countries in Central America; 16.8 percent versus a regional average of 13.5 percent over 2003-05. While Honduras's tax ratio is somewhat lower than the overall LAC average, it continues to be significantly higher when the non-Spanish speaking countries are omitted from the calculations. Note that these findings are consistent with the earlier findings in Chapter 111, which also showed the revenues for the Non-Financial Public Sector in Honduras to be higher thanthe averageobservedfor other countries inthe region.52 4.26 Eventhough Honduras ranks fairly high inthese comparisons with other Latin American countries, the tax revenues collected inthe region as a whole are modest compared to developed and developing countries outside of the region. One indicationis given by the muchhigher tax ratios collected in the three non-Spanish-speaking countries included in the sample shown in Table 4.7. Accordingly, a recent study by the IDB identifiedvarious opportunities for reforming the tax system in Honduras to generate further revenue increases as well as improving the efficiency, transparency, and equity of the tax system.53 Implementation of these measures i s projected to raise total revenues by 2.1 to 2.2 percent o f GDP inthe short term, and by close to 4 percent of GDP inthe long run, relative to the 2002 tax ratio, which was around of 16 percent o f GDP. Since then, central government tax revenues increased to about 17.0 percent of GDP in 2005. This suggests that there i s room for additional tax revenueincreases of around 3.0 percent of GDP inthe longrun. ~ 52. These findings would have to be reconsidered to the extent that Honduras's GDP may be undervalued. Using the findings of recent household survey data, Annex D estimates that the GDP of Honduras could be about 16 percent higher than the official national income accounts. In that case, the proportion of central govenunent tax revenues to GDP declines to 14.5 percent. This would still put Honduras above the Central America average and equal to the Spanish-speaking L A C average. 53. G6mez Sabaini, Juan Carlos (2005), "Honduras: Hacia un sistema tributario m8s transparente y diversificado", in Agosin, Manuel R., Albert0 Barreix and Roberto Machado, eds. Recaudar Para Crecer: Bases para la reforma tributaria en Centroamkrica (Inter-American Development Bank, Washington, DC.) 51 Table 4.7 CentralGovernment Tax RevenuesinLatinAmerica & the Caribbean 2003 2004 2005 Average (2003-05) Bolivia 17.8 20.7 25.7 21.4 Brazil 15.9 16.2 17.2 16.4 Columbia 16.9 17.7 17.3 17.3 DominicanRepublic 14.9 15.1 16.8 15.6 Ecuador 10.3 10.5 11.6 10.8 ElSalvador 11.2 11.5 11.6 11.4 Guatemala 10.3 10.3 9.6 10.1 Guyana 28.9 33.1 32.9 31.6 Honduras 16.3 17.1 17.0 16.8 Jamaica 26.9 27.5 28.4 27.6 Nicaragua 15.0 15.5 16.4 15.6 Paraguay 10.2 11.9 11.8 11.3 Peru 13 13.3 13.8 13.4 .-__-____----------_----------------. Suriname _____. 27.8 _28.5_ _ _ _ _ _ _ _ _ _ _29.6_ _ _ _ _____. 28.6 Regional Averages LatinAmerica & Caribbean 16.8 17.8 18.6 17.7 excl. non-Spanish-speaking 13.8 14.5 15.3 14.6 Central America 13.2 13.6 13.7 13.5 Source: IMF, Article IV Consultations StaffReports and World Bank. TheScopefor Expenditure Switching 4.27 Broadly speaking, there are two ways o f generating more fiscal space on the expenditure side o f the public sector budget: one way i s by reducing discretionary public expenditures in non-prioritized areas and the other way is by improving the efficiency o f public expenditures in general. With respect to the first of these, the most obvious candidates to consider for expenditure cuts are the public sector subsidies provided in the energy, transport, water, agriculture and telecommunications sectors. These subsidies amount to approximately 3.5 percent o f GDP and many o f them are badly targeted and, thus, wasted. As discussed inChapter VIII, the government could save between 0.6 and 2.0 percent o f GDP by improving the poverty targeting o f these subsidies. These fiscal savings could help-to finance the expansion o f priority expenditures by an equivalent amount. TheScopefor Improving the EfJiciencyof Public Spending 4.28 Improvements inthe efficiency o f public spending may turn out to be the most promising source o f potential fiscal savings to finance increased priority spending. As indicated inAnnex A, the overall efficiency o f public spending appears to be quite low in Honduras, considering that its efficiency scores rank below the average and median scores inthe LatinAmerican region, both from an input and output-oriented perspective. Improvements in the efficiency o f public spending have a direct effect on expanding fiscal space by enabling the public sector to achieve the same sector outcomes with lower budgetary outlays. They are also likely to have an indirect impact in terms o f facilitating the expansion o f the resource envelope. That is, unless the government can show that its resources are being spent in a reasonably efficient manner, it may be difficult to persuade tax payers and donors to provide more resources. 52 4.29 One way o f estimating the potential room for improving efficiency is by comparing the efficiency of public spending in Honduras to the efficiency level exhibited in the most efficient countries world-wide. By this comparison, one could argue that Honduras can potentially increase the efficiency o f public spending by roughly 100percent ineducation andby roughly 50 percent in health, using the input-oriented measures o f efficiency.54 (This result is derived by comparing Honduras's average efficiency score o f 0.50 ineducation spending and 0.68 inhealth to the maximal possible efficiency score o f 1.0 achieved by the country on the efficiency frontier.) This estimate may be too ambitious, however. An alternative estimate of the scope for improving efficiency in Honduras is by using the average efficiency score in Latin America as the point o f reference; Table 4.8. This comparison suggests that Honduras could improve its efficiency in education spending by about 16 percent in education, and by 0 percent in health, given that it i s already close to the L A C average. These estimates, in turn, do not appear ambitious enough inlight o f the shortcomings inHonduras's social indicators. Table 4.8 Average EfficiencvScores for Public hending inEducationandHealth RegionalAverages Latin Sub- East Asia 8 EastJCentr EasV North South Middle Honduras i,'America America 8 Saharan Caribbean Africa Pacific ai Europe Africa INPUT-oriented Measures Education 0.50 I 0.57 0.58 0.50 0.60 0.60 0.48 0.55 Health 0.68 j 0.65 0.69 0.65 0.73 0.62 0.74 0.70 OUTPUT-oriented measures Education 0.52 0.54 0.65 0.38 0.66 0.73 0.56 0.47 Health 0.92 ij 0.97 0.90 0.67 0.84 0.93 0.89 0.77 Source: Basedon datafrom 1996-2002from Herreraand Pang (2004); see AnnexA. The efficiencyscores refer to single inpuffsingle output measures,usingthe Data EnvelopmentAnalysis method. 4.30 While seeking to achieve a position on the efficiency frontier i s certainly something to strive for, it may be more realistic to assume that a country may be able to raise its efficiency score through a concerted reform effort to the level attained by countries that rank at the 75h percentile o f the world or regional distribution o f efficiency scores, or half-way between the average country and the most efficient country. By implementing measures that raise its efficiency scores to this level, Honduras could potentially increase its efficiency o f public spending by 24 percent in education and by 5 percent in health.55 With education sector spending currently representing 7.2 percent o f GDP and health sector spending representing 3.2 percent o f GDP, an improvement in the efficiency o f public spending by the indicated amount would yield fiscal savings o f approximately 1.9 percent o f GDP (1.7 percent ineducation and 0.2 percent in health). If it i s assumed that the scope for improving the efficiency o f the rest o f the 54. The efficiency o f public spending can be measured, either by comparing the amounts o f sector spending to achieve a given sector outcome across countries (input-oriented measure) or by comparing sector outcomes for a given spendinglevel (output-oriented measures). Since this discussion focuses on the fiscal savings generated by an increase inefficiency, the input-oriented measure o f efficiency i s the more directly relevant measure. 55. These potential efficiency gains are calculated on the basis o f Table A3 in Annex A, which presents the efficiency scores pertaining to the 75th percentile o f the world-wide and o f the L A C distribution o f efficiency scores. 53 central government's non-interest spending i s similar in magnitude, then the total fiscal savings potentially obtainable through efficiency improvements would be close to 4 percent of GDP.56 4.31 It is important to recognize, that these calculations of potential fiscal savings accruing from efficiency improvements are likely to be over-estimating the actual gains that Honduras can realistically expect to achieve inthe short and mediumterm. Inprinciple, fiscal savings of this magnitude might be possible ifthe authorities were free to re-allocate resources any which way. In practice, however, the authorities are limited by constitutional constraints (e.g., mandatory municipal transfers), international treaties (e.g., governing debt service payments) and strong unions (e.g., among doctors and teachers), which severely constraint their flexibility in re- allocating resources more efficiently. So, even though these simulation results are not meant to serve as a basis for concrete policy recommendations, they are useful for heuristic purposes and to motivate debate. Table 4.9: Potential Sourcesof Fiscal Savings for the Central Government Actual Levels Target Potential (2005) FiscalSavings as percent of GDP Expandingthe ResourceEnvelope - 3.0 through: HigherTax Revenues 17.0 20.0 3.O LargerPrimary Deficit 2.6 2.6 0.0 Reducing Expenditures 4.6 -6.0 through: Reduction of Public sector subsidies 3.5 1.5 -2.9 0.6 -2.0 Efficiency Gains inEducatiodHealthspending 10.4 8.5 1.9 Efficiencv Gains inother non-interest ewenditures 11.5 9.4 2.1 Source: paras. 4.22-4.33 of maintext. 4.32 Table 4.9 summarizes the main sources of potential fiscal savings to finance additional MDG-related expenditures by the central government as discussed previously. It identifies potential fiscal savings of between 7.6 and 9.0 percent of GDP, of which 3.0 percent are associated with an expansion of tax revenues, between 0.6 and 2.0 percent are associatedwith a reduction inpublic subsidies and 4.0 percent are associated with improvements inthe efficiency of public expenditures. These fiscal savings would helpto close most of the gap of 10.4 percent of GDP in additional public spending required by 2015 vis-a-vis 2004, in order to achieve the MDGs. Although this represents a significant reduction in the overall resource gap, the remaining gap still poses a major challenge, considering that it is close to the maximum levels o f foreign grants obtained by Honduras in the past,57and appears at a time when Honduras's per- capita income level is risingabove the eligibility thresholds for concessional assistance. 56. Total non-interestcentral government spending amountedto 22 percent of GDP in2005. The potential fiscal savings fiom improvements in efficiency in education and health spending were calculated as 1.9 percent of GDP, out of a combinedtotal spending on educationand healththat amountedto 10.4 percent of GDP. The total potential fiscal savings figure indicated inthe maintext was derived by applying the same proportionof fiscal savings to total spending calculated for education and health (1.9A0.4) to the entire amount of non-interest central government spending. 57. During the post-hurricane Mitch reconstruction period (1998-2002), annual inflows o f official grants never exceeded2 percentofGDP; see IMF, Staff Report for the 2003 Article IV Consultations. 54 D. MainConclusionsandRecommendations 4.33 Honduras has made significant progress inimproving its social indicators since 1990, and may hopeto achieve several ofthe MDGs (e.g., people living on less thanUS$1per day; under-5 mortality rate) by 2015 ifthe current pace of advance is maintained. Nevertheless, to maintain that pace of advance, MDG-related expenditures will have to grow faster than the 3.9 percent real GDP growth projected for Honduras over the medium and long term. The simulations performed with a dynamic general equilibrium model that links public expenditures to social outcomes indicate indicate that inorder to reachthe MDGs by 2015, either total public spending inMDG-related areas wouldhaveto increaseabovethe 2004 levelby around 10percentof GDP by 2015, or the overall efficiency o f public spending would have to increase by roughly 89 percent inthat period. 4.34 Faced with such high expenditure requirements, this chapter has analyzed the potential fiscal space available to the Honduran authorities for increasing public spending inMDG-related areas. This analysis focused on four broad alternatives for financing such an increase in public spending: (i) more public borrowing, which implies increasing the primary fiscal deficit, (ii) raising the level of tax revenues, (iii) eliminating untargeted subsidies to yield fiscal savings that can be switched toward MDG-related areas, and (iv) generating fiscal savings by improving the efficiency of public expenditures. The main findings from that analysis are summarized below: 4.35 The scope for increasing the fiscal deficit. There does not appear to be any significant room for increasing the primary fiscal deficit in Honduras over a prolonged period, without jeopardizing macroeconomic stability or violating over public policy objectives. A debt sustainability analysis presentedinAnnex C concludesthat ifnet concessionalfinancing inflows from abroad as well as GDP growth remain at recent historical levels, then Honduras may afford a primary fiscal deficit as high as 2.3 percent of GDP, before its (non-concessional) public debt ratio exhibits an unstable trajectory. IfHonduras were also preparedto contemplate a significant increase in domestic inflation rates for the sake of maximizing seigniorage revenues, it could raise the sustainable primary deficit a little further, to perhaps 3 percent o f GDP. It turns out, however, that this maximal sustainable deficit level i s not far removed from the actual primary deficit exhibited by the combined public sector during 2001-06, when it averaged 2.9 percent o f GDP. 4.36 The scope for raising tax revenues. A recent study by the IDB identified various opportunities for reforming the tax system in Honduras to generate further revenue increases as well as improving the efficiency, transparency, and equity of the tax system. Implementation of these measures i s projected to raise total revenues by close to 4 percent of GDP in the long run, relative to the 2002 tax ratio of around of 16percent of GDP. Since then, central government tax revenues increased to about 17.0 percent o f GDP in 2005, leaving an additional margin of 3 percent of GDP. Looking beyond this margin, the scope for further tax revenue increases appears limited, considering that Honduras already exhibits the highest tax ratio in Central America and among the highest inLatinAmerica. 4.37 The scope for expenditure switching. The public sector subsidies currently being provided in various sectors and amounting to approximately 3.5 percent of GDP represent obvious candidates to consider for expenditure cuts. Many of these subsidies are badly targeted and, thus, wasted. It is estimatedthat the government could save between0.6 and 2.0 percent of 55 GDP by improving the poverty targeting o f these subsidies. These fiscal savings could help to finance the expansion o f priority expenditures by an equivalent amount 4.38 The scope for improving efficiency. Honduras generally ranks below the median in regional comparisons o f public spending efficiency indicators, which suggests that there is considerable scope for improvement on this front. By implementing measures that raise its efficiency scores to the level o f countries located at the 75' percentile o f the distribution o f efficiency scores, the preceding analysis found that, Honduras could potentially increase its efficiency o f public spending by 24 percent in education and by 5 percent in health. Weighing these potential efficiency gains by the amount o f public spending in these two sectors (a little over 10 percent o f GDP), yields a projected fiscal savings o f approximately 1.9 percent o f GDP. Ifefficiency gains inasimilarproportionwere also achievedwithregardto the restofthe central government's non-interest expenditures, the overall fiscal savings would amount to roughly 4 percent o f GDP. 4.39 The preceding debt sustainability and international benchmarking exercise yields a combined fiscal savings o f 7.6 to 9.0 percent o f GDP. This would reduce the overall resource gap to finance the additional expenditures needed to achieve the MDGs to approximately 1.4 to 2.8 percent o f GDP. Although this represents a significant reduction, the remaining resource gap still poses a major challenge. Even so, the authorities would be well advised to at least aim to achieve these potential fiscal savings in order to come closer to reaching the MDG targets than would be the case under baseline assumptions. 4.40 Accordingly, the main recommendations that result from this analysis are: 0 D o not allow the primary fiscal deficit to increase above the level reached in 2006 (2.6 percent o f GDP), inorder to avoid destabilizing growth inthe public debt. 0 Implement the remaining tax reforms identifiedinthe IDB studies. 0 Reduce or eliminate the badly targeted subsidies currently offered in the electricity, fuel, transport andtelecommunications sectors; as recommended inChapter VIII. 0 Implement efficiency-raising reforms, especially in education, where Honduras appears to exhibit the greatest shortcomings inefficiency. (Specific measures to enhance the efficiency in the education, health and transportation sectors are provided in the respective sector- specific chapters.) 56 CHAPTERV EDUCATIONDEMAND AND TEACHER SALARIESINHONDURAS 5.1 Chapter I1found that public sector salaries in Honduras are extremely high by regional standards, and that much of that is due to the rapid rise in the wages of teachers, who are extremely well organized and politically powerful. The agreement reached with the teachers union in August 2006 is poised to boost teacher wages even further. Chapter 111, inturn, found that the high level of public wages in Honduras represents one of the most important fiscal obstacles to economic growth. These findings argue strongly in favor o f containing the growth of teacher wages in Honduras. Honduras i s also a country with rapid demographic growth, however, and with very low average schooling attainment levels (another key growth determinant). To satisfy the rapidly rising demandfor education, it will be necessary to expand the supply of education services by a commensurate amount. This will necessarily require an increase inteacher services, which puts the authorities in a difficult bind. This chapter seeks to sort out how much additional public spending on education i s warranted by the estimated increase inthe demand for education estimated over the next decade, and how much additional spending is likely to result from the recent teacher agreement, independento f any changes inthe supply or quality of teaching services. It concludes by recommending a set of measures that could mitigate the fiscal impact of the August 2006 teacher wage agreement and improve the quality of teaching inHonduras. A. Sector Overview 5.2 Honduras has registered steady improvements in key education indicators since the end of the last decade (Table 5.1). Between 1998 and 2006 the average years of study in the population aged 10 or more increased from 6.1 to 6.4 years, the illiteracy rate decreased from 15.2 to 14.1 percent, preschool attendance increased by almost 44 percent and the primary attendance of 9-year olds rose by 2.9 percentagepoints. Gross primary education coverage rose slightly in the period 1999-2004 to 113 percent, while net coverage (ages 7-12) rose to 90.6 percent.58 Gross secondary coverage also has grown inthat period, from an estimated 35 percent to 65.5 in 2004, while coverage in higher education grew from 14.4 percent of the population aged20-24 in 1999to an estimated 18.7percent in2004.59 Table 5.1: Trends inEducationalOutcomes 1998 2002 2006 Total Urban Rural Total Uhan Rural Total Urban Rural Average yews of study, ppln.Aged lo+, % 6.1 6.0 6.3 5.0 6.6 3.5 6.4 7.7 5.0 IlliteracyRate, % 15.2 15.6 13.1 16.4 8.7 25.6 14.1 7.5 22.0 P r i m qAttendance, children aged5 and6, % 28.7* 28.1* 32.2* 31.3 40.2 25.5 41.2 45.4 38.5 PrimaryAttendance, childrenaged9,% 93.3 92.7 97.0 94.4 96.4 92.5 96.2 96.9 95.5 Source:InstitutoNacionalde Estadistica.EPHPM.May. * Datafrom1999. 58. World Bank, edstats. 59. Unesco (2006), Informe sobre la Educaci6n Superior en America Latina y el Caribe 2000-2005. 57 5.3 With 7.2 percent of GDP devoted to public education in 200660,Honduras spent much more on this sector than other Central American countries (Table 5.2). This contributed to Honduras' relatively high primary enrollment rate. In spite of this relatively strong funding effort, however, Honduras' education outcomes are not superior to those of its neighbors. Table 5.2: EducationOutcomes in Centraland LatinAmerica Honduras CostaRica ElSalvador Guatemala Nicaragua Average LAC AdultLiteracyrate, 2000-2004(age Ut),% 80.0 94.9 85.0'' 69.1 76.7 90.2 Primqcompletionrate, 2004, % 79.4 92.3 85.7 70.2 73.5 96.0 NetPrimqenrollment,2004, % 90.6 8a.P 92.3 93.0 87.9 95.3 NetSecondqenrollment,2004, % 36.1 40.8n 48.1 34.0 40.7 65.5 Publicedu.Spending % GDP,2004 7.3fi 4.9 2.8 1.7" 3.1 4.3 Source:Unesco andWorldBank edstats. "Year1990.' Year1995. 'CentralGov.Expenditureineducation."Year 1995. 5.4 Key shortcomings remain in terms of completion rates, secondary coverage and education quality. Grades 1-2 still have high repetition rates that set childrenback, and about 30 percent still fail to graduate from 6thgrade (the percentage of children completing the 6th grade has remained stable at 70 percent since 1994). Inequities inenrollment and completion rates are persistent across urban-rural areas and socioeconomic strata. 5.5 Figure 5.1 shows that the enrollment in lower secondary education in Honduras is lower than countries with equivalent per capita income. Behind this low enrollment there is a low transition to seventh grade. Figure 5.2 provides a graphical illustration of the survival rate to grade 11, using a survivor function plot, which projects the expected probability of survival of the current 7-year old cohort to each grade. It shows that students drop out continuously throughout the educational cycle (primary and secondary cycles), but there is a particularly low transition from grade 6 to 7. Figure5.1: The Secondary EducationGap V I Figure 5.2: Survivor FunctionPlot by Years of Edul ion 1 - . I I 'I 1: P " 44:: P 8 s II i G 5.6 In terms of education quality, standardizedtest scores in Spanish and Mathematics, at third and sixth grade, have shown no improvement since 1997. The average percent of correct 60. This correspondsto the CentralGovernmentExpenditureinEducation. 58 answers inthe standardizedUMCE test of 2004 are below 50 percent in Spanish or Mathematics at third and sixth grade (considered low levels of achievement in the scale used by UMCE). Only 20 percent o f the students reached the sufficiency level (attain more than 60 percent of correct answers in the test). Additionally, there is a significant gap in terms of performance betweenrural andurbanareas. Urbanscores inthe 2004 UMCEtest were five percentage points higher than rural scores. In terms of international assessments, even after allowing for Honduras' low per capita income, Honduran students perform below almost all other Latin American countries in the 1997 Latin-American Laboratory Assessment (LLECE). This poor performance is consistent with the country's high repetition and drop-out rates and may be reflecting in part the lack of time devoted to learning activities by students (school closures, teacher absenteeism andtardiness, etc.), insufficientperformance and accountability by teachers, text availability, lack of adequate teacher training for textbooks and knowledge o f subjects that characterizeHonduras' public education system. Figure 5.3: Quality: National Student Assessment Results (2004) Ian I B. The Demand for Schooling 5.7 Table 5.2 describes the evolution of education in Honduras over the past half-century.61 Mean schooling attainment has been rising steadily for at least half a century. It went from 3 years for the cohort born around 1939 to 6.7 years for the one born around 1981. Population growth, together with the doubling in school attainment has meant a very rapid rise in the country's overall average attainment. Furthermore, the growth inschooling appears to havebeen concentrated at the lower levels. Among the oldest cohorts two fifths received no schooling at all and half had only some primary education. These cohorts had an average of about 3 years of schooling, but this average came mostly from a higher level of schooling for an elite group that 61. Table 5.2employs the Lamand Levinson (1992) method for inferring secular change fi-om cross-sectional data. The data were grouped into 15 3-year cohorts representing some 2.3 million people between the ages o f 22 and 66. Average educational attainment in 2004 for this group stood just below primary completion level, at 5.6 years o f formal schooling. Since all o f the data were gathered in2004, there is a simple correspondence between age and year of birth. Furthermore, because most people enter the school system at approximately the same age, differences across age groups can be interpreted as representing the development o feducational policy outcomes over time. 59 accounted for less than 10 percent of the population.62 Among the youngest cohorts, only about 9 percent have no education, two-thirds have some primary schooling, with half of these (one- third) having completed all six years. Another 25 percenthas somepost-primary education. 5.8 Itis well establishedthat the most important factor inexplaining income inequality is the distribution of education.63 The reduction inHonduran schooling inequality can be seen directly by examining the measures shown incolumns (2) and (3). The standard deviation and variance of attainment peaked for the generation born between 1948 and 196964and then declined steadily. One reason for non-monotonic change inthese measures of dispersion i s the rapid rise inmeaneducation. The coefficient ofvariation, amean-adjustedmeasureofinequality, shows a fairly steady decline ininequality over the four decades shown. Table 5.3 A HalfCenturyof SchoolingEvolutioninHonduras:Inequalityand Attainment ---->EducationalInequality<== == >EducationalAttainment <=== Mean Cohort Estimated (1) (2) (3) (4) (5) (6) (7) (8) (9) Cohort Age Cohort Birth Size Mean St. Coeff. Of Some Year Schooling DeviationVariance Variation None Primary 7-9 10-12 13+ 1981 22-24 386,479 6.71 3.96 15.70 0.590 8.76 65.76 10.21 12.8 2.47 1978 25-27 3 10,797 6.60 4.15 17.22 0.629 9.86 61.74 10.34 13.15 4.91 1975 28-30 271,153 6.20 4.29 18.38 0.691 10.63 62.87 9.22 11.45 5.83 1972 31-33 225,627 5.86 4.21 17.75 0.719 12.28 63.16 7.78 11.11 5.67 1969 34-46 236,997 6.07 4.55 20.66 0.749 13.48 57.3 1 8.7 12.78 7.73 1966 37-39 201,750 5.91 4.64 21.52 0.784 14.26 57.53 8.48 12.28 7.45 1963 40-42 186,939 5.86 4.75 22.58 0.811 14.75 56.06 7.71 14.43 7.05 1960 43-45 181,321 5.24 4.77 22.74 0.911 20.10 56.07 4.25 12.36 7.22 1957 46-48 151,597 4.98 4.48 20.05 0.898 19.49 58.97 5.35 9.78 6.41 1954 49-51 137,806 4.69 4.48 20.07 0.956 20.73 59.68 4.89 7.85 6.85 1951 52-54 129,83 1 4.54 4.70 22.09 1.035 24.46 53.54 3.23 12.19 6.58 1948 55-57 101,293 4.13 4.49 20.17 1.087 25.93 56.58 3.79 8.48 5.22 1945 58-60 88,897 3.39 4.22 17.79 1.245 35.95 50.22 2.59 7.54 3.70 1942 61-63 72,172 3.08 4.16 17.31 1.349 40.93 46.97 2.05 6.18 3.87 1939 64-66 68,858 3.01 3.87 14.99 1.285 39.20 52.55 0.29 5.03 2.93 Total 37.9 2,751,517 5.6 4.49 20.16 0.802 16.73 59.03 7.16 11.4 5.68 Source: see Footnote 1. 5.9 Honduras is in a very special stage of educational development. Age-appropriate (net) school coverage has increased dramatically, meaning that more and more children are entering school at the appropriate age. But at the present time, not only is a higher percentage of young children going to school, older people are catching up on the education they missed. Gross enrollment rates are very high. In fact the MECOVI data show that 15.2 percent of the population betweenthe ages o f 21 and 30 are enrolled. Less thanhalf ofthese (48 percent) are in tertiary programs. About 38 percent are in secondary school programs and another 15 percent 62. Note an inherent bias in this method exists if there i s a correlation between educational attainment and longevity. It may be true that more educated people have better nutrition and health care as well as less dangerous lives. If so, a cross section o f the population will over-estimate the percentage o f educated people inthe original cohort for the older age groups. The results reported in Table 5.2may therefore under-estimate the improvement both inmean education and education distribution. 63. For instance, Lamand Levinson (1992) andAlmeida dos Reis, et al. (1991). 64. They should have enrolled inschool at age 7, between 1955 and 1976. 60 are inprimary school programs! Thus the data in Table 5.2 actually understate the educational progress being made by the younger cohorts. The FutureDemandfor Education 5.10 Figure 5.4 depicts Honduran education today. The top line shows the percentage o f each age group enrolled - whatever the grade level. The line below it indicates the percentage that i s net-enrolled. Among those who are enrolled, the variable "net-enrolled'' takes on the value o f 1 if(Age-attainment-7) <2. Inother words, a studentis consideredto benet-enrolled ifitiswithin two years o f the appropriate age for the grade they are in.65The two values are equal only at age 7, where 7-year old students in first grade can not have repeated.66 After that a gap develops between net and gross enrollment. Peak overall demand for education emerges betweenthe ages o f 7 and 12. This six-year span is sufficient to finish primary school, but fiom the difference in these two curves it i s evident that much o f this demand i s due to repeated attempts to pass the lower grades. Another noticeable pattern i s that by age 16, less than half o f the population i s in school. Given Honduras' highage o f initial enrollment and highrepetition rates, this means that only a small minority o f students is completing secondary school. Figure 5.4 Honduras (2004) SchoolEnrollment by Age a9 I , : \ u o a7 a6 a5 ....... a4 a3 a2 a1 0 5.11 Demand for education will grow markedly over the next decade. Several trends conspire to bringthis about: Honduran income per capita is growing and young parents have at least twice as much education as did their own parents. Income and parental education are the two most important determinants o f demand for s~hooling.~' Increased demand for education will 65. The number 7 appearsbecausethis is the age legally requiredfor first grade enrollment. 66. Note that the value must be less than two. This allows for negative values, as for a 7-year old student with 2 years o f school completed (value=-2). This method counts students that are worse-off, treating those doing better thantheir cohort as the same inthe sensethat they are "no worse-off". 67. There are two reasons parents will want their children to enroll at a younger age than before. One i s that they want more education for their children and will now be able to afford it. Another i s that becausethe new generation 61 manifest itself in several ways that have distinct impacts on the structure of the educational system. Most importantly, these include the age of initial enrollment, the quality of education, andthe duration inschool. 5.12 The demand for education quality will also grow. Educated parents will have a better understanding of when learning is or is not taking place and they will be more assertive in demanding it. At the same time, more educated parents will better reinforce school-based learning by providing an appropriate home environment and helping with homework. As , Honduras continues to more clearly define curriculum objectives and adopt objective criteria for passing and failing grades, higher quality education will be manifestedin lower repetition rates. This will reduce the proportion ofthe student population inthe first and second grade of eachthe three levels (primaria, ciclo comzin, diversiJicado)and increasedemandfor the higher grades. Figure5.5 Honduras-Survival in the School System, by Grade(percentages) 100.00 90.00 80.00 \ 5 50.00 +Probability Will Enroll InThis Grade 30.00 . \ 20.00 . 10.00 . 0.00 - I I , , , I I 5.13 Finally, demand for grade attainment will rise, as studen s will I ay in school for more grades. There are three reasons for this: first, earlier enrollment will have a direct impact on school attainment. This i s because for a given age of withdrawal from school, the earlier a child enrolls, the more grades it will be able to attain. Consider a boy and a girl withdrawn from school at age 14, the legal age of employment in Honduras. Suppose further that the girl was enrolled at age 10 andthe boy when he turned 7. With equal abilities this will meanthat the boy could receive 7 years of schooling, while the girl could attain at most a fourth grade education. Secondly, higher school quality will meanthat time spent inschool i s more rewarding and that - with lower repetition rates -it will be less costly to attain more grades. Thirdly, higher parental incomes and education will have direct effects on quantitydemandto the extent that education i s a normal good and to the extent that at any given income level more educated parents tend to allocate more of their time and money to the education of their children. of mothers is more educated they will be able to earn higher wages, so more ofthem will want to work. Inorder to be able to work they will search for a safe and rewarding alternative to maternal childcare. School provides it for fiee. 62 5.14 Table 5.4 shows the combined effect of all of these forces on estimated schooling demand from 2006 to 2017. Initially, enrollment growth is even, but after a decade it becomes apparent that there i s a marked difference across grade levels. Forecast enrollment growth i s inverse U-shaped, highest in the middle, significantly lower in grades 1-3 and lowest of all for grades 11and 12. Thus, while significant growth can be expected at all levels of schooling, the biggest change will be due to higher transition rates from primary to secondary schooling. The weighted average forecast by sector is 40.4 percent growth at the primary and 43.8 percent growth at the secondary level for the decade after 2006. Table 5.4 PrqjectedEnrollmentGrowth2007 to 2017, by Grade(MECOVI 2004) Grade Year 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 2006 100 100 100 100 100 100 100 100 100 100 100 100 2007 103 103 103 104 104 104 104 104 103 103 103 103 2008 106 107 106 107 107 108 109 108 107 107 106 106 2009 110 110 109 111 111 112 114 112 110 111 108 108 2010 113 114 112 115 115 116 118 116 114 114 111 111 2011 116 117 115 118 119 120 123 121 117 118 114 114 2012 120 121 119 122 123 124 128 125 121 121 117 117 2013 123 124 122 126 126 129 134 130 125 125 119 119 2014 127 128 125 130 130 133 139 134 128 129 122 122 2015 131 132 128 134 134 138 145 139 132 133 124 125 2016 134 136 132 138 139 142 150 144 136 136 127 127 2017 138 140 135 142 143 147 156 149 139 140 129 129 Source: Calculations by John Edwards. The Demandfor Teachers 5.15 The demand for teacher services depends on student demand for education, but also on several other factors, including class size and curriculum. While the data that were available for this study did not report class size, it is unlikely that increases in class size can contribute significantly to absorbing education demandwithout seriously compromising education quality. For instance, Edwards et al. (1997) report that at the primary level average class size was 40.7. Also, in 2004, 17.9 percent of rural primary school students report that they are attending a school withjust one teacher for all the grades. Ifclass size and curriculum are held constant and projected enrollment demand growth is accommodated, then the number o f teachers must grow in proportion to the number of students over the next decade. Namely there must be a 40.4 percent increase at the primary and a 43.8 percent increase at the secondary level. If the pay structure i s also kept intact then this o f course translates directly into teacher salary expenditure increasesofthe same proportions. Optionsfor dampening expenditure increases 5.16 Increased Class Size. The MOE data show that all public school positions with the title of "Maestro/a" require 36 hours of instructionper week. Since there are 32,757 primary school jobs, this amounts to 1.2 million hours of instruction per week nationally. If classroom size i s 63 kept the same, this will have to grow to 46,089 teaching positions and 1.66 million weekly hours o f instruction inten years. 5.17 At the secondary level there are 25,416 public school positions with the title o f "Maestro/a" or "Profesor/a" providing a total o f 362,670 hours o f instruction per week. This amounts to an average of only 14.3 hours per week perjob, but the range goes from 2 to 56 hours per week, with 54 percent of the jobs requiring 10 hours or less per week, 40 percent requiring between 11 and 36 hours, and only 5 percent requiring more than 36 hours. 5.18 MECOVI-based enrollment estimates suggest about 1.1 million primary and 0.3 million secondary school students in Honduras in 2006. This would mean that the average student teacher ratio i s currently about 33 at the primary and 14 at the secondary school level. These numbers again make it evident that it is not advisable to use classroom size expansion as the major cost-containment policy instrument at the primary level. It may be possible to scrutinize the school system and find some sparsely populated primary schools, but in general increased classroom size is likely to adversely affect school quality at this level. 5.19 At the primary level, the student teacher ratio translates directly into students per class. At the secondary level this is not true because subject specialization means that students are in front o f more than one teacher each day. Thus ifthe average student teacher ratio i s 14 and two teachers are required, then class load is 28. With three teachers it would be 42, and so forth. So increasing class size is not likely to be an effective way o f obtaining efficiency gains at the secondary level either. 5.20 Curriculum Change. Another possibility is to alter the curriculum in a way that reduces the need for teachers. This will not be possible at the primary level. Here, students typically have only one teacher for the entire school day. Also, total yearly hours o f instruction in Honduras are already low by international standards. 5.21 It may be possible to consolidate teaching posts at the secondary level. Careful review might identifyways to tighten up the curriculum without harmingthe quality o f education inkey basic areas. For instance, in El Salvador in the 1990's the national curriculum contemplated optional instruction in woodworking, acting, cooking, make-up, and accounting. An ordinary secondary school teacher i s not capable o f teaching these subjects, so special teachers had to be hired. Not all schools were able to offer the optional subjects. Typically it was only the larger, urbanpublic schools in well-to-do neighborhoods that could do so. Savings can be realized if this sort o fpattern is found inthe Honduran system too. Specialized classes that are not part o f the core curriculum, that are not offered at all schools can be moved out o f the regular curriculum. They can be placed into an after school program financed by user fees. This will free up classrooms and salaries to accommodate more students incore curriculum education. 5.22 Private Sector Alternatives. Private schools can provide an alternative to public sector expansion. If there is excess capacity in the private sector, the additional cost o f admitting students will be below current average cost. This means that the government can negotiate with private schools to make space available at preferential tuition rates that may well also be less than what it would cost the government to provide the spaces publicly. Once a list of quality- 64 certified private institutions is developed,68vouchers or other mechanisms can be usedto direct the expenditures to targeted groups. 5.23 Once again, this i s not an alternative at the primary level, but it may work at the secondary level, especially inthe upper 3 grades. The MECOVIhousehold survey shows that in 2004 less than 1 percent of Honduran primary school students were in private schools. So whatever private school excess capacity may be at this level, there i s no way that existing private schools could absorbthe expected40 percent expansion intotal primary enrollment over the next decade. 5.24 The private sector plays a much more substantial role at the secondary level. Private schools account for 15 percent of enrollment in grades 7-9 ("Ciclo Basico") and 37 percent of enrollment in grades 10-12 ("Diversificado"). However, we do not have information on their capacity nor on the quality of private secondary education. Budgetary Impact of the IncreasedDemandfor Education 5.25 Ifthe public sector maintains its share ofthe projected expansion instudent enrollment, without changes in wages or in average studentlteacher ratios, projected increase in teacher demandwill have an enormous fiscal impact. This would have beentrue even without the very large increase in teacher salaries agreed to in August o f 2006. Table 5.5 shows current educational wage expenditures stand at 6 billion Lempiras -or about USD $318 million. This amounts to 3.8 percent of 2005 GDP. To put this inperspective, consider that total expenditure and lending of the consolidated non-financial sector was 32 percent of GDP. Teacher salary expenditures are more than 10 percent of the total. Even if recent economic performance is maintained for the next 10 years at the 2.3 percent mean yearly GDP growth level attained over the 2000-2005 period, expenditures on teaching as a share of GDP will have to increaseby 12.7 percent to 4.3 percent of GDP by 2017 to keep pace with increased demand under the existing structure. Table 5.5 Honduras:ProjectedIncrease inPublicSector Teaching Expenditures(Aug. 2006 Lempiras) July 2006, Status Quo Sector Number Mean Monthly Total Monthly Current (2006) Estimated 2016 YOChange of Jobs Salary Wage Bill Annual Wage Bill Annual Wage Bill 2006-2026 (14 Months) (14 Months) Primary 32,868 8,895 292,367,466 4,093,144,53 1 5,746,774,92 1 40.4% Secondary 29,5 19 4,610 136,097,202 1,905,360,827 2,739,908,869 43.8% Total 428,464,668 5,998,505,357 8,486,683,790 41.5% 5.26 Evenif teacher compensation were to remain constant inreal terms, accommodating the projected growth ineducation demandwill strain the Honduran fiscal system. It will be seen that the recent salary accords betweenthe government of Honduras and the national teachers unions threatento derail recent economic progress and stability. 68. Inaddition to the issue o f quality, there may be constitutional issues involved inemploying public fundsto pay for private schools that have a religious affiliation. 65 C. The Structure of Teacher Salaries 5.27 This section begins by examining public educational manpower allocation and expenditures on primary and secondary schooling in Honduras as they existed up to the August 12PASCE accord. Multiplejobs 5.28 Table 5.6 shows that there were 71,139 jobs listed on the MINED payroll (panel 1) and 49,727 employees held that these. Most of the analysis in this section will concentrate on the primary and secondary sectors, whichjointly account for 88 percent of alljobs and 86 percent of all personnel. A striking difference betweenthese two sectors emerges immediately: the primary sector accounts for 63 percent of all personnel, but for only 46 percent of all posts, while the secondary sector accounts for only 22 Table 5.6 percent of the personnel, but for 41 (1) All Jobs by Level (2006 MINED) percent of all posts. Number Percent Cumm. YO Educaci6nBBsica 2388 3.36 3.36 5.29 MOE personnel include EducacidnMedia 29519 41.49 44.85 administrators, teachers, and teachers' Educaci6nPre-BBsica 5 146 7.23 52.09 aides of various sorts. In fact, in Educaci6nPrimaria 32868 46.2 98.29 addition to teachers, the MOE databases Educaci6n a Distancia 483 0.68 98.97 record 103 different non-teaching Educaci6n de Adultos 735 1.03 100 positions. However, most of them are held by a very few people. Teaching (2) Teachers by LevelTaught inPrincipalJob (2006 MINED) positions69 account for just over 91 Number Percent Cumm. YO percent of all positions held. Educaci6nBBsica 1678 3.37 3.37 Furthermore, at the primary level only 4 EducacibnMedia 11162 22.45 25.82 distinct positions are recorded, namely Educaci6nPre-Bhsica 4994 10.04 35.86 Director, Sub-director, Maestro/a and Educaci6nPrimaria 31334 63.01 98.88 Pofesor/a whereas almost 100 Educaci6n aDistancia 170 0.34 99.22 administrative, auxiliary, and Educaci6n de Adultos 389 0.78 100 Total 49727 specialized teaching jobs -such as lab assistant, shop supervisor, and vocational councilor -arise at the secondary level. So, while the primary structure is very simple, the secondarystructure canbevery complex. 5.30 Honduras had about 47,139 teachers, but 66,632 teachingpositions. The definition of "teacher earnings" inthis environment i s not an easy task. Teachers can hold more than onejob. They can hold two or more teaching positions, they can hold teaching and administrative positions, and they can hold more than one administrative position. Ministry o f Education (MOE) databases record teacher data by "job" or "post", while INPREMA and the Ministry of Finance databases record it by employee. 5.31 Table 5.7 shows MINED data for all public schools in August of 2006, aggregated to makethemteacher-based. They record 49,727 public sector teachers ina total of 71,139 "jobs", or posts. The fact that some teachers hold more than one job is often articulated along with concern that it could affect teacher performance or result in some teachers being unfairly overpaid. For instance teaching quality and teacher absenteeism could be affected adversely if 69. Those withjobs labeledas "Maestrola" or "Profesorla". 66 jobs are far apart. Also, multiplepostings could signal administrative anomalies too, for instance teachers that have changedjobs but continue being paid twice, even though they are really only working at the new job. Until now most evidence has been anecdotal. There has been little substantiation of how widespread or how problematic this practice is. Table 5.7 All PublicSchools MultipleJobs Held (MINED2006) ## of Jobs with # of Teachers with YOof all Cumulative% # ofJobs YOof All Jobs Cumulative% Salary this manyjobs Teachers of all Teachers of all Jobs 1 39181 78.79 78.79 39181 55.08 55.08 2 5397 10.85 89.65 10794 15.17 70.25 3 2391 4.81 94.45 7173 10.08 80.33 4 1316 2.65 97.1 5264 7.40 87.73 5 719 1.45 98.55 3595 5.05 92.79 6 348 0.7 99.25 2088 2.94 95.72 7 177 0.36 99.6 1239 1.74 97.46 8 96 0.19 99.79 768 1.08 98.54 9 44 0.09 99.88 396 0.56 99.10 10 28 0.06 99.94 280 0.39 99.49 11 18 0.04 99.98 198 0.28 99.77 12 5 0.01 99.99 60 0.08 99.86 13 2 0 99.99 26 0.04 99.89 14 1 0 99.99 14 0.02 99.91 15 1 0 99.99 15 0.02 99.93 16 3 0.01 100 48 0.07 100.00 Total 49727 100 71139 100 5.32 Multiple posting of teachers generates statistical confusion about hours worked and total earnings, Teacher databases mainly serve the purpose tracking payroll and manpower distribution across schools and they do this best by tracking each job as a separate entry, regardless of who holds it. However, to understandthe incentive environment and education- quality consequences it is necessary to understandthe total number of hours each teacher work, hertotal incomeandaveragehourly compensation. 5.33 For this study, MOE and INPREMA data were reorganized and aggregated by teacher. These data are summarized inTable 5.7. It shows that there are indeedteachers with more than onejob. Buti s this a serious problem? Within the Honduran educational environment it is often heardthat "teachers hold 10 or morejobs". The data show this is true, infact there are teachers holding as many as 16jobs, but the number of teachers with 10jobs or more i s so small as to be insignificant (less than one-tenth of one percent). More than 80 percent of the teachers have only onejob and 99 percent of the teachers hold fewer than6jobs. 5.34 In August of 2006 1,204 people on the payroll (2.4 percent of total) held only non- teaching jobs. They were earning an average of 11,980 Lempiras per month from all sources, with a bi-modal and somewhat right-skewed distribution (Figure 5.6a). This compares to an average monthly income of 9,193 Lempiras for 91 percent o f people on payroll (45,168) who hold only teaching positions with the title of "Maestro/a" or "Profesor/a". The remaining 3,353 MOE employees heldboth teaching and administrative jobs. They earned an average o f 17,413 Lempiras per month and constitute the highest average income category of teacher. This 67 category also has the highest salary spread, as can be seen by comparing the kernel density estimates of Figure 5.6~with those in 6a and 6b. The salaries among the 91 percent that only teach are the most systemically skewed(6b). The very noticeablemode that appears early inthe distribution, at about 5,000 Lempiras, represents the earnings of the bulk of new teachers with basic qualifications and no seniority. The skew is by systemic design of the incentive system that existedupto August of 2006, which purposefully assignslarge rewardsto academic degrees, hardship postings, and seniority. It will be seen below that the seniority reward structure accounts for most of the linear rightward skew in teacher salary distribution evident in Figure 5.6b. .0001 Density .00005 - 0 - I I I I I 1 ,0002 ,00015 Density .0001 .00005 0 0 10000 Salary from alljobs by idente 20000 30000 FIGURE 5.6b: Salary Distribution, NonAdministrativeTeachers .OW15 1 Density .0001 ,00005 0 - 0 10000 Salary fmm alljobs: by idewe 20000 30000 40000 FIGURE 5 . 6 ~Teaching & Non-TeachingSalary Distribution : 68 5.35 The phenomenon o f multiple job holdings appears to be more o f a problem for the teachers than for the Honduran educational system. It is much more common at the secondary school level, where teachers that specialize in certain subjects like chemistry or physics have a difficult time finding a post at a single school that completely fills their schedule. Table 5.8 shows that at the primary level in August o f 2006 there were a total o f 31,228 persons on the MOE payroll assigned to public primary education. Altogether they held 32,786 jobs, or approximately 1.05 jobs per person. While there are as many as 10 posts held simultaneously, almost all primary school teachers (96 percent) have only one post. Table 5.8 PublicPrimary Schools Multiple Jobs (MINED 2006) # of Jobs # of Teachers with Percentof all Cumulative YO# of Jobs YOof All Cumulative YOof with Salary this manyjobs Teachers of all Teachers Jobs all Jobs 1 30081 96.33 96.33 30081 91.75 91.75 2 914 2.93 99.25 1828 5.58 97.33 3 130 0.42 99.67 390 1.19 98.51 4 48 0.15 99.82 192 0.59 99.10 5 32 0.1 99.93 160 0.49 99.59 6 15 0.05 99.97 90 0.27 99.86 7 4 0.01 99.99 28 0.09 99.95 8 1 0 99.99 8 0.02 99.97 9 1 0 99.99 9 0.03 100.00 10 2 0.01 100 Total 31228 100 32,786 100 5.36 Job structure at the secondary level is entirely different than at the primary level. The databases list 8,948 secondary school employees holding a total o f 25,364 jobs. This is an average o f nearly 3 jobs per person. More than three-fourths o f the employees hold at least two jobs (78 percent) and nearly half (47 percent) hold 3 or more. The last three columns show the same data fiom the "jobs" perspective. O f all public secondary level jobs only 8 percent are the only job held by somebody. About 70 percent ofjobs are held by people with 3 or more and half by people with 4 or more jobs. In summary, it is not generally true that secondary school teachers hold a large number o f jobs -say 10 or more. But it is relatively uncommon for a teacher to hold only onejob. 5.37 What are the issues raised by these results? Implications for teacher efficiency should be examined closely. Perhaps in most cases this is just an accounting quirk. For instance the finding that one teacher holds two posts could turnout to meanthere is someone teaching 9' and loth grade math every day at the same school, but the accounting scheme shows this as two separate part-time posts. That would not be a problem. On the other hand, if the same data reflect a person teaching 9' grade math at one school and 10' grade math at another one on the same days with separately coordinated schedules, then that could be a problem. 5.38 Another issue is the payment o f the bonuses to teachers assigned to different posts. Are seniority, education, and hardship-posting bonuses lump-sum amounts to be paid once to each teacher or once for each job? Should they be prorated by the hour, incorporated into an hourly 69 pay andpaid regardless of how many hours are worked? Should itbe some combination of these policies: pro-rated into base pay up to a ceiling? Table 5.9 All Public SecondarySchools Multiple Jobs (MINED2006) # of Jobs # of Teachers with YOof all Cumulative% # ofJobs YOof All Cumulative with Salary this manyjobs Teachers of all Teachers Jobs YOof allJobs 1 1985 22.18 22.18 1985 7.83 7.83 2 2709 30.27 52.46 5418 21.36 29.19 3 1814 20.27 72.73 5442 21.46 50.64 4 1122 12.54 85.27 4488 17.69 68.34 5 639 7.14 92.41 3 195 12.60 80.93 6 323 3.61 96.02 1938 7.64 88.57 7 166 1.86 97.88 1162 4.58 93.16 8 91 1.02 98.89 728 2.87 96.03 9 43 0.48 99.37 387 1.53 97.55 10 26 0.29 99.66 260 1.03 98.58 11 18 0.2 99.87 198 0.78 99.36 12 5 0.06 99.92 60 0.24 99.59 13 2 0.02 99.94 26 0.10 99.70 14 1 0.01 99.96 14 0.06 99.75 15 1 0.01 99.97 15 0.06 99.81 16 3 0.03 100 48 0.19 100.00 Total 8,948 100 25,364 5.39 For instance, consider the hardshipbonus. One way to look at it i s to recognize that there exist regions with school-age populations where teachers would rather not live. From this perspective, the bonus can be seen as a lump sum amount intendedto cover the cost of leaving the region a couple of times per year. Once the teacher has moved to the undesirable region, the purpose of the bonus is accomplished. It need not be paid more than once. There i s a second way to look at it though. Ifthe area is undesirable for educatedpeople because formal education i s a rarity inthe region, then local remuneration (wage) for educatedlabor may be higher. From this perspective, the "bonus" is just a wage adjustment that recognizes regional differences in labor market returns to education. More hours worked means more pay. Such issues need to be thought through andcarefully andtransparentlynegotiated with the teachers andtheir unions. 5.40 Tables 5.9 and 5.10 examine the impact of multiple postings on total hours worked and teacher compensation. The MOE data for August, 2006 were once again ordered by teachers in order to examine the length o f the workweek, monthly compensation, and average hourly compensation. 5.41 The legal definition of "full time employment" inthe public school system is 156 hours per month. Table 5.10 shows that the 96 percent ofprimary school teachers that hold only one job hold it as a full-timejob, with exactly the requirednumber of hours. The big impact on total hours worked shows up in going from one to more than one job. Mean total monthly hours jumps from 156to 267. This is more thana 71 percent increase. 70 Table 5.10 PublicPrimary Schools Compensation & Number of Jobs (MINED 2006) # ofJobs with YOof Teachers with Mean Monthly Mean Hours Per Mean Hourly Rate Salary this manyjobs Gross Income(Lps.) Month 1 96.3 8,73 1 156.0 55.97 2 2.9 16,209 267.1 60.69 3 0.4 17,587 287.3 61.22 4 0.2 17,363 281.7 61.65 5 0.1 17,087 284.2 60.12 6 0.0 17,319 292.1 59.30 7 0.0 18,020 291.4 61.83 8 0.0 20,702 312.0 66.35 9 0.0 15,843 273.0 58.03 10 0.0 19,355 305.5 63.35 Table 5.11 Public Secondary Schools MultipleJobs Held (MINED200) # of Jobs with # of Teachers with this Mean Hours Per Mean Monthly Gross Income Mean Hourly Salary manyjobs Month (Total) (Lempiras) Rate 1 22.2 145 8,397 57.84 2 30.3 212 13,007 61.46 3 20.3 218 12,984 59.49 4 12.5 229 13,204 57.62 5 7.1 226 12,627 55.87 6 3.6 223 12,112 54.31 7 1.9 227 11,900 52.44 8 1.o 236 12,717 53.78 9 0.5 227 11,338 49.85 10 0.3 240 12,766 53.08 11 0.2 258 14,091 54.55 12 0.1 226 12,020 53.14 13 0.0 201 12,696 63.01 14 0.0 299 16,557 55.37 15 0.0 321 17,040 53.14 16 0.0 230 12,524 54.53 5.42 At the secondary school level teachers with only one job are not fully employed (Table 5.11). The jump inhours worked from obtaining a second position is 46 percent. It rises from 145 with one to only 212 hours per month with the second job. A regular teaching week i s supposed to last 36 hours. All primary school teachers with the title of "Maestro/a" or "Profesor/a" are in 36-hour posts. But this is true for only 7.7 percent o f their secondary school colleagueswith the same designation. Infact 87 percent of secondary schooljobs are described as requiringless than a full workweek. 5.43 The impact of having more than two jobs on total hours worked i s far less significant, especially at the secondary level where it i s most prevalent. There is also no strong link between holding more than two jobs and total hours worked. It begins to look like these roving teachers 71 are just trying to fill up their schedules. Notice that gross monthly income i s virtually unchanged as the number of postings rises. To the extent that teachers with multiple postings are able to relatively unchanged.7 r "get more work" the do so at a lower hourly wage, leaving their total monthly earnings TheSalary Structure before August 2006 5.44 Before the agreements of August 2006, the salary structure for Honduran teachers was determinedmainly by a series of statutory regulations jointly known as the Estatuto del Docente. These determine a base hourly wage ("hora clase") of 41.45 Lempiras per hour of class, full- time employment of 36 hours per week, and a month of 4.33 weeks. This translates into a base monthly salary ("sueldo base") of 5,216.64 Lempiras. 5-45 As discussed below, existing legislation also stipulates substantial bonuses for such things as formal education (69 percent), seniority (up to 120percent), and service inthe states of Gracias a Dios and Islas de la Bahia, areas with a higher cost of living (100 percent). These are stipulated as multiplesof the base monthly salary. 5.46 According to the 6 sectors Table 5.12 Average Earningsby Postand Sub-sector differentiated within the MINED data Level Monthly SalaryStd. DeviationNumber bases, the current distribution of teaching Educaci6n Bhsica 6,187.54 2,465.36 2,387 posts, average earnings, and dispersion of Educaci6n Media 3,556.24 3,621.03 25,415 Educaci6n Pre-BBsica 8,183.16 2,461-43 4,98 1 earnings shows (Table 5.12) Educacidn EducacionPrimaria 8,902.05 2,839.44 32,758 Media and Educacidn Bdsica accounting Educaci6naDistancia 2,499.23 1,598.89 481 for over 84 percent of all teaching posts. Educaci6n de Adultos 2,282.10 614.89 528 Educacidn Bdsica and Pre-Bdsica Total 6,610.55 4,034.62 66,550 account for another 11 percent , This leaves distance education and adult education with 1.5 percent of all posts and an insignificant share o f the overall salary budget. Both within and across sub sectors, mean salaries vary mainly because of variation in the total number of hours linkedto each post. Monthly salaries inthe primary sector are therefore higher because nearly all primary school teachers have full time jobs. This also accounts for the difference in inequality within sectors. Inequality is much higher at the secondary level ("Media") because of most teachers do not have full timejobs and there i s a very high variation inthe total number of hours worked. The standarddeviation is actually larger thanthe mean. 5.47 Table 5.13 shows the mean Table 5.13 Public SchoolTeacher ProfessionalTraining (%) wage distribution by teacher Maestro EmpiricoTitle All Levels Primary Secondary qualification. The ministry of Maestro 0.32 0.32 0.51 45.35 57.39 8.16 education recognizes a total of 5 Profesor Empirico 2.73 0.25 14.4 formal schooling categories for Profesor Titulado 24.42 11.26 61.9 teachers. The Maestro Titulado is SinTipo de Profesional 8.11 7.46 13.94 a more common fixture at the Tdcnico Universitario 19.06 23.31 1.09 70. INPREMA data that also includes compensation and working conditions in the private sector not analyzed here shows that the work life o f these teachers i s even more complex. About 15 percent o f public secondary school teachers also teach inthe private sector. 72 primary level, where they account for nearly 60 percent of the public school teaching staff. At the secondary level, the Profesor Titulado makes up a similar percentageof the total. 5.48 Overall, these data are evidence of success in efforts to improve the quality of education inHonduras since the mid 1980's. The ``Ticnico Universitario" has auniversity degree andnow accounts for nearly one-fifth of the total teaching staff. The "Empiricos" -working teachers without the requisite skills -were once a common feature of the Honduran primary school system. They now make up less than ?4of one percent of all teachers. The "Profesor Empirico" accounts for 14.4 percent of secondary school teachers, but they are probably substantively different. At the primary levelthe designation of "empirico" usually a signalsteacher with fewer than 9 years of formal schooling. At the secondary level it does not generally mean the same thing. Here, where skill differentiation is an important aspect of academic staff composition, it may signal teachers that have formal schooling -not in education - but in some other needed field like music, chemistry, or computer programming. 5.49 Additional compensation is also awarded for performing ancillary and administrative services. Table 5.12 shows the distribution of teachers across sub-sectors and mean remuneration by job. The primary and secondary sub-sectors jointly account for more than 87 percent o f all public sector teachers. It is clear that at the secondary level ("Media") and inadult education and distance learning, jobs are not designedto be full time. At the primary, basic and pre-basic levels mean salaries are substantially higher than the minimum. But average salaries for distance learning and adult educationjobs are well belowthe 5,216.64 base salary. 5.50 The salary for a full-time teacher with Table 5.14 FullTimeTeachersEarning no administrative responsibilities should Less than 40 Lempiras/Hour therefore range from the minimum of 5,216.64 Qualifications Mean Hourly Wage Numbers Lempiras to more than four times that - Maestro Empirico 21.80 174 Maestro Titulado 36.27 5,313 accounting for all bonuses. It is therefore ProfesorEmpirico 32.99 448 puzzling that MINED August 2006 data in fact ProfesorTitulado 26.86 109 show earnings well below this range. Monthly Sin Tipo de Profesional 33.44 4,580 earnings for full-time jobs (36 or more hours Tdcnico Universitario 33.79 55 per week) range from 3,022 to 20,938 Total 34.58 10,679 Lempiras. In fact, Table 5.14 shows that out of the 66,550 MINED teaching jobs, 15 percent (10,382) are assignedto 9,994 persons that earn less than 40 Lempiras per hour. These are not ancillary positions. They are designated as teaching jobs that requireat least 36 hours per week. They are mostly inprimary education, but there is a substantial number inthe secondary sector as well. Furthermore, these are not "second jobs". Nearly all are the only job held by the teacher. About 45 percent of the teachers involved are classified as having "no professional category", more than half are formally trained as teachers ("Maestro Titulado") and should thus be receiving substantially more thanthe basepay. TheLabor Market as a Salary Benchmark 5.51 A paper by Liang (1999)'' compares teacher pay to labor market alternatives in 12 Latin American countries. Using household surveys from the mid-90's she finds that total salaries are 71. "Teacher Pay in 12 LatinAmerican Countries", LCSHD Working Paper Series No. 49, World Bank. 73 lower for teachers than for people with similar education and experience who work inthe private sector. However once hours worked and summer vacations are accounted for it turns out that teachers' hourly remuneration is significantly higher. In fact Honduras stood out because teachers earned a 48 percent premium above what they could expect in the private sector. This was the highest premium inLatin America. 5.52 The more recent MECOVI 2004 survey includes 326 primary and secondary teachers. More recent estimates o f log-wage equations based on this data show a very highrate o f returnto education, the usual u-shape for experience, 22 percent lower wages for women, but no significant wage difference for teachers. Note that while this adjusts for hours worked per week, it does not adjust for the fact that teachers receive 2 to 3 months off inthe summer. Ifthis is accounted for there i s actually a significant pay premium (16 to 25 percent) for teachers. In summary, labor market analysis suggests that Honduran teachers earn an amount comparable to what they could earn inthe private sector, but they have more vacation time. On an annualized hourly basis they are paid significantly more. So inthe market sense, teachers are at least as well off as they would be ifthey quit teaching and went to work inthe private sector. Infact they are better remunerated because they receive about the same yearly pay but work 2 or 3 months less each year. 5.53 On the other hand, it was also seen Table 5.15 OLS Log-WageRegressionEstimates that base pay and bonuses are in some cases (t-statistic) well below the amounts that are stipulated by Years of Formal Schooling 0.15 law. It is this sense o f a mis-match between (63.3) Labor Market Experience 0.061 legal agreements and actual pay that forms (26.57) the basis o f a widespread perception among ExperienceSquared -0.00084 teachers that they are "underpaid". It i s not in (-17-25) the economic sense that compensation is Dummy Variable Indicating Female -0.22 below the market-determinedopportunity cost (-10.81) o ftheir time. Dummy Variable Indicating this i s a Teacher 0.0080 (.017) 5.54 As the events o f August 2006 showed Constant 6.03 all too well, Honduran teacher unions are (194.41) politically powerful. They are capable o f R-square 0.42 negotiating very lucrative pay packages - Observations 6,719 packages so lucrative in fact, that past Source: John Edwards, using data on hourly wages fiom MECOVI2004 governments have found it difficult to honor them. Public default on the agreements exacerbates the sense o f entitlement and mistrust and sets the stage for the nextwave o f strikes, demonstrations, and demands for reparations. D. ThePASCEAgreement and itsEstimatedFiscalImpact 5.55 The PASCE72agreement o f August 12, 2006 implies a further, very large increase in teacher compensation. The official agreement touts a total budgetary impact o f 7.2 billion 72. The acronym stands for "Propuesta Salarialy ProgramaDe Ajuste Social Y Calidad Educativa ",or "Proposal for Salary Reform, SocialAdjustment, and Education Quality". 74 Lempiras-equivalent to USD $381.8 million at the September 2006 exchange rate. MOE payroll data record a total salary bill ofjust over 488 million Lempiras for the month of August 2006. Assuming 14 salaries are paidper year,73the annual wage bill is inthe order of 6.8 billion Lempiras. Innominalterms the PASCE pay raise i s therefore roughly equivalent to the country's entire educational salary budget for one year. Table 5.16 Statutoryand Actual Seniority Bonus(668.04) Actual Bonus Statute 5.56 Program details are shown in Years of Service Lempiras YOof Base Salary YOof Base Salary lowest Table 5.17. salary a teacher can earn and The base salary -the 5 664.04 12.7% 15 10 1328.08 25.5% 30 earned by 6,700 teachers inAugust of 15 1992.12 38.2% 45 200674-is scheduledto increaseby 24 18 2656.16 50.9% 60 percent, 19.3 percent, and 16.2 21 3320.20 63.7% 75 percent in 2007, 2008, and 2009, 24 3984.24 76.4% 90 respectively. The agreement does not 27 4648.28 89.2% 105 30 5312.32 101.9% 120 mention what will happen to those roughly 10,000 teachers who are working full time and earning less thanthe minimum. Another noteworthy fact, i s that the 2006 bonuses are all below their statutory level. This is one of the reasons for recent unrest. The main statutory bonuses for academic qualifications, university technicians, and seniority each rise by 6.1 percent in all 3 adjustment years. Beginning in 2010, the PASCE agreement mandates a returnto piggybacking base salary increases on national minimumwage adjustments, as laid out inArticle49 ofthe "Estatuto del DocenteHondureAo". 5.57 Inaddition, a compensatory supplement of 3,124.94 Lempiras would be paid out every trimester for 16 consecutive trimesters. These payments totaling 50,000 Lempiras (12,499.74 yearly) will be made to all 48,141 teachers on record at the time the agreement was signed. On the one hand, the agreementmakes it clear that this is a one-time payment that it is not to become part of teachers' regular pay. By fixing the total compensation to be paid and the number of teachers eligible it is also implied-though not explicitly stated-that this payment is not intended for new teachers. That is because the rationale for this compensatory supplement is that it i s reparation for breaches of previous salary agreements. On the other hand, the agreement states that "the continuity" o fthis supplement shall be evaluatedin2010. 5.58 An important aspect of the PASCE plan is that all increases are spread over time insuch a way that inflation and real GDP growth will somewhat reduce its real budgetary impact. The compensatory PASCE supplement is spread out over 4 years75and the rest of the program over three. With the exception o f the base salary all of the increases are inequal amounts andthey are fixed in terms of current (nominal) Lempiras. The bottom panel of Table 5.18 illustrates the 73. Local custom dictates that the yearly compensation be divided into 14 rather than 12 payments. Inaddition to regular 12 monthly payments an "aguinaldo" or 13thand "decimocuarto" or 14thpayment are made inthe month o f December. 74. The base salary is earned by teachers without the academic qualifications, with no universitydegree, and with fewer than 5 years o f teaching experience. In August o f 2006 there were 6700 teachers earning the base salary. Among these 4,547 were at the primary level and 565 were at the secondary level. 75. There appears to be an inconsistency, or error in the table o f the PASCE accord signed (Appendix 1). The accord notes that the total compensatory package will amount to 2,406.4 million Lempiras, paid to 48,141 teachers. This amounts to 50,000 Lempiras per teacher. But, as in Table 5.17, the PASCE agreement document shows 12,499.74 Lempiras paid out over just 3 years, which would leave the total (nominal) payment one year shy. The text of the agreement makes it clear that there are to be 4 payments. 75 likely real impact under the assumptionthat inflation and GDP growth remain both remain near 2000-2005 averages of about 10percent Adjustment, and Education 3 0 - h k Quality 30-J~n 30-Sep 30-NOV 5.59 Continued moderate inflation will bring about a US$70 million erosion of PASCE's real fiscal impact. The last three rows of Table 5.18 show the current, constant, and U.S.dollar cost of PASCE for the first three years. The 2006 present value of the PASCE agreement can be found by summing the bottom two rows of Table 5.18 and then adding the present value of the fourth installment of the compensatory supplement (about 411million Lempiras, or 21.8 million dollars). The present value o f the entire PASCE agreement amounts to 5.89 billion Lempiras, or about US$311 milli~n.'~ Notwithstanding this implicit inflationary financing o f the government's new commitment, the burdenof PASCE remains substantial. It will still increase national education salary expenditures for each of the adjustment years by somewhere between one-quarter and one-third. 5.60 Table 5.18 also shows that under these inflation (7c)and GDP growth assumptions only the base salary will grow inreal terms. All three bonuseswill actually fall at arate o f 3.9 percent per annum for 3 years, for a cumulative decline of 11.2 percent. Since at the end of the 3-year adjustmentperiodthe base salary will be 32.3 percent higher, PASCE signals amajor move towards equalizing teaching salaries. It also means there will be concomitant erosion of the incentives that the bonuses were originally designedto provide: an inducement for higher formal teacher education and for the retention of experiencedteachers. Furthermore, itwill reduce savings of any early retirementbuyouts. 76. The flows are discounted at the 10 percent rate of expected inflation. Ifthe rate of real GDP growth is added to the discountfactor to represent average opportunity cost, the present value would fall to US$274.5. 76 ConstantSept. 2006 USD% (Millions) $ 106.0 $ 96.4 $ 87.6 5.61 These points highlight another problemthat will arise whenthe 3-year adjustment period ends in2010. It is not entirely clear how things should proceed thereafter. The agreement only specifies that that the need for continuing the compensatory supplement will be re-evaluated at that time and that the base salary i s to be adjusted at the same rate as the national minimum wage, as established in Article 49 of the Estatuto de D ~ c e n t e s . ~This signals a return to ~ "normality". Yet there is a problem. 5.62 Other laws ofthe Estatuto linkbonus payments to the base salary. Specifically, e Article 50 states that the bonus for having a degree inteaching secondary education or a university degree7*shall be 69 percent ofthe base salary e Article 51 establishes bonuses for job title, seniority, academic degrees, professional merit, and hardship location. These are all explicitly linked as fixed percentages of the base salary. e Article 52 links awards seniority bonuses ranging from 15 percent to 120 percent of base Pay 3 e Article 1 ofDecreto number 174-2000 establishes that the bonus for working inthe states of Graciasa Dios and Islas de la Bahiawill be 100percent ofbase pay. 77. All bonuses that are not explicitly mentioned in PASCE will be increasedat the same rate as the minimum wage throughoutthe 2007-2009adjustmentperiod. 78. Originally this was only for undergraduate university degrees in pedagogy, but subsequent amendments qualify any university degree alongwith a certificate inpedagogyfiom a qualified university. 77 5.63 Then again, simple calculations based on Table 5.18 show that ifthese ratios are close to holding in 2006 they will be far from the legally specified ratios by the end of the PASCE agreement, precisely because the bonuses are not scheduledto grow at the same rate as base pay. For instance The Academic Qualifications bonus i s currently about 59 percent o f base pay (in 2006), but will fall to less than 41 percent by 2009. PASCE does not stipulate what will happen to these bonuses after the adjustment period. It states that from 2010 on bonuses will be adjusted as stipulated inArticle 49. But Article 49 only applies to the base salary and does not mention bonuses. The PASCE agreement does not explicitly nullify Article 50, 51,52, or any other law linking bonuses as fixedpercentages of the base salary. 5.64 The percentages stipulated by these laws may be too high. The PASCE agreement certainly seems to indicate that the government andthe teachers' unions agree on this point. But ifthe Estatuto del Docente remains intact teacherswill go into the next decade feeling that they have not beenpaid what the law stipulates they are due. And they will be right. This sense of unsatisfied entitlementwill set the stage for another confrontation like the one that took place in August of 2006, a confrontation that was costly for teachers, students, the government, and the general public. Paragraph 17 of the PASCE agreement contemplates sending to Congress a bill that will turn these agreements into law. It would be helpful for this bill to explicitly uncouple all of the bonuses from the base salary. PASCE 's Impact on Future Budgets 5.65 When the supposedly temporary compensatory supplementi s subtractedfrom the PASCE agreement it contemplates an impact of 4.8 billion Lempiras on the teacher wage bill spread evenly over the coming three years. This will eventually amount to a nominal increase o f 70.7 percent over the current wage bill. However this implicitly assumes that the number of teachers will remain constant. If increased student demand is accommodated under existing quality parameters, it will mean an additional 14.1 percent increase at the new (PASCE) rates. This brings the total wage bill estimate in 2010 to 13.2 billion Lempiras, up 94.7 percent from the annualizedbillof 6.8 billion in2006. Ifnothingelse changes the wage bill will nearly double in the next three years and it will increaseby 141percent by the year 2017. Can Early Retirement Reduce PASCE 's Fiscal Impact? 5.66 When there are significant seniority bonuses -as is the case in Honduras - early retirement programs can sometimes be a source of fiscal savings. However, the PASCE agreement, coupled with the way that the teacher pension retirement scheme is structured, will make it difficult to induce teachers to retire early. As shown in Annex H, these will instead induce a three- to seven-year delay in retirement as older teachers wait for the PASCE raise to work its way into their retirement benefits. The teacher pension system, INPREMA, also functions as a bank, and many teachers may continue to work beyond their retirement age because of cash flow problems arising from debts owed to INPREMA. In those cases, debt restructuring is one possibility that holds promise as an inexpensive early retirement policy handle. 78 5.67 Nevertheless, even if a way could be found to induce early retirement, this option does not offer any fiscal relief for the public sector as a whole, and it would endanger the financial solvency of the INPREMA retirement system. Annex H shows that even though the early retirement of senior teachers andtheir replacementby younger ones would reducethe Education Ministry's total payroll costs, it would increase the deficit of the teacher pension fund by a similar amount and, thus, would not yield any fiscal savings for the overall public sector. Also, such a measure i s likely to result inan early collapse of INPREMA, which already appears to be actuarially insolvent. E. Will PASCEBuyIncreasedEducationQuality? 5.68 The Honduras Poverty Reduction Strategy outlined a number of goals for the Honduran educational system, targeting increased coverage at the primary and secondary levels as well as higher transition rates; Table 5.19. These goals seem reasonable within the context of what has been achieved in the last two decades, but it is not yet clear exactly how they will be attained programmatically or financially. Table 5.19 PovertyReductionStrategyGoals inEducation Base Line Estimates Goal Indicator(base line year) Obs Est 2006 2010 2015 2005 2005 Double net preschoolcoverage Preschool education coverage rate (1999) 34.3 36.5 38.6 42.0 60.0 through 5 years 95% net coverage for the first 2 Net coverage rate (excl. Over-age) for lSt 89.4 and 91.7 92.5 93.2 95.0 cycles o fprimary education 2"dcycles (1999) 70% net coverage for the 31dcycle Coverage rate (excl. over-age) for the 3Tdcycle 39.0 37.3 39.7 51.1 70 (middle school) ofprimary o f primary education (middle school) (1999) education 50% o fthe emergingpopulation Net coverage (excl. over-age) o f Diversified 22.7 22.3 24.2 33.4 50.0 completes secondary education Secondary cycle (10-12 grade) (1999) Source: PRPS Update Report 2006. Calculations based on EPHPM- INELabor Surveys. For more information please visit wwwsiemhn. 5.69 The Education for AllEast Track Initiative (EFABTI) that was designed subsequently had more ambitious goals and provided more detail on how to meet those goals; Table 5.20. For instance, while the PRS calls for 60 percent preschool coverage by 2015, the EFABTI is aiming for 99 percent coverage by the same year. Also, while the PRS calls for 95 percent net primary coverage by 2015, EFNFTI targets require that 100 percent of 12-year olds graduate, requiring not only perfect internal efficiency, but also that the age of first enrollment be reduced from age 7 to age 6. This table suggests efficiency gains will come from higher test scores and lower repetition and dropout rates. One way to stretchthe resourcesthat are available for coverage and quantity improvements i s to increase the quality or intensity of the learning experience. If studentscan be madeto learnmore per year ofteacher salary, unitcostswill necessarilyfall. 5.70 Unfortunately, it does not appear that the PASCE is a good instrument for generating improvements in teacher quality. The reason is that the salary raises agreed under the PASCE 79 agreement were not motivated as a payment in exchange for better performance. Rather, the agreement was premised on the notion that previous salary promises made to teachers had not been fulfilled, and so needed to be raised. As described earlier, the agreement consists o f two parts for fulfilling the promise. One part is the 3-year schedule o f adjustments that gradually builds up future salaries to agreed levels. The second part targets past salaries by paying out 50,000 Lempiras to the 48,141 teachers who were on staff when the agreement was signed in August 2006. This is in compensation for having been paid less than their presumed legal entitlement for a number o f years before the PASCE accord. Table 5.20 EFA/l?TIGoals Goal 2003 2005 2010 2015 Coverage Pre-school 35.0 43.0 76.0 99.0 Enrollment of 7-year old children on 51.0 65.0 93.O 98.0 1" grade (goal 3) Enrollment of children younger than 7 65.0 years old on 1" grade (goal 4) Quality Academia PerformanceinMathand 43.0* 50.0 61.0** 70.0 Spanish(31dand 6" grade) (goal 5) Efficiency Graduatedfrom 6" grade with multiple 68.5 75.0 95.0 100.0 ages (graduatedfrom 6" grade/l2 years-old population) (goal 1) Graduatedfrom 6" grade with 12 3leg*** 32.0 65.0** 85.0 years (goal 2) Repetition 1" grade (goal 7) 18.5 13.0 3.9 0.8 2"dgrade (goal 8) 10.8 1.5 2.3 0.5 31dgrade (goal 9) 8.1 5.6 1.7 0.3 4" grade (goal 10) 5.2 3.6 1.1 0.2 5" grade (goal11) 3.5 2.5 0.7 0.1 6" grade (goal 12) 1.o 0.7 0.2 0.0 Dropout lSt-31dgrade (goal 6) 4.0 2.2 0.6 0.0 4'-6' grade (goal 6) 2.0 1.1 0.3 0.0 Notes: *2003 data, ** 2011data, ***2001 data. Source:MoE Informe del ProgresoEducativo de Honduras2005. 5.71 Nevertheless, the government has suggested that the newest round o f wage agreements be linked to quality and performance. Three observations argue in favor o f linking increases in teacher salaries to better performance: one i s that teachers were already earning at least as much as they could earn elsewhere in the economy before PASCE and that - once the vacations are accounted for - they earned substantially more. Another i s that performance improvements are long overdue, as past increases inteacher pay were almost entirely de-linked from improvements inteacher performance. A third issue is that the limitedfiscal resources will make itdifficult for the government to meet the higher salary commitments, so the only hope o f being able to raise 80 additional internal or external financing would be on the premise that the government would be obtaining something additional inreturn." 5.72 Since the signing of PASCE, it has been proposed that the payment of compensatory bonuses be conditioned on quality andor performance outcomes. Attempts to link it to indirect indicators of educational quality considered inthe EFNFTI such as repetition and dropout rates were resoundingly rejected by teachers unions. The effort currently under way has "tripartite" commissions -consisting of one parent, one ministry official, and one teachers' union official - scheduled to make school visits to verify teacher identity, absenteeism, and lesson-plan preparation. The compensatory bonus would then be withheld if the teacher is absent or the person found teaching is not the one officially assignedto the job. It is also proposed that ifthe teacher has not submitteda lessonplan, thenpayment ofthe bonus should be delayed untila plan i s turned in. 5.73 Linking the 4-year 50,000 Lempiras in PASCE's compensatory bundle to quality improvements is complicated by the following aspects: i)Entitlement:Teachersseethissalarybonusasanentitlement. Intheireyesitis money owed to them for past work. They will therefore be certain to challenge any newpreconditions, suchas increasedeffort. ii)Coverage: All teachers - including the new ones (but excluding PROHECO teachers) - will see their salaries grow substantially under the new agreement. However the compensatory bundle does not apply to everyone. It only covers the 48,141 regular teachers who were on staff when the agreement was signed inAugust 2006. It excludes any teacher hired after that date and it excludes all PROHECO teachers. iii)Duration: Thecompensatorybundleisaone-timepaymentspreadoverfouryears. Itwas not designedto bepart ofan already very generouswage bill. 5.74 Ifthe compensatorybonus is linked to quality improvements, its impact will dependon whether the agreement is implementedas written or not. If it is implemented as written, new teachers and PROHECO teachers will be left out of the quality enhancementprogram. Also, any quality enhancement brought about by this incentive will end infour years. The alternative is to extend the 12,500 Lempira payment to all of the teachers and thento extend it beyond the agreed 4 years. This will amount to a further, permanent wage increase -on top o f the very generous one already agreed to. Teachers would then permanently have a 2-part salary, effectively recognizing that once appointed to a teaching post, they are entitled to the base salary whether they show up at work or not and whether they preparetheir classes or not. Then in addition to this guaranteedbase salary, they could earn another 12,500 Lempiras per year ifthey prepare a lessonplanand show up at work. ~~ 79. External donors, inparticular, are generally limitedintheir capacity to fmance recurrent spending that does not entail any permanent improvements inperformance. 81 F. What To Do Instead? 5.75 A decade ago, the primary school dropout rate stood atjust under 10percent. This seems much improved now, with 2004 estimates showing a primary school dropout rate o f less than 4 percent among children betweenthe ages of 7 and 13. Repetition rates in 1993 were found to be extremely high at the primary school level. Half of all repeaters age 7 to 13 were infirst grade. Data from the 2004 MECOVI survey reveal that first grade repetition rates are now at about 29 percent. Though far from satisfactory, it i s an improvement. It is still the case that 12percent of all primary school students are repeating the year they are in. Earlier studies of Honduran primary education quality have suggested that repetition was often arbitrary. When learning objectives and curriculum are poorly defined, passing a grade more often than not depends more on each teacher's individual assessment of student "readiness" than on a set of objective, nationally defined criteria. This obviously leads to wastage, as students end up taking the same classes two or three times andthen drop out. Figure5.6 Primary School RepetitionRates, by Grade in2004 1 2 3 4 5 Grade 5.76 Another issue i s the total hours of classroom instruction. In 1993 the average primary school student reported receiving about 5 hours of classroom instruction." This situation has not improved, and may even have gotten worse: in2004, primary school students reported receiving an average of only 4.6 hours per day! 5.77 At the core of the needed changes is the effort to convert the country investment in education into learning time and education quality. Frequent school closures (strikes, weather, teacher training, extra holidays), teacher absenteeism and tardiness, reduces the class time devotedto any learning task. According to the school census, 40 percent o fteachers inHonduras missed one month or more o f school in 2000 (World Bank, 2005). Also problematic is that the 80. Edwards, Fuller, andParandekar (1997),p. 8. 82 gravity o f the absenteeism problem may not be fully acknowledged, as few principals in Honduras (a mere 7 percent) consider high absenteeism to be a problem in a school survey conducted by UMCE (2003). Table 5.21 Measuresof Teachers Efforts Teachers' Absences Teachers' yearly absences (days) 16 Teachers with no month away (%) 60 Teachers with 1monthaway (%) 26 Teachers with 2 or more monthsaway (%) 14 Average hours Der week 30 Average declaredhours per week according to householdsurvey Primary and Secondary 33 Primary 31 Secondary 38 Source: World Bank (2005) 5.78 Usinga 42 normal weeks per year and 21 hours of weekly teaching hours; an estimated losses of 4 weeks due to teachers' absences and 3 weeks of school closings due to non-teacher absence reasons, World Bank (2005) obtained an estimated 735 hours o f effective number of class hours per year which corresponds to 83 percent of the 882 notional class hour per year (or 61 percent o fthe 1,200 notionalhours ofEurope). 5.79 Another important problem identifieda decade ago that persists inthe systemto this day i s the lack of national testing. At present, teachers have little to go on when it comes to national standardsfor passing from one gradeto the next. Also, inthe absence ofregular nationaltesting, it is almost impossible to track any butthe grossest quality changes inthe educational system. 5.80 There is no "culture" of testing in Honduras. The UMCE carried out some testing on student learning, and data is available for 1997, 2002, and 2004 for third and sixth graders. However these tests are based on national sample of schools and not on a census of schools and cannot be compared from year to year. Now there is a new curriculum, the so-called "Curriculum Nacional Basico" introduced by the government in 2003, but the old exams do not test for itg1 through the MIDEH project, has helped to produce a set of standards for USAID, teachers based on the new curriculum. This is a simple declaration designed for grades 1 through 6 and aimed at helping to define more clearly a set of knowledge standards (Le., a declaration of what students should know and know how to do at each grade). It includes a month-by-month program of subject matter to be covered for each grade. The same program is now developing sets o f standardized tests which are to be administered to all Honduran students. These are not intended as high-stakes tests, but rather for having the results sent back to the teachers as diagnostic tools. An end-of -year test which can be usedto evaluate student learning i s expected to be developed during 2007 by MIDEH. Two issues will have to be addressed by the government: first, how to make this evaluation truly external to the Ministry of Education, andsecondly, howto assurethe sustainability ofthe evaluation efforts. 81. UMCEhas not conductedtests since 2004. 83 Secondary School Coverage, Cost and Effectiveness 5.81 Secondary education problems clearly abound as well. Since primary enrollment legally begins at age 7, middle school ("ciclo ~omlin'~) enrollment should begin at age 13 and high school ("ciclo diversifzcado") at age 16. Since each level lasts for 3 years, roughly one third (33 percent) of students at each level should be one o f three different ages. Infact, in2004 only 2.1 percent of middle school studentsand about 1.6 percent of high school students start their levels at the appropriate age; Table 5.22. Part of the problem is a carryover of primary school issues, including late initial enrollment and grade repetition. There are problems within these levels as well though. For instance, the overall middle school repetition rate stood at nearly 8 percent in 2004 andthat for highschool at about 4 percent. Table 5.22 SecondaryEnrollmentby Age and Level; 2004 Ciclo Comrin CicloDiversijicado Age YO Cum. YO YO Cum. YO 12 0.5 0.5 13 2.1 2.6 14 6.4 9.0 15 5.7 14.7 0.3 0.3 16 17.2 31.8 1.6 1.9 17 10.2 42.0 3.6 5.5 18 16.8 58.8 8.9 14.4 19 13.1 71.8 16.4 30.8 20 5.6 77.4 14.6 45.4 21 4.5 81.9 9.8 55.2 22 1.9 83.7 8.8 64.0 23 2.5 86.2 6.2 70.1 24 3.2 89.4 3.3 73.5 25 1.5 90.9 3.8 77.3 26 1.o 91.8 4.4 81.7 27 1.6 93.4 4.4 86.1 28 0.7 94.1 2.7 88.8 29 0.6 94.8 0.5 89.3 30 & Over 5.3 100.0 3.4 100.0 5.82 One potentially big issue i s the curricular design of the secondary system. The Honduran secondary system lasts 6 years. These are divided into two segments of equal length,the Ciclo Comlin and Ciclo Diverszjicado. There are a very large number of tracks or "carreras" that can be chosen in the Ciclo Diversifzcado. The MECOVI data show that in 2004 there were about 200,000 students enrolled at this level and that they were in40 different tracks. This implies an average of about 5,000 students per track. Actually though, some tracks are far more popular than others. The four most subscribed tracks account for 81 percent of enrollment -about 166,000 students -leaving the other 36 tracks to divide up the remaining 34,000 students. In fact, the 20 least popular tracks account for less than 3 percent of total enrollment. 5.83 This system originates from a viewpoint that merges vocational training with academic education. It assumes high school is a terminal degree and that having a "professional" title like commerce, hotel and tourism, or mechanics enhances job prospects. Problems with this approach can originate in fragmentation of the basic curriculum and the early-age tracking it 84 entails. In systems like these it is never the case that all 40 tracks are offered by every school. Instead, schools tend to specialize, offering 3 or 4 tracks, and often offering only one. Some tracks - like mechanics or wood-working - can be very capital intensive. Others -such as commerce andcommercial secretary-may requirevery little capital investment. 5.84 The national curriculum may well requirethat math, language, science, and social studies be taught in all schools, regardless of the track chosen. Yet in fact it i s often the case that schisms develop betweengood and bad schools, private and public sectors. Some schools will specialize inthe "academic" track -seen as necessary for entrance to a university. These will be the "good schools" that emphasize math, language, and the sciences. They will draw their teachers from the academic community -often seeking teachers with university ties. A school specializing in "hotels and tourism" or metal-working will tend to skimp on the basic subjects, preferring instead to emphasize the "practical skills". Proficiency in the basic subjects will be seen as more of a bureaucratic obstacle to be overcome than as a goal in its own right. In fact unless there i s some sort of national exit examination, there may be a metamorphosis of the curriculum itself.82 Maintaining a secondary system with 40 tracks can be expensive and an administrative nightmare. It is likely to erode quality of the core curriculum. G. MainConclusionsandRecommendations 5.85 Population growth, together with the doubling in school attainment has meant a very rapid rise inthe country's overall average education attainment. Mean schooling attainment has been rising steadily for at least half a century. It went from 3 years for the cohort born around 1939to 6.7 years for the one born around 1981. Schooling growth was concentratedat the lower levels. This has lead to lower educational inequality and to a concomitantly lower income inequality, Age-appropriate (net) school coverage has increased dramatically, meaning that more and more children are entering school at the appropriate age. Gross enrollment rates are very high. Infact the ENCOVI data show that 15.2 percent of the population between the ages of 21 and 30 are enrolled andthey are mostly enrolled inprimaryand secondary grades. 5.86 Several trends conspire to fundamentally alter the current educational demand structure. Honduran income per capita is growing and young parentshave at least twice as much education as did their own parents. Increased demand for education will manifest itself in several ways that have distinct impacts on the structure of the educational system. Most importantly, these include the age of initial enrollment, the quality of education, and the duration in school. The major forces at play in reconfiguring education demand inHonduras will therefore be (i) higher initial (net, or age-appropriate) enrollment rates at earlier ages, (ii)greater educational attainment, and (iii) eliminationof the over-age (gross) enrollment bubble. 5.87 Forecast enrollment growth i s inverse U-shaped, highest in the middle, significantly lower ingrades 1-3 and lowest of all for grades 11and 12. Thus, while significant growth can be expectedat all levels of schooling, the biggest change will be due to higher transition rates from 82. In the 199O's, before extensive reforms were introduced, one o f neighboring El Salvador's largest and best secondary schools prided itself in offering over a dozen distinct tracks. When asked how he coordinated the basic curriculum requirements across the traces, the director answered that "All o f our students take mathematics. It's just that the beauticians take math for beauticians, the carpenters take math for carpenters, and the academic track gets what they will needfor the university." 85 primaryto secondary schooling. The weighted average forecast by sector is 40.4 percent growth at the primary and 43.8 percent growth at the secondary level for the decade after 2006. Ifthe public sector maintains its share of the projected expansion in student enrollment, without changes inwages or in average studentlteacher ratios, projected increase inteacher demandwill have an enormous fiscal impact. Expenditures on teaching as a share of GDP will have to increase from about 3.4 percent of GDP in 2006 to 4.3 percent in 2017 to keep pace with increaseddemandunder the existing structure. This would have been true even without the very large increaseinteacher salaries agreedto inAugust of 2006. 5.88 At the primary level it is unlikely that increases inclass size can contribute significantly to absorbing education demand without seriously compromising education quality. At the secondary level, there may be some room for savings by consolidating schools and simplifying the curriculum, but this requires further study. 5.89 The study of teacher payroll data indicates significant differences between the primary and secondary sub-sectors. Over 96 percent of primaryschool teachers hold a singlejob and this job employs them full time. At the secondary level only 22 percent of teachers are single jobholders. This is largely because the vast majority of secondary schooljobs do not provide full time employment. At this point it is not clear whether this is an insignificant accounting technicality or an aspect of secondary schooljob design that might be modified to reduce costs, or improve quality, or both. 5.90 Private schools may be able to provide a partial alternative to public sector expansion at the secondary level. This is especially likely in grades 10-12 where the need for teacher specialization and expensive physical infrastructure i s greatest. This would require a study of existing private schools to determine quality, excess capacity, and cost. If found feasible, implementation could be accomplished with avoucher system. 5.91 A labor market analysis indicates that Honduran teachers are better remunerated than similarly qualified counterparts outside the education sector. The legal structure governing public teacher pay includes significant rewards for education, seniority and hardship posting. However actual rewards have lagged behind legally stipulated ones. Claims that teachers are underpaid culminating inrecent teacher union unrest stem from this discrepancy betweenlegally mandated and actual remuneration. "Underpayment" does not imply teachers are paid less than the opportunity cost oftheir time outsidethe teachingprofession. 5.92 The PASCE agreement of August 2006 seeks to redress these underpayment grievances. It commits the government to a 7.2 billion Lempira increase in teacher compensation over the next4 years. This very large increaseis roughly ofthe same order of magnitude as the country's entire education salary budget for one year. Nevertheless, PASCE is unclear inseveral important areas. One of these i s the role of PASCE in determining new teacher salaries. The agreement raises base salary significantly more than the bonuses for seniority, education, and hardship postings. This will have the effect of equalizing teacher pay and also of reducing the impact of these incentives. Perhaps most worrisome i s that under the current Estatuto del Docente law all of the bonuses piggy-back as fixed percentages of base salary. PASCE stipulates that in 2010 there will be a return to this law. But since PASCE increases base salary far more rapidly, the bonuses will be even further from their legally stipulated percentages in 2010, setting the stage for further teacher union unrest basedon claims ofunderpayment. 86 5.93 Even without any further increases in 2010, when PASCE salary increases are coupled with projected enrollment demand the educational wage bill stands to increase by 141 percent over the next 10years. 5.94 When there are significant seniority bonuses -as i s the case in Honduras -early retirementprograms can be a source of savings. However the PASCE agreement coupled with the INPREMA retirement scheme will induce a 3 to 7 year delay inretirementas older teachers wait for the PASCE raise to work its way into their retirementbenefits. INPREMA functions as a bank as well as a pension system. Debtrestructuring i s one possibility that holds promise as an inexpensive early retirement policy handle from the viewpoint of the Education ministry. As argued earlier, however, any early retirementprogram would endanger the financial solvency of the INPREMA retirement system and is unlikely to generate a net fiscal savings for the public sector as awhole. 5.95 Addressing;the Fiscal ImDactof PASCE. Increasededucational demandis coming and it will be large. The rapid wage increases called for under the PASCE agreement mean that meeting that demandwill be expensive. Public school teachers are not underpaid incomparison to other professionals with similar qualifications elsewhere in the labor market. The current political environment nevertheless does not permit much leeway for a more gradual adjustment of teacher wages. Therefore, the recommendations inthis section concentrate on chipping away at the financial problems inthe education sector, rather than seekingto eliminate them inone fell swoop. These recommendations are: 0 Study the secondary sector to clarify why most teachers are working multiple jobs and why most jobs are less than full time. Consider consolidating schools to fully employ teachers and save on fixed costs. Consider secondary curriculum reform to eliminate nonessential subjects or move them into after-hours optional user-fee financed programs. 0 Consider using the private sector to accommodate some of the expected increase in secondary enrollment demand. This would require a background study of private school quality, slack capacity, and cost. Ifthe unit cost is found to be less thanthe marginal cost of public expansion develop a voucher systemto finance targeted low-income enrollment inqualified private schools. 0 Publicize teacher remuneration to clarify inthe national debate that though teachers have not been paid the amount negotiated by the unions, they are well paid in relation to similarlyqualifiedpersonsworking inother sectors ofthe economy. The focus should be on the basis of an hourly wage computed as total annual compensation (including bonuses) divided by total hours worked per year. Length of workday, yearly vacation time,job-security, andhealthandpension benefits canalso behighlighted. 0 Re-examine qualification for the education bonus. It is reputedly paid to people of very differenteducational backgrounds. 0 Insist that the compensatory component of the PASCE agreement not be extended beyond the stipulated 4 years and that since it i s compensatory, it not be paid to new teachers. 0 Ensure that PASCE-enabling legislation explicitly overturns the Estatuto del Docente provisions that fix pay bonuses as set percentagesofthe base salary. Interpret legislation governing all bonuses to mean that they are paid at most once per person, pro-rated per hour up to the full time equivalent, but not beyond. Thus a person employed half time would receive 50 percent of the bonus and a person with 3 jobs with 87 hours totaling 150 percent o f full time would receive 100 percent o f the seniority, education and other bonuses.83 5.96 Raising the Oualitv o f Education. In order to achieve the education goals set in the Poverty Reduction Strategy and Education For All initiative, additional efforts will need to be made. Given the already elevated level o f public expenditures inthe education sector, increases insystem-wide efficiency will be a crucial component of any strategy that seeks to significantly raise coverage at the preschool and secondary levels. There are readily identifiable quality and efficiency problems inthe Honduran educational system, o f which the most prominent include, 0 low teacher attendance and shortened class days, 0 highgrade repetitionrates, the absence o f clear educational standards and testing, the absence o f objective criteria for transition between grades, 0 low secondary education coverage, and an overly complex upper-secondary educational system. 5.97 Enforcing teacher attendance and class preparation are worthwhile goals, but more i s needed. Even here, the PASCE agreement does not provide a good instrument, as it does not cover all teachers, is scheduled to expire in four years, and threatens to inflate teacher salaries further by turning what was intended as a one-time reparation payment into a permanent salary bonus. Teacher attendance and preparation for class should be a pre-requisite for receiving the regular salary, not rewarded as extraordinary behavior. A possible solution here would be to turn control o f salary payment over to parent associations, rather than relying on costly, centrally organized inspectionmissions. 5.98 Information on school closures, teacher absenteeism and tardiness, should be collected and reportedfor each school to improve quality through continuously updated schools' registers. Duly completing the school register should be made compulsory. This can be complemented by a monitoring system o f teachers' absences by the communities themselves. They could record teacher absences and report them to district offices, which should then take some disciplinary action. 5.99 High repetitionrates can be at least partly addressed by developing a clear set o f content standards and national testing. The efforts begun by the MIDEHproject inthis regard should be supported, continued, and expanded to all secondary grades. The goal should be for all teachers to know what their students are expected to know and that they have an objective set o f criteria for determining if they know it. Once these criteria are set and the instruments are in place, a longer term goal would be to set national, local, and even school-specific performance target. A 83Another point that was initially nuclear was the treatment o f the one-time compensation portion o f the PASCE agreement fiom the viewpoint o f the teacher insurance system. It appears that this portion does not form part o f the base salary usedby INPREMA. This is important from a fiscal savings viewpoint and means three things: first, that the teachers would be paid the full Lps.12,500 amount -without the 7 percent INPREMA deduction all four years. Second, it means that this does not enter the base for the government's additional 12percent retirementcontribution. Third, it also means that these Lps.12,500 do not form part of the last 3-years salary used to compute pension benefits. 88 sustainable system of (external) assessment to evaluate student learning should be implemented by the government andthe results should be readily available to parents andthe community. 5.100 Secondary education is, by nature, more expensive thanprimary education, as the greater depth needed in each subject requires specialized teachers and is more intensive in the use of libraries, equipment, and computers. In addition, however, the Honduran upper secondary curriculum is too complex. It needs to be streamlined to concentrate on competencies in mathematics, writing, and the sciences. The same standards inthese subjects should be applied to all students. This will improve the quality of education and reduce administrative and equipment overhead costs. 89 90 CHAPTERV I REVIEW OFHEALTHSECTOR SPENDING 6.1 This chapter examines the progressmade inthe health sector inHonduras since the 2001 PER, with a view to looking into health sector performance interms of sector outcomes, access and equity, efficiency, and quality together with trends in financing of health services. It also discusses certain key institutional issues and provides an overview of the methodology and preparation process o f the health sector expenditure framework using a results-based planning andbudgetingtool with the aim ofpreparinga costed sectoralplanfor 2008. A. Summary of MainFindingson Healthinthe2001PER 6.2 The 2001 PER notedthat healthoutcomeshave greatly improvedover the last 20 years in Honduras. The relative effectiveness of primary health care programs invaccination, control of diarrhea and acute respiratory infections, and more emphasis on maternal and child health and expansion o f public sector health network inthe 1990s have contributed to a reduction in infant mortality and an increase in life expectancy. Service expansion was also characterized by a shift of ambulatory consultations from hospital outpatient clinics to lower cost ambulatory facilities which indicates an improvement inthe referral system and service network rationalization. 6.3 Public spending declined in real terms over the 1990s. Majority of public health expenditures were spent on hospitals (in terms o f program costs) and personnel (in terms of expenditure groups). Health spending across regions was not strongly correlated with incomes, suggesting that resources are not being allocated systematically to wealthier areas. While the expansion of service coverage has beenreaching the poor, important differences remaininterms of effective access betweenpoor andnon-poor households. 6.4 The report made recommendations infive major areas: (1) increasing the effective access of the rural poor to health services (for example, extension of regular working hours, availability of medicines); (2) improving the nutritional status o f young children, especially in poor, rural households; (3) encouraging institutional reforms that promote quality improvements and enhance the efficient use of public resources by promoting the use of private resources for non- primary health care treatments; (4) institutional strengthening efforts in MOH planning and regulation and (5) reforming the Social Security Institute (IHSS) which was confronted by both serious financial and quality of service crises. As will be gleaned in the following sections, progressvaried for each of these recommendations. B. HealthSector Overview Evolution of key sector outcomes 6.5 Since the PER (2001), Honduras has continued to make notable progress inits key health, nutrition, and population indicators. Table 6.1 highlights a number of these achievements including: a significant reduction in under-five mortality (U5M) from 44 from 1996-2001 to 30 in2001-2005; a decline inthe infant mortality rate (IMR) from 34 in 1996-2001 to 23 in2001- 2005, as well as a 7 percentage point drop inneonatal mortality (first 28 days after birth) from 21 to 14 duringthe same period-- after being stagnant from the late 80s to early 2000. Prenatal care 91 coverage increased from 83 percent to 92 percent, while the number o f women who received postnatal care has more than doubled from 34 percent to 73 percent. Chronic malnutrition for children below 5 years of age has also declined from 29 percent to 24.7 percent. These improvements can be attributed largely to the following factors: the continued expansion o f prevention, promotion, and health care programs and services with an emphasis on Maternal and Child Health Programs, as well as the use of alternative models of service provision with increased participation of communities; improvements in access to potable water; and higher educational levels of Honduran women. (ENDESA 2005-06). Table 6.1: Health Outcomesand Outputs,2001 and 2005/06 Indicator* 2001 2005106 Life Expectancyat birth (years)** 70.7 72.1 InfantMortality Rate (per 1000live births) 34 (1996-2001) 23 (2001-2005) Under 5 Mortality rate (per 1000live births) 44 (1996-2001) 30 (2001-2005) Total fertility rate 4.4 3.3 Percent of women using a family planning method 61.8 65.2 Percent of women using amodem contraceptivemethod 50.8 56.4 Percent of women using a traditional contraceptive method 11 8.9 Prenatalcare coverage 83 92 Womenwho receivedat least 4 prenatalcontrols N.A. 81(87 - urbanand 76 - rural) Percent ofwomen who receivedtheir 1'' prenatalvisit before4" 56 69 month ofpregnancy Percent of institutional births 52 67 Percent ofwomen who delivered attendedby a skilled health - 67 (90 inurbanand professional 50 inrural) Percent ofwomen who receivedpost natal care 34 73 Children 12-23 months, filly immunized 78 75 Chronic malnutrition (stunting or low height-for-age) 29 25 Underweight (low weight-for-age) 17 11 Sources:ENDESA 2006*, MOH StatisticalBulletin 2005** 6.6 Some indicators, however, have worsened and fluctuated during this period. For example, while immunization rates for young children for most vaccine types remain above 90 percent each, the percentageof fully immunized children 12-23 years of age has slightly decreasedfrom 78 percent to 75 percent (DHS 2005-06). Also, although malaria prevalence significantly decreased from 370 in 2001 to 206 in2005, malaria prevalence was much lower at 149 in 2003 (WHO, MOH 2006). 6.7 Inaddition, although the total fertility rate has declined from 4.4 in2001 to 3.3 in2005, Honduras i s secondto Guatemalainterms of having the highest fertility rate inCentral America. About 17 percent of women have an w e t needfor family planning (8.5 percent o f women ina relationship would like to limit the size of their families and another 8.5 percent would like to space the birthsof their children). 6.8 AIDS remainsanissue, with Honduras having the highest incidence (2.1 percent in2005) of HIV/AIDS in 6 Central American countries. The reported number of HIV/AIDS cases in the country increased by 37 percent from 2003 to 2005 (UNAIDS 2005), also resulting in a corresponding increase in the number o f HIV/AIDS orphans by 25 percent from 2001 to 2005 (UNICEF 2004, UNAIDS 2005). 92 6.9 Significant gaps persist interms o f access and outcomes between rural and urban areas, regions, and poor and wealthy segments o f the population. For example only 50 percent o f rural women have had their deliveries assisted by a skilled health professional compared with 90 percent inurbanareas; and 32 percent o f children inrural areas are stunted relative to only 13.7 percent in urban areas. Regional differentials exist, for example, San Pedro Sula has a chronic stunting rate o f only 13.5 while Copan and Lempira have stunting rates o f 42 and 49.5, respectively. Interms o f differentials based on income levels, only 33 percent o f mothers inthe lowest income quintile received skilled professional assistance while giving birth compared to 98.5 percent o f women inthe richest income quintile. Moreover 43 percent o f the children inthe poorest households are stunted compared to only 5 percent o f the children in the wealthiest quintile. Progress inMeetingthe Millennium Development Goals andGovernment Targets 6.10 Taking into account, Honduras's recent rate o f progress inkey maternal and child health and nutrition indicators, and assuming that this pace is maintained until2015, then the likelihood o f achieving the MDGs and Government goals o f reducing infant and child mortality rates and malnutrition (stunting) is high. While the previous trend o f declining maternal mortality rates i s also promising, there i s no recently updated information to confirm whether Honduras i s still on track in achieving the target.. This contradicts the results presented in chapter 4Moreover its proxy indicator, i.e. the percentage o f women who deliver attended by a skilled health professional still remains relatively low at 67 percent, mainly because only 50 percent o f women inruralareasobtainthistype ofassistancewhile givingbirth. Figure6.1 Progresstowards meetingthe MillenniumDevelopmentGoals and GovernmentHealth and Nutrition Targets, Honduras; 1990-2005 Child Malnutrition Maternal Morlallty, 100,000 1986-1990 1991-96 2001 20054 2015 93 Figure6.1 (Continued): Progresstowards meetingthe MillenniumDevelopment Goals and Government Healthand Nutrition Targets, Honduras, 1990-2005 I Under 6 mortality11,OOO Infant Mortality11,OW "1 1 45 40 35 30 25 20 n a 5 0 lS?&tsgl 1991.96 m1 2W-6 2015 Note: MOH and ERP end goals were the same for IMR and U5M rates but ERP goals were lower (more ambitious) for malnutrition (18 for ERP vs. 20 for MOH) and MMR (70 for ERP and 73 for MOH) so these were usedfor the analysis. Source: ENDESA 2005/06, ENESF 2001, MOH statistical Bulletin and National Policy Documents, ERP Progress Report, 2007. Health careprovision by type ofprovider 6.11 Health care in Honduras is provided by the MOH, IHSS and the private sector, and nongovernmental organizations (Table 6.2). These providers generally function independently from eachother although bothMOH and IHSS have beenincreasingly contracting out services to certain private providers; IHSS has also beenrecently contracting some services from the MOH. There is also a lack of systematically updatedand consistent informationabout the private sector althoughthis situation is expectedto changeprovided the MOH improves its capacity to perform its stewardshiprole. Table 6.2 NumberofHealthFacilities(by Type of Facility andProvider); 2005 Type MOH IHSS" Private-for Private-notfor profit** profit** Hospitals 28 2 99 8 HealthCenters (urban& rural) 1324 Clinics (CMIs and CLIPERs) 53 8 760 35 Total 1405 10 859 43 Note: IHSS also subcontracts services of 19ambulatory clinics and 11private hospitals * Source:MOHHealth Bulletin2005 and Census ofHealthEstablishments2002** 6.12 MOH documents indicate that the potential health care coverage by provider is: 60 percent and an additional 4 percent for outreach services for the MOH, 18 percent by the IHSS, and 5 percent by the private sector, with about 13 percent of Hondurans with no access to health care. However, while actual utilization rates based on recent household survey data (ENCOVI and ENDESA) confirm that the Ministryof Health (MOH) continues to be the main provider of health services inHonduras, it identifies the private sector as the second most utilizedprovider o f health care, followed by IHSS (the section on household utilization of health services will discussthis inmore detail). 94 6.13 From 1990 to 2005, M O H facility expansion has continued with a focus on primary health care facilities. Table 6.3 shows that the numberof MOH tertiary facilities increasedby 47 percent from 1990 to 1999 but has remained constant since 1999 while the numbers of health posts and maternal and infant clinics have increased. Rural health posts, inparticular, increased by 22 percent from 1999 to 2005 andby 105 percent from 1990 to 2005. By 2005, the Ministry had 28 hospitals and 1,377 ambulatory facilities (1,058 rural healthposts, 266 urbanhealthposts, 50 maternal and infantclinics, and 3 peripheral clinics). Table 6.3 TrendsinMOHFacilityExpansion; 1990- 2005 Facility 1990 1999 2005 YOchange since 1999 YOchange since 1990 NationalHospitals 6 6 6 0 0 Regional Hospitals 6 6 6 0 0 Area Hospitals 7 16 16 0 128 Sub total, Hospitals 19 28 28 0 47 UrbanHealthPost (CESAMO)' 177 241 266 10.3 50.3 Rural health Post (CESAR) 516 867 1058 22 105 Mother and child clinics 0 15 50 233 500 CLIPERS (peripheral clinics) 0 3 3 0 300 Sub totalrambulatory facilities 693 1126 1377 22.3 98.7 Note: Centros de SaludCon Medico y Odontologo or health centers with doctor and dentist. 'RuralHealth Center 6.14 IHSS has continued to expand its services through both its own facilities and by subcontracting. As shown in Table 6.4, it has two national hospitals, and is offering hospital services to its affiliates through an additional 11 private hospitals located in different urban centers in the country and via a M O H hospital in the city of Juticalpa. Interms o f ambulatory care, it has 8 clinics (three clinics inTegucigalpa and five clinics inother cities) and subcontracts the services of 19 private ambulatory clinics (MOH Bulletin2005). Prior to 2003, the Institute provided services in 9 cities. From 2002-2005, it has extended its coverage to cover 7 more cities. The Company Medical System (Sistema Me'dico de Empresa, SME) functions alongside the IHSS, providing services through 136 clinics. The SME was developed in the early 1990s because of employers' concerns about the amount of work-time lost by workers obtaining IHSS care, mainly by the maquilas located in the Sula Valley whose workers had to travel to San Pedro Sula to obtain their care. Under the SME scheme, employers establish clinics at the work-site and contract medical staff, while IHSS provides drugs and medical supplies and manages referrals made by the clinics to IHSS facilities for both inpatient and outpatient care (Fiedler 2004). Table 6.4 IHSS facilitiesby Type ofProviderin2005 Type of Service Own MOH Private SME Total Hospitals 2 1 11 14 Day hospital or Center for the Elderly 1 1 Ambulatory (clinics) 8 19 136 163 Rehabilitation Centers 2 2 Dental Center 1 1 Source: M O H statistical bulletinand IHSS statistics. 95 6.15 The private sector is also an important provider of services-comprised by both independent providers of services and entities contracted out by either the MOH or IHSS. According to a 2002 censuss4of facilities, the private sector i s composed o f 107hospitals (99 for profit and 8 non-profit) and 902 clinics (760 for profit and 35 non-profit). The private sector, like IHSS, i s concentrated in the two major urban centers. Only 74 (25 percent) of Honduras' counties have one or more private health facilities, and only 38 (13 percent) of them had more thantwo private facilities. 6.16 As part of its efforts to expand health coverage, the MOH has also embarked on two major strategies that engage the private sector, predominantly NGOs: From 2001 to 2005, it contracted out services through NGOs to provide a basic package of services to communities who, otherwise, would not have had access to basic health services mainly because o f geographical reasons. This undertaking covered 290,000 persons in three years, including 1,100 communities in 14 departments with a cost of US$21 per person. However, after external funding stopped, these services did not continue because of the absence of a strategy to sustainthem as part of the regular service network. Ongoing implementation of new management models in which providers are contracted by their communities or other NGOs to provide health care to specific geographic areas, using a system of payment related to performance indicators (for example, per institutional delivery). These new models are being applied in three maternal and infant clinics (CMIS) and 11 communities with services oriented towards primary and ambulatory attention at the first level, covering a total of about 175,000 persons. These models will be discussed infurther detail inthe quality section of this chapter. C. Health ExpenditureAnalysis TotalHealth Expenditures 6.17 Honduras' total health expenditures as a percentage of GDP are lower than the average for Central America.*' It also has the lowest health expenditure per capita (US$184) in international dollars or purchasing power parity terms in 2003. However, when only public spending is considered, Honduras' spending on health is slightly higher than that of Guatemala's andNicaragua's (Table 6.5). 6.18 Honduras's government spending andprivate expenditures as a percentageof total health expenditure at 56.5 percent and 43.5 percent are more or less similar to the CA averages of 56 percent and44 percent, respectively (Table 6.6). 84. The census identified the names and location o f the facilities, but did not collect any information on the number or mix o fstaffor beds, nor any data about service provision. 85. Work on updating the National Health Accounts (NHA) is currently underway. At present, the only available NHA for Honduras was preparedin 1998. For that year, healthexpenditures inHondurastotaled 1.3 billion Lps and constituted 5.6 percent o f Gross National Income or 6.2 percent o f GDP (MOH, 2002). Households contributed the largest share o f financing, 45 percent, followed by the central government which accounted for 33 percent o f the total, external sources contributed 1lpercent, and private companies accounted for about 9 percent o f total health financing. Slightly more than 50 percent o f financing went to private providers. 96 Table 6.5 Population,GDP,and HealthExpendituresin CentralAmerica Population, Per capita Total health Average per Percapita Public 2005 in GDP expenditures capita health Expendituress6on millions in as a expenditures healthin International percentageof in International US%,2004 GDP,2003 International US%,2003 (Yo) US%,2003 Costa Rica 4.3 8,949 7.3 616 486 ElSalvador 6.8 4,793 8.1 378 174 Guatemala 12.6 4,486 5.4 235 93 Honduras 7.2 2,748 7.1 184 104 Nicaragua 5.4 2,832 7.7 208 101 Panama 3.2 8,103 7.6 555 368 Average 7.2 362.6 221 Source: WHO Core Indicators, 2006. Table 6.6 Total Healthexpenditures and privatehealthexpendituresby financingsource (in %) Government Private expenditure Prepaid plans as YO OOP expenditureon expenditureon health on healthas YOof of private healthas YOof private as YOof total health total health expenditure on expenditureon health, expenditure, 2003 expenditure,2003 health, 2003 2003 Costa Rica 78.8 21.2 2.1 88.7 ElSalvador 46.1 53.9 6.3 93.5 Guatemala 39.7 60.3 3.9 91.9 Honduras 56.5 43.5 7.1 85.8 Nicaragua 48.4 51.6 4.1 95.7 Panama 66.4 33.6 17.8 82.2 Average 55.9 44 6.8 89.6 Source: WHO. Core indicators, 2006. 6.19 While the proportion of private health expenditures financed through out of pocket payments (OOP) in Honduras is slightly lower (85.8 percent) than the CA average o f 89.6 percent, it i s above the most recently available LAC average (75.3 percent). Other countries, such as Bolivia and Jamaica, that have similar GDP per capita to Honduras, have higher shares of public health expenditure (64 percent inBolivia and 51 percent inJamaica) and lower private health expenditures financed by OOP (79 percent and 65 percent respectively). Total Public Health Expenditures 6.20 From 1999 to 2005, total public health expenditures increased inreal terms by 94 percent and in nominal terms by 217 percent (Figure 6.2). During this period, per capita public health expenditures in US$increased in real terms by about 83 percent. Also, the respective shares of public health expenditures out of total government expenditures and public health spending out of GDPhave increased by almost 55 percent and 59 percent, respectively (Table 6.7). 86. Public health expenditure includes both recurrent and capital spending fiom government (central and local) budgets, external borrowings and grants (including donations fiom international agencies and nongovernmental organizations), and social (or compulsory) health insurance fimds. 97 Figure6.2 Trendsin publichealthexpenditures, (constant 1995 and current Lps), 1999-2005 8oooooo.o7 7000000.04-1- Epublichealth expenditures (real) <.-E6000000.0 8 5000000.0 Epublichealth 4000000.0 expenditures 5 (current) B 3000000.0 n 3 2000000.0 1000000.0 0.o 1999 2000 2001 2002 2003 2004 2005 years Source:MOF statistics andauthor's calculations. Table 6.7 Publichealthexpendituresas sharesof total publicexpenditures - and GDP, and per capitapublichealthspending 1999 2005 YOchange from 1999 to 2005 %public healthspending/Total Government Spending 11.9 18.4 54.6 %public health spending/GDP 2.9 4.6 58.6 Per capita public health spending inLps (constant 1995) 198.4 318.5 78.5 Per capita public health inLps (current) 335.4 998.2 166 Per capita public health spending inUS$(constant 1995) 12.4 22.7 82.6 Per capita public health spending inUS$(current) 26.1 52.5 101 Source: MOF, MOH, and author's calculations. Note: figures inthis table are higher than those in Chapter 2 because these also include spending for IHSS, PRAF, and FHIS. 6.21 Among the four main government agencies involved inhealth, the MOH invests the most on health, its spending share averaging 77 percent of total public health expenditures from 1999 to 2005 (Table 6.8). Since 2002, and mainly as an aftermath of the social security reform o f 2001, IHSS has progressively increased its share of total public health expenditures from 17 percent in 1999to 30 percentby 2005. 6.22 The government has consistently been the main source of public health funds contributing, on average, about 86 percent while external sources provided 14 percent. In real terms, The former has increased fairly consistently while loans and grants tended to fluctuate from 1999to 2005 (Figure 6.3). 98 Table 6.8 HealthExpendituresand Sharesby Government Institutions, 1999 2005 - Total Health Expendituresby GovernmentInstitution(in constant1995 Ips) Year M O H IHSS PRAF FHIS Total 1999 204,843.75 10,067.5 20,545.9 1,197,986.5 2000 1,962,529.18 140,292.2 244,196.0 11,175.7 17,538.5 1,413,202.3 2001 1,325,580.0 234,063.2 11,397.3 45,928.7 1,616,969.2 2002 1,301,371.1 328,510.8 10,878.1 17,489.2 1,658,249.3 2003 1,5 19,175.1 409,397.0 10,952.0 10,729.9 1,950,254.1 2004 1,477,150.4 488,959.5 17,312.4 6,395.7 1,989,818.1 2005 1,570,121.7 687,759.8 23,024.3 11,615.4 2,292,521.1 Health Expenditures Shares by Institution(YO) Year M O H IHSS PRAF FHIS Total 1999 80 17 1 2 100 2000 81 17 1 1 100 2001 82 14 1 3 100 2002 78 20 1 1 100 2003 78 21 1 1 100 2004 74 25 1 0 100 2005 68 30 1 1 100 Source: MOFand author's calculations Figure6.3 Trends in publichealthfinancingby fundingsource(National and externalfunds), 1999-2005, inconstant 1995 Ips 2,500,000,000.00 2,000,000,000.00 I 500,000,000.00I 1999 2000 2001 2002 2003 2004 2005 year Sources: MOF and author's calculations 6.23 From 1999 to 2005, both recurrent and capital expenditures increased inreal terms by 84 percent and 198 percent respectively from 1999 to 2005. Recurrent expenditures have consistently comprised the major share of public health expenditures (about 93 percent on average), although its share has decreased to 88.6 percent by 2005. Capital expenditures have increased in 2003 to 2005 primarily because of the infrastructure investments undertaken by IHSS (Table 6.9). 99 Table 6.9. PublicHealthRecurrentand CaDitalExDenditureShares. 1999-2005 Year YORecurrent YOCapital 1999 92.7 7.3 2000 94.3 5.7 2001 89.9 10.1 2002 96.6 3.4 2003 94.8 5.2 2004 94.0 6.0 2005 88.6 11.4 Source: MOF 6.24 The following section will focus on MOH and IHSS expenditures because these are the two major publicproviders of health care inHonduras. Analysis of Ministry of Health (M0H)Expenditures 6.25 FundingSources ofthe MOH. The NationalTreasury has beenthe major funding source for the MOH. It financed, on average, about 85 percent of M O H expenditures from 1999 to 2005, followed by external credit (10 percent), and HIPC (3 percent). M O H financing from own resourceshas beenvery minimal (Figure 6.4). Figure6.4. MOH Sources of Financing(% contribution), 1999-2005 100% 90% 80% 70% /.HIPC g!m 60% 0 Grants S 50% 0 External Credit v) $ 40% rn Own Resources 30% ENational Treasury 20% 10% 0% 1999 2000 2001 2002 2003 2004 2005 years Source: MOF statistics. 6.26 MOH exDenditure trends. From 1999to 2005, total MOH real expenditures increasedby about 64 percent while its per capita expenditures increased by four percent in real terms. In 2005, MOH per capita spending was US$36 and, taking into account its official coverage rate of 60 percent, it spent approximately US$60 per personcovered by its services. 6.27 Recurrent expenditure shares during this period increased by 3 percent, averaging about 95 percent over the 7 years. Capital expenditures shares have been relatively low, peaking at 8 percent in2001 and thendropping to 3.3 percent in2005. 100 Table 6.10 ChangesinMOHtotaland per capita expenditures, recurrentand capitalexpenditure shares, 1999i o 2005 1999 2005 YOchange 1999 to 2005 Total MOHexpenditures (inconstant 1995Lps) 962,528.6 1,570,070.8 64 MOHper capita expenditures incurrent US $ 21 36 72 Recurrent expenditure share 93.6 96.7 3.3 Capital expenditure share 6.4 3.3 -48 Sources: MOF, MOH, and author's calculations 6.28 Program spending. Until2004, the M O H had allocated its spending to the following five main programs: (1) central activities, (2) ambulatory care; (3) environmental sanitation and health promotion; (4) hospital care; and (5) unassigned. In2005, it added a category of common activities and removed environmental sanitation and health promotion as a separate expenditure category. 6.29 Hospitals continue to have the largest share of total MOH expenditures, decreasing slightly from 42 percent in 1999 to almost 38 percent in 2005. On the other hand, ambulatory care's share declined by almost 50 percent from 29.6 percent in 1999 to 15.6 percent in 2005 (Table 6.11). Table 6.11 MOHPrograms: Trends inExpenditureShares 1999-2005 (in %) Program 1999 2000 2001 2002 2003 2004 2005 Central Activities 23.8 7.8 16.9 21.5 33.5 23.6 26.2 CommonActivities 5.8 Ambulatory Care and Epidemiological Control of Illnesses 29.5 27.6 31.3 22.6 19.4 19.5 15.6 Environmental Sanitation and Health Promotion 4.6 4.9 4.3 4.4 3.8 3.6 Hospital Care 42.1 48.4 39.5 44.8 32.8 40.7 37.7 Unallocated 11.2 7.9 6.6 10.5 12.6 14.7 100 100 100 100 100 100 100 Source: MOF statistics 6.30 Inrealterms, total MOHexpendituresincreasedby about 63 percent from 1999to 2005, while central activity expenditures increased by 80 percent, and expenditures on hospitals increasedby 46 percent. Expenditures for ambulatory programs, however, declinedby about 14 percent (Table 6.12). 6.3 1 Expenditure category allocations. Personnel costs comprised the majority of M O H total expenditures fiom 1999to 2005 (Table 6.13), averaging about 53.7 percent across the sevenyear period, followed by transfers (16.3 percent), medicines (10.7 percent), and materials and supplies (10.5 percent). 6.32 Personnel costs, however, comprised a larger percentage (67 percent) of M O H hospital expenditures, followed by materials andsupplies (16 percent) andmedicines (10 percent). 101 Table 6.12 Change inMOHExpendituresby Program Category: 1999-2005 (in 000 constant 1995 Lps) Program 1999 2005 YOchange 99- `05 Central Activities 228,822.0 412,128.0 80 CommonActivities 90,893.4 Environmental Sanitation & health prom. 44,442.8 Ambulatory Care & Epidem. Control 283,996.8 244,592.9 (14) Hospital Care 405,266.9 592,146.8 46 Unallocated 230,309.8 Total 962,528.6 1,570,070.8 64 Source: MOF statistics and author's calculations Table 6.13 M O HExpenditureGroup Shares, 1999-2005(in %) Personnel Non-personnel Materials and Medicines Goodsand Transfers* services Services Supplies Equipment 1999 47.1 4.8 12.9 12.2 6.4 16.6 2000 50.6 4.3 10.7 15.8 5.5 13.1 2001 54.2 3.1 9.3 13.4 8.4 11.6 2002 56.7 5.4 10.2 14.1 1.7 12.0 2003 58.1 3.4 8.3 2.2 0.7 27.3 2004 56.7 6.6 9.3 6.9 4.6 16.0 2005 52.7 7.3 8.2 10.8 3.3 17.8 Average 54.7 4.9 9.8 10.7 4.0 16.3 Source: MOH; *transfers to SANAA, NGOs, and National Universities. 6.33 Interms ofrelative increases, real expendituresonnon-personnelservices have increased the most (150 percent) from 1999 to 2005, followed by spending on personnel (82 percent), transfers (75 percent), drugs (45 percent), materials and supplies (1 percent). Spending on goods and equipment declined by 16 percent. However, during this period, MOH spending on goods and equipment was complemented by FHIS expenditures which were channeled towards construction, repair, and Table 6.14 M O H budget rehabilitation of the MOH's health centers and maternal and ExecutionRates,1999-2005 MOHBudget Execution infantclinics. Rates(YO) Budget execution rates 1999 67.3 2000 89.0 6.34 From 1998-2005, MOH's annual budget execution 2o02 2001 79.3 rates averaged about 78 percent, ranging from 67 percent in 74.3 1999 to a high of 89 percent in 2000, and declining to 82 2o03 74.8 2004 82.9 percent in2005;Table 6.14. 2005 82.4 6.35 The relatively low budget execution rates are mainly Source :MOF, due to planning and implementation constraints at the central and regional health administration levels. Plans and budgets have been traditionally prepared based on historical trends and have tendedto been done separately. The MOH is taking steps to address these issues, partly by (a) establishing resource allocation criteria to regional health administrations and by (b) agreeing to 102 look into a results-basedplanning and budgeting approach (this will be discussed inmore detail inthe section on sectorexpenditureframework preparation). Analysis of IHSSExpenditures 6.36 From 1999 to 2005, IHSS spending as a proportion of total government spending and GDP increased by 175 percent and 180 percent, respectively. Its spending per beneficiary has also increasedinbothreal andnominalterms (Table 6.15). Table 6.15 IHSS Expendituresas a Percent of GovernmentSpending and GDP, spending per beneficiary,1999to 2005 1999 2005 YOchange from 1999 to 2005 YOIHSSspendinglTota1Government 2 5.5 175 Spending %IHSSspendinglGDP 0.5 1.4 180 SpendingBeneficiary inLps (constant 1995) 256.4 447 74.4 SpendingBeneficiary inUS$ (constant 17.8 23.5 31.7 1995) SpendingBeneficiary inUS$(current) 33.7 72.9 116 Sources: MOF, IHSS, and author's calculations 6.37 Salaries as a percentage of IHSS total spending ranged from almost 66 percent to 73.6 percent in 1999 to 2001. After the 2001 IHSS reforms, it declined considerably to about 40 percent by 2005 because IHSS increased its spending on construction and equipment. IHSS hospitals spent a slightly higher amount on salaries (43 percent), followed by goods (26.4 percent) and medicines (12.7 percent) in2005. Table 6.16. MSS: expendituresharesof expenditure categories, 1999to 2005 1999 2000 2001 2002 2003 2004 2005 PersonnelServices 65.8 65.8 73.6 52.0 53.5 47.0 35.9 Non-personnelServices 5.9 5.4 7.4 5.1 6.1 9.4 10.2 Materials and Supplies 6.3 5.0 5.0 3.8 2.8 3.7 3.1 Medical. Surgical, and laboratory equipment 8.0 1.2 1.6 5.0 1.8 1.0 16 Medicines 5.6 5.7 2.7 11.7 6.5 13.0 8.9 Various transfers 7.2 16.8 9.1 20.0 9.5 17.5 130 Recurrent Total 98.8 99.8 99.4 97.6 80.2 91.5 72.6 Office equipment 0.3 0.2 0.3 0.2 0.9 0.4 2.1 Medical andlaboratory equipment 0.7 0.0 0.2 1.5 13.5 0.9 13.5 Vehicles and others 0.1 0.0 0.0 0.1 1.4 2.4 0.2 Construction 0.2 0.1 0.1 0.5 4.1 4.7 11.5 Total Capital 1.2 0.2 0.6 2.4 19.8 8.5 27.4 Grand Total 100 100 100 100 100 100 100 Sources:MOF and IHSS 103 6.38 Table 6.16 shows that recurrent expenditures decreased from 2002, while capital expenditures significantly increased from less than 0.4 percent in 1999 to 27.4 percent by 2005. IHSS spending on medical equipment increased from less than 0.2 percent in 1999 to 13.5 percent in2005 while infrastructure costs have significantly increased from less than 0.5 percent to 11.5 percent. D.Access andEquity Household Analysis 6.39 Utilization. Both the ENCOVI and DHS surveys conducted in 2004 and 2006-2006 indicatethat: Majority o f all Honduran households sought care inpublic health facilities (59 percent inDHS and 57 percent inENCOVI). The richest households sought care inprivate facilities (62.5 percent inDHS and 55 percent inENCOVI) while households inthe lowest income quintile consulted CESAMOs and CESARs. 6.40 About 30 percent o f households reported an illness and this was fairly equal across income quintiles, but higher for rural households (35 percent) than urban ones (27 percent). Of those households who reported an illness, only 47.5 percent sought care. More than twice as many households (65 percent) in the highest income quintile sought care compared to only 29 percent o f the households inthe lowest quintile. More urbanhouseholds (55 percent) also sought care compared to rural households (42 percent). Table 6.17 also shows a direct relationship between householdincome level and households going for non-illness related consultations. Table 6.17 HouseholdHealth SeekingBehaviorduringPreviousMonth Consulted Got Ill Consulted FacilityConsulted When Ill last Last WhenIII MOH IHSS CESAMO CESAR Private Other month Month Hospital Hospital Facility but not Ill Income Quintile Lowest income (Ql) 0.8 30.2 29.2 10.3 0.2 39.2 29.9 6.5 14 42 1.2 31.0 41.1 19.1 1.4 42.7 17.3 13.9 5.5 43 2.6 30.7 48.1 21.1 6.1 34.1 13.3 19.2 6.2 44 4.7 33.3 52.7 20.5 9.4 27 5.2 31.8 6.1 HighestincomeQS 8.5 31.4 65.5 19.9 10.2 10.1 1 54.8 4.1 Location Urban 2.4 27.2 55.3 22.9 10.5 22.6 1.1 38.0 4.9 Rural 4.8 35.2 41.9 15.3 2.7 33.0 19.7 21.5 7.8 All 3.6 31.3 47.5 19.0 6.5 27.9 10.7 29.5 6.4 Source: WE3 2006 basedon ENCOVI2004 6.41 Reasons for Not Seeking Care. While 30 percent o fhouseholds reported that they did not seek care because it was not necessary and 24 percent resorted to home remedies, approximately 20 percent o f households cited lack o f funds for treatment. About 27 percent o f those in the poorest quintile and in rural areas mentioned funding constraints compared to only 9 percent o f therichest households and 18.4 percent ofthose inurbanareas(Table 6.18). 104 Table 6.18 Reasonsfor Not SeekingCare amongThose Who said Theywere Ill2004 Reasons Not Home No Not Too Just ill Clinic Other Necessary remedy money serious far closed Income quintile Poorest Q1 24.5 19 26.9 9.7 7.2 3.1 2.5 7.1 42 24.3 26.9 24.2 11.7 2.6 2 1.7 6.6 43 31.6 26 18.7 12.9 1.8 3.5 0.6 4.9 44 36.6 23.7 15.6 16.1 0.8 2.1 0.8 4.3 Q5 40.2 25.5 9 16.1 0.9 2.5 1 4.8 Location Urban 40.3 18.4 16.4 15.4 0.5 2.6 0.6 5.8 Rural 24.6 26.8 22.3 11.5 4.5 2.7 1.9 5.7 All 30.2 23.8 20.2 12.9 3.1 2.7 1.4 5.7 Source: ENCOVI data cited inWE3 PovertyAssessment (2006) appendix tables. 6.42 Problems Encountered in seeking, care. According to the DHS (2005-06), women encountered four main problems in accessing health care: lack o f availability of drugs (81 percent), no one will attend to them (74 percent), lack of funds (58 percent), and transport and distance to facility (47 percent). Table 6.19 Problems Citedbywomen in AccessingHealth Services Obtaining Obtaining Distance Transport Does Thinks Think Thinks permission h d s not there will not there will to go for want to will be be be no treatment go alone no attended available female to medication staff Quintile 1 30.1 79.9 75.1 75.1 61.4 62.3 86.6 92.6 2 22 71.4 64.8 65.5 53 55.9 82.4 89.9 3 18.7 61.3 49.6 50.8 42.9 47.8 77.3 84.9 4 15.6 52.9 34.5 37.9 38.4 40 71.1 79.7 5 12.5 37.8 24.5 26.1 34.7 29.5 58.7 64.6 Urban 15.4 49.1 30.7 33 37 37.1 66.6 74.3 Rural 22.8 68.9 65 65.7 53.1 54.5 81.8 88.4 Total 18.8 58.3 46.5 48.1 44.4 45.2 73.7 80.8 Source: ENDESA 2005-06 6.43 Waiting time in private facilities (0.7 hour or 42 minutes) was less than half of the waiting time in M O H and IHSS hospitals and CESARS (Table 6.20). People waited longest (about 2 hours) in M O H hospitals. The average waiting time for households in the top quintile was less than the households in other quintiles although there was no significant difference in waiting times across the four other income quintiles and between urban and rural areas; Table 6.20. 105 Table 6.20 Average waitingtime per consultationby facility, incomequintile,and urbadrural Average total waiting time (for MOH IHSS CESAMO CESAR Private all facilities) hospital hospital facility Quintile 1.4 1.7 1.6 1.3 1.8 0.7 1.5 2.1 2.4 1.3 1.7 0.6 1.5 2.2 1.6 1.4 1.8 0.7 1.5 2.2 1.9 0.5 1.8 0.9 5 1.1 1.8 1.8 2.8 1.7 0.7 Urban 1.5 2.1 1.9 2 1.9 0.7 Rural 1.4 1.9 1.5 1.3 1.7 0.7 Total 1.4 2.0 1.8 1.3 1.8 0.7 Source: INE2006. 6.44 Household Health Expenditures. Household health per capita expenditures in 2004 amount to about 1,688 Lps or US$88 per year. While percentage shares do not vary between rural and urban households, actual amounts spent do, with urban households spending twice as much (2,352 Lps annually or US$124.4 per capita) compared to 1,044 Lps (US$55.2) for rural household^.^^ Table 6.21 Sourcesof financingfor consultation,medicines, and examinations by incomequintile, urbadrural(by percentageshares) Category Own Only Only Only Only Own Own Own Own Others resources MOH IHSS Insurance donated by funds funds funds & funds & facility- and & insurance facility NGO clinic MOH IHSS Quintile PoorestQ1 17 5 0 0 3 72 0 0 0 0 42 17 8 1 0 2 67 1 0 1 1 43 28 7 5 1 2 50 3 1 1 1 44 36 5 7 1 1 39 4 1 1 1 Q5 50 4 7 3 2 21 4 3 1 1 Urban 38 6 6 1 1 37 4 2 4.2 1 Rural 21 5 2 1 2 63 1 0 3.7 0 Total 29 6 4 3 2 50 2 4 1 1 Source: INEbasedon ENOVI2004 data. 6.45 Household sources o f financing for health care. Only 7 percent of those who sought medical care had some form of insurance (3 percent private insurance or 4 percent through IHSS) to pay for their health care costs during their last consultation, while 8 percent benefited from exemptions (6 percent from a M O H facility and 2 percent from the NGO facility). About 29 percent o f households that sought medical care paid with their own resources while 50 percent paid by combining their own funds with M O H support. There i s a direct relationship between payments fiom own resources and level o f income, with significantly fewer households from the poorest quintiles paying solely fiom their pockets compared to households in the highest income quintile; more households inurban areas also pay with their own funds relative to households in rural areas. About 72 percent of the poorest quintile financed their health care using both own and M O H funds compared to 21 percent of households in the richest quintile; Table 6.21. 87. WB PA estimates basedon the ENCOVI data indicate that householdper capita monthly spending on health i s about 139Lps or US$7.3 monthly and householdtotal monthly spending per capita is 1,424 Lps or US75.3. 106 6.46 Table 6.22 indicates that among the households who benefited solely from M O H financing, 31 percent of them were households in the two richest quintiles, indicating cross- subsidiesthat exist inthe system. More households inthe third (24 percent) and second quintiles (26 percent) tend to benefit more from M O H financing than households in the lowest income quintile(18 percent). Table 6.22 Percentageof householdsfinanced bythe MOHonly and by acombinationof own resourcesand MOHfunds duringtheir last consultationby incomequintile Income Quintile M O H financingonly M O H + own resources 1 18 28 2 26 26 3 24 21 4 18.7 15.6 5 12.1 7.2 Didnot declare income 1.4 1.9 Source: INEbasedon ENCOVI2004 6.47 There was hardly any difference across quintiles and betweenurbanand rural households interms ofhavingtheirhealthcarecostsfinanced byanNGO facility or solely bythe MOH. 6.48 Benefit-Incidence o f Public Health Spending;. The WB Poverty Assessment Report indicates that, assuming uniform quality of public medical services, social spending for health in Honduras tends to be slightly focused on the poorest quintiles. In particular, the two poorest quintilesreceive about their share ofthe population inhealth spendingor about 41 percentwhile thetwo richest quintilesreceive 36.8 percent. 6.49 The MOH is the main provider of consultation services for poor households, attending to almost 78 percent of each o f the bottomtwo income quintiles, compared with only 27 percent of the richest quintile. The bottom two quintiles account for 74 percent and 54 percent of consultations inCESAMOs and CESARs, respectively while the richest quintile has significantly lower consultation shares for CESAMOs (3 percent) and CESARs (6 percent). For M O H hospital consultations, however, the third and fourth quintiles have the highest shares, while the poorest quintile and richest quintile have lower shares; the richest quintile uses mainly private and IHSS facilities. 6.50 As expected, 67 percent of IHSS' provision is concentratedinthe top two quintiles (the urbanformal sector). Similarly, private suppliers provide about 58 percent of medical attention to the richest quintile (ENCOVI 2004). 6.51 In terms of programs, the Government's Primary Health Care and Epidemiological Control of Diseases programs which consists of vaccination program has a pro-poor character. There is no gap between income quintiles and between urban and rural areas in terms of vaccination rates. A greater percentage of children from the poorest quintiles and rural areas are vaccinated in CESAR health centers. However, the hospital care program does not appear to have a clear distributional incidence, appearingto benefitthe third quintile the most. Depending on the assumptions used, this program is either (a) neutral at best when uniform donations/subsidies are assumed, or (b) not pro-poor, when survey reported donations are taken into account inthe analysis. 107 Geographic allocation of resources 6.52 The average health spendingper capita across regions in2005 was about 28 Lps, ranging from 11-38Lps in Santa Barbara to 122.3 Lps in Gracias a Dios. Table 6.23 suggests that there seems to be no direct relationship betweenhealth expendituresper capita andthe level of poverty andpopulation size ineachregion. For example, Atlantida spent about 31Lps, whichis about 70 percent more than what El Paraiso spent although it has a similar population size and higher poverty rate. Also, Lempira, which has the highest poverty rate among all the regional departments, spent only about 17.4 Lps per capita; this is much lower than the health spending per capita of regions with lower poverty rates such as: Ocotepeque, Olancho, Valle, La Paz, Choluteca, El Paraiso, Colon, Comayagua, Islas de la Bahia, Atlhtida y Francisco Morazh. On the other hand, Graciasa Dios spent 122Lpsandits poverty rate is 72. Table 6.23 RegionalHealth Spending,Population,andPovertyIndex,2005 RegionalHealth Health Population Health Poverty Index Administration Expenditures expenditures per in Lps capita Tegucigalpa 13,225,266.04 983,179 13.45 32.2 San Pedro Sula 8,393,807.1 1 622,657 13.48 44.9 FranciscoMorazan 21,294,362.91 1,179,768 18.05 49.9 Cortes 18,858,078.84 1,365,497 13.81 53.0 Atlantida 11,532,112.29 372,532 30.96 53.5 Islas de la Bahia 1,928,793.93 43,018 44.84 59.2 Valle 5,477,459.61 160,346 34.16 61.9 Yoro 6,856,226.90 530,886 12.91 66.2 Comayagua 8,582,154.46 390,643 21.97 67.2 Colon 6,147,876.10 266,775 23.05 68.3 ElParaiso 6,958,443.74 383,565 18.14 69.4 Choluteca 10,799,966.93 420,350 25.69 69.9 LaPaz 5,087,246.68 173,731 29.28 70.4 Olancho 11,154,057.28 458,365 24.33 71.5 Gracias a Dios 9,333,888.13 76,278 122.37 71.8 Ocotopeque 4,736,726.69 118,558 39.95 72.1 Santa Barbara 4,192,844.14 368,29 8 11.38 72.5 copan 8,608,408.45 320,562 26.85 75.0 Intibuca 6,520,968.26 202,140 32.26 77.2 Lempira 4,852,668.14 277,910 17.46 80.9 Source: MOFandMOH E. EfficiencyofPublicHealthSpending 6.53 Chapter 4 of this report includes an updated analysis that compares the efficiency o f public spending inhealth in Honduras based on the analytical framework developed by Herrera and Pang in 2004 confirms that further efficiency gains can be made inthe sector. By relating total publichealth expenditures andtwo other inputs (adult literacy andprivate expenditures) and four output indicators referring to life expectancy at birth, DPT (Diphteria, Pertussis and Tetanus), Measles immunization and the Disability-adjusted life expectancy (DALE) index, 108 which takes into account both mortality and morbidity,88 results indicate that Honduras has an input-oriented single input/output measure of 0.68, which i s higher than the CA average of 0.65 and slightly lower thanthe LAC average of 0.68. Its output oriented efficiency score of 0.92 is slightly higher than the Central American average of 0.91 and LAC average of 0.90. A similar picture emerges when using multiple inputs and outputs with Honduras having similar scores to the two regional averages. Inother words, while the Honduran health sector's efficiency score is fairly similar and at times, slightly better than the CA and LAC averages, there i s still room for increasing efficiency by 32 percentage points when using an input-oriented approach and by 8 percentage points when an output efficiency measure is used. 6.54 Machado (2006) extends the preceding analysis by measuring the operational efficiency of public spending from 1999-2003 in terms o f a country's poverty and over-all development. His findings indicate that Honduras's health sector performance index (measured by life expectancy at birth and infant mortality rate) is 0.95, which is lower than the CA average of .99 andthe average for 19 LatinAmerican countries of 1.0. Figure6.5 Relationshipbetween Child StuntingLevelsand Per Capita HealthExpendituresinPPP Terms* % usstunting Child stunting & per capita health expenditure 60% 50% 40% 30% - . - 20% NlCaQEWa ti> aivaaor 1 .Argentina 10%. Cost Rica 0% + Chile 0 200 400 600 800 1000 1200 Per capita health expenditure (in international dollars, PPP, 2003) Note: *Basedonlatestcomparabledataacross countries Source: WHO CoreIndicators, 2006, Honduras2003 figures estimatedbasedon 2001 data andDHS2005/06. 6.55 A study by Evans et al. 2006 estimates the efficiency of the health system of 191 countries by comparing levels of health outcomes and health inputs (health expenditure per capita and education level). The study finds that it is difficult for systems to be efficient with annual healthper capita expenditures lower thanUS$SO. Honduras' per capita health expenditure i s higher than US$80, but it can still improve the efficiency of health spending in terms of some key indicators such as chronic malnutrition for under-five year old children (Figure 6.5). 88. Herrera, Santiago and Gaobo Pang, (2004), "Efficiency of Public Spending in Developing Countries: An EfficiencyFrontier Approach", WorldBank Working Paper, December. 109 Honduras has a slightly higher stunting rate than Bolivia although it spends more per capita on health in international $ terms (PPP). Its stunting rate is also much higher than Nicaragua's although their per capita healthexpendituresinPPP terms is relatively small. 6.56 Recent ENDESA results, however, indicate that Honduras has made notable progress in reducing neonatal mortality, infant mortality, under-five mortality rates. Nevertheless more can be done to improve nutritionespecially inrural and poor areas. Honduras can also improve how efficiently it i s using its resources, especially in terms of its resource allocation emphasis on curative care andhospitals insteadof preventive care and its management of humanresources. Human resources. 6.57 Numberanddistribution. In2005, the Ministryof Health had 16,024 employeesofwhom 31 percent are administrative and 69 percent (11,078) are clinical staff. O f the latter, 90 percent on curative care and injuries (atencion de daAo) and only about 9.7 percent are dedicated to prevention of risks and healthpromotion (MOH 2006). 6.58 Inthe same year, the Medical College of Honduras listed 6,309 doctors, of whom 2,033 are specialists and 4,276 are general practitioners. 2,080 doctors work for the SS and 600 doctors work for IHSS on either a permanent or contractual basis.89 On the other hand, professional nurses (with a university degree) remain disproportionately fewer than doctors in Honduras. There are a total of 1,089 professional nurses inthe public sector, 886 of them work for the SS and 203 are with the IHSS. There are 6,677 auxiliary nurses inthe public sector, of whom 5,834 work for SS and 843 for IHSS. 6.59 According to international standards (WDR 1993), public health and minimum essential clinical interventions require about 1 physician per 10,000 people and 2 nurses per 10,000 persons. Honduras exceeds this minimum standardfor physicians. Inthe case o f nurses, it only meets this standard if both professional nurses and auxiliary nurses are taken into account. However, if only professional nurses are considered then there are only 1.5 nurses to 10,000 persons inthe country. While there are disproportionately more doctors than professional nurses inHonduras, it still has a lower doctor per 10,000 personsratio of 9 which is less thanhalfofthe most recently available average of 19 for the LAC region. 6.60 In principle, the MOH knows how many professionals they have but the official designation of a health worker is frequently not a good indicator o f actual staff distribution across regions because a doctor might be working inTegucigalpa although she i s being paid by another regional department. 6.61 Moreover, the M O H decides the number of staff neededbased on historical budgets, and thus health personnel deployment has not beentraditionally based on the existing needs, risks, and opportunities. 6.62 Personnel costs. Staffing costs amount to a substantial share of the M O H budget. Since 1996, personnel cost increases have been driven by the Estatuto Mbdico, a law which sets physicians' salaries as a multiple of the legal minimum wage. The base salary of MOH general 89. The MOHhas 2,043 doctors working on a permanent basis while IHSS has 124 generalpractitioners and 252 specialists; IHSS has 135 nurses and 562 auxiliary nurses working on a permanentbasis. (MOH and IHSS statistics 2005) 110 practitioners and specialists are pegged at 12 and 14 times the minimum wage, respectively. Thus whenever the government raises the minimumwage, MOHphysicians also receive a wage increase. The Estatuto defines a 6-hour workday as the norm for all M O H physicians, which, together with other provisions, has inflatedpersonnel costs. 6.63 From 1996 to 2006 physicians' average base salary grew faster than that of all other M O H personnel categories (Table 6.24). While the rest of MOH staff-person's average salary grew by a factor of about 4.3 over this period, general practitioners' average salary grew by a factor of 6.5. Table 6.24. MOHStaffSalaries: Trendsfrom 1996to 2006 Type of Staff and Average Annual Salary 1996 2000 2006 A. InNominal Lps Physicians-GeneralPractitioners 55,087 216,359 358,648 Nurses 32,183 105,000 140,488 Nurse Auxiliaries 21,532 52,005 92,613 All Others 21,549 78,879 91,768 B. Relativeto thePosition's 1996level Physicians-GeneralPractitioner 100 393 651 Nurses 100 326 436 Nurse Auxiliaries 100 242 430 All Others 100 366 426 C. Relative to Physician-GP salav (averageper year) Physicians-GPs 100 100 100 Nurses 58 49 39 Nurse Auxiliaries 39 24 26 All Others 39 36 26 Sources: MOH statistics, Fiedler 2004, andauthor's calculations 6.64 As shown inFigure 6.6, the average monthly salary of aphysician inHonduras is higher, ranging from 1.5 times to 3.8 times that of average monthly salaries o f physicians in five other Central American countries. 6.65 Moreover, Honduras's average yearly physician salary as a percentage o f GDP per capita i s 1,476 percent in 2006, which i s at least three times the ratios of El Salvador, Guatemala, and Nicaragua (MOH statistics 2006, PAHO, and WB 2007). However, Honduran doctors have shorter working hours (six hours per day; overtime would need to be paid beyond these working hours) thanthose inother countries. 111 Figure 6.6 Average General Practitioner Salaries- comparison among 6 CA countries I country Source: Morales et. al. (2006) based on ECLAC& WHO; Honduras data from MOH. 6.66 While it may not be realistic to expect changes in salary levels for doctors in Honduras because of its potential political economy implications and the strength of the unions, the M O H can take steps to: (a) ensure a better allocation of doctors across the country, instead of being concentrated in urban areas especially in Tegucigalpa and San Pedro Sula; (b) train and retain more nurses and ensure that they are also well-distributed across regional health administrations; and (c) explore the possibility of hiring future personnel on a contractual basis based on performance instead of continuing with the current practice of having most workers on permanent contracts which makes it difficult to apply sanctions and/or fire them for poor performance. Procurementof Pharmaceuticals, 6.67 The purchase and distribution of pharmaceuticalsis mainly a Central level responsibility, although the Regional Health Administrations are entitled to purchase medicines. For many years the MOH has struggled, due to problems with institutional capacity and corruption, to ensure that hospitals and health centers obtain their medicine requirements. Shortages o f medicines were reported to be a recurrent problem. In order to ensure transparency in the purchase of medicines the M O H delegated this task to UNDP. However this practice had to be discontinued because o f the long period required for UNDP to conduct this process, its transaction fees and an accusation of engaging in some corruptive practices. In July 2006, the President of Honduras createdthe Inter-Institutional Commission comprised ofrepresentatives o f the Catholic Church Archdioceses, the Medical Association (Colegio Medico), the National Council of Anticorruption, the Forum for AIDS and the Forum for National Convergenceto lead the procurement process for medicines (CAN 2006). This Commission has facilitated the procurement of medicines and by December 2006, about 95 percent of all medicines requirementswere purchased(La Prensahn.com2006). 6.68 The Inter-Institutional Commission also identified the following as the main bottlenecks inthe procurement of medicines: (i) of commitment of certain MOHpersonnel to expedite lack the procurement process; (ii) deficiencies in the planning and programming of the medicines required; (iii) inadequate proceduresto issue purchase orders and provider payments orders; and 112 (iv) high level o f inefficiencies inthe process of reception, storage and distributionof medicines due to insufficient communication channels within the MOH and inadequate logistics ( for example, lack of equipment and vehicles to transport the medicines). 6.69 In order to address the above problems, the Commission recommends: (i)major organizational changes in the MOH, (ii) have M O H operations subject to audits by both public and private sector agencies, (iii) the SOHpersonal who hinder or do not comply with the punish procurement procedures.; (iv) upgrade the MOH's model for central storage with external aid and to review, (v) every three years, update the Basic Medicines Profile according to international standards, and (vi) empower the National Health Council (Consejo de Salud or CONSALUD inSpanish)" as an agency that couldprovide technical oversight. 6.70 It would be important to see how this processevolves inthe near future and whether the above recommendations yield improvements not just inthe timely purchase of adequate drugs at the national levelbut also inhowthese translate into actual availability andlower prices of drugs inhealthfacilities. Productivity 6.71 Yearly production trends from 1999 to 2005 by type of provider and facilities are only available for public providers (MOH and IHSS). During this period, the MOH and IHSS increasedtheir ambulatory consultations by 30 percent andpercent, respectively. 6.72 From 1999-2005, the M O H provided, on average, about 83 percent o f the public sector ambulatory consultations in the country. In 2005, its coverage declined slightly to 78 percent. This i s partly because o f the progressmadeby IHSS inincreasing its affiliates and expanding its service coverage, As a result, IHSS' share of average public sector ambulatory consultations provided increasedfrom 16percent 1997to 21 percent in2005 6.73 M O H and IHSS have very similar patterns interms of shares of ambulatory consultations provided by type of facility. Both provided, on average, about 27 percent of their ambulatory consultations inhospitals and 73 percent inclinics andhealth centers. 6.74 From 1999 to 2005, the expansion of the MOH ambulatory network has resulted in an increase in ambulatory consultations per 1000 persons by 11.3 percent while hospital consultations per 1000 persons increasedby about 4.7 percent. While the share o f consultations of hospitals has declined by 3.4 percent during this period, available data in Table 6.25 indicate that a number of the ambulatory consultations provided in hospitals deal with basic issues that could have beentreated inambulatory facilities. 90. CONSALUD is a high-level inter-institutional organization responsible to provide counseling and advice to the Government inHonduras inthe reform and consolidation of the national health system and social security, and to coordinate consultation with different agencies o f the health system. By conducting these tasks, CONSALUD is expected to create the conditions for a continuous increase o f the efficiency, efficacy and rational use o f the resources o f the health system, while at the same time respecting the principles o f equity, universalism and solidarity. CONSALUD i s comprised o f 3 1members from the public andprivate sector such as the sector ministries, corporation boards, councils o fprivate business and academia. 113 Table 6.25 Ambulatory Consultationsper 1000 personsby type ofMOHfacility Facility 1999 2005 YOchange YOchange in 1999-2005 shares of consultations NO. Yo NO Yo Hospitals 317 29 332 28 4.7 -3.4 Ambulatory facilities 780 71 868 72 11.3 4.2 Total 1097 100 1200 100 9.4 Source: MOH statistics 6.75 Ambulatow Facilities. Interms of the average daily medical attention provided, CESARs increased their output by 20 percent from 1999 to 2005 while CESAMOs increased theirs by only 3 percent. Despite these increases, the average daily output for the period i s 12 for CESARs and 44 for CESAMOS. This is still relatively low taking into account that the national standardis about 36 consultations per doctor per day and that these facilities-especially the CESAMOS also have other personnel suchas nurses who are also supposedto provide medical care. Table 6.26 Average Daily MedicalCare AttentionProvidedby Type ofHealth Center 1999 2000 2001 2002 2003 2004 2005 Average %change'99to'05 CESAR 10.5 10.2 10.9 12.8 12.1 12.6 12.6 12 20 CESAMO 41.7 39.1 40.5 48.8 48.1 49.6 42.9 44 2.9 Source: MOH statistics andauthor's calculations 6.76 The low utilization rates of ambulatory facilities can be attributed to issues related to geographical access, logistics, lack o f medication, limited hours of service, and personnel according to studies undertakenby the Access Project (Proyecto Acceso). This is also confirmed by the ENDESA results which indicate that about 80 percent of women cited lack of medicines as a problem in accessing health services, followed by lack of attention, insufficient funds to obtain treatment, andproblems oftransport and distance to facilities. 6.77 Hospitals. From 1999 to 2005, hospital production indicators did not demonstrate a consistent performance pattern; with markeddecreases observed in2000 and 2001 (Table 6.27). Nevertheless, during this period, the number of hospital beds increased by 7.6 percent while discharges increasedby 15.7percent. Total annual dischargesper 1000persons declined slightly by 2.7 percent from 1997to 2005. Nationalhospital dischargesper 1,000 persons declinedby 14 percent during this period, those of regional hospitals increased by 9.1 percent and there was no change in area hospital discharges. The average hospital bed occupancy rate during this period was 70 percent, increasing by 2.8 percent from 1999 to 2005. The number o f births attended in public hospitals increasedby 2.7 percent from 1999to 2005. 6.78 Despite these general improvements, hospitals continue to operate with certain inefficiencies: hospitals tend to procure their equipment individually thereby resulting in higher prices. These purchases are also usually not based on nationally set standards, underscoring the need for the MOH to provide guidance on equipment standards and establish a system of coordinated procurement to benefit from lower prices. Also, majority of their personnel work mainly in the mornings and therefore result in inefficient use of existing facilities such as surgical rooms, diagnostic services and equipment, resulting inlong waiting lists and delays. The 114 share of cesarean operations out of total number of births has also increased from 12 percent in 1999to 20 percent in2005, the latter is slightly higher thanthe recommendedrate o f`15. Table 6.27 M O HHospitalProductionIndicators, 1999 to 2005 Hospital Production Average from YOIncrease Indicators 1999 2001 2005 I999 to 2005 1999 to 2005 Number o fbeds 4,093 4,076 4,403 4,163 7.6 Bedper 1000persons 0.68 0.65 0.62 1 -8.8 Total Discharges (000) 228,623 204,433 264,487 236,058 15.7 Nationalhospital 85,992 74,736 88,406 86,227 2.8 Regional hospital 65,474 63,936 84,971 71,410 29.8 Area hospital 76,226 65,294 91,470 77,911 20.0 Total bed-days (000) 1,493,945 1,449,000 1,588,727 1,506,565 6.3 Average length o f stay 4.6 4.5 4.4 5 -4.3 Occupancy rate 71 64 73 70 2.8 Surgeries 66,550 64,645 94,650 77,342 42.2 Numbero fbirths 92,058 75,254 94,555 87,8 11 2.7 Cesareansections 11,373 11,278 18,453 14,005 62.3 Hospital discharges per 1,000 persons national hospitals 14 12 12 13 -14.3 regional hospitals 11 10 12 11 9.1 area hospitals 13 10 13 12 0.0 Total 38 32 37 36 -2.6 Source: M O H statistics, author's calculations 6.79 Maternal and Infant Clinics. Table 6.28 indicates the low production rates of CMI (Clinicus Maternus-ln~untiles).The average number of discharges per month i s about four over the course of 7 years, dropping to a low o f two per month in2005.91 Only about 2.4 percent of nationwide births take place in these facilities. According to the Direccion de Regulacion Sanitaria, 96 percent of the CMIS have adequate infrastructure but only 39 percent have the necessary medicines. Also, about 20 percent of these facilities are not open 24 hours per day and only 39 percent of them have adequatepersonnel while the rest dependonly on auxiliary nurses. Other reasons cited by the study are geographical and economic access, as well as cultural barriers especially among ethnic groups and inrural zones where there is scarce promotionof the clinic's existence and benefits to communities. Table 6.28 C M I productivityindicators, 1999to 2005 1999 2000 2001 2002 2003 2004 2005 Ave. % change 99-05 Consultations/CMI 8,008 9,806 9,749 9,243 8,081 8,769 5,197 8,408 -35.1 Discharges/ C M I 62 68 20 34 69 94 25 53 -59.7 NumberofCMIs 15 23 23 28 28 32 5Q 28 233.3 Source: M O H Statistics 91. It is not clear why rates dramatically dropped for 2005. These could be preliminaryestimates. 115 6.80 A WB 2006 survey of 18 CMIs carried out in 2006 also underscores limitations in personnel and medicines. Forty four percent of the CMIs do not have the required number of personnel, andall but one CMIhad inadequatestocks of medicinesand other inputs (WB 2007). TheHondurasSocial Security Insitute (Instituto HondureEo SeguridadSocial - IHSS) 6.81 The 2001 PER noted that the IHSS faced serious financing and efficiency issues: The annual contribution per worker to the IHSS was US$51 in 1999, of which 33 percent was nominally for long term (pension) insurance or its Disability, Old Age, and Death Regime (Invalidez, Vejezy Muerte, IVM) and 67 percentwas for health insurance, known as the Illness- Maternity regime (Enfermedad-Maternidad, EM) (World Bank, 2001). Contributions were also the same, regardless of earnings and the contribution ceiling was frozen for 40 years. Thus the IHSS health system had dealt with rising costs and the declining value of its real revenues through deteriorating service quality, which caused insured workers to increasingly use both MOH and private sector facilities. IHSS hospitals were criticized for being generally overstaffed and inefficientlyrun. 6.82 In 2001 the Social Security Law was amended, increasing the maximum possible contribution by raising the maximum salary level to which the tax could be applied. In June 2001, the cap increasedto 2,400 Lps, and additional hikes were made, up to 4,800 Lps inJune of 2003. The process of organizationally and operationally restructuring IHSS in order to separate the accounting and financing functions ofthe pensions and healthregimes has beeninitiated and i s still ongoing. A third regime was also added inJuly, 2005 to cover professional risks with the objective of covering the costs of accidents and labor-related illnesses that usedto be covered in the illness and maternal regime. 6.83 Inaddition, the revised law called for the Government to pay its debt to IHSS because it was in arrears on its 0.5 percent of payroll contribution to each o f the Regimes. In 2003 and 2004, the Government allocated 6.25 million lempiras to be paid to the Institute as part of this debt-repaymentprogram (Fiedler 2004). 6.84 Analysis of IHSS Performance. From 1999 to 2006, IHSS realized about a 65 percent increase in the number o f its affiliates (from 362,346 in 1999 to 599,068 in 2006) and a 50 percent increase in its total number of beneficiaries ( from 800,785 in 1999 to 1.2 million in 2006). The significant increase in beneficiaries has also resulted in a substantial increase in ambulatory consultations anddischargesfrom 1999to 2005. 6.85 From 1999 to 2005, other indicators such as discharges per affiliate and beneficiary, as well as number of consultations per affiliate and per beneficiary have also improved. On the other hand, while the absolute number of beds increased by 13 percent, the number of beds per 1,000 persons decreased by 23.6 percent. However, this number does not take into account the beds inthe facilities subcontractedby IHSS. 6.86 As shown inTable 6.29, IHSS hospital occupancy rates were relatively low from 1999to 2001, averaging 64 percent. After 2002, this indicator has steadily improvedreaching a high of 81 percent in2005. There have also beenimprovements inother key indicators such as number of medical attentions per affiliate and beneficiary and number of surgeries performed. 116 Table 6.29 M S S productionindicators, 1999 to 2005 1999 2001 2005 YOIncrease YOincrease 1999-2001 2001-2005 Number of Affiliates 362,346 352,950 525,845 -2.6 49 Number of beneficiaries 800,785 780,020 1,162,118 -2.6 49 Number of beds 437 424 495 -2.9 3.1 Beds per 1000 0.38 0.37 0.29 -2.6 -21.6 Discharges 47,164 57,346 97,086 21.6 11.1 Average length o f stay 3.3 2.9 3.2 -12.1 1.4 Occupancy rate 68 61 81 -10.3 5.8 Surgeries 10,607 11,493 24,500 8.4 16.3 Number o f births 13,725 12,332 14,620 -10.2 3.5 % ofbirths which are Cesareans 19.8 23.1 25.4 16.7 5.5 Attentions per affiliate 3.37 3.43 4.43 1.9 2.9 Attentions per beneficiary 1.05 1.07 1.38 1.9 19.6 Source: IHSS Statistics 6.87 Beneficiary assessments undertaken by the WB Health Reform Project also indicate that the quality o f IHSS services has improved over time. Inparticular the waiting time to be seen by a health staff and/or to purchase medicines at the IHSS pharmaceutical stores has been reduced. A medical appointments system has also been established to minimize the likelihood that patients get to "jump" waiting lists for surgeries and other medical consultations based on their personal connections. 6.88 However, certain issues still need to be addressed. For example, the share o f births comprised o f cesarean operations has increased to 25 percent in 2005, which is higher than the recommended rate o f 15 percent. Available statistics also show that the number o f stillbirths has also increased which could indicate issues inprenatal control and quality o f assistance provided during delivery. Moreover, the bed per 1000 persons ratio is low at 0.29 in 2005. Aside from increasing the number o f beds, the management o f beds would need to be improved. According to the CHC findings, in some rooms, there are two women who share a bed while there are other wards with empty beds. Also, about 70-80 percent o f IHSS discharges are gynecological- obstetric in nature and this should be taken into account in the allocation o f rooms, beds, and investments inequipment. 6.89 Costs o f selected services in IHSS hospitals compared to MOH hospitals. A comparative analysis o f costs of the two major hospitals of the MOH ( Medical School or Hospital Escuela and El Mario Carino Rivas) and IHSS (Specialidadesin Tegucigalpa and Hospital de Regional Del Norte in San Pedro Sula) indicate that the former have lower costs per bed and per discharge compared to the latter. Inparticular, Figure 6.7 shows that IHSS' Hospital de Especialidades has a cost per bed that i s five times more than that o f the Medical School (both are comparable national level specialized hospitals) while IHSS' Regional Hospital o f the North has a cost per bed that is 6 times more than the comparable MOH hospital (El Mario Carino Rivas). 117 Figure 6.7 Cost per bed for MOHand IHSS hospitals Hospitales Source: MOHand IHSS 6.90 On the other hand, IHSS hospitals have three to four times as many discharges per bed than the MOH hospitals. However each discharge costs more in IHSS than M O H hospitals; Figure6.8. Figure 6.8 Cost per Discharge for MOHand IHSS hospitals 2005 /- -. 20000 15000 10000 5000 0 Hospitales Sources: MOHand IHSS 6.91 Financial Viability o f the IHSS health insurance regime. In order to ensure the sustainability o f its health insurance regime, IHSS would need to increase its contributions and/or lower its costs As Figure 6.9 indicates, the regime has enjoyed a surplus from 2002 to 2004 after operating in a deficit situation from 1999 to 2001. However, the gap between revenues and costs has been narrowing and the surplus declined to only 5 percent in2005. 118 Figure6.9 Revenueand ExpenditureTrends of IHSS HealthInsuranceRegime 1,800,000,000.00I I 1,600,000,000.00 1,400,000,000.00 C 1,200,000,000.00 ::1,000,000,000.00 I Q Q +Revenue -a- Expenditure! m 800,000,000.00 3a 600,000,000.00 C E Q 400,000,000.00 200,000,000.00 0.00 -I 1999 2000 2001 2002 2003 2004 2005 year Source: IHSS statistics; Fiedler2004; and author's calculations 6.92 SANIGEST (2004) recommends increasing contributions to ensure the financial viability o f the health insurance regime. In moving forward, it would be important to determine exactly the content and coverage o f the guaranteed health insurance plan, including the policies o f the Institute, the structure o f the supply o f services to be offered, the structure o f demand based on age, sex and geographic distribution of the beneficiaries together with registered morbidity and mortality rates, and the sustainability o f the financing With regard to coverage, it is critical to detail any exclusions, and how highly complex cases and/or prolonged treatments with low incidence will be financed because these have a large cost impact on the institution. 6.93 Otherways to decrease the costs o f service delivery could be: (a) Focusing on contracting out services instead o f continuing to build its own infrastructure. This would need to be accompanied by improvements in the capacity o f the MOH to monitor service providers, as well as its capacity to accelerate the process o f licensing andaccrediting privateproviders. (b) Strategically invest in the equipment that is needed to address majority o f the types o f cases handledwhich tend to be primarily surgeries o f low complexity (c) Improving rationale use o f surgical rooms and surgical services because the hours assigned to each do not reflect needs and actual use. The percentage o f cancelled surgical services amount to about 36 percent per year indicating organizational and personnel issues, as well as insufficient pre-operation preparationo f patients (CHC 2006). (d) Enhancing its procurement system. Drugs are reported to be bought at high prices. According to a November 2006 interview with the IHSS director and his team, IHSS has requested technical assistance from UNDP to improve and reorganize its procurement management system. (e) Improving its referral and counter-referral system to ensure that beneficiaries use hospitals for more complex services insteado f bypassing ambulatory facilities. 119 (f) Investing in preventive rather than focusing mainly on curative services in order to be more cost effective. IHSS has already developed a basic package o f preventive andhealth promotion services for families and i s inthe process o f hiringa consultant firm to assist it inits implementation. (g) Separating functions o f pensions, health insurance, and services provision. If undertaken in Honduras this would allow the IHSS to procure services from the MOH and from private sector suppliers, as well as from the existing IHSS hospitals, which would be placed under a separate administration from the health insurance fund. The latter would then be able to choose its suppliers on a competitive basis, in order to get the best value for the money for the insured population. This should eventually lead to a rationalization o f the hospital sector, where at present there is a wasteful overlap between the IHSS and MOH, especially in the major cities (WBb 2006). The new government i s considering promoting this initiative. (h) According to IHSS management, private providers do not always deliver services o fhigh quality. In order to get value for money, supervision needs to be improved to ensure a certain level o f performance. M O H would also need to improve its capacity in order to accelerate the licensing and accreditation o f facilities. Performance based contracts could also be introduced. (0 Finally although its information system i s reported to be working well in Tegucigalpa, delays inexpanding the systematization work in other areas, have made it difficult to keep track o f beneficiaries and ensure that they use the system properly. There is therefore a need to accelerate systematization and also to have a more integrated system to verify information across facilities. Systematic data collection andassurance o f consistency would also be important. F. Quality of Services 6.94 As part of the ongoing health reforms, the M O H has been implementing a licensing and accreditation process to ensure that a certain minimumstandardo f services is provided by health facilities. As o f 2006, it has only officially approved licenses for 162 health centers, indicating that about 88 percent o f the facilities did not qualify for accreditation. The M O H has also provided licenses to two public hospitals and 8 private ones. Most o f the main problems encountered during the licensing process were related to lack o f norms and equipment, followed by human resources issues, and then inadequate infrastructure (Direcci6n General Regulacih Sanitaria, SdS). 6.95 The growth and quality o f outpatient service delivery has been significantly affected by the Ministry's human resource management issues. Approximately 30 percent o f ambulatory facilities are closed due to personnel shortages (MOH 2007 evaluation workshop report). Moreover a study undertaken by the WB Health Sector Reform Project on 100 health centers covering 81,137 persons in 288 communities indicates that 72 o f these facilities have not been operational. The other 28 centers report staffing issues ranging from personnel who are on the payroll but do not work inthe facility, persons who work but do not get paid, insufficient and/or untrained staff, and having to depend on social service workers due to lack o f regular staff. Many o fthese facilities are registered as M O H facilities althoughthey have been constructed by other public institutions such as FHIS, communities or NGOs without MOH approval. This situation further emphasizes the importance o f strengthening the MOH's stewardship role to 120 ensure quality of services, together with the need for additional resources to support the facilities that are unable to comply with quality standards because o f lack o f personnel and/or sufficiently trained staff andequipment. 6.96 The results of a 2006 survey o f 100 users o f 10 health public primary care units on quality of service delivery indicate that public health facilities are providing acceptable services, but that they can still be improved. As Table 6.30 shows, while 93 percent o fthe surveyed users were satisfied with the service they received, only 75 percent stated that they would go back to the same health center. Of the 25 percent who would go to another facility for their next treatment, 15 percent of them would go to another public facility while the remaining 10 percent would go to a private one. 6.97 Sixty one percent o f the users had to wait for more than an hour to be seen by a health staff. Almost all of the users reported that they received explanations about their diagnosis and 97 percent received instructions on how to use their prescribed medicines. Also, 99 percent o f the users state that the health units' staff treated them courteously. The survey, however, reveals that only one o fthe healthunits (10 percent) showed the services price list to the users. 6.98 Despite the efforts made by the M O H to improve staff and medical supplies availability, the survey also indicates that shortages o f personnel and medicines still constrain the capacity o f primary health centers to provide high quality services. For example, only 30 percent o f health units are open 7 days a week, 50 percent operate withjust one shift (7am to 3:3Opm) while 20 percent close their units during lunch time. Moreover, 60 percent o f these health units have to close their health units when the staff needs to do community work. None o f the health units interviewed reported that they have 80-100 percent o f their required supply o f drugs while only sixty percent have between 60-79 percent o f their required stock andthe rest have between40 to 59 percent. Only 60 percent o f the health units have a nearby pharmacy. Only 61 percent o f the equipment in these health centers were functioning properly and none o f their vehicles were rated as working well by the staff. 6.99 Compared to the public health facilities, the surveyed decentralized units do not seem to face personnel problems, but they also face some drug shortages. All o f the 5 CESAMOs interviewed have their required physicians and all but one have at least two auxiliary nurses. However, 80 percent o f these health centers work only one shift, and 20 percent o f them are closed when the staff does community work, About 40 percent o f them have within 80 to100 percent o f their required medicines while 50 percent have 60 to79 percent o f their required medicine stock. Only 16.3 percent o f users had to wait more then three hours compared to 28 percent o f users o f centralized facilities. In addition, 98 percent o f users said that the staff explained to them how they should take medicines, and 92 percent o f the users could get the prescribed medicines at the health centers. About 83 percent o f the equipment in the decentralized units was rated as working well. The decentralized units also have more vehicles and they also have fewer vehicles with functional problems. About 92 percent o f users reported that they are satisfied with the quality o f services, and 83 percent indicated that they would return to these health centers compared with 75 percent o f the users o f centralized facilities (WB BNPP study 2007). 121 Table 6.30 Selected Characteristics a ators by Provide] rype (2006) Variable 11thUnit (HU) Decentralized TOTAL -No. Y O No. YO No. Y O Number of days Per week HUis oDen 5 days 7 70.0 1 10.0 8 40.0 6 days 0 0.0 6 60.0 6 30.0 7 days 3 30.0 3 30.0 6 30.0 TOTAL -10 100.0 10 100.0 20 00.0 Office Hours for External Consults 7:OO AM a 4:OO PM 1 10.0 0 0.0 1 5.0 7.00 AM a 3:30 PM 5 50.0 0 0.0 5 25.0 8:OO Ah4 a 4:OO PM 2 20.0 5 50.0 7 35.0 7:30 AM a 4:OO PM 0 0.0 1 10.0 1 5.0 7:30 AM a 3:30 PM 0 0.0 1 10.0 1 5.0 All 24 hours 2 20.0 3 30.0 5 25.0 TOTAL -10 100.0 10 100.0 20 00.0 Do thev close to do community work? Yes 6 60.0 2 20.0 8 40.0 No 4 40.0 8 80.0 12 60.0 TOTAL 10 100.0 10 100.0 20 00.0 Percentage of existing medicines in the HU 40 -59 60 -79 4 40.0 1 10.0 5 25.0 80 - 100 6 60.0 5 50.0 11 45.0 TOTAL 0 00.0 4 40.0 4 20.0 10 100.0 10 100.0 20 00.0 Were the medicines obtained in the HU? Yes, all 62 68.1 76 91.6 138 79.3 Yes, some 26 28.6 7 8.4 33 19.0 No 3 3.3 0 0.0 3 1.7 TOTAL -91 100.0 83 100.0 174 00.0 Waitinp time to receive service? Less than 1hour 39 39.0 49 50.0 88 44.4 1a 2 hours 33 33.0 33 33.7 66 33.3 3 hours or more 28 28.0 16 16.3 44 22.2 TOTAL -100 100.0 98 100.0 198 100.0 Was the staff courteous? Yes 99 99.9 98 98.0 197 99.5 No 1 1.o 0 0.0 1 0.5 TOTAL -100 100.0 98 100.0 198 100.0 Was the user satisfied with the service received? Yes 93 93.O 93 94.9 186 93.9 Regular 6 6.0 5 5.1 11 5.6 No 1 1.o 0 0.0 1 0.5 TOTAL -100 100.0 98 100.0 198 100.0 Where will the user seek care next time? Tothe same healthHU 75 75.0 81 82.7 156 78.8 To a private facility 10 10.0 4 4.1 14 7.1 To anotherpublicfacility 15 15.0 13 13.3 28 14.4 IndifferentI no preference 0 0.0 0.0 0 0.0 TOTAL -100 100.0 98 100.0 198 100.0 Source: WB-BNPP survey 2007. 122 6.100 While these models have been demonstrating positive results and the M O H intends to expand services to 50 additional municipalities, they are also being increasingly met with resistanceby unions who perceive them as precursorsof "privatization". G. InstitutionalandGovernanceIssues ResourceAllocation and Program Implementation 6.101 The MOH's reorganization of Regional Health Departments (RHDs) in 2004 and its implementation of the new budget allocation criteria to RHDs since 2005 (Annex H discusses these aspects in more detail) have resulted in important challenges for the RHDs for the following reasons: (a) The new allocation criteria reduced the weight given to population size and RHDs state that their resources have been seriously diminished. As a result, some o f them are "resisting" the new allocation criteria ((Orstrom et. a12005). (b) Although the MOH is trying to correct the serious imbalance in distribution of personnel across the regions, the current Estatuto Medico allows physicians to move with their posts to another RHD. This hinders the MOH's control of personnel deployment andto some extent, weakens the links betweenplanning, implementation, andresults. (c) Budget allocations to the RHDs do not include funds from external aid, yet 39 percent of the resources fkom program 11 (ambulatory care and health promotion) come from this source of funds. This could generate potential imbalances among the RHDs if there are important differences between the donors' and MOH allocation criteria. Sustainability issues could arise ifthis TA were requiredafter the external aid is finished because the RHDswould have to finance itwith their own funds. 6.102 The RHDsalso faces some challenges inallocating and implementing their funds due to gaps or inconsistenciesinthe decentralization policy. For example, there i s a lack of regulation of the decentralized funds. Health facilities can receive transfers from municipalities to complement M O H funding and charge user fees. The volatility of municipal funding couldbe preventedifinstitutional arrangementsare introduced that ensure a regular source of funds for the health units. Inthe case of user fees, although the RHDs are entitledto set the fee levels, they do not use clear and systematically applied criteria (Fiedler et. a12000). 6.103 RHDsoperational responsibilities are also restrictedto carry out supervisiontasks, and they only have very limiteddiscretional authority to reallocate funds thereby limitingthe benefits ofpromoting aprogram-basedbudgeting approach. Moreover, bothM O HandRHDsalso have limitedinstitutional capacity to planarealistic budget; resulting intheir approved budgetsbeing subject to many modifications during the fiscal year. Also, the RHDs supervisiontasks are constraineddue to central laws such as the Estatuto Me`dicowhich limit their authority to hire and fire physicians. Also, given the inefficiencies inthe distribution of medicines by the central level as discussedin section E o fthis chapter, the RHDscannot be heldfully accountable for not providingthe services programmedinthe operationsplansor the quality o ftheir services. 123 H. Preparingthe HealthSector ExpenditureFramework 6.104 The MOH recently finalized its National Plan 2021 and established its institutional goals that remain to be costed. InMarch 2007, the Government agreed to a WB-UNICEF proposal to pilot in the Marginal Budgeting for Bottlenecks Model (MBBM), a results-basedplanning tool that serves to (i)guide a country-based process of identification of bottlenecks to successfully deliver high impact interventions, (ii) calculate the incremental cost of removing systemic and intervention-specific bottlenecks to achieve effective intervention coverage, and (iii) estimate the potential impact of health interventions on health outcomes, such as rates of child and maternal mortality, HIV/AIDS, nutrition, and malaria-associated mortality (an issue in a few regional departments inHonduras). Ifadditional resourcesare not available or insufficient, itwill provide options for improving sectoral efficiency. The MBBM framework is plannedto be operational intime for thepreparationofthe 2008 budget. 6.105 Annex J provides a detailed description ofthe MBBMframework and the next steps to be followed indeveloping the MTEF and2008 budget. I.ConclusionsandStrategicImplications 6.106 Performance of Kev Indicators. Honduras continues to demonstrate progress in most national level health, nutrition, and population indicators. It has succeeded in significantly reducing neonatal mortality which had been stagnant for almost two decades, infant and child mortality, and, to some extent, chronic malnutrition. Based on the DHS 2005/06 results and assumingthat the health sector is able to maintain its recentrate of progress and invests adequate resources on primary health care then it is highly likely that it will reach the MDG and Government nutrition and child mortality goals Nationally representative information would need to be updated to determine if Honduras is still on track in achieving the MDG goal for maternal mortality. However, the number of births attended by skilled health professionals (a proxy for reduction in MMR) i s only 67 percent. Without significant changes in addressing supply and demand side factors, it may be difficult to reach the target of 100 percent by 2015, mainly because only 50 percent o f women inrural areas avail ofthis type of health service. 6.107 Significant gaps remain in terms of access and outcomes betweenrural and urban areas, regions, and poor and wealthy segments of the population. For example while the percentage of stunted children who belong to households in the lowest income quintile has declined from 50 percent in 2001 to 43 percent in 2005, this number is still very high and has repercussions on poor children's health andfunctional capabilities. 6.108 Coverage of Services. Physical coverage of services has improved although service providers largely continue to function independentlyfrom each other, resulting in fragmentation and duplication of services - although some changes are taking place. Both MOH and IHSS have been increasingly contracting out services to private providers and IHSS has also been formally contracting some MOH services. Nonetheless more formal coordination i s needed and cost recovery mechanisms should be clearly establishedbetweenM O H and IHSS. Coordination would improve efficiency, and appropriate cost recovery would ensure that both institutions will be adequately reimbursed for services provided. This is especially important for the MOH because of its lower overall level of financing per capita-even though it has a higher coverage rate-compared with IHSS andto avoid cross-subsidizing the non-poor. 124 6.109 There i s also a lack of systematically updated and consistent information about the private sector. The last formal private facility census undertaken was in 2002 and this only included limited information. More effort should be undertakento regularly and systematically gather information about the number, coverage, and basic production statistics of the private sector. 6.110 ExpenditureEfficiency. Health spending increased inreal terms from 1999to 2005, but remains low relative to LAC and CA average spendinglevels. While it would helpto have more fiscal space that would allow increased investments in health, there i s also much room for improving efficiency by (a) channeling more funds to preventive andhealth promotionactivities, and(b) improvingthe managementofhealthpersonnelandpharmaceuticals. 6.111 Curative health care bias. Hospitals continue to have the largest share of total M O H expenditures although their share decreased slightly to 38 percent in 2005. However, primary health care's share declined by 50 percent, comprising only about 16 percent of total M O H expendituresin2005. Reduction in some key indicators (full immunizationrates and exclusive breastfeeding of young children) further underscores the need for Honduras to allocate more resources to preventive and health and nutrition education and promotion services; otherwise it might risk losing the recent gains it has made inreducing infant and childmortality rates, as well as malnutrition. It i s also advised to channel more efforts towards meeting the unmet family planning needs of 18 percent o f Honduran women and in reducing the further spread of HIV/AIDS. 6.112 To improve efficiency inthe use of resources invested in hospitals the current system of individual purchases per facility should be improved. The M O H needs to establish and implement national standards across all facility levels (including guidelines on preventive maintenance and repairs) and a coordinated procurement system in order to benefit from lower prices. The MOH should also identifyways to improve the efficient use of underutilized rooms and equipment which tend to be used mainly in the mornings because majority of health personnel work during this period. A detailed study o f hospital service delivery and costs i s highly recommendedto identify strengths, opportunities, andkey areas for reform. 6.113 Personnel Management. A considerable portion of the budget remains devoted to personnel (about 54 percent of total MOH expenditures and 67 percent of M O H hospital expenditures). The maintenance of relative compensation levels that are non-strategic in nature, and that have beengoverned mainly by the Estatuto Medico has resultedinhighpersonnel costs, with doctors being paid disproportionate amounts relative to other health staff. This has created disincentives, especially in the nursing profession, leading to a low nurse to doctor ratio. The Estatuto Medico also limits the ability of the MOH and RHDs to hire and fire health staff, especially doctors. Health personnel deployment traditionally has not been based on existing needs. Health personnel, particularly doctors remain inequitably distributedacross regions. 6.114 While it may not be realistic to expect changes in salary levels because of union strength andpotential politicaleconomic repercussions,the MOHcantake concrete steps to (i) the update mapping of healthworkers that was done inJanuary 2006, relative to where they are supposedto be working and receiving their salaries and address the discrepancies, (ii) improve the regional allocation of doctors across the country, (iii) implementstrategiesto train andretain more nurses and ensure that they are also well-distributed across regions, (iv) provide incentives for health staff to work in rural and remote areas because these are currently not being offered, and (v) 125 explore the possibility o f hiring future personnel based on performance based, fixed-term contracts instead of continuing to hire health personnel on permanent contracts that make it difficult to apply sanctions for poor performance. 6.115 Pharmaceuticalsprocurement. While the government has taken steps in2006 to improve the national-level procurement o f pharmaceuticals through the Inter-Institutional Commission (IIC), it is important to monitor the implementation o f the following key I I C recommendations to see whether they do result in actual availability and lower prices o f drugs inhealth facilities: (i) major organizational changes in the MOH, (ii)subjecting MOH operations to audits by both public and private sector agencies, (iii) punishing M O H staff who hinder or do not comply with the procurement procedures.; (iv) upgrading the MOH's central storage with external aid; (v) updating the Basic Medicines Profile every 3 years according to international standards, and (vi) empowering the multi-sectoral National Health Council (Consejode Salud) to provide technical oversight. Also, Honduras i s encouraged to examine the reforms inGuatemala and Ecuador that allow local decentralized health facilities to determine the amount o f pharmaceuticals they need and purchase them directly from providers, based on lower prices the MOH has negotiated centrally. More assessment and policy work will be needed inthis area. 6.116 Productivity and Ouality o f Services. Despite some improvements inproductivity, such as increases in ambulatory consultations and bed occupancy rates, certain productivity indicators such as average number o f daily consultations/medical attention provided remain below standards. Quality o f services in terms o f availability o f adequate staff, equipment and medicines also remains an issue, with only 12 percent o f health centers and 8 hospitals being accredited. Thirty percent o f ambulatory facilities remain closed due to lack o f personnel. 6.117 A strategic expansion o f services i s needed to reachunderserved communities. To ensure adequate quality and efficiency in the use o f resources, however these communities need to strategically address the nature o f the mechanisms and the timing o f implementation based on need and availability o f resources. Criteria exist for determining the location and need for a health facility but sometimes political considerations over-ride them. In moving forward it is recommended that: (a) priority should be given to staffing and equipping the existing facilities that remain unaccredited, but are needed instead o f continuing to construct new facilities and (b) the Government consider having NGOs or communities manage a number o f the existing nonfunctional facilities, similar to what was done ina pilot involving seven decentralized models in 273 communities. Available evidence regarding contracting out of services to NGOs and communities indicate that this strategy seems to be working well. Surveys indicate an increase in institutional births and beneficiary satisfaction with services and that these facilities are better equipped and staffed than centralized facilities. The MOH plans to expand these types o f services to 50 additional municipalities. However, there is a growing resistance against these decentralized models by unions who equate them with "privatization" and as a possible threat. Encouraging beneficiary communities to express their views about the benefits o f these facilities could help counteract this opposition. The Ministry could also consider having citizen type report cards which rate and publicize the performance indicators and service satisfaction for health facilities inorder to address quality issues. 6.118 Financial Protection. Access to funds for treatment continues to be an issue. About 20.7 percent o f persons who reported they were sick but who did not seek treatment cited lack of funds as the bindingconstraint while 58 percent o f interviewed women cited financing as one of their mainproblems in seeking care. Only 7 percent o f households had some form of insurance 126 to pay for their health care costs during their last consultation, while 8 percent benefited from exemptions. The coverage of IHSS is concentrated disproportionately among households in higher income quintiles and geographically in the capital cities and urban areas, primarily because it coverage to formal workers. Honduras has expressed interest in the case o f Ecuador where the Government i s launching the Aseguramiento Universal en Salud (AUS). This initiative aims to increase coverage to the poorest 40 percent who are not covered by any other type o f insurance, by coordinating through existing social health insurance schemes, as well as introducing a publicly subsidized health insurance called Programa de Aseguramiento Universal en Salud. In Honduras, a possible first step towards starting a social health insurance scheme could be linking contracts with providers to expand health service provision andthe IHSS. 6.119 The performance o f the Honduran Social Security Institute (IHSS) has improved as a result o f the reforms but it continues to operate at high cost. It would need to lower its costs and increase its revenues in order to ensure the financial sustainability o f the Illness and Maternal Regime. It would be important to determine the type and coverage o f the guaranteed health insurance plan, including the structure o f the supply o f services to be offered, the demand structure based on age, sex and geographic distribution o f the beneficiaries together with registered morbidity and mortality rates, and the sustainability o f financing. Coverage policies should define exclusions, and how highly complex cases and/or prolonged treatments with low incidence will be financed because these have large cost implications. Other ways to decrease the costs o f service delivery couldbe: a. Focusing on contracting out services instead o f continuing to build own infrastructure. This would needto be accompanied by improvements inthe capacity o fthe MOHto monitor service providers. b. Strategically investing mainly in equipment needed to address the majority o f the types o f cases handled, which tend to be primarily surgeries o f low complexity. Improving rational use o f surgical rooms and surgical services because the hours assigned to each do not reflect needs and actual use. c. Reviewing and improving its procurement system because drugs are reported to be bought at highprices. IHSS has requested UNDP assistance with drug procurement and the results o f this action need to be reviewed to see if it translates into lower prices for drugpurchases. d. Improving its referral and counter-referral system to ensure that beneficiaries use hospitals for more complex services instead o f bypassing ambulatory facilities and that people utilize the level o f service appropriate to their needs. e. Investing in preventive care rather than focusing mainly on curative services in order to be more cost effective. IHSS has already developed a basic package o f preventive and health promotion services for families and is inthe process o f hiring a consultant firm to assist it inits implementation. f. Separating the functions ofpensions administration, healthinsurance, and services provision. If undertaken, this would allow the IHSS to procure services from the MOHand from private sector suppliers, as well as from the existing IHSS hospitals, which would be placed under a separate administration from the health insurance fund. Suppliers can then be chosen on a competitive basis, in order to get the best 127 value for the money for the insured population. This should eventually lead to a rationalization of the hospital sector, where at present there is a wasteful overlap betweenthe IHSS and MOH, especially in the major cities (WB 2007). The new government is considering promoting this initiative. g. Accelerate systematization of information and ensure a well integrated system to verify information across facilities. This would help to keep track of whether beneficiaries are using the system properly. Systematic data collection and monitoringwould be important to keeptrack of progress. 6.120 ResourceAllocation to the Regions. In2005, the MOH introduced newbudget allocation criteria to promote a more equitable resource allocation. First, while this is a welcome step, its implementation process and results needto be evaluated. For example, it would be important to check if the weights given to population need to be increased based on feedback from more populated Regional Health Departments in case that the lower weight given to population has had a negative impact on their budgets. Second, donor funds especially technical assistanceneed to be quantified and accounted for in RHD budgets. Third, introducing institutional arrangements that ensure a regular source of funds for the health units could minimize the volatility of municipal funding availability. Inthe case of user fees, RHDs needto establish clear and systematically applied criteria to ensure optimal fee levels based on the particularities of each health unit's situation. Doing this would not only improve the regularity o f this source of funds, butalso preventpotential inequities inthe population's access to financial protection. 6.121 As part of the overall decentralization process and the national move towards program based budgeting, it is important to assess whether RHDs operational responsibilities can be expanded together with their discretional authority to reallocate funds. Their capacity to plan, budget and implement also needs to be strengthened. The MOH's capacity to perform its stewardshiprole and follow-up and support sectoralreforms needs to be enhanced. 6.122 Preuaring a Medium-TermExpenditure Framework. The M O H has recently finalized its NationalPlan 2021 and establishedits institutional goals that remain to be costed out. InMarch 2007, the Government agreedto a WB-UNICEF proposal to pilot inthe Marginal Budgeting for Bottlenecks Model (MBBM), a results-based planning tool that serves to (i) guide a country- based process of identification of bottlenecks to successfully deliver high impact interventions, (ii)calculate the incremental cost of removing systemic and intervention-specific bottlenecks to achieve effective intervention coverage, and (iii)estimate the potential impact of health interventions on health outcomes, such as rates of child and maternal mortality, HIVIAIDS, nutrition, and malaria-associatedmortality (an issue ina few regional departments inHonduras). If additional resources are not available or insufficient, it will provide options for improving sectoral efficiency. The MBBM framework is planned to be operational in time for the preparation of the 2008 budget. 6.123 Recommended areas of future study are: (a) an expenditure tracking survey from the central level to health care units, (b) an assessment of progress and lessons learnt regarding the implementation of the 2005 established resource allocation criteria to RHDs (this could be a subset of (a) or a separate study), and (c) a detailed study of hospital costs, efficiency, and quality of services. 128 CHAPTERVI1 REVIEW OF PUBLIC SECTOR SPENDINGONTRANSPORT 7.1 Transport has historically been one of the most important sectors in the national economy, representing one of the largest public investment portfolios. It is also one of the areas where the private sector plays a major and increasing role. The Ministry of Public Works, Transport, and Housing (SOPTRAVI) i s the central government entity responsible for coordinating and directing the sector's managementand investment policies, and simultaneously i s represented on the boards of directors of various organizations involved intransportation, such as the National Port Enterprise (ENP) and the Honduran National Railways Enterprise (FNH). Moreover, it serves as a contracting party in the concessioning to the private enterprise InterAirports S.A. (IASA), which has been managing the country's four international airports since 2000, and also chairs the Technical Committee that directs the activities and programs of the RoadFund. Therehave beensome recentreforms inthe institutional arrangementsfor sector regulation, notably the creation of the Superintendency of Concessions and Licenses. Its activities in the transport sector have however to date been restricted to the regulation of the airport concession. 7.2 Overall public expenditures in the sector, broken down by year for each sub sector, are summarizedinTable 7.1. Table 7.1 Summaryof TransportSector Expenditures,2002-06 (in Lps millions) Sub Sector 2002 2003 2004 2005 2006 Roads 792.9 1001.5 1071.5 1262.5 1113.2 RoadMtce. 483.5 650.7 587.3 737.1 669.1 Rail 31.6 35.1 33.1 13.2 15.1 Port 998.3 1080.5 1391.9 1408.1 1237.9 Air 71.4 75.4 79.1 81.3 81.3 Total 2377.7 2843.2 3162.9 3502.2 3116.6 Source: SOPTRAVI, RoadFund, FNH,ENP, Superintendencyof ConcessionsandLicenses. 7.3 Roads and ports account for more than 95 percent of public expenditure intransport. The share of road sector expenditure allocated to recurrent (maintenance) has been within the range 32-37 percent with no discernable trend in the periodsg2 As a percentage of GDP, transport sector expenditures have averaged 2.2 to 2.3 percent during this period, which i s relatively low by regional standards. This also compares unfavorably to the average of about 2.9 percent of GDP that prevailed inthe 1990s. 7.4 At the macro level, we can identify severalproblem areas inthe sector, each generating a series of inefficiencies: 0 The lack of uniform legal rules covering all activities in the sector and serving as an organizing and regulatory framework for all sub sectors. 92. In 1999maintenance expenditureswere about 53 percentoftotal 129 The absence of a single body or entity with sufficient capacity to take responsibility for performing the technical work of multisectoral planning at the Central Government level. The absence of an entity in the road sector with the technical capacity to carry out sector planning tasks. (As a result, each mode of transport and even each office responsible for any given financing reaches its own decisions, doing so from a short-term perspective rather than inresponseto problems ofgreater scope andcomplexity atthe central level.) 0 Informationon the sector is scattered about and lacks homogeneity. Information is available from many sources, but there is a lack of reliable statistical data in critical areas such as safety (there i s little data on accident incidence) and, more generally, there is no clear overview ofthe capacity andcurrent condition of the transport system. 7.5 The preceding inefficiencies limit the management capacity of SOPTRAVI, which i s reflected in the institution's poor budget execution performance. Over the last six years, SOPTRAVI only has managed to execute on average 73 percent of its total approved budgetary funds; Figure 7.1. Figure 7.1 Budget Execution Ratio in SOPTRAVI 3,000 .- 2,500 E 2,000 Q) -I IC 0 1,500 -.-5 u) 1,000 _I 500 0 2000 2001 2002 2003 2004 2005 2006 Source: SOPTRAVI-SEFIN. 7.6 As of 2006, budget execution is being measured in accordance with an innovative appraisal technique known as "results-based management" used by the Technical Support Unit (UNAT) assigned to the Ministry of the Presidency. SOPTRAVI received a 65 percent rating, which places it in the "regular" execution category (or yellow light range under the "traffic light" methodology introduced in the G~vernment.'~This appraisal method also takes into account the scope of the goals pursued and the size of investment of each state agency. The UNAT Director himself confirmed that there is insufficient managerial capacity in the various ministries. However, because of the change of government that occurred in late January 2006, the official budget was not approved until mid-year, which induced a considerable lag and slowed down activity during the months in the season that are generally the most active. As a 93. Below 55 percent, execution is ranked as poor (red), while above 80 percent it is ranked as good (green). 130 result of this record, SOPTRAVI is having its budget cut by 15 percent, in2007, relative to 2006, although the Road Fundbudget is being increasedby 10percent. 7.7 There is thus an opportunity for substantial progress to be made in regard to managing SOPTRAVI operations and to capitalizing on any prospects there are to improve efficiencies. During the early 2000s, there has been some attention devoted to organizational restructuring - including the creation of new structures for planning and programming (UPEG) and for social and environment management (UGA). This has been backed by investments in equipment, information systems and capacity building. These efforts now need to be supplemented by actions to develop and sustain a core staff with competences inthe areas of policy and regulation to address the evident gaps in these areasag4The other key consideration will be developing efficiency and productivity indicators so as to provide an objective basis for assessing SOPTRAVI's future performance. These measures should be disclosed through the SOPTRAVI web site. The indicators would cover inter alia the performance of SOPTRAVI inrelation to its contract management functions which are likely to remain of significant importance in the road sector inthe future. 7.8 Current estimates (most likely scenario) of sector expenditure through 2010 are shown in Table 7.2. The forecast calls for expenditures equivalent to not less than 2.7 percent of GDP. For roads, the balance of expenditures moves progressively in favor of maintenance, 39 percent of total by 2010. O f these total expenditures, roughly 22 percent are projected as recurrent as against 78 percent for capital." Table 7.2. ProjectedPublicSector Expenditures inthe TransportSector (in Lps millions) Sub Sector 2007 2008 2009 2010 Roads 3294.6 3123.6 2259.1 1981.7 RoadMaintenance 1193.2 1271.1 1369.9 1476.3 Railways 11.4 11.4 11.4 13.3 Ports 1235.0 1368.0 1406.0 1558.0 Airways 38.0 47.5 57.0 74.1 Total 5772.2 5821.6 5103.4 5103.4 Note: Projectedexchangerate is US$1to Lps. 19 Source: Bank estimates, Soptravi, ENP, FNH 7.9 Investment expenditures in the road sector are initially very high due to lumpy expenditures associated with the "corredor logistico" upgrading. Expenditures could be as high inthe outer years iffinance canbemade available inatimelymannerto fundother economically viable projects. This will also be dependent on the success with which implementation capacity andmanagementefficiency inSOPTRAVI canberaised over the coming years, regardless ofthe contracting out of works and of most services. In the ports sector there are likewise viable investmentsthat could be implemented earlier if the private financing can be mobilized. 96 The projection for rail investment is predicated on the continuity of the current restricted network operation. However as the report has recommended shutting down the system and opening up 94. To notethat even ifthere were to be createda separate regulatory agency, it will be necessary to sustain policy andregulatory oversightcapacity within the ministry 95. With maintenanceworks includedunder capital 96. This forecast only includesabout $100 million equivalent of investment spending 131 the possibility of future investment in this mode by the private sector, projected investments could differ from this forecast. 7.10 Inthe sections that follow, each sub sector will be examined separately interms of its recent capital and operating expenditures, its problem areas, andpossible steps to be taken with a view toward improving future performance. A. The RoadsSub-sector Overview 7.11 The official road network of Honduras is under the responsibility of SOPTRAVI and the Road Fund. It is 13,603 kilometers in length, of which 2,777 are paved (20 percent) and 10,826 unpaved(80 percent). Other State entities, such as the State Forestry Agency AFE-Cohdefor, the Honduran Coffee Institute, andthe Honduran Telecommunications Enterprise HONDUTEL, also build or maintain roads, but these are not consideredpart of the official network managedby the Central Government because they meet special needs relevant only to the entity concerned or were not built inaccordance withthe standardsand specifications usedby SOPTRAVI. 7.12 Each year about 13 million metric tons of land cargo are transported in Honduras, of which over 60 percent i s domestic inorigin anduse. O fthe cargo carried, containers account for 22 percent and fuel and petroleum derivates 15 percent. Each day there are about 105,000 intercity trips in light-duty vehicles, moving roughly 280,000 persons and consuming approximately 30 million gallons o f fuel per year. Urbantransport consumes about 85 percent o f the fuel usedinthe sector. Generally, there i s a shortage of adequate road corridors for existing traffic levels. The road network is characterized by slow speeds and increasing deterioration owing to the lack of timely and sustainedmaintenance. 7.13 Inrecentyears, investments on roads have beenfinanced mainly with external resources (70 percent). The most important funding agencies have beenthe World Bank, IADB, CABEI, MCC, USAID, and the Nordic Development Fund. The main bilateral donors active in this sector have been Taiwan, Sweden, Spain, Japan, and Mexico. Current and operating expenditures for the sub sector historically have been around 15 percent of the amount spent on works and projects (construction and consulting services). Efficiency Issues 7.14 Key issues andbottlenecks that meritattention inthe road sector are: The inadequate allocation of resources for road conservation, which makes it possible to attendto only about 65 percent o fthe network that requiremaintenanceeachyear. 0 Cost overruns on works execution in general. Examination of a sample of the projects carried out in the past 15 years indicates that cost overruns generally tend to exceed 30 percent of the initial value. These cost overruns are mainly attributable to (i) excessively an high level of unforeseen expenditures or additional tasks identified during works execution and not contemplated in the original design, (ii) changes from the original designs, radical 132 introduced by the supervisor concerned owing to differences o f opinion or because a greater and more detailed understanding o f the project is achieved once work begins at the site, (iii) limited coverage o f the original engineering designs, driven almost always by the lack o f sufficient resources and the extremely short period o f time allowed for preparing them; (iv) considerable lags between the dates of design preparation and o f work execution, (v) lack o f a detailed and far-reaching review of the studies by the owner o f the works, inturn reflecting the State's lack o f qualified specialists, and (vi) inefficient prequalification, leading to inappropriate final selection o f construction contractor^.^^ 0 The sample of contracts examined also indicates that road project execution times are, on average, extended by a factor o f two, that is, the project completion period tends to double in length, the equivalent o f average contract extensions o f 11.6 calendar months. 0 Incontrast, the foregoing problems are minimalinthe RoadFund, where the contracts have a more clearly delineated scope, term, and cost, which are not subject to alteration except in highly extraordinary circumstances. 7.15 Total road expenditures inthe period 2002-2005 have averaged $91.5 million equivalent per annum, or about 1.3 percent of GDP, low by regional standards. The road service index (kilometers per thousand population) is 0.45, compared to the Central American average o f 0.63 and the density index (kilometers per thousand square kilometers area) is 29 compared to the Central American average o f 55. These figures are indicative o f a relatively non-extensive network for an economy o f the size o f Honduras. However the quality o f the road network in Honduras is towards the upper end o f the scale for Central America - 40 percent o f all roads in good condition o f which 90 percent o fpaved roads are in good condition. 7.16 In terms of the cost of road maintenance and investment operations, Honduras is comparable, though generally a little higher, to its neighboring countries and to other countries o f a similar income level in the Latin America region. From the perspective of investment expenditure, much o f this can be related to the above-mentioned inefficiencies in the ways in which SOPTRAVI works and i s organized. The costs o f routine maintenance o f the paved network have been traditionally high in relation to regional comparators - average o f US$2,200 per kilometer per year o f which about US$1,360 i s accounted for by the costs o f the micro- enterprises. With the increasing scale o f activities and greater experience in the Road Fund - including increased efficiency inthe way in which works are supervised (see para. 8.24) - these costs are expected to come down over time. 97. A similar investigationwas conductedin 1997, coveringthe 1987-1997period, whenthe Master RoadPlanof 1989was beingupdated. Inthis instance, it was foundthat the increases for works amountedto 38 percenton average. The increase in the level of unforeseen expenditure was even more pronouncedfor the emergency works resulting from Hurricane Mitchin 1998, this because of the extreme speed with which the engineering studies and their corresponding support information and budgeting were prepared owing to the many unplannedtasks which arose during execution. The study also showed that, on average, approximately 11.3 percent of the amount designated for the complete work was added as Price Escalation or Adjustment, using the automatic escalatorclausemethodologiesappliedby SOPTRAVI. This ledto the findingthat the provision for this contract section might be 10 to 15 percent in the corresponding budgets. In some cases, expenditure under this headingwas as highas 17 percent, probably resulting from the considerableincrease inthe contract deadlines. 133 MultiyearPlanfor the2003-2006Period 7.17 Inaneffort to maintainresponsibleandadequate expenditurelevels, the Government had prepareda multiyear budget covering 2003-2006 that sought to reflect the likely availability of revenues, particularly tax revenues, and was based on a realistic assessment of its external borrowing capacity and international aid. This is shown in the first two columns of Table 7.3. The second set of columns in this Table show the total amount of resources required to adequately maintain and improve the network, as identifiedina 2002 Road Plan. Comparing the requiredinvestment to the State's limitedresources for the road sector yields a budget gap for the 2003-2006 period, which i s identifiedinthe last column o f Table 7.3. Desirable Budget Total Required Budget Gap Year Maint. Invest. Maint. Invest. Maint. Invest. Total 2003 565.7 904.1 717.7 1215.0 152.0 3 10.9 462.9 2004 551.6 873.9 712.9 1451.4 161.3 577.5 738.8 2005 579.5 917.5 766.4 2140.9 186.9 1223.4 1410.3 2006 608.5 963.3 866.7 1734.1 258.2 770.8 1029.0 Total 2305.3 3658.8 II3063.7 6541.4 II 758.4 2882.6 II3641.0 Note: Exchangerate assumedUS$1to Lps 17 (rate applying inJanuary 2003). Source: Soptravi,Multiyear Plan. 7.18 Although the multiyear program had sought to cover the entire network, the allocations finally grantedmanagedto cover only 65 percent of the network's needs, inparticular as regards the road conservation activities of the Road Fund. This led to a premature deterioration of at least 35 percent of the road system. Consequently, rational resource utilization had to be supported by careful prioritization of projects on the basis of impact indicators associated with factors such as the "Rural Population Served" in the case of the unpaved network, and "Economic Benefits" in the case of paved roads. The financial gap and the gap in coverage of the network are expected to be reduced gradually beginning in 2007, the first year of planned, agreedincreases inthe annual RoadFundallocation for periodic androutine maintenance. New Multiyear InvestmentPlanfor the 2005-2009 Period 7.19 In line with the vision emerging from the Poverty Reduction Strategy (PRS), the Government of Honduras structured a new short- and medium-termInvestmentPlan that can be monitored, adjusted, and reassessed periodically. Inthis context, the Road Investment Plan aims to improve the condition of the network, directly supporting production and production factors interrelated with competitiveness and enhanced efficiency. This objective clearly takes private sector participation into account. Furthermore, since the early 199Os, the conservation of the road network has been delegated in its entirety to private companies, with the aim o f achieving greater efficiency in the use of the scarce resources. The multiyear budgets prepared for 2005- 2009 takes into consideration not only the objectives of the PRS, but also the continuity and complementarity that must exist between the projects executed to date and the portfolio of new 134 efforts to be undertaken in the short and medium term. Accordingly, investment prioritization has beencarried out inlight ofthe overall available ceilingto the sector ineach fiscal year. 7.20 The new Plan was preparedby the LAVIAL and Regioplan firms in2004, inaccordance with prioritization procedures consistent with the World Bank's HDM-4 Model, which focuses on the orderly prioritization of the projects in SOPTRAVI's historical portfolio that appear to have been the most profitable, benefiting the largest number of individuals and serving the greatest traffic volumes. As the lead entity in this area, SOPTRAVI has embarked on a strong campaign to bolster and broaden this initiative through concessioning or through Public-Private Projects (PPPs). This new effort has been conceived with the vision that, owing to its economic characteristics and the relatively low traffic levels, Honduras cannot afford the luxury of concessioning works to create new roads that runparallel to existing ones, and that instead, this participation should be oriented primarily toward the more efficient preservation and improvement of the existing network. Inthis connection, the previous administration entrusted various technical and financial feasibility studies on engaging such concessioning to recognized international firms. These studies have mainly focused on the so-called "Logistical Corridor," (or "Inter-Oceanic Corridor `I). 7.21 More recently in2007 this Plan is inthe processof being further updated as a result of an analysis of the status o f the principal road corridors in Honduras. This analysis has allowed for the recosting of the proposed improvements of the principal corridors as well as to reconfirm their economic viability. In line with the government's broader objectives to promote growth through increased competitiveness in the rural economy, the prioritization o f the corridor improvements is beingrevisited. RoadMaintenance 7.22 The Road Fund (FV) is the agency responsible for ensuring the financial sustainability and ongoing maintenance service for the official network of Hond~ras.'~It is a deconcentrated Central Government entity, with administrative, technical, and financial independence. Its board of directors, known as the Road Technical Committee, is the highest management body, and is made up of five representatives from the related government sectors (the SOPTRAVI minister, who chairs the committee, the Director-General o f Roads, SEFIN, the Ministry of Industry and Commerce, and the Municipality Association of Honduras-AMHON) and three from the private sector (Civil Engineers' Association, Honduran Private Enterprise Council, and National Transport Council to represent carriers). The majority presence of the government has led to some degree of politicization of activities and decision-making, thus diminishing the high technical profilethat the Road Fundshould have inorder to maintain its credibility at all levels. 7.23 The Road Fund prepares the annual maintenance programs, draws up budgets, and implements the maintenance plans and strategies. By law, it is financed exclusively with domestic funds from fixed fees and taxes as a percentage on sales of fuel. Most important i s the fee referred to as the "Road System Contribution"; Table 7.4. Inpractice however, the relevant percentage mandatedby law has never been passed forward to the Road Fund. Historically, the 98. The RoadFundwas established by NationalCongress Legislative Decree No. 131-93 of 1993, as subsequently amended by Decree No. 286-98 of 1998. Taken together, the two decrees constitute the Road FundLaw. 135 allocations have scarcely amounted to half o f the amount mandated by law, resulting in the severe shortcomings inattending to the nationalnetwork (Table 7.5) Table 7.4 Tax Revenue from Petroleum Derivatives, 2002-2005 (in Lempiras) Item 2002 2003 2004 2005 Tax on the Production and Consumption 42.493,108.15 259,420,694.87 0.00 0.00 of PetroleumDerivatives Road System Contribution 2,553,656,619.07 3,466,148,761.39 4,539,750,853.50 4,915,584,944.33 Tax on PetroleumImports 688,215,248.16 175,840,077.98 139,522,616.05 0.00 Total 4,084,364,975.38 3,901,409,534.24 4,679,273,469.55 .4,915,584,944.33 Source: RoadFund. Expenditurefor the auditingfirm is estimatedfor eachyear. Revenuefrom Allocation granted YOof Total YOthe Road Shortfall from Fuel Sales to Road Fund Revenue Fund should the allocation (Lps. Millions) (Lps. Millions) granted receive by law called for by law 2002 2,553.66 483.30 19 35 408.58 2003 3,466.15 650.01 19 38 658.57 2004 4,539.75 587.37 13 40 1,225.73 2005 4,915.58 738.02 15 40 1,228.89 Total 15,475.14 2,458.70 16 3321.77 Note: *These taxes havebeenconsolidatedinto a single tax. Source: SEFIN. 7.24 The current expenditures o f the Road Fundhave been relatively low. Its personnel roster only has 40 employees, which together cost about as much as the external auditing firm responsible for the physical and financial monitoring o f the Fund's performance. While current expenditures are low, the number o f employees has been rising in 2006 and the cost of supervision i s relatively high; see para. 7.19. Under the law establishing the Road Fund, up to 2.5 percent o f expenditures can be allocated to administration. The Road Fundhas sta ed within thislimitwhich is consideredto bereasonable onthe basis ofinternational experience. r-! 7.25 There will be some planned increases inthe RoadFundbudgets, owing to the stipulations and conditions in formal agreements set forth in Conventions with the World Bank, the Millennium Challenge Account (MCA), and the IADB ( at minimum an increase in the annual budget by 10 percent per annum beginning in 2007. Accordingly, the Road Fund budget is scheduled to increase to about US$45 million in2007, then to US$50 million in2008 and US$55 million in 2009, and it is anticipated that it will continue increasing and not decrease at any 99. Typically, road fund legislation may allow a percentage in the range 2 percent to 5 percent to cover administration and overhead. 136 point. With this partial improvement, it will be possible to broadenthe coverage o f the network, and it is estimated that about 75 percent of the network will be covered as compared to the 60 percent that currently is. This issue should be examined and monitored on a regular basis through the so-called "Cooperation Round Table" o f representatives from the aforementioned institutions as well as the Central American Bank for Economic Integration (CABEI) and the Japanese International Cooperation Agency (JICA). Table 7.6 OperatingExpensesof the RoadFund (in Lps.Millions) Total Investment Administrative Expenses External Audit Expenses YOof Total Investments 2000 499.3 14.00 14.00 5.6 2001 547.26 14.90 14.00 5.3 2002 483.29 14.90 14.00 5.9 2003 650.01 14.90 14.00 4.4 2004 587.37 14.90 14.00 4.9 2005 738.02 16.90 14.00 4.2 2006 669.47 20.54 14.00 5.2 7.26 Although all the works executed by the Road Fund are carried out through private construction and consulting firms, it i s estimated that the cost o f supervising the various Fund projects could be drastically reduced ifthe sectors could be reduced innumber and be distributed in accordance with geographical traits. This would make it possible to achieve economies of scale and improve spending efficiency, inasmuch as field inspections cost up to 25 percent o f the amount o f the construction. This i s considered extremely high by international standards. The excessive atomization of works has also resulted in a proliferation o f makeshift firms, especially in consulting, were about 140 firms compete for the roughly 50 works supervision contracts offered annually by the Fund. It is therefore advisable to scale back these lists considerably on the basis o f performance in projects already executed and the quality o f the results obtained. Furthermore, the time period allowed for execution o f conservation works i s extremely long, and encourages low capacity firms with extremely poor monthly execution rates, on the order o f $30- 40 thousand per month, resulting ina highcost o f supervision which has to be inplace duringthe entire contract period for the construction. In addition, the deficiencies o f the Road Fund are attributable fundamentally to a limited allocation o f resources and an inadequate form o f contracting for works, and not to the fact that the country is buildingnew roads. Indeed, since 1990 there have been virtually no new projects carried out, and resources have been used exclusively on the rehabilitation, reconstruction, improvement, paving, and conservation o f existingroads inkeeping with a government policy on the use o fresources adopted at that time. 7.27 The decentralization o f conservation tasks is expected to continue to move forward gradually, as greater responsibilities are shifted to the municipalities. It i s envisaged that the road maintenance program should eventually be expanded to provide Road Fundresources -possibly on a cost share basis - to help muncipalities maintain a "red basica municipal". One initiative already linked to this is the Rural Infrastructure Project (PIR) financed by the World Bank through the Honduran Social Investment Fund (FHIS). This Project was launched in August 2006 to support road, electrification, drinking water, and basic sanitation programs in various rural areas o f Honduras. Inthis project, the benefiting municipalities are grouped into 10 pools in order to bolster their capacities and obtain economies o f scale in works implementation and futuremaintenance. 137 7.28 The program on RuralRoadMaintenance Micro enterprises, initiatedin2001 with World Bank support, has beena resounding success and resulted inmarkedimprovement inthe groups and communities selected. It has had a major positive impact on the standard of living of its beneficiaries, in addition to imparting a genuine entrepreneurial spirit among these groups. In addition, there has been a radical improvement in the condition of roads as regards the clearing of rights of way and drainage, which will surely contribute to extending the useful life of roads and to enhanceduser safety.loo However, this program i s focused solely on routine maintenance, mainly including the clearing of rights of way and drainage, and consequently its impact on the conservation and state of the road network is limited. Other maintenance activities are contractedout to medium scale firms. B. The Ports Sub-sector 7.29 TheNationalPort Enterprise (ENP) is the centralized Government agency responsible for administering the five ports currently operated under its auspices: Puerto CortCs, Tela, La Ceiba and Puerto Castilla, on the Atlantic, and Henechn (or San Lorenzo), on the Pacific. It currently functions with a total of 1,338 employees, 993 of whom are permanentand 345 temporary. It i s projected that these figures will drop in 2007 to a total of 1,276 (950 permanent and 326 temporary). The dollarized fees chargedby the ENP have remained unchangedsince 1998. 7.30 Inconjunction withthe entry into force ofthe free trade agreementwiththe UnitedStates of America, the ports of San Lorenzo, Puerto Castilla, and Puerto CortCs received their corresponding certifications issuedby the National Port Protection Commission (USA), an entity which uses parameters from the International Maritime Organization (IMO). This process led to substantial improvement in the safety, organization, and control measures used in the current operations of these ports in particular. A breakdown of ENP expenditure over 2002-2006 is shown inTable 7.7. Table 7.7 Total Expendituresof the NationalPorts Authority (EM) (in Lps millions) Current Expenditure Capital Expenditure Financial Uses Total Executed Expenditure 2002 645,074 235,430 217,080 997,583 2003 791,369 106,449 187,449 1,085,266 2004 900,072 154,895 340,199 1,395,167 2005 938,083 140,403 334,224 1,412,709 2006 886,925 3 10,370 40,747 1,238,042 Source: ENP PlanningDirectorate. 7.31 Capital expenditures are broken down into machinery/equipment and the projects or works carried out in the past five years, from which it is possible to deduce that ENPs current expendituresrepresented 66 percent of total expenditures. ENP has historically shown profits, mainly as a result of operations at Puerto Cortds, which has resulted in sizable annual transfers from the enterprise to the Central Government. In recent years, these transfers amounted to 100. 36 such micro enterpriseshave been foundedwith support fiom the previous Credit 2458-HO and the Credit IDA 3432-HO fiom the World Bank, and 28 groupings have been establishedwith domestic finding As of . February 2007 all (now 70) micro enterprises are funded nationally and cover the entirety of the main paved network. 138 around Lps. 200 million. Inaddition, ENP also has been responsible for a series of transfers to other decentralized State entities not associated with the sector, such as the Honduran Institute for Drug Addiction and Pharmacological Dependency (IHDFA), the Youth Council (CONJUVE), and other subsidies grantedto social or nonprofit organizations. 7.32 Inthe 2000-2006 period, cargo shipmentsby port increasedgradually by anaverage of 7 percent a year, with roughly 81percent of the cargo handlednationally inthe country's six main ports being concentratedinPuerto CortCs. Annex J shows the amount of cargo handledin each port in the period under review, followed by the traffic projected by the ENP for the five ports currently operating under its auspices (Puerto CortCs, Tela, La Ceiba, Puerto Castilla, and San Lorenzo), itbeing recalledthat Roatin is now managedby the private sector. The Ports Master Plan 7.33 This 1994 study conducted under a grant from Japan is still used today as the primary guide for the institution. It indicates that all the ports require a wide range of investments in equipment and facilities, valued at US$800 million to US$1,300 million, by a time horizon endingin2010. The largest ofthese, other thanthose already mentioned, are the coastal shipping pier at La Ceiba, andpetroleum facilities at SanLorenzo, Tela, and Cortes. For Puerto CortCs in particular, the study further recommends the construction of three berths for containerized cargo with an initialpier 185 meters inlength, a granular cargo terminal, a coastal shipping terminal to accommodate 25 vessels, and the provision of an alternate road approximately 1,380 meters in length.A port crane is recommended at each berth, Also recommended for Puerto Cortes is a new container terminal, thereby making it possible to use the existing Pier No. 5 exclusively for general cargo and discontinue the filling and emptying of containers at Shops No. 3 and No. 4. Those operations would be performed inthe new terminal, thereby increasing the cargo handling capacity andefficiency o fthe port ingeneral. 7.34 It also recommends providing greater incentives for private participation innewprojects and in certain operational activities once the works have been completed. The rate o f return on all these projects was estimated at 22-23 percent, dropping to 18 percent with a sensitivity analysis of reduced benefits and increased costs. Furthermore, the study suggested a strengthening of port monitoring and of the surrounding environment, in particular as regards water quality, along with personnel training in this regard and the establishment of an environmental control system. 7.35 Table 7.8 shows the investments suggested under the Master Plan through 2010, which remains inforce today. 139 Table 7.8 TotalCost of PortsInvestmentsContemplatedinMaster Plan (inLps. Thousands) Port Primary Facilities Equipment Total Investmentthrough2010 (estimation) a. CortCs 161,000 -256,000 366,080 -320,040 527,080 - 576,040 322,000 -419,000 488,580 -444,520 810,580 - 863,520 b.Tela 40,000 8,200 57,200 c. LaCeiba 48,000 4,220 52,200 d. Castilla 51,000 -97,000 20,130 -25,730 71,130 - 122,730 e. Lempira 3,000 1,750 4,750 f.Roath 15,000 2,020 17,020 Note: The top lines for PuertoCortCs andPuerto Castilla correspondto Scenario No. 1, while the bottom lines correspondto Scenario No. 2. Source: MasterPlan Port Productivity and EfJiciency 7.36 While ENP remains profitable - and indeedreturns resources to the government's central treasury - its performance lags significantly interms of productivity and efficiency inrelation to comparator ports. The specific dimensions include principally the following: (1) Labor productivity: very low due to overstaffing especially inthe adminsitrative grades. One estimate is that under a concesssion arrangement staffing could fall to about 200;'O' (2) Operational performance: Puerto Cortes lags international standards in regard to: waiting times; cargo volume handled; and tonnage per meter of dock All these factors contribute to higher user costs and the likley loss of market share to competitor ports, especially inguatemala. ENP is unlikely to be able to meet continued demands for port services if the constraints to efficiency improvements are not addressed. 7.37 Government policy inregard to addressingthese structural problems inport management has been unclear and inconsistent in the past. The deficiencies in the current arrnagements are recognized, but it has proven politically to be quite difficult to confront the issues. International experience is strongly supportive of the view that the increased investment and higher productivity necessary to sustain the position of Honduran ports would best come through the private sector. There are a variety of arrangements short of outright privatization which could secure these benefits. the net fiscal impact of a policy infavor of a PPP, which is one option, in the sector would be positive with the government continuing to enjoy royalty/license payments equivalent or better to what i s currently available and not having to invest scarce public resources infuture investment. C. The Airports Sub-sector 7.38 The airport sector in Honduras underwent a radical change in October 2000 when a 20- year concession - awarded after a competitive process - was granted to the INTERAIRPORTS S.A. (IASA) firm to manage all four of the country's international airports (Toncontin in Tegucigalpa, Villeda Morales in San Pedro Sula, Golos6n inLa Ceiba, and Juan Manuel Galvez on RoathnIsland). These airports hadpreviously beenmanagedbythe Civil Aviation Directorate 101. PPIAF -Private Solutions for Infrastructure inHonduras-A Country Framework Report, July 2003 140 (DGAC), a unit of SOPTRAVI, and immediately prior to the concessioning, by a nonprofit Administrative Board. To date, IASA has been administered by three different private entities, and, since February 2006, has been managed by the national group known as TERRA, which receives advisory assistance from the Chilean-Swiss group UNIQUE. The improvements made to date are basically confined to the terminal buildings, particularly inTegucigalpa andLa Ceiba, leaving aside the repair of runways and steps relating to air safety. Air navigation issues will continue to be handledby the DGAC. 7.39 The concessioncontract underwentan initial modification inNovember 2003, extending the deadline for completing the urgent improvements and some new construction works, including the extension of the runway at Tegucigalpa. These improvements are necessary to meet B service level, an international standard defined by IATA. Subsequently, in December 2005, the headof SOPTRAVI signed a secondmodification, which has yet to be approved by the National Congress of the Republic. Among other things, it deals with various contentious issues of national interest, including exempting the concessionholder from all types of taxes and fines andenabling itto modify rate structures at will. 7.40 By way of illustration, overall passenger traffic at the four airports is on the order of 1.6 million passengers a year, which confirms that the unique form of having been able to concessionthem to a private operator as a group was necessary, inthat it is normally maintained that an airport is profitable individually when it handles at least one million passengers a year. Passenger traffic is distributedas follows: San Pedro Sula accounts for 38 percent, Tegucigalpa 33 percent, La Ceiba 15 percent, and Roatiin the remaining 14 percent. On the cargo side, San Pedro Sula handles the majority, at about 60 percent, while Tegucigalpa accounts for 35 percent. The La Ceiba and Roath airports handleonly marginal amounts of air cargo, primarily local in nature. Average productivity in the airports is close to the upper end of the international range.''* Table 7.9 Summaryof Gross Revenueand Royalties Paid by InterAirports S.A., 2000-2006Period (in Lps) No. Year Gross Revenue ConcessionRoyalty Royalty Percentage (YO) 1 2000* 61,969,134.00 24,588,670.71 39.6789 2 2001 240,002,933.80 95,230,524.10 39.6789 3 2002 276,286,628,65 109,627,495.10 39.6789 4 2003 306,052,341.00 121,438,202.33 39.6789 5 2004 319,071,474.02 112,711,961.73 34.4 6 2005 276,096,950.1 1 122,054,773.13 34.4 7 2006** 292,665,847.00 100,5 18,056.00 34.4 1,772,145,308.58 686,169,683.10 Note: 2000* covers the October-December period. 2006** covers the January-September period. Source: Superintendency of Concessions and Licenses. 102. PPIAF -Private Solutions for Infrastructure inHonduras -A Country Framework Report, July 2003 141 Table 7.10 U S A Investmentsin ConcessionedAirports Investmentin Terminal Buildings Other Investments Total Total (Construction Costs and (Works outside the Investment Investment Airports Equipment) inUS$ terminal) inUS$ inUS$ (Lps )* Tonconth 13,327,417.44 611,066.00 13,938,483.44 265,214,493.65 Ramon Villena 827,008.57 940,386.00 1,767,394.57 33,629,100.18 Morales Juan Manuel 3,897,898.38 558,029.00 4,455,927.38 84,785,158.22 GBlvez Goloson 3,733,409.26 783,325 .OO 4,5 16,734.26 85,942,161.13 Total 18,785,733.65 2,892,807.00 21,678,540.65 412,488,432.22 Note: *US$1 = I-INL 19.0275 Current rate 2005-2006. Source: Superintendency of Concessions and Licenses. The breakdownof current expenditure i s not available. Table 7.11 PASSENGERVOLUME YEAR I I I I I 2006 ANNUAL ~ AIRPORT 2001 2002 2003 2004 2005 Projected AVERAGE Tegucigalpa 3,752,800 3,603,341 3,681,002 4,102,297 4,894.722 5,185.000 4,203,194 San Pedro Sula 6.779.719 7,158,340 7,857,306 8,929,771 7,300,087 6,827.000 7,475,371 TOTAL 10,632,619'10,761,681 ' 11,538,308 13,032,068 12,194,809 12,012,000 l - " . " _ i 11,678,664 I i- _ - _ J - _ _ _ _ I _I^_.I -N-oe-:-The!e-is no-+gnifiIcant cargo tcafic atfh-e_La-Ceib-a_and Roa-*n airPo-&_ - - __, - II Source: Superintendency of Concessions and Licenses 7.41 There is some degree of dispersion inrespect of decision-making and control inthe sub sector, as indicated by the fact that the concession holder is audited by the Superintendency o f Concessions and Licenses, the contracting party i s SOPTRAVI, and the Directorate-General of Civil Aviation (DGAC) regulates matters pertaining to air navigation but only up to an altitude of 19,500 feet, beyond which limit the regional entity COSESNA regulates and controls flights, absorbing the sizable flow of revenue thus generated. It would be timely and appropriate to review, and probably downsize, the role played by DGAC - which should focus on air transport policy only in future. Some budgetary savings could be made. Management of air space could bedevolved to COSESNA. D. The Railways Sub-sector 7.42 The Honduran NationalRailways Enterprise (FNH)is an autonomous public entity o fthe Government, createdby Decree-Law No. 48 approved by the National Congresson May 7, 1958, as an Autonomous State Body with separate legal standing, its own assets, and of indefinite 142 duration. Its board of directors is chaired by the SOPTRAVI Minister. The institution is State- owned and there is virtually no private sector involvement inits operation andmaintenance. 7.43 Ingeneralterms, the condition of the institutionhasbeenlargely unchangedover thepast decade, that is, it has a fixed operating deficit, its roster of personnel have scant activities, there has been continuing deterioration in its railway property and equipment, and thefts of metal track, accompaniedby the loss of its real property assets through squatting on landthat it owns, including the right of way. Under the total control of the State and with a 100 percent national budget, it currentlyi s operating scarcely 109kilometers of rail lines, distributedas follows: 7.44 The transport of bananas historically has been the largest component of the FNH's business. This has been totally discontinued because both multinational banana companies (StandardFruit Co. and Tela RailRoad Co.) and the independentproducers now transport all of their output by road in vans or containers, principally to the Trujillo and CortCs ports. Decree No. 132-89, which dates from 1989, authorized the FNHto carry the property and general cargo of all State institutions so as to generate volume. This provision was only fully complied with between 1990 and 1994, given that in subsequent years not even the State agencies themselves moved their cargo by rail, possibly owing to a lack of confidence in prompt and efficient shipmentbecauseof the chronically poor condition ofthe real lines androlling stock. 7.45 A sideline activity for the FNHhas been the Customs Agency that usedto operate inside the institution. Among its most important clients were the National Electricity Enterprise (ENEE) and the Honduran Telecommunications Enterprise (HONDUTEL). The agency has beenclosed since November 2004, but the current managementhas plansto reopen it inorder to generate greater revenue. 7.46 Labor relations at the FNH improved when, in late 2004, workers' benefits and entitlements were distributed to 125 employees, representing a distribution of about Lps. 9.8 Million (US$0.53 million) in benefits. At present, the institution is assigned only 36 positions (29 full time, 3 are vacant, and 4 are temporary employees) in four main units, which in turn have 6 departments, but there is still a sizable fixed financial burden in supporting 106 retirees who, during their working lives, contributed absolutely nothing to the system and collectively seek approximately Lps. 2.2 Million (US$116 million) annually in compensation and medical expenses. 7.47 The financial situation of the FNH from 2002 through the third quarter of 2006 demonstrates that the institution i s still running an operating deficit that has fluctuated between Lps. 4 and 18 million annually (or US$ 0.2 to US$ 1.O million) during the last five years. The bulk of FNH expenditure has gone toward the payment of wages and office arrangements, leaving virtually nothing for other operational expendituresand maintenance. This situation can only be sustainedthrough the subsidies provided by the Central Government. 7.48 The current Administration has identifieda number of short, mediumandlong term goals at the level of ideas. It i s highly unlikely, however, that the FNH will be able to obtain fresh funding from the Government or from a private source to carry out these ideas even if first having made a number of major investments to modify its current state of widespread deterioration. Another problem is the absence of a strategic planto help FNH's top management visualize the institution's role in the short or medium term (even though various consultants' efforts have been oriented toward providing support to this end), as well as the absence of a specific Investment Plan for the years ahead. At this stage, FHN is merely subsisting on a 143 regular subsidy from the Central Government, while its assets continue to deteriorate. The idea of building an Inter-Oceanic Railway, which is frequently mentioned, seems highly remote and impossible to achieve in the absence of clear Government policies with respect to such a route andthe identification of probable sources of financing. 7.49 There would be no justification for the government to continue to subsidize the operations of FNH, as the operation i s unviable as it stands. Current traffic, limitedas it is, could be carried by alternative modes of transport. Strategically speaking there is a potential role for rail to play inthe transport systemof the country, with particular regardto bulk cargo movement to and from the Caribbean ports. But for that to happen, a new operation properly capitalized and managed, either privately owned or privately operated under a long term concession. A recent review of Latin America experience with railways reform'03 suggests that rail concessioning can increase market share and improve productive efficiency for rail particularly for freight transport, though muchless so for urban and inter-urbanpassengertransport. E. Transport Subsidies 7.50 The subsidy to the transport sector is basically extended to urban passenger transport in the capital city, and is provided to the private carriers operating the buses serving traditional routes. The payment o f this indirect subsidy is by bus, with an allocation of Lps. 575.00 (US$30.40) per day, assuming that each unit completes four complete route cycles. For microbuses the subsidy i s Lps. 320.00 per day. The consequences of this measure have been rather negative for the system because, owing to the manual type o f inspection carried out in order to verify the completion of route cycles, there has been a stimulus to fraud and imprecision. Nor has it preventedthe use of aging vehicles that have virtually exhausted their useful life, thus swelling the vehicle fleet and consequently increasing the levels of traffic congestion, pollution, and the demand on roads, thereby contributing to sustained deterioration in service and in the environment. Accordingly, the current subsidy system is deficient and promotes the inadequate andunjustallocation of resources. Table 7.12 BudgetExecutionof Expenditurefor Granting Subsidies (Lpsmn) Year Initial Budget RevisedBudget ExecutedBudget 2005 114,000,000.00 114,000,000.00 113,539,375.00 2004 114,000,000.00 111,673,060.00 116,782,105.00 2003 114,000,000.00 113,605,594.00 113,595,285.00 2002 114,000,000.00 114,000,000.00 111,426,690.00 Total 456,000,000.00 453,278,654.00 455,343,455.00 7.51 A recent study (CPC-Regioplan 2005) -financed with resources from World Bank Credit 3432-HO - recommended the introduction of a mechanism for replacing the vehicle fleet inthe capital, makinguse of the same funds as the subsidy, which amounts to HNL 114million (US$6.03 million) annually. The fund would be used to guarantee the new purchases that the system urgently requires in light of the fact that this significant amount of annual funding granted in the form o f subsidies over the past 10 years has failed to contribute to improving 103. "Results of Railway PrivatizationinLatin America" by Richard Sharp (World Bank Transport Papers TP-6, September 2005 144 service and, to the contrary, complaints and safety lapses are much greater in number and frequency. There has been virtually no renewal o f the fleet (the average age o f the buses in the capital is 23 years) andthe attentionto passengers continues to be o f very poor quality.lo4 7.52 These recommendations however address only in part the issues o f service quality and reliability. The larger issue i s reforming the system o f route designation and their allocation through a competitive and transparent process, the idea being to address head on accessibility which arises as a problem because many o f those who should benefit form a transport subsidy are not currently served on any routes. The competitive allocation o f routes, preferably on the basis o f least subsidy, which would provide for value for money but would also recognize the social justification o f providing some support to low income passengers. Direct income subsidy to these passengers would be better still in principle, but would be probably too complex to implement. 7.53 The Government also grants a direct subsidy to the roughly 125,000 students in public secondary education in the major cities o f Honduras who use the urban transport service. The respective payment o f Lps. 400.00 per year (US$21.OO/year) to each student i s made by transfers to the Ministry o f Education, which passes the subsidy on directly to the students on the list o f officially registered pupils for each academic year. The traditional delays occurring inproviding these funds regularly generate complaints from students, as there are extreme cases inwhich the payment is not made until the ninthmonth o f the year. According to information provided by SEFIN, the direct subsidies allocated each year remained steady, at around Lps. 14 million during2000-04, butthen doubled to Lps.30 millionin2005 andLps. 50 millionin2006. F. MainConclusionsandRecommendations Sector Policy, Regulation and Institutional Framework 7.54 There are three overarching problems inthe transport sector. They are (i) absence o f the an overall sector framework to guide planning and implementation, (ii)The absence o f an institution with the capacity to manage the sector multi-modally, and (iii)the absence o f sufficient planninginthe roads sector, which has the consequence o f bringingtoo much focus on the short term. These problems impact adversely the efficiency and effectiveness o f the delivery o f public services to the transport sector and to the efficiency and effectiveness o f agencies and ministries inthe sector. 7.55 The first measure recommended to address these overarching problems is to pass a "Ley de Transporte", with the objective o f reorienting the public sector around regulation and facilitation, and o f bringing more private sector capital into investment in infrastructure. lo5 104. The NationalTransport Census, conductedin2005, bearswitnessto the poor condition ofthe vast majorityof passenger transport vehicles inparticular, and also notes the disorderprevailingas regards the legality of the OperatingPermits and other recordsthese units must have to operate. There was evidence of many duplicate or triplicateregistrations with the same permitnumbers, as well as fraud with regardsto engine numbers and vehicle serial numbers. 105. This recommendationis consistentwith the reformstrategy that hadbeen agreed in2004 under PRSC 11, but for which the progressbenchmarkshave not beenmet. 145 Regarding transport services, the focus o f government attention would be on assuring safety and efficiency. The latter would require addressing specific weaknesses in both institutional structures and in data, inparticular regard to transport services and road safety (accidents). The second measure would be to review the current institutional framework for carrying out transport regulation with a view to increasing its effectiveness. The third measure would comprise the working out o f a specific policy statement in regard to private sector participation in transport, given that this to likely to be critical to increasing the extent and the efficiency o f future investment inroads and ports and that private sector investment is the only viable option for the resuscitation o fthe rail sector. 7.56 Notwithstanding all the above measures, there will still be a need for supporting actions to increase capacity to plan and execute programs inthe public sector - with particular regard to SOPTRAVI andto the road sector. It is notable that the rate o f program and budget execution in SOPTRAVI has shown no improvement at all since 2000, inspite o f the reforms implementedto date and despite the technical assistance provided.lo6 A major adverse factor here has been the frequency o f change in key personnel over the political cycle. Also, an important number o f positions are filled on a consultancy basis due to the difficulty o f attracting qualified staff for the offered salaries. There i s an urgent need to develop specific indicators for access, quality and efficiency improvements and to ensure that achievements against these indicators are effectively monitored. Fortunately there have been recent institutional improvements carried out in SOPTRAVI and investments have been made in planning systems and the supporting software andtraining which give SOPTRAVIthe capacity inprinciple to respond to this requirement. Sector Expenditure 7.57 Total sector expenditures are summarized in Table 7.13. Roads and ports account for more than 95 percent o f expenditure. The share o f road sector expenditure allocated to recurrent (maintenance) has been withinthe range 32-37 percent with no discernable trend inthe period.lo7 As a percentage o f GDP, transport sector expenditures have averaged 2.2 to 2.3 percent during this period, which i s relatively low by regional standards. This compares unfavorably to the average o f about 2.9 percent that prevailed inthe 1990s. Table 7.13 Summary of Transport Sector Expenditures,2002-06 (in Lps millions) Sub Sector 2002 2003 2004 2005 2006 Roads 792.9 1001.5 1071.5 1262.5 1113.2 Road Mtce. 483.5 650.7 587.3 737.1 669.1 Rail 31.6 35.1 33.1 13.2 15.1 Port 998.3 1080.5 1391.9 1408.1 1237.9 Air 71.4 75.4 79.1 81.3 81.3 Total 2377.7 2843.2 3162.9 3502.2 3116.6 106. The rate of execution actually fell from 83 percent in 2000 to 65 percent in 2006. In absolute terms, the comparable figures are Lps. 2.1 billion and Lps. 1.6 billion 107. In 1999maintenance expenditures were about 53 percent oftotal 146 Roads 7.58 The public investment program for the road sector has been updated in 2005. Appropriate principles have been followed in terms of identifying viable projects. All projects that meet target rates of returnshould be financed. Availability of funding - 70 percent of which has traditionally come from external grant of concessional loan resources - and implementation capacity are practical constraints on program implementation. Specific actions that need now to betaken would be: factoring inthe needto support (rural) competitiveness inthe prioritizationof any future road corridor upgrading; actively encouraging further road concessions - one only exists at the moment - with the first priority given to the operations and maintenance of the "corredor logistico." 7.59 Total road expenditures in the period 2002-2005 have averaged US$91.5 million equivalent per annum, or about 1.3 percent of GDP, low by regional standards. The road service index (kilometers per thousandpopulation) is 0.45, comparedto the Central American average of 0.63 and the density index (kilometers per thousand square kilometers area) i s 29 compared to the Central American average of 55. These figures are indicative of a relatively non-extensive network for an economy o f the size of Honduras. However the quality of the road network in Honduras is towards the upper end of the scale for Central America - 40 ercent of all roads in good condition of which 90 percent of paved roads are ingood condition.". Unitcosts for both investment and maintenance in the road sector are on the high end of the scale in relation to comparators. 7.60 SOPTRAVI reforms need to advance further to increase efficiency in the road sector however. Only 55 percent of planned expenditures on road investments could be implemented since 2002. High costs and quality of works are still an issue. The consistent application of agreed design standards i s also a problem. Even so, the establishment of UPEG is a plus as well as the associated training and systems development, including HONDUSAP (bridge management system), which have beensupported by the Bank andthe IADB. Progresshas been made in the improvement o f social and environmental management through UGA. The current focus is on the Department of Transport which has ultimate responsibility for public service vehicle inspection and licensing as well as axle loadandweight controls. Weaknesses have been identifiedinthese areas and an action planhas beendrawn up which has already attracted some support from the Bank. RoadMaintenance 7.61 The creation o f the Road Fund is seen as a largely positive step. On average, a higher volume of resources in real resources are going to road maintenance than before. All works are being contracted out to the private sector, and a significant portion of this is going through the micro-enterprises (largely for paved road maintenance). Road Fund projects have typically experienced fewer delays and cost overmns than on average for the road sector. However the relatively high cost of supervision services has been identified as an issue to be addressed. 108. This compares to 37 percent overall and 47 percent paved ingood condition in 1999 147 Administrative overheads have been held within the mandated ceiling o f 2.5 percent o f expenditures. 7.62 Specific issues inroad maintenance include the following: Road Fundrevenues still do not meet 100percent o f target. That is, all maintainable public roads. Mostly external pressure has helped to get the amounts allocated increased, now by 10percent per annumthrough the rest o fthe decade.lo9 Government is broadly inagreement that existingmaintenance expenditures be prioritized to make sure main roadnetwork gets first attention. Ofthe three scenarios presented to 2010, that which would secure 85 percent bythe end o fthe period seems reasonable as it will ensure that all roads with viable traffic levels will be covered -that i s all maintenance projects which meet or exceed the target rate of economic return. Government i s not following the spirit o fthe law interms o fthe transfers to be made. Budgettransfers are always less than what is requested, which inturndoesn't necessarily require all the transfers that should legally be made (because o f capacity constraints). By 2005, only 42 percent o f the funds nominally assigned to the Road Fundwere so transferred. Delays are generally not a problem, but there were transitional problems with change o f government, andwith the introduction o f SIAFIin2006. The most serious issue here i s the misalignment o f future SEFINfundingtargets for road maintenance with what i s requiredto secure the objectives o fthe PMV. These expenditures would not raise the percentage o f roads maintained above what i s currently achieved - around 65 percent. The quality andindependence o fRoadFundmanagement is inquestion. The CEO is a political appointee, the chairman is the Minister. There appears to be little effective user participation (only 3 out o f 8 Boardmembers are from the private sector), added to which there have been frequent changes inkey personnel inthe administration. There i s some recent evidence that procurement procedures are not being followed to the letter. Therefore the Road Funddoesn't meet standards o f a "second generation" road fund. Its legal status and institutional structure ought to bereviewed. Ports 7.63 With the exception o f Roatan (concessioned out in 2004), the ports are under ENP, a publicly owned company with 1,338 employees and a Lpsl.2 billion. ENP makes a surplus most o f which is returned to government rather than being plowed back into port improvements (Lps 468 million in 2006). Traffic has been increasing at 7 percent per annum since 2000 - 81 percent o f the traffic uses Puerto Cortes, whereas Puerto Castilla i s expanding most rapidly. Recent developments include port certification by T S N I M O which improves the business prospects for Honduras in the sub-regional context. Puerto Cortes in particular is on course to expand, ifthat is the current capacity constraints can be lifted. Government has suggested it put in $150 million, and defer consideration of any further private sector participation at this stage. Government does however: (i) recognize that high costs and low efficiency are a problem that could undermine port competitiveness; and (ii) remains open to the idea o f PPP inthe port sector 109. The Plande Mantenimiento Vial (PMV) 148 (concessioning of activities within a public landlord port rather than of the port as a whole), on which a number of training and dissemination activities have taken place. A clear strategy needs to be adopted on ports and the role of PPP. New investmentsshould be undertakenby the public sector only with clear evidence of a lack o f private sector interest (highly unlikely if PPP is properly designed). Airports 7.64 Itwould betimely and appropriate to review, andprobably downsize, the role played by DGAC - which should focus on air transport policy only in hture. Some budgetary savings could be made. Management o f air space could be devolved to the regional organization, COCESNA which i s already in place. Management of international airports has already been contracted out since 2000 on a 20 year concession to a private consortium. This is managed by the Superintendency. There have been some problems and the ownership and management of the consortium has gone through some adjustments. There have also been some delays in the completion of capital improvements, but a clearly established plan exists for this to take place. Some $21.7 million has already been so invested, with a further $11 million expected over the next five years. The model adopted for the management of the airports appears fundamentally to bethe correct one. Railways 7.65 The sector is now making a very small contribution to cargo and passenger transport. There are 109 kilometers o f track, 36 employees, and an annual turnover o f around Lps. 6-7 million. The sector i s run by a publicly owned company, FNH. There have been some reductions in operational losses, but there i s no future for rail without infusion of fresh capital. Effective and efficient rail services cannot be provided inthese circumstances. FNH should be closed down and all subsidies to operations should cease, though the government will still have to meet some obligations (pensions, and costs of 125 employees recently separated). An effort should be made to attract a private investor, but there are few assets to be had right now (depreciated even further since the Bank funded studies were carried out in 2004, when they were valued at $21 million). There is a potential economic role for rail inthe deliveryof certain cargo in relation to the least cost, multi-modal development of Puerto Cortes and its associated roadcorridors. SubsidizationPolicy 7.66 The system inplace evidently doesn't work, but the proposal made to replace it may not make things much better. (This observation emanates from the recommendations of the Tegucigalpa UrbanTransport Study.) The current bus system, while low fare, is o f poor quality and reliability. This does tend to target the poor, as the better off have other transport options. However as access is considered to be more important than affordability, a problem with the current set-up is that many marginal urban areas are not served. So while the proposed scheme has the advantage o f encouraging existing owners to replace depreciated, environmentally 149 unfriendly buses for new, it will not allow the development o f new routes for which service providers could be competitively selected. Projected Sector Expenditure 7.67 Current estimates (most likely scenario) o f sector expenditure through 2010 are shown in Table 7.14. The forecast calls for expenditures equivalent to not less than 2.7 percent o f GDP. For roads, the balance o f expenditures moves progressively in favor o f maintenance, 39 percent o ftotal by 2010. Table 7.14 ProjectedPublicSector Expendituresinthe Transport Sector (in Lps. millions) Sub Sector 2007 2008 2009 2010 Roads 3294.6 3123.6 2259.1 1981.7 RoadMaintenance 1193.2 1271.1 1369.9 1476.3 Railways 11.4 11.4 11.4 13.3 Ports 1235.0 1368.0 1406.0 1558.0 Airways 38.0 47.5 57.0 74.1 Total 5772.2 5821.6 5103.4 5103.4 RoadMtceAlt 1'lo 1597.9 1622.6 1675.8 1761.3 RoadMtceAlt 2'" 855.0 950.0 1045.0 1216.0 Note: Projectedexchange rate is US$1to Lps. 19 7.68 Investment expenditures in the road sector are initially very high due to lumpy expenditures associated with the "corredor logistico" upgrading. Expenditures could be as high inthe outer years iffinance canbemade available inatimely mannerto fundother economically viable projects. In the ports sector there are likewise viable investments that could be implemented earlier if the private financing can be The projection for rail investment i s predicated on the continuity o f the current restricted network operation. However as the report has recommended shutting downthe system and opening up the possibility o f fbture investment in this mode by the private sector, projected investments could differ significantly from this forecast. 7.69 In regard to road maintenance, there are a number of possible scenarios. What is contained inthe table above i s the preferred scenario representing the least amount o f resources that should be made available. As mentioned, the SEFIN forecast is significantly less than this and should be adjusted upwards with immediate effect. The rate o f increase in required road maintenance expenditure i s likely to increase beyond 2010 as a result o f the completion o f the corridor upgrading prior to 2010. Recourse to management and operations concessions however i s likely to hold any increase inpublic expenditure incheck. 110. High case scenario - 100 percent. Note that 2001 PER quotes a SOPTRAVI estimate of $76.9 million to cover all roadmaintenanceneeds. 111. Basedon SEFINPlanPluriannual 112. This forecast only includes about $100 million of investmentspending 150 CHAPTERVI11 REVIEW OF SOCIAL PROTECTIONPROGRAMSAND SUBSIDIES A. ConceptualFramework 8.1 This assessment of Honduras' Social Protection (SP) system is based on the social risk management fiamework developed by the World Bank.'13 It comprises (i) assistance programs that are part of the country's social safety net such as transfers in cash or kind, subsidies, and workfare, and (ii)social insurance programs such as pension, health, disability and unemployment insurance. The concept of managing social risk comes fiom the notion that certain groups insociety are vulnerable to unexpected shocks which threatentheir wellbeing. SP focuses on the poor since they are more vulnerable to the risks and normally do not have the instruments to manage adequately these risks. This prevents the poor from taking more risky - butalso more profitable -activities that could helpthemovercome their poverty situation. 8.2 Social risk management consists in the choice of appropriate risk prevention, mitigation and coping strategies to minimize the adverse impact of social risks. It aims at reducing key risks, and breaking inter-generational cycles of poverty and vulnerability. In general, it is less costly to society to prevent risk than to cope with it afterwards; for individuals and householdsit i s often better to self-protect, rather than turn to the government for assistance that may not be forthcoming. The great majority of the poor in Honduras, however, do not have the means (assets and income) to be able to self-protect, thus this group should be the focus of public assistance programs. 8.3 Under the social risk management framework, SP should provide: (i)a safety net, particularly for the poor that are likely to "fall in the cracks" of established programs; (ii) a "springboard for the poor to bounce out of poverty" by building the poor capabilities; (iii) support to help the poor access basic social services, avoid social exclusion, and resist coping strategies with irreversible negative effects during adverse shocks; and (iv) opportunities to the poor to adopt higher risk-returnactivities, and avoid inefficient and inequitable informal risk sharingmechanisms.' l4 8.4 The Honduran poor are exposed to several risks, including those resulting from: (i) external shocks and consequent variations inemployment, income, and consumption, (ii) natural disasters, and (ii) microeconomic conditions that expose the poorest families to a series o f vulnerability situations. It is possible to reduce the risks associated with external schools by pursuing a strategy that involves market insurance, self-insurance and self-protection. For example, to minimize the adverse impact of a terms o ftrade deterioration, it is possible to engage inself-insurance mechanisms such as commodity price stabilization funds, the diversification of external trade, or futures contracts. Inthe face o f volatility in private capitals, it is possible to resort to contingent lines of credit. To weather fluctuations in economic activity, it is important to count on sufficient fiscal room to increase spending during economic recessions, including well structured employment programs (food for work), without riskingdestabilization. 113. Holzmann, Robert and Steen Jorgensen, "Social Risk Management: A New Conceptual Framework and Beyond". World Bank, February 2000. 114. Holzmann, Robert and Steen Jorgensen, "Social Risk Management: A New Conceptual Framework and Beyond". World Bank, February 2000. p3. 151 8.5 As is also true with other Central American countries, Honduras is very vulnerable to natural disasters, particularly hurricanes and floods. To diminish the risks associated with natural disasters, it is generally necessary to count on good civil defense organizations, to invest in reducing the vulnerability of the national infrastructure, and develop and implement sound landuse and natural resources management policies. Inorder to prevent or mitigate these risks, it is sometimes necessary to resettle populations living in high risk zones or inadequate housing, and to implement civic education and disaster preparedness programs. When a disaster strikes, it is necessary to provide emergency relief to the victims and reconstructionassistance. 8.6 This assessment focuses on microeconomic risks, namely the risks facing poor households because o f the conditions in which they live, rather than on risks related to macroeconomic fluctuation or natural disasters. These microeconomic risks are related to the poor households' access to basic services and to sources o f employmentlincome as well as to other conditions that can affect certain groups in society such as exclusion. It is important to note that existing microeconomic vulnerabilities may magnify the vulnerability resulting from macroeconomic disturbances or natural disaster. At the same time, the strategies pursued by individuals and families to face the risks to which they are exposed directly, often also help to diminish their vulnerability to the risks associated with changes in economic activity or natural disasters. Taking a non-agricultural job by a member o f a fanner's household (diversification o f the production), opening a savings account (diversification o f assets), participating in a training course (investment in human capital), and buying insurance (by contributing to a pension plan) all constitute self-insurance, self-protection and market insurance strategies that help individuals face the risks that may affect the community or the whole country. 8.7 The following analysis is organized aroundthe risks faced by each age group (life-cycle), risks that affect households in general, and some particular risks faced by the indigenous and black population. For each group, the main sources o f risk are identified and the population-at- risk is q~antified,"~and the principal public programs in place to address each risk are identified, with special attention to the public subsidies. The chapter concludes with an overall assessment o fthe social protection system and recommendations. B. Social Risks and Groups-At-Risk 8.8 This section discusses the risks that face the different age groups, the risks that face households in general, and the specific risks facing the indigenous and black population. The review o f the risks i s not comprehensive but focuses on the key risks that, if not properly addressed, may contribute to perpetuate the intergenerational transmission o f poverty. Vulnerable Group: 0-5 Year Olds 8.9 The two main risks that face poor children under five years old are inadequate diet and lack o f early stimulation, both o f which impair their development and perpetuate their poverty. 115. This analysis updates a previous more comprehensive evaluation of the Honduran SP system conducted in 2002. See Jose Silvdrio Marques "Honduras: Vulnerabilidad, Riesgos y Pobreza". Study prepared for the World Bank, 2002. 152 Problems may begin with poor prenatal care leading to low birthweight and be compounded by lack ofbreast-feeding, adequatediet and stimulation during the first years of child development. 8.10 Malnutrition in children. Low weight at birth due to inadequate food intake by mothers may lead to premature death or poor development in the early years. According to special tabulations of the ENCOVI 2004 prepared by UNAT, it is estimated that 7.1 percent of all newborns (67,450 children) had low birthweight, with a higher prevalence registered among the poor (7.9 percent) and inrural areas (7.7 percent).116 Table 8.1 PrenatalCare inPrevious5 Years, 2004 (Women 12-49 years) Has had Facility More Frequently Consulted children Had Prenatal in the checkups No. of Last 5 during checkups MOH IHSS Private years pregnancy received Hosp. Hosp. CESAMO CESAR Facility Other All 100.0 89.2 5.6 12.8 6.6 43.1 18.3 17.1 2.1 Extreme 25.4 84.8 4.6 5.4 0.2 47.8 40.2 3.6 2.8 Poor Moderately 28.8 89.3 5.3 13.4 3.1 52.2 20.8 8.4 2.1 Poor Non-Poor 45.8 91.5 6.3 16.2 12.0 35.0 5.4 29.4 2 Urban 47.1 89.8 6.1 16.6 12.0 41.6 3.1 25.4 1.3 Rural 53.0 88.6 5.2 9.4 1.8 44.4 31.9 9.6 2.9 Note: CESAMO (Health Center with doctor and dentist); CESAR (Rural Health center) Source: "Honduras Poverty Assessment". World Bank2006. Vol. 11, Table 48, p. 147 8.11 Low weight at birth reflects, among other things, poor prenatal care. Prenatal checkups during pregnancy have increased in recent years, but there are still a considerable number o f women that do not receive care during this critical period. About 15 percent o f the extreme poor pregnant women did not have any checkups during pregnancy; Table 8.1. For those that had some controls, the most visited facilities were the M O Hfacilities; bothfor the poor andnon-poor women. A large percentage of non-poor women also visited M O H hospitals (17 percent) and CESAMOs (35 percent). 8.12 Chronic malnutrition(height for age) affects over one-third o f the children that attend the first grade of primary school (ages 6 to 9 years). According to the Progress Report on the Millennium Development Goals (MDG) of 2003, malnutrition in this cohort had increased slightly from 34.9 percent in 1991to 36.2 percent in2001, making it difficult to reach the MDG target of 17.4 percent in 2O15.ll7 But the problems of malnutrition start earlier. For children under five, estimates based on the ENCOVI indicate that in2004, chronic malnutrition affected one of every five children (20.8 percent); Table 8.2. That is, about 198,000 children suffered from this condition. As expected, chronic malnutrition is more prevalent inpoor children (34.1 percent) and among those living in rural areas (39 percent), though one o f every four poor children inurbanareas also suffers from malnutrition. 116. Low weight at birth is defined as having less than 5.5 lb. Estimates fiom the Ministry of Health for 2005, show 10 percent of newborns with low weight at birth. Source: Department of Statistics, Ministryof Health, basedon "Boletin Ambulatorio de Estadistica". 117. Honduras-InformeSobre las Metas del Milenio". UNDP 2003. p.21. 153 Table 8.2 Malnutrition and DiarrheaAmong ChildrenUnder 5 Years of Age, 2004 Chronic No of Children Diarrhea Malnutrition * Suffering (last 15 days) (YOincidence) Malnutrition (YOincidence) All 20.8 198.038 17.1 Extreme Poor 46.6 134;868 17.9 Moderately Poor 24.3 65,493 20.3 Poor 34.1 190,597 19.1 Non-Poor 8.8 34,599 14.3 Urban 13.1 53,827 15.0 Rural 33.4 180,766.1 18.7 Note: a/ Malnutrition -Height for age (stunting). Source: ENCOVI2004. 8.13 Malnutrition i s caused by many factors, not only lack o f access to food. Poor health and diarrhea contributes to malnutrition as it impairs the retention o f foods and nutrients. The correlation between malnutrition in children and the level o f education o f the mother i s well established, and Honduras i s not an exception: while one-half o f children o f mothers without formal education suffer from malnutrition, less than 8 percent o f children suffer from this condition when the mother has 7 or more years o f education.' l8 8.14 Lack of early stimulation and low preprimary enrollment. Early stimulation is as important as adequate nutrition during the first years o f children development. Children born with low weight will be able to catch up in their development to children born with adequate weight if they receive adequate nutrition & stimulation. With just adequate nutrition, but without stimulation, the development o f those children will continue to lag those that receive both. Stimulation o f infants i s a task to be performed by parents at home and/or by specially trained staff in kindergartens. Early enrollment in preprimary is also important for continued child development. There is ample evidence that preprimary education increases children's cognitive ability when they enter the first grades o f primary education as well as over the longer term. Repetition rates in primary school are lower for children who have attended preprimary andoverall educational attainment is increased by preprimary attendance. 8.15 Very few infants attend kindergartens in Honduras. Indeed, the ENCOVI reports only one observation; Table 8.3. If mothers stays at home and take care o f the children then, in principle, there is no problem, as mothers are generally the best guardians o f their children. However, poor mothers often cannot stay at home to take care o f their children; because they need to work. Therefore, infants are often left at home under the care o f older sisters, who should be inschool. 8.16 At the preprimary level, only one o f every five children attends school; Table 8.3.'19 There are one-half million children that do not attend school at this level. Attendance is lower for the extreme poor children and inrural areas. 118. "Investigaci6n Acerca del Estado Actual de la Primera Infancia en las Areas de Educaci6n y Salud en Hondura". p. 18. CINDE, Octubre 2006. Enbase a la Encuesta Nacional de Epidemiologia y Salud Familiar, Ministerio de Salud, 2001. 119. Inthis studythe structure o f school levels and corresponding children ages are those adopted officially by the Ministry of Education as posted in the Government o f Honduras's official web site: http://www .gob.hn/portaVpoder-ej ecutivo/secretarias/educacion/estructura/ 154 Table 8.3 Kindergartenand Pre-PrimaryEnrollmentRates,2004 Attendance at Attendanceat Number of preprimary Kinder Preprimary age childrenthat do not (less than 3 years) (3-5 years) attend Total 0.0 20.5 502.264 Extreme Poor 0.0 13.7 177,990 Moderately Poor 0.0 17.9 149,992 Poor 0.0 15.6 327,981 Non-Poor 0.1 28.3 174,283 Urban 0.0 25.3 199,977 Rural 0.1 17.0 302,287 No of observations 1 718 Expandedobservations 174.6 129573.6 Note: NE'SHousehold Survey- (May 2006) indicatesthat 41.2% of children 5-6 years old attend school. . Source: ENCOVI 2004 ~ Vulnerable Group: 6-17 Year Olds 8.17 Dejcient primary education. Although (net) primary enrollments have increased in recent years and now are estimated at 81.5 percent, there are still 200,000 children aged 6-11 years that do not attend school; Table 8.4.120 Enrollment rates for the extreme poor children are 10 percentage points below those o f the non-poor. Efforts to increase coverage (more on this on Section D, below) inrural areas have narrowed the coverage gap between rural and urban areas to about 4 percentage points. Table 8.4. NetPrimary and Secondary EnrollmentRates,2004 Primary Secondary p/ No. of primary age No. of secondary age (6-11 years) (12-17 years) childrenthat do not childrenthat do not attend Attend Total 81.5 35.2 199,805 677,832 Extreme Poor 75.8 7.1 75,549 239,070 Moderately Poor 81.4 22.3 59,645 227,204 Poor 78.6 15.2 135,194 466,275 Non-Poor 85.5 57.4 64,610 211,557 Urban 83.7 52.6 74,294 235,849 Rural 79.8 19.5 125,510 441,981 No. of observations 4770 2387 Expandedobservations 877369.1 368505.9 Note: a/Refersto the third cycle of basic educationand "educaci6n media". Source: ENCOVI2004 8.18 The quality o f education remains a problem. According to the PRSP Progress Report 2005, standardtests conducted with sixth graders show an average attainment o f 36 percent in mathematics and 44 percent in Spanish.'*l 120. The 2005 Progress Report on the Poverty Reduction Strategypreparedby UNAT (June 2006) reports ahigher net primary enrollment of 89.4 percent based on PermanentHousehold Survey (EPHPM) data (may 2005). Household Survey (May 2006) indicatesthat 89.6 percent of children 7-12 years old; 39.4 percent of children 13-15years old; and24.1 percentof children 16-18years old, attendschool. 121. "Estrategia de Reducci6n de la Pobreza, Informe de Avance de 2005". UNAT, June 2006. p. 25 155 8.19 The efficiency of the Table 8.5 EducationIndicators education system remains low. The Primary Mean years of schooling repetition rate in primary school i s RepetitionRate (10-19 years) more than twice as high for All 11.7 5.0 extreme poor children than for non- ExtremePoor 17.6 3.4 poor children; Table 8.5. Data Moderately Poor Non-Poor 10.7 4.5 from the 8.1 6.2 Survey (May Source: "Honduras Poverty Assessment". WorldBank, 2006. Vol I, 2006) indicates that the repetition Table 1.7, p.20 rate is 19.9 percent in the first grade, but declines to 12.8 percent inthe third grade and to 5.9 percent inthe sixth grade, being higher in rural areas compared to urban areas. Drop out rates are about 3.6 percent in the first grade and decline to 2.4 percent in the third grade and 1.9 percent in the sixth grade. Lower attendance and higher repetition and drop out rates for the poor, lead to low level of schooling. The average years of schooling for extreme poor children age 10-19 years old is only 3.4, about one-half the averageyears of schooling for the non-poor; Table 8.5. 8.20 The mainreasonfor not attending school is affordability; Table 8.6. Forty two percent of children aged 7-11 years old in extreme poverty gave "money" as the main reason for not attending school. One inevery two childreninrural areas gave the same reasonfor not attending school. Family or health problems are the second reason given for not attending school. Note that distance is not an important reason for not attending school inrural areas, a reflectionof the effort to increasecoverageofthe education systeminthe last severalyears. Table 8.6 ReasonsFor Not AttendingSchool, 2004 (7-llyear olds) All Extreme Moderately Non- Urban Rural Poor Poor Poor Finishedstudies 0.8 1.6 0.0 0.0 1.4 0.0 Not interested 9.8 5.9 15.5 11.1 9.7 9.9 Needto work 2.6 1.5 3.5 5.1 3.6 1.2 Domesticduties 0.6 1.o 0.3 0.0 0.9 0.3 Distance 4.9 7.2 3.2 0.0 7.0 1.8 Family/healthproblems 20.8 17.9 25.3 21.8 20.9 20.7 Moneyreasons 40.9 42.2 40.1 38.2 35.7 48.7 Too oldto study 0.0 0.0 0.0 0.0 0.0 0.0 Insecurity 0.0 0.0 0.0 0.0 0.0 0.0 Enrollmentclosedfull 5.2 6.5 1.4 9.0 2.9 8.6 Levelnot offered 1.5 0.9 3.1 0.0 2.5 0.0 Got marriedor got pregnlant 1.3 0.0 0.0 0.0 0.0 0.0 Didnot liketeacher 11.5 0.0 3.9 0.0 1.1 1.5 Other 0.1 15.5 3.7 14.9 14.4 7.4 Source: "Honduras Poverty Assessment". World Bank 2006. Vol. I1Table 16, p. 126basedonthe LSMS 2004. 8.21 Low secondary enrollment.122 Enrollment at secondary level is very low: 35.2 percent overall and 7.1 percent for the extreme poor children; Table 8.4. As expected, secondary enrollment inrural areas (19.5 percent) is muchlower than inurbanareas (52.6 percent). About 678,000 children and teenagers aged 12-17 years do not attend secondary school. 122. Inthis study we used the UNESCO defmitionof the secondary levelwhich includes the third cycle of basic educationand"educaci6nmedia". 156 8.22 Vulnerable childredteens. Among the most vulnerable are children that do not attend school becausethey needto work or because of premature pregnancies, as well as children that are inactive. Childrenthat work may be employed inhazard activities and/or may have no time to attend school. Without schooling they are likely to remain poor. According to INE's Household Survey data of May 2006, there are about 17,660 children aged 5-11 years (1.2 percent of the cohort) that work and study, and another 6,000 that only work (0.4 percent of the cohort); Table 8.7. Ifteens betweenaged 12-18 years are added, there are 115,750 children and teens (4.1 percent o f the cohort) that work and study and another 266,280 that only work (9.5 percent of the cohort). O f the group that work and study, 69 percent are boys and 31 percent girls, while ofthe group that only work 82 percent are boys and28 percent are girls. Table8.7 ChildrenandTeenagersOccupations,2006 Age Total Works Only Neither Works Only Neither and Works studies and Works Works Study nor Studies nor works studies 5-11 1,454,623 17,662 6,012 157,789 1.2 0.4 10.8 12-15 803,695 54,384 91,053 115,547 6.8 11.3 14.4 16-18 550,796 43,704 169,219 132,676 7.9 30.7 24.1 Total 2,809,114 115,750 266,284 406,012 4.1 9.5 14.5 Source: XXXII EncuestaPermanente de Hogares. ME, Mayo 2006. 8.23 The majority of children aged 13-18 years that work do so in agriculture (60 percent). Commerce accounts for another 17 percent; and the remaining 23 percent i s distributed across the other sectors, including industry, construction and mining.'23 The average income for this group i s below the minimum wage and increases with age. The average number of years of schooling is about the same for the group 13-15 years old that work and study and for those that only work (4.8 years). For the group aged 16-18 years, however, the average schooling is 7.2 years for those that work and study, comparedto 5.2 years for those that only work. 8.24 Poor teenage girls that become pregnant usually have to leave school as they must work to raise their children. In 2004, there were 21,600 poor girls aged 12-17 years that became pregnant or had a child in the last five years (8.6 percent of the cohort), with the prevalence being higher for the moderately poor group (10 percent) and in rural areas; Table 8.8.124 A study conducted by the IDB in 2002 found that 30 percent of the teenagers interviewed had sexual relations, which exposed them to early pregnancies and also to sexually transmitted diseases, such as HIV/AIDS.'25The likelihood that these teens and their babies remain poor is very high. Indeed, teen pregnancy is a major cause of the intergenerational transmission of poverty. 123. "Trabajo Infantil en Honduras". INE.Encuesta de Hogares de Septiembre, 2002. 124. For girls 10-19 years old in 2005, the Ministry o f Health reports 32,162 pregnancies o f a total o f 117,048 (27.5 percent). Source: Department of Statistics, Ministry of Health, based on "Boletin Ambulatorio de Estadistica". 125. "Caracterizacih de la Adolescencia en Hondura". INE.Octubre 2003. Based on the "Estudio de 10s Valores en 10s J6venes de Tegucigalpa y Choluteca". BID, 2002. The same study found that 25 percent o f the teens hadtaken alcoholic drinks and 10percent had experienced other drugs (marihuana andcocaine). 157 Table 8.8 Incidenceof Teen Pregnancies,2004 Girls that Got Pregnant No. of Girls Y O (12-17 years old) (12-17 years Old) All 37,223 515,217 7.2 ExtremePoor 7,211 109,310 6.6 Moderately Poor 14,421 143,066 10.1 Poor 21,632 252,376 8.6 Non-Poor 15,591 262,84 1 5.9 Urban 14,798 259,126 5.7 Rural 22,425 256,091 8.8 No of observations 182 2,850 Expandedobservations 37,223 515,217 Note: Girls age 12-17years old that got pregnantor hada child inthe last five years. Source:ENCOVI 2004. 8.25 Inactivity is another major risk for teens. In Honduras 406,000 children and teens aged 5-18 years, or one of every six childrenand teens, neither worked nor studied; Table 8.7. O fthis total 65 percent are girls (263,670) and 35 percent (142,330) are boys. It i s likely that many of the girls are not inactive inthe true sense of the word, but are at home doing chores and taking care of their young brothers and sisters, which prevents them from being in school. As for idle boys, an increasing number join gangs and enter a life of crime. In Honduras, 62 percent of teenage deaths are related to violence. Homicides cause 25 percent of total teenage deaths and are six times more prevalent among boys than girls.'26 Vulnerable Group: 18-64 Year Olds 8.26 The major risk facing the poor working population in Honduras is that their income is low and variable because they have poorly paid and insecurejobs. Low paid jobs often result from low productivity, which in turn is associated with low educational attainment and lack of skills. 8.27 Low and unstable income. The main income generating asset of the poor is their labor. The rates of unemployment and underemployment are good indicators of the capacity o f this cohort to generate income. In2006, about 96,400 persons (3.5 percent of the labor force) were unemployed and near 800,000 persons (30 percent) were invisibly underemployed, that i s they worked more than 36 hours but received less than a minimum salary; Table 8.9. Therefore, having ajob inHonduras is no necessarilyaticket to exit poverty. 126. "Caracterizacih de laAdolescencia en Hondura". NE. Octubre 2003 (p. 12). 158 Table 8.9 Unemployment and Underemployment,2006 Total Urban Rural Total Urban Rural Number (order than 10years) YOlabor force OpenUnemployment 96,414 71.034 25.381 3.5 5.2 1.8 InvisibleUnder&nployment 799270 267215 532055 29.6 20.7 37.8 Visible underemploymentb' 135845 66772 69073 5.0 5.2 4.9 Notes: aiOccupiedwho work less than 36 hoursa week andwishto work more.bi Workedmorethan36hours a week butreceivedless thanminimumsalary Source: XXXII EncuestaPermanente de Hogares.INE, Mayo 2006. 8.28 Invisible underemployment is particularly pervasive in rural areas (37.8 percent), where low productivity activities, particularly in traditional agriculture, yield very low incomes to farmers. Low productivity is often associated with low educational attainment. Illiteracy in Honduras is still high particularly among the very poor (40 percent for the first income quintile) and inruralareas (26 percent); Figure 8.1. Overall illiteracy levels are slightly lower for women (17.3 percent) than for men (17.6 percent). Women have higher illiteracy rates than men in urban areas, but lower inrural areas. Figure8.1: Illiteracy Ratesfor the PopulationOlder Than 15 Years, 2006 50 n 40 In f 0 30 0 2 n 20 Q 10 0 Total Q1 Q2 Q3 Q4 Q5 Urban Rural Poorest Richest Men worne en Source: "XWIIEncuesta Permanente de Hogares". INE, May 2006. 8.29 Overall average education attainment is only 5.6 years; Figure 8.2. It i s slightly higher for women (5.8 years) than for men (5.4 years). While the very poor (Ql) have 4 years of schooling on average, the rich (Q5) have more than twice that amount (9.7 years). Average school attainment is lower inrural areas (4.3 years) than inurbanareas (7 years). 159 Figure 8.2: Average Years of Schoolingfor the Population Older than 15 Years, 2006 years 10 8 6 4 2 0 Total Q1 Q2 Q3 Q4 Q Urban Rural Poorest Richest Source: ``XYXIIEncuesta Permanente de Hogares". INE,May 2006. Vulnerable Group: 65-t Year Olds 8.30 The principalrisk facing senior citizens is that they reachthe retiring age with no pension or medical insurance, and must continue to work or depend on family or charity for their survival. 8.3 1 Lack of pension and medical insurance. According to the ENCOVI, only 5.2 percent of the population older than 65 years (the age of retirement for meninthe Honduran Social Security Institute, IHSS) has a pension, with the poor faring worse (less than one percent) than the non- poor (8.7 percent).'27 Only 1.6 percent of the seniors that live in the rural areas have a pension compared to 9.5 percent of those that live in urban areas; Table 8.10. The coverage of medical insurance is similarly low. As a result, there are about 280,000 seniors that have no pension. O f thistotal, 129,000 are poor. Table 8.10 Population of 65 Years and Older With Pension and Health Insurance, 2004 With Number With Pension Without Medical NumberWithout (YOcohort) Pension Insurance MedicalInsurance All 5.2 280,366 6.5 276,601 Extreme Poor 0.7 55,193 0.4 55,328 Moderately Poor 0.9 74,202 1.9 73,472 Poor 0.8 129,395 1.3 128,799 Non-Poor 8.7 150,97 1 10.6 147,802 Urban 9.5 123,168 11.8 120,034 Rural 1.6 157,198 2.0 156,567 No of observations 1,593 1,551 Expanded observations 280,366 276,601 Source: ENCOVI2004. 127. IHSS's retiring age is 65 for menand 60 for women with a minimumof 180 quotas. 160 Vulnerability of the Population in General 8.32 For the population in general, the major risks they face is not having access to health services, living in inadequate housing, or having no access to basic services such as potable water and electricity. 8.33 Health. According to the National Health Plan 2021, the Ministry of Health (MOH) covers about 60 percent of the population with regular health services and 4 percent through itinerant services which delivers a the minimumpackage of services; the IHSS, 18 percent; the private sector 5 percent; and the remaining 13 percent "does not have access to any health services". 12* Consequently, according to the M O H there are 900 million Hondurans that do not have, at least, adequate access to medical services. 8.34 IHSS reports that in 2006 it had 554,855 direct affiliates equivalent to 20 percent of the labor force and, including spouses and children up to 11 years old, it had a total membership o f 1,226,230, equivalent to 16.6 percent of the total population. This population i s concentratedin the 18 most populated cities o f Honduras. For 2004, IHSS reported 458,844 direct affiliates and total membership of 981,635 or 14 percent of the total population. This implies a large (21 percent) increaseinaffiliates intwo years. 8.35 Estimates based on the ENCOVI indicate that in 2004 there were 322,788 members of the labor force that had a healthinsurancewith IHSS, equivalent to 12percent ofthe labor force. About 85 percent of the labor force does not have health insurance, as coverage concentratedon the non-poor andinurban areas; Table 8.11. Table 8.11: Members of the Labor Forcewith Health Insurance,2004 (Percentage oftotal labor force) Not covered Covered IHSS Private Other All 84.5 15.5 12.0 2.5 1.o Extreme Poor 99.0 1.o 0.8 0.2 0.1 Moderately Poor 93.9 6.1 5.0 0.9 0.2 Poor 96.1 3.9 3.2 0.6 0.1 Non-Poor 75.0 25.0 19.2 4.1 1.7 Urban 75.1 24.9 19.6 3.9 1.3 Rural 94.5 5.5 3.9 1.o 0.6 Source: ENCOVI 2004. 8.36 When ill,less than one-third o f the poor visit a facility to seek treatment; see Table 6.17 inChapter VI. Ofthe poor that seek help, most visit the CESAMOs or CESARs. Very few visit IHSS hospitals because they are not affiliated. About 45 percent o f the non poor seek help in M O Hfacilities when illand another 41percent seek helpinprivate facilities. 8.37 One of every four poor people that does not seek care when illgives "no money" as the reason; Table 6.18 inChapter VI. For the extremepoor, 28 percent o f those that do not seek help give the same reason. Note that the reason the facility is "too far" is given by 4.5 percent o f 128. Plan Nacional de Salud 2021, Secretaria de Estado en el Despacho de Salud, Tegucigalpa, Diciembre 2005 (P. 30). 161 those living in rural areas, where physical access to medical care is, as expected, more problematic. 8.38 To summarize, there are a large number of Hondurans without adequate access to health care services. Access to health insurance is even quite limitedamong the non-poor. When ill, two thirds of the poor do not seek help mainly because they cannot afford it. O f the third that seeks help, most visit MOH facilities. Almost one-half of the non-poor also use MOH facilities. Table 8.12 Inadequateand OvercrowdedHousing,2004 (% of households) Inadequatehousing '` Overcrowded Housing b' All 37.8 15.6 Extreme Poor 76.1 31.5 Moderately Poor 50.4 20.7 Poor 61.4 25.3 Non-Poor 19.6 8.0 Urban 18.0 13.6 Rural 58.5 17.6 Note: a/ Inadequate walls, floor, or ceiling. b/ More than 3 persons per room. Source: "HondurasPoverty Assessment".World Bank 2006. Vol 1, Table A. 1.4. p. 86 8.39 Housingc Housing conditions in Honduras are extremely poor, a situation that has been aggravated by natural disasters, such as hurricane Mitch in 1998. According to the ENCOVI, over one third of existing housing i s inadequateinterms o f materials usedinthe halls or ceilings, or because of unfinished floors; Table 8.12. Two thirds of the extreme poor households live in inadequate housing, with this problem being more severe in rural areas (58.5 percent) than in urbanareas (19.6 percent). Moreover, there are about 200,000 householdsliving inovercrowded conditions (more than three people per room). O f the poor families, one of every four lives in overcrowded conditions. Overcrowding is more prevalent in rural areas (17.6 percent) than in urbanareas (13.6 percent), probably because ofhigher incidence ofpoverty inrural areas. Table 8.13 HouseholdsWithout Basic Services, 2004 (% of Households) All Extreme Moderately Non-Poor Poor Poor Without Safe Water a/ 19.9 41.1 27.1 9.3 Urban 6.5 23.4 14.1 3.8 Rural 34.1 43.9 34.4 24.4 Without Sanitation b/ 18.0 44.5 24.1 6.5 Urban 5.2 28.1 11.5 2.3 Rural 31.4 47.0 31.1 16.6 Without Electricity c/ 33.0 81.3 49.0 10.4 Urban 7.0 50.0 16.0 2.0 Rural 60.0 87.0 65.0 31.0 No. of Households 1,424,498 266,862 357,419 800,2 17 Note: The Household Survey (May 2006) estimates:16.2 percent housing units without adequate water services; 21percent without sanitation; and 30.2 percent without electricity. Source: "HondurasPoverty Assessment".World Bank2006. Table A2-H.l- p. 133-135. 8.40 Basic Services. The coverage of essential services such as potable water and sanitation and electricity is still quite limitedand the quality of services inmany areas is poor. There are 283,500 households (20 percent) without safe water, 256,400 (18 percent) without sanitation and 470,100 (33 percent) without electricity; Table 8.13. Forty four percent o f extreme poor 162 households inthe rural area have no safe water and 87 percenthas no energy. Overall, about one of every four poor families has no water or sanitation and one in every two households has no electricity. 8.41 The quality of water services is very deficient. In Tegucigalpa, there are many neighborhoods that receive water only a few hours a day during few days of the week. The coverage of public systems is limitedanda large proportion ofthe water services are provided by the private sector. Of the families that have piped water, 61 percent is connected to a private systemand 39 percentto a public system; Table 8.14. Ofthose that havepipedwater, 58 percent have the pipe outside their housedand only 42 percent have it indoors. Table 8.14 Households with Access to Water Services, 2004 (% of households) Has Piped Water Other Locationof Pipes Total o/w to o/w to Type of Within Outside Public Private Access a/ Home Home System System All 80.1 39.4 60.6 19.9 42.1 57.9 ExtremePoor 58.9 10.0 90.0 41.1 9.2 90.8 Moderately Poor 72.4 27.0 73.0 27.6 19.6 80.4 Non-Poor 90.7 50.2 49.8 9.3 57.2 42.8 Urban 93.8 62.6 37.4 6.2 54.6 45.4 Rural 65.8 5.0 95.0 34.2 23.4 76.6 Note: a/Public faucet or tank, well with rope, well with pump, rivers/stream/lake/lagoon/spring, car with cistern, pickup truck with barrels, other.b'Totalaccess multipliedby coverageofpublic system Source: "HondurasPoverty Assessment". World Bank 2006. Vol 11, Table 75 8.42 O f the householdsthat have electricity (67 percent of the total), the majority is connected to a public system. Private systems and own generators account for very little of the overall coverage; Table 8.15. A World Bank study reported that in 2004 the incidence, length, and impact of power outages in Honduras were worse than in other Central American countries and comparable to Bangladesh. It is mher indicated that the circuits in Tegucigalpa and San Pedro Sula "suffer @om overloads and low voltage, whichfurther reduces the quality ofpower myflies and increases the risk of additional cost to consumers @om motor burnout,for example)". Table 8.15 Households with Access to Electricity, 2004 (% of households) Public Private Public & Own With Gas Other system collective Private Generato Electricity lamp a/ (1) system (3) = (1)+(2) r (5)=(3)+ (6) (7) (2) (4) (4) Total 65.0 1.2 66.2 0.8 67.0 16.3 16.7 ExtremePoor 17.4 0.2 17.6 1.1 18.7 41.4 39.9 Moderately 50.8 0.8 51.6 0.3 51.9 23.3 24.8 Poor Poor 36.6 0.6 37.2 0.8 38.0 31.0 31.0 Non-Poor 87.1 1.7 88.8 0.8 89.6 4.8 5.6 Urban 92.2 0.6 92.8 0.5 93.3 2.4 4.3 Rural 36.6 1.8 38.4 1.1 39.5 30.8 29.7 Note:a/ Solar energy, candles, other. Source: "HondurasPoverty Assessment". WorldBank 2006. A2-H.l- p. 133-135 129. "Honduras- InvestmentClimate Assessment". WorldBank,November2004, p.52 163 8.43 Table 8.15 indicates that of the one-third of the households that do not have electricity, about one-half uses gas (lamps) and another one-half uses other sources (solar energy, candles, etc.). Of the 62 percent of the poor households that do not have electricity, 31percent use lamp fuel with gas and another 31percent use candles and other materials. Vulnerable Group: Indigenous and Black Population 8.44 The indigenous and black population of Honduras represents about 7 percent of the total population, distributed by 9 main ethnic groups descendants of the Mayans, Aztecs, Hokan, Sioux and Afri~an.'~' The Garifimas and Lencas comprise nearly two thirds of the total indigenous and black population; Table 8.16. The main risks that this population faces are low education attainment because o f lack of intercultural bilingual education, high prevalence of diseases suchas chagas, malaria and leishmaniasis, andexclusion. Table 8.16 Indigenousand BlackPopulation Ethnical Estimated YO Location/ Departament Groups Population Garihdas 200,000 40.77 Atlhntida, CortBs, Colon, Gracias a Dios Lencas 110,000 22.42 Intibuch, Lempira, LaPaz, FranciscoMorazhn, Valle, Santa Biirbara Islefios 80,000 16.31 Islasda Bahia, Colon, CortCs, Atlhntida Chorti 68,000 13.86 Ocotepeque, Cophn Miskitos 64,000 13.05 GraciasaDios, Colon Tolupanes 25,000 5.10 Yoro, FranciscoMorazan Pech 2,900 0.59 Olancho, Colon, GraciasaDios Tawahkas 1,353 0.28 Olancho, GraciasaDios Nahoas 1,300 0.27 Olancho Total 490,553 100.00 Source: Gloria LaraPinto "PerJil de 10s Pueblos Indigenas y Nepos". RUTA,2002, Annex 7, p. 120. 8.45 The UNDP's Human Development Report 2006 indicates that the indigenous people fares worse than the rest of the population in key indicators: illiteracy, 22.8 percent for the indigenous population comparedto 17.5 percent for the rest o f the population; no access to piped water, 26.8 percent of the indigenous population versus 19.2 percent for the other groups; no sanitation, 28.2 percent versus 17.4 percent. It further indicates that the majority of the indigenous population has inadequate housing. Low education attainment reduces the changes for the indigenous population to exit poverty; deteriorated housing conditions and lack of access to services, makesthe indigenous population more vulnerable to disease. 8.46 Indeed, according to a UNICEFhJational Institute for Children and Family (IHNFA) study, the ethnic minorities of Honduras face serious healthproblems. 13'The life expectancy is "36 years for the men and of 42 years for the women, one-half the national average of 67.5 and 72.6, respectively. Chronic undernourishment affects 80 percent children under 7 years". And there is a clear overlap between the areas where the ethnic minorities live and the rates of poverty andunsatisfied needs. 130. The XVI Population Census of 2001 (based on self-identification) estimated the indigenous and black population at 440,323 or 7.2 percent oftotal population. 131. "Informe sobre la Situacion de la Naez con Relacion a las Metas de la Cumbre Mundial a favor de la Infancia". IHNFA /UNICEF", marzo 2001 (p. 7). 164 Table 8.17 PrincipalDiseasesthat Affect the IndigenousandBlackPopulation Disease Mostly Affected Number of Cases Incidence Groups Nationwide2005 o/ Chagas Tolupanes, Chorti, Lencas 160 Malaria Garifimas, Islanders, Miskitos and Tawahkas 16,007 Leishmaniasis Pech 969 HIV/AIDS Garifimas, Islanders b/ Note: a/MinistryofHealth. 802 cases nationwide reportedby Ministry ofHealth. '' ``Secretariade Salud" y OPS/OMS. Source: Department of Statistics, Ministry of Health, based on "Boletin Ambulatorio de Estadistica". "Salud de 10s Pueblos Indigenas y Negros de Honduras, Interculturalidady Procesos de Convergencia".Secretaria de Salud" y OPS/OMS, Febrero de 2001. 8.47 The indigenous areas have a high prevalence of transmissibleand vector diseases such as chagas, malaria and leishmaniasis; Table 8.17. A study by the Ministry of Health indicates that the "Garifima, Islanders, and Misquito towns that are located in low ground where water stagnates (coastal zones and lagoons) suffer from high incidence of malaria. The Toluph register high incidence of chagas disease of and acute respiratory infections particularly tuberculosis; the Pech, high leishmaniasis incidence; the Lencas and Chorti, acute respiratory infections pneumonias.; and the black towns (Garifima and Islanders) are being affected increasingly by the VIWSIDA. O f each 100 indigenous children who are born, 68 die of infected-contagious diseases9, .132 8.48 Finally, the indigenous population also faces the risk of exclusion, both because o f physical isolation and because of a lack of opportunities to participate fully in the political, social, cultural and economic affairs of their country. C. The Principal SocialProtection Programs 8.49 This section reviewsthe principalpublic programs that are being implementedto address the risks identified,followed by a detailed analysis ofthe mainpublic subsidizationprograms. It focuses only on key public programs and does not cover programs that are financed and/or managedbyNGOs. Overview 8.50 As mentioned at the outset, social protection programs comprise social assistance (including subsidies) and social insurance. InHonduras, the main institutions that executed the social assistance programs are the PRAF, FHIS, IHNFAandthe line ministries. 8.51 The FamilyAllowances Program(PRAF) is a deconcentratedagency ofthe Presidency of the Republic created in 1991. The Executive Director of the PRAF i s a member of the Social Cabinet, the government's senior coordinating body for social policy, and participates in the Social Safety Net Roundtable. Inthe last 15 years the PRAF has executed a number o f projects for the most vulnerable segments of the population including conditional cash transfer (CCT) 132. Salud de 10s Pueblos Indigenas y Negros de Honduras, Interculturalidad y Procesos de Convergencia" Secretaria de Salud" y OPS/OMS, febrero de 2001, pagina 11. 165 projects financed mostly by the IDB. Among the CCT projects are the maternal infantile and school vouchers. According to the Ministry o f Finance, PRAF budget in 2005 was Lps. 334 millions equivalent to 0.9 percent o f total Central Government expenditures. 8.52 Honduras' Social Investment Fund (FHIS) is another key player in social protection. It finances the country's basic social infrastructure. It was created in 1990 as a deconcentrated agency of limited duration to finance small-scale infrastructure projects in poor communities. The government has extended and substantively modified the institutional mandate o f FHIS several times since 1990, allowing the organization to respond to emergencies (such as post- hurricane Mitch reconstruction) and extend its operations to more communities. The most recent Congressional reforms o f FHIS (in 1999) extended the duration o f FHIS through 2012 and marked a new direction in its institutional orientation. This legislation calls for FHIS to support the process o f decentralization and municipal capacity building, whereby FHIS is to eventually transfer its operational, technical and financial functions to municipalities. In 2005, FHIS executed a budget o f Lps. 1, 246 million, or 3.4 percent o f total Central Government expenditures. 8.53 The National Institute for Children and Family (IHNFA), i s an autonomous institution with responsibility for formulating and overseeing public policies in the areas o f children and families and helping address the issues facing vulnerable groups. It was created in 1997 with the mission o f protecting children and the family, within the framework o f Constitution o f the Republic, the Code o f the Childhood and the Adolescence, the Family Code, the Convention on the Rights o f the Children and other related conventions subscribed to by Honduras. Its programs include centers to care for children at risk, the placement o f children in foster homes, and adoptions. IHNFA executed a budget was Lps. 139 million in2005, or 0.4 percent o f total Central Government expenditures. 8.54 With regard to social insurance, there are three main public pension systems: the Honduran Social Security Institute (IHSS) that covers the employees o f private enterprises, IMPREMA that covers teachers, and INJUPE that covers other public employees. The coverage, contributions and benefits o f these intuitions are discussed below (and in Chapter 11). The other two public pension systems - the military pension fund (IPM) and the national university staff fund (IMPREUNAH)- are smaller and not reviewed. Vulnerable Group: 0-5 Year Olds 8.55 To address the main risks facing this group, the Government o f Honduras relies on the following key programs. The Ministry o f Health has several maternal-infant programs, including conditional cash transfers through PRAF (one supported by the IDB and the other financed directly by the Government), which require mothers and young children to attend periodic controls at health clinics. Prenatal care has improved and an increasing number o f women have institutionally assisted deliveries. Still, only one-third o f all deliveries by extremely poor women are attended by a doctor (Table 20) and 7 percent o f all newborns have low weight at birth. 166 Table 8.18 Place ofDeliveryandBirth Attendance, 2004 Place of Delivery Birth attended by: M O H IHSS Private Maternal Home Other Doctor Midwife Nurse Other hospital hospital Hospital Clinic All 55.4 7.2 6.1 2.2 28.3 0.8 67.1 25.3 4.4 3.2 ExtremePoor 36.4 0.3 1.7 0.8 59.9 0.9 35.5 52.8 4.7 7.0 Moderately 60.1 3.4 2.6 3.2 29.4 1.2 63.6 26.6 6.3 3.5 Poor Non-Poor 62.7 13.3 10.7 2.3 10.5 0.6 86.5 9.5 3.1 0.9 Urban 65.0 12.8 8.6 1.8 11.1 0.7 85.1 9.8 3.6 1.4 Rural 46.7 2.1 3.8 2.6 43.8 1.0 50.9 39.2 5.2 4.7 Source: "HondurasPoverty Assessment". World Bank 2006. Vol. I. A.2 F.19, p. 129. Table 8.56 Vaccination coverage of children under 5 years old in Honduras has improved significantly across regions and income groups in recent years. In 2004, over 93 percent of all children under five received the BCG, Polio, and measles vaccinations. Nonetheless, as indicated in the Report on the Millenniums Development Goals Progress Report for Honduras (UNDP 2003), the infant mortality rate and the mortality rate for children under 5 years remain at 34 per 1,000 and 45 per 1,000, respectively, making it difficult to reach the target of 13 per 1,000 and 18 per 1,000 in2015.'33 Table 8.19 Vaccination in ChildrenUnder 5 Years Old, 2004 (Percentages) BCG Sabin/Polio DPT MMR All 95.5 94.0 80.1 93.7 ExtremePoor 95.2 94.0 81.7 93.1 ModeratelyPoor 95.6 94.0 79.3 94.3 Non-Poor 95.6 93.9 79.5 93.7 Urban 96.0 94.2 79.5 94.0 Rural 95.0 93.8 80.5 93.4 Note: Against tuberculosis (BCG), polio (Sabin), diphtheria,whoopingcough, tetanus, andhepatitis-meningitis (DPT), andmeasles, mumpsandrubella(MMR) Source: "Honduras Poverty Assessment". WorldBank2006. Vol. 11, Table 55, p. 151. 8.57 There are several programs to address malnutrition in children, including the Comprehensive Community Child Health Services (AIN-C), a recently initiated program that concentrateson childrenunder 2 years old.134 The Ministryof Health is implementing aprogram with support from the World FoodProgram for vulnerable groups includingpoor children under 5 years at risk of malnutritionand pregnant and lactating women inmunicipalities with very poor health services coverage. IHNFA runs 39 centers of Integral Attention to Children (CANES), which provide nutrition attention 3,600 children. Despite these and other efforts, one of every five children under five (198,000 children) still suffers from chronic malnutrition. 8.58 Early stimulation remains a challenge. There are very few community-based centersthat provide support to infants when mothers are at work. IHNFA has Community Centers for Attention to Children, Community Homes for Attention to Children and Madres de Guias (or GuidingMothers) less than 6 years, but these programs cover only a few thousand childrenwhen there are over 700,000 poor children under five. Some municipalities also support kindergartens, but coverageis small. 133. "Honduras-Informe Sobre las Metas delMilenio". UNDP2003. pp. 34,35. 134. The AN-C strategy i s supportedby the WorldBank andfollowsup on a previousUSAIDprogram. 167 8.59 The coverage of preprimary school, either formal or informal, has increased in recent years and now covers 190,000 children. Among the programs that are supporting the expansion of preprimary education are the Education for All, Community Education, PROHECO, and Foundation FEREMA programs.135 They support the establishment of Community Centers for Prebasic Education (CCEPREB) in rural and peri-urban communities where there i s no preprimary education of any type (formal or informal). These centers are managed by voluntary per~onne1.l~~ In2005, EFA, trained 903 voluntary educators, organized 2,543 CCEPREB, and enrolled 13,242 children. It seeks to enroll all children 5 years old inpreprimary education and the in primary education, so that all children complete 6 grades of primary school by 2015. Despite these efforts, one of every five children in the ages of 3 to 5 years (500,000) do not attendpreprimary school. VulnerableGroup: 6-17 Year Olds 8.60 To address the main risk of this group there are several programs that promote the participation of children in schools including Conditional Cash Transfers through PRAF, programs to reduce desertion and repetition (such as Let U s Save the First Grade with the support of the Organization of American States) and programs to extend the coverage o f preprimary and primary school inrural, more isolated areas (such as PROHECO which enrolled 122,000 children in 2,136 schools in2005). The Healthy School Program (EscuelasSaludables) offers a school lunch and medical attention to over one million students in 13,662 education centers, and provides incentives for studentsto remain inschool and contributes to keep students healthy and better prepared to learn. The school lunch, which is financed by the Government and the World Food Program, distributesa packageof rice, beans, maize, oil, soy beans and CSB (codsoy blend) inrural areas and a nutritivebiscuitinurbanareas. 8.61 At the secondarylevel, efforts to increasecoverageincludes the IDB-supported 3rd Cycle program that seeks to take the teaching of grades 9-12 to rural areas (enrolling 60,000 students in 2005), and the distance education program TELEBASICA, which also focuses on grade 9-12 in rural areas and peri-urban areas (enrolling 5,707 students in 65 Centers of Basic Education, 90 percentof them inrural areas). 8.62 There are several scholarship programs which benefit over 28,000 children in 2005. There are scholarships for those who excel in basic, secondary and higher education, for poor studentswith good grades at basic and secondary level. Inaddition, the Government pays for the graduation costs of about 60,000 studentsthat graduate from preprimary education inrural areas and secondary education. For children that needspecial education there are very few facilities or programs. In 2005, IHNFA provided special education to 280 children and special training to another 120 children. It should be mentioned that beginning in2006 students that attend public schools (preprimary, primary and secondary education) are not required to pay registration (Matricula Gratis), apromise madeby PresidentZelaya during his campaign. 135. The Community Education Project supports de PROHECO schools. In 2005, it fmanced 1,052 schools and 1,674 teachers in 11Department. 136. CCEPREB are community centers for prebasic educationattendedby a trained volunteer for usually up to 20 students. CEPREB are center for prebasic education in or outside the school attendedby a basic education teacher for up to 35 students. There are also CCEPREB iteractive (PREPI) and other preprimary informal education centers. The total number of CCEPREB, CEPREB and CCEPREB PREPI are about 3,200 with 23,000 students. 168 8.63 Several programs are directed at vulnerable children and teenagers. INFHA's Social Intervention Program helps children that have been abandoned or sexually exploited. INFHA's Social Education Program helpsteenagers that have trouble with the law. PRAF offers a CCT to teenagers that are out of school to enroll in training (Youth Voucher and Friendly Hand). The FHIS finances projects developed at local levels directed at vulnerable groups, such as children andfamilies inextremepoverty through the World Bank-supported Development andAssistance Innovation Fund (FIDAS). The Ministry of Education, the Ministry of Labor, and two NGOs (Save the Children UK and Plan International) are implementing a program to increasethe level o f schooling of 1,221 children involved in the worst forms of youth labor. Despite all these efforts, many children remain that do not attend primary school (200,000), secondary school (678,000), and children that do not study or work (406,000 age 5-18 years). VulnerableGroup: 18-64 Year Olds 8.64 There are two major literacy programs for teenagers that have dropped out of school and for adults: EDUCATODOS and PRALEBAH. EDUCATODOS is supported by USAID and use iterative methods to offer basic education to youth at risk, female heads of households, migrants and indigenous population, mostly between 13 and 35 years of age. In2005, it covered 76,300 persons. PRALEBAH is financedby the Government and focuses on the population aged between 14 and 39 years that lives in rural or peri-urban areas and that did not complete basic education. It also offers vocational training courses, enrolling 37,000 people in2005. 8.65 The National Vocational Training Institute, INFOP, offers several training courses to the working age population. Its mobile training courses in different skills reached 3,800 persons in 2005. INFOP also provides training for youth in the rural areas in seven Centers for Family Development as well as training for agricultural workers. 8.66 The PRAF has a program called Family Voucher (Bono Familiar) that seeks to give poor families access to formal credit from Community Banks and Rural Credit Unions and help them develop productive projects. In2005, the program supported 2,000 families. It also supports the development of local community shops (SAP) managed by women; about 4,000 SAPS were establishedin2005. 8.67 Finally, the World Food Program manages a food-for-work program that is activated when a disaster occurs. The characteristics of the programs (duration, benefits, etc.) are tailored to the needs ofthe population affected. Vulnerable Group: 65+ Year Olds 8.68 The public pension plans administered by IHSS, INJUPE and IMPREMA consist of obligatory defined-benefit pay-as-you-go schemes with partial collective capitalization funded through compulsory contributions. These systems only cover 20 percent of the labor force, which i s low compared to the 51 and 49 percent coverage o f the Panamanianand Costa Rican systems, but similar to the coverage achieved in the other Central American countrie~.'~'The pensionspaidby IHSS are low, averaging Lps. 891 per month (US$47). This i s mostly the result of very low contributions and from the maximumsalaries to whichthe contribution applies, Lps. 137. Contributors relative to the labor force. See Caroline Crabble, editor, "A Quarter Century o fPension Reform inLatinAmerican and Caribbean: LessonsLearnedandNextSteps", IDB,2005, (Table 1.12, p. 36). 169 4,800 (Table 8.20). The pensions received under the other two systems are much better: INJUPE Lps.4,165 and IMPREMA Lps.7,795. Thepensioncannot be belowthe minimumwage incase of INJUPE nor can be greater than 90 percent of the average wage in case o f INJUPE and IMPREMA. 8.69 Less than 6 percent of Honduras' seniors have a pension, so many are forced to work or depend on family or charity. PRAF has a voucher program that provides US$31.75 per year (less than 10 cents per day) to seniors inpoverty. Local governments participate inthe selection of beneficiaries. In 2005, the program covered 36,000 seniors of an estimated 129,000 poor seniors without pension. Table 8.20 Honduras-PensionPlans,2005 '` Affiliate Total Employee Employer/Govt. Retirement Age Average Benefi Contri- Contribution Pension ciaries bution (LpsJmonth) IHSS 525,845 1,162, 1%of salary Employer:2% of 65 years man; 60 891 118 a/ b/ salaryb/ years women; Govt.:0.5%" I minimum180 quotas INJUPE 55,493 N/A 7% ofsalary 11%of salary 65 years; minimum 4,165 10years of service IMPREMA _ _ N/A 7%of salary 11%-12%of 50 years; minimum 7,795 salary ' 10years of service Note: a/Excludesde pensionplanofthe military (IPM) andofthe staff ofthe national university (INPREUNAH). ` With a maximum of Lps. 4,800. " Additional contribution to the health insurance are: 2.5 percent employee; 5 percentemployerand0.5 percentgovernment. There is also a contribution from the employer to professionalriskof 0.2percent.'11percent for privateor semi-officialschools; 12percent. Source: IHSS, IMPREMAandINJUPE Vulnerable Population in General 8.70 The MinistryofHealth hastwo important programsthat seek to reachthe 900,000 people (13 percent of the population) without adequate access to medical services. The Ministry of Health is providinghealth services to people in five Commonwealths and ineight of the poorest and most isolated Departments (Lempira, Olancho, El Paraiso, Gracias de Dios, Santa Barbara, Ocotepeque, Colon and Oro), with support fiom the Swedish Agency for International Development. It also delivered a minimum package of health services to families without regular access to the health system (65,000 people in 2005), with the support of IDB under the PRIESS program (Institutional Reorganization and Extension of Health Services) and the cooperation of NGOs. [This program i s coming to a close, but a new program "Health Improvement Program" (US$ 17 million for 2006-08) was initiated in 2006 that targets women and childrenunder4 inthe poorest Departmentsofthe country.] 8.71 There are several programs to help reduce the poor housing conditions, including overcrowding which affects over 200,000 households. Two programs supported, by Germany's KfW, provide loans and direct subsidies to low income families, with priority given to female heads of households. In 2005, the Urban Housing Program (PRIMHUR) benefited2,900 poor families; and the Rural Minimum Housing Program (PVMR) benefited another 1,500 poor households. Families with income less than four minimum wages are eligible for a subsidy. The Government also has a low income housing program-Housing for the People- that i s managedby BANHPROVI and provides direct subsidies andcredits. The credits can beusedfor 170 the purchase of a plot of land or a housing unit, or for construction and housing improvements. Eligible families are those earning less than 4 minimumsalaries. In2005, this program benefited 1,100 families. The IDB recently approved a major housing program (US$ 31 million) to be managedby the Directorate General of Housing and UrbanDevelopment in SOPTRAVIand that seeks to rationalize the housing financing system andunifythe criteria and mechanismsto award direct subsidies. 8.72 The Honduran Social Investment Fund(FHIS) is the major provider of financing for the development of basic infrastructure inHonduras. Its programs cover both rural and urbanareas, and it is supported by the World Bank, IDB and many other donors. Among other things, the FHIS seeks to reducethe existing basic infrastructure deficits ofthe 283,000 householdswithout safe water, 256,000 without sanitation, and 470,000 without electricity. In2005, FHIS financed US$30million ininfrastructureprojects. Vulnerable Group: Indigenous and Black Population 8.73 There has been some progress in intercultural bilingual education (EIB) during the last few years. The Community Education Project (supported by the World Bank) has helped to improve the quality of preprimary and primary education in rural areas including indigenous communities. It has helpedto adapt and complement the Basic National Curriculum to EIB, and has financed the preparation o f EIBteaching guides andtexts for preprimary andbasic education (Integral Guides for Prebasic and Basic Education and the Guide for reading and writing for Prebasic and first grade in seven languages of the nine indigenous people). The National Education Program for Indigenous and Afro-Antillean Ethnic Groups in Honduras (PRONEEAAH) initiated in 1997 trains teachers in EIB and has helped establish 172 Community Centers for Prebasic Education (CCEPREB) in indigenous communities that enroll 2,739 children. The PRAEMHO Program (Support to Secondary Education Teaching in Honduras) financed by the EC has provided 44 scholarships for teachers o f EIB from the Departments of Olancho, Atlhtida, Colon, and Mas de la Bahia, and another 489 scholarships for teachers of EIB under the modality of "Work for Study". The Ministryof Education is also training teacher in EIB with the support of regional project (PROEIMCA) financed by the Government o f Finland. 8.74 Japan and Canada support a program executed by FHIS that is seeking to combat the Chagas disease that affects mostly the Tolupanes, Chorti, and Lencas. In 2005, the program reached 3,450 families. Other programs to combat vector-transmitted and infectious diseases, such as malaria and VIWAIDS are implementedby the Ministry of Health with the support of various donors. D. The MainSubsidizationProgramsinHonduras 8.75 The Government of Honduras maintains a series of subsidies that seek to make public services more affordable to the population. Some of these subsidies are direct, while others are indirect or hidden. This section reviews the electricity, fuel, transport, water, agriculture, and telephony subsidies.138 138. This section does not consider the subsidies given to senior citizens (those above 60 years) by public and private companies, which usually involve a discount over the price charged to the general public. For 171 Electricity subsidies 8.76 Most electricity subsidies involve ENEE. This is a state owned enterprise that has the monopoly for the transmission and distribution of energy inHonduras. ENEEproduces about 25 percent of the energy it distributes and buys the remaining 75 percent from private generators under longterm contracts. Prior to 2000, its tariffs were set by the NationalEnergy Commission using a formula that took into account (i) the domestic prices of energy purchases and (ii) a margin for the cost of transmission and distribution. ENEE's underlying tariffs were frozen in 2000. Since then a factor of adjustment has been introduced to pass through to consumers the changes ininternational fuel prices and exchange rate adjustments. 8.77 There are five subsidies in the electricity sector:'39 (1) subsidies for households that consume less than 300 kwwmonth, (2) a generalized implicit subsidy for all customers when ENEEsets tariffs below cost (partial cost adjustment subsidy), (3) a cross-subsidymandatedby Article 46 of the Electricity Subsector Framework Law, (4) Bono 80, a subsidy to compensate urbanpopulation for the increase in transport costs that is provided through the electricity bill, and(5) the exoneration oftaxes onfuel oil anddiesel usedto generateelectricity. 8.78 The 300 Kwh Subsidy (also called direct subsidy). InOctober 2006, ENEE had 862,103 residential clients. Of these, 755,453 (88 percent) consumed less than 300 kwh and received a subsidy paid by the Government. Those consuming less than 100 kwh received on average Lps. 14.6 in subsidy per month, while those consuming between 100-300 kwhreceived Lps. 48 per month; Table 8.21. In2005, those consuming less than 100kwhreceive only 22 percent of the total amount of subsidy (Lps. 60 millions); in contrast 78 percent of the subsidy or Lps. 215 million went for those consuming between 100-300kwh. Table 8.21 Beneficiariesof the 300KwhSubsidy, December2005, October 2006 Consumption No. Consumption Amount Billed Subsidy Subsidy kwh Beneficiaries Kwh Lps. Lps. Per (Lps.) Client Kwh December2005 Less than 100 373,729 15,051,802 10,996,266 5,563,846 14.89 0.37 100-300 358,304 64,653,68 1 60,428,301 17,248,329 48.14 0.27 Total 732,033 79,705,483 71,424,567 22,8 12,175 October 2006 Less than 100 366,380 14,3 15,934 10,503,479 5,343,659 14.59 0.37 100-300 389,073 71,538,977 67,194,995 18,680,277 48.01 0.26 Total 755.453 85.854.91 1 77.698.474 24.023.936 Source:ENEE example, in the case of the electricity company, ENEE, this involves a discount of 25 percent on the electricitybill, whichcosts the companyL24 million a year (US$ 1.3 million). A similar benefit is givenby the SANAA,apublicwater company, cinemas, pharmacies,etc. 139. The Government does not get any return on the equity it holds on ENEE. This constitutes a major hidden subsidy that is not considered in what follows. This "subsidy" is estimated at Lps. 2,021 million in 2005, basedona stockof equity of Lps.20,208 million anda 10percentopportunitycost ofcapital. 172 8.79 Partial Cost Adiustment subsidy. ENEE's underlying tariffs have been frozen since 2000. By law, ENEEis supposedto adjust what it charges its customers when its costs move up or down by at least 5 percent, calculated according to a formula that reflects changes in the international prices of petroleum and the exchange rate. ENEE has made this adjustment only partially, generating an implicit subsidy for all of its costumers, as well as substantial losses. For instance, in December 2005 the change in international prices and the exchange rate movement dictated an adjustment inENEE's prices o f 124.5 percent (the adjustment by the formula), but it only implemented a partial adjustment of 53.5 percent (the average adjustment during January/October 2006). ENEE's losses reachedLps. 1,506 million; Table 8.22. Inthe first nine months of 2006, the adjustments continued to lag the underlying change in costs and have averaged only about one-half of what they should have been according to the formula. The shortfall incost adjustment in2006 has beenrunning at Lps. 250 million per month. Therefore, it is likely that ENEE's overall lossesin2006 are 50 percenthigher thanthe projection received fiom ENEE. 8.80 Article 46 Subsidy. According to Table 8.22 ENEE's IncomeStatement, 2005-2006 Article 46 of the Electricity Subsector (Lps. million) Framework Law, ENEE is to charge 110 Projected 2005 2006 percent of its total cost (TC) of supplying Total Income 8,008 9,140 electricity to residential consumers over 500 Electricity Sales 7,825 8,904 kwh, 100 percent or more of TC to Other Income 183 236 consumers between 301-500 kwh; 80 Total Costs 9,514 11,258 percent or more of TC to consumer of 101- Losses -1,506 -2,119 300 kwh, and 45 percent or more of TC to Source: ENEE consumer between 0-100 kwh. So, for example, for customers that consume 100 Kwh per month, the cost of the kwhbilled (without subsidy) is Lps. 0.73; for those that consume 200 kwh per monthis Lps. 0.96; and for those that consume 300 kwhis Lps. 1.03. This represents a cross subsidy betweenENEEcustomers. 8.81 Bono 80. The Bono 80 (Voucher 80) is a subsidy given since 2005 to the urban population to compensate for increases intransportation costs. It cost the Government Lps. 198 million in2005 and Lps. 550 million in2006. To qualify for the Bono, the householdmust be up to date with its electricity payments and have a functioning meter. The Bono is paid through ENEE's electricity bill. The subsidyof Lps. 80 permonthwas originally meantto betargetedon the poorest consumers (those consuming less than 100 kwh per month), but the National Congress extendedthe benefitto more affluent consumersby increasing the amount o f the Bono to Lps. 100 for those households that consume between 101-240 kwwmonth, to Lps. 110 for those that consume between 241-260 kwwmonth, to Lps. 125 for households that consume between 261-280 kwwmonth, and to Lps. 135 for households that consume between 281-300 kwwmonth. Since the amount is discounted directly from the electricity bill, ENEE finances the subsidy if the Government does not reimburse it. During January-October 2006, ENEE paid Lps. 462 million in subsides, but was only reimbursed Lps. 275 million by the Ministry of Finance, resulting ina shortfall of Lps. 187million. 8.82 During January-October 2006, an average of 457,624 households received Bono 80 every month; Table 8.23. O f these, 199315 (44 percent of those receiving the Bono) were households consuming 100Kwhor less which received a total subsidyo f Lps. 160million (37 percent of the total subsidy), or Lps. 800 per household (Lps. 80 per 10 months). The other 268,764 173 households (56 percent) received Lps. 269 million in subsidy (63 percent), an average o f Lps. 1,041 per family. Of the household consuming 100 kwh or less in October 2006, only one-half received Bono 80 compared to 65 percent o f those consuming between 100-300 kwh that received the Bono. It appears to be more difficult for the poorest consumers to qualify for the Bono. Table 8.23 Bono 80 Subsidy, 2006 Cosnumption Value of January September Kwhtmonth Subsidy Beneficiaries Subsidy " Beneficiaries Subsidy " Lps. (000) Lps. million (000) Lps. million 100 80 209.2 16.7 184.9 14.8 101-240 100 213.1 21.3 204.4 20.4 241-260 110 18.7 2.1 18.8 2.0 261-280 125 15.6 1.9 16.7 2.1 281-300 135 13.8 1.9 14.5 2.0 Total d a 470.4 43.9 439.4 41.3 Note:a/Excludesthe commissions chargedby the banks. Source: ENEE 8.83 Tax exoneration on fuel oil and diesel used to generate electricity. The Government provides a subsidy to consumers o f electricity by exonerating private generators and ENEE from taxes on fuel oil (bunker) and on diesel used for electricity generati~n.'~' The cost to the Treasury o f this tax exoneration was Lps. 1,757 million in 2005 and is projected to cost Lps. 1,795 million in2006; Table 8.24. Table 8.24 FiscalCost of FuelTax Exonerationsfor GeneratingElectricity,2005-2006 (Lps. millones) 2005 2006 Projected Jan-Oct 2006 ENEE 56.4 1,471.9 N/a Private Sector 1,700.2 24.1 N/a Total 1,756.6 1,495.9 1,795.0 Source: Ministry of Finance 8.84 Table 8.25 illustrates how these subsidies would work out for hypothetical consumers o f 100kwh, 200 kwhand 300 kwhper month. First, note that ifthey obtain Bono 80, consumers o f 100 kwh or less might end up paying no electricity or very little, depending on the actual cost adjustment made inthat particular month. Next, consumers in the 200 kwh and 300 kwh range would pay only about one-third and one-half, respectively, o f their electricity bill. Finally, while the underlying tariff per kwh increases with the level o f consumption (cross-subsidy), the absolute value o f the subsidies also increases with the level o f consumption (300 kwh subsidy). The Partial Cost Adjustment subsidy received by the 300 kwh consumer is 4 times that received bythe 100kwhconsumer. 140. This is a specific tax ofUS$0.6106/gallon for diesel and US$0.4267/gallon for bunker "C". 174 Table 8.25 Subsidesto HypotheticalElectricityConsumers,2005/2006 (Lps. and percentages) Monthly Consumption/Payment 100kwh 200kwh 300khw Monthly bill without adjustment (Lps.) 73.00 191.30 309.60 Adjustment by formula (124.5%) (Lps.)a/ 90.80 238.10 385.40 Billwith adjustment (Lps.) 163.80 429.40 695.00 Total Subsidies(Lps.) 168.00 284.90 403.90 300 kwh subsidy (Lps.) 36.20 49.10 49.10 Partial CostAdjustment subsidy (53.5%) (Lps.)a/ 51.80 135.80 219.80 Bono 80 (Lps.) 80.00 100.00 135.00 Net payment (Lps.) -4.10 144.50 291.10 Tariff without adjustment(Lps./kwh) 0.73 0.96 1.03 Note: a/Note that the adjustment estimated by the formula and the actual adjustment implemented varies every month. The adjustment estimated by the formula of 124.5 percent (international price changes and exchange rate movement) is similar to the actual inDecember2005. The partial cost adjustment of 53.5 percent is similar to the average ofthe actualadjustment implementedduring January-October2006. Source: ENEEand staff estimates Fuel Subsidy 8.85 There are two fuel subsidies. One i s to cover the price differential o f fuels andthe second involves a reduction or elimination o f fuel taxes to reduce the price paid by consumers. 8.86 Price Differential. This subsidy was Table 8.26 Fuel Subsidies2005,2006 (Lps. million) introduced on September 8,2005, following Subsidies Actual Estimated a hike in the international prices o f 2005 2006 petroleum products. The Government froze Price Differential a/ b/ 159.4 507.0 the prices o f the different types o f fuels G a s o l i n e ~ e r 67.2 213.7 (gasoline super and regular, diesel, Gasoline regular 7.7 24.4 Diesel 54.0 171.8 kerosene, and liquid petroleum gas, LPG) 6.9 22.0 andpaid the petroleum companies (insome Kerosene LPG 23.6 75.2 cases with a delay) the difference between Reduction intaxes NIA 134.3 the market price (calculated in accordance Total 159.4 641.3 withan establishedfornula) and thefrozen Note: Excludes services stations. b/Distribution in 2005 price. The new Government again froze the Source: basedon actualdistribution of 2006 of Finance prices on April 9, 2006, adjusted them slightly the following two weeks, and then left them frozen again. The Ministry o f Finance estimates that the cost o f this subsidy was Lps. 159.4 million in 2005 and Lps. 507 million in 2006; Table 8.26. 8.87 With the subsequent decline ininternational fuel prices, the price differential between the frozen price and the market price disappeared after September 2006 for gasoline, diesel and kerosene, and has declined substantially for LP; Figure 8.3. Indeed, in November 2006, the market price o f gasoline was 12 percent below the frozen price; and it was below by 8 percent and 6 percent for diesel and kerosene. 175 Table 8.27 Market Price of Fuel, November, 2006 Type of MarketPrice FrozenPrice Difference Nov 5-18,2006 April 30,2006 Absolute. Relative Lps./gl Lps./gl Lps./ gl (YO) Gasoline super 63.52 71.07 -7.55 -11.9 Gasolineregular 56.49 63.44 -6.95 -12.3 Diesel 52.27 56.45 -4.18 -8.0 Kerosene 40.76 43.17 -2.41 -5.9 LPG domestic 36.17 36.03 0.14 0.4 Source: Ministry of Finance Figure 8.3: Differential between Market and Frozen FuelPrices April 9-Nov 18,2006 Lps. per Gallon " -- April 9-22 June 4-10 July 23-29 Sept 10-16 NOV5-18 +Gasolinesuper -IC-Gasoline regular Diesel Kerosene +K- LPGdomestic Source: Ministry of Finance 8.88 Reduction in Fuel Taxes. On February 9, 2006, the National Assembly of Honduras reduced the tax on gasoline from US$1.15 per gallon to US$O.99 (equivalent to Lps. 3 per gallon) and inMarch 16, 2006, it reduced the tax on LPG on containers of 25 lbs inLps. 8, and the tax on keroseneinLps. 5, which was equivalent to eliminatingit. The cost ofthis reduction intaxes is estimatedat Lps. 134.3 millionin2006; Table 8.28. Table 8.28 EstimatedFiscal Costs ofReduction in Fuel Taxes, 2006 Type of Fuel FiscalCost Gasolineregular 79.2 Kerosene 31.9 LPG domestic 23.2 Total 134.3 Source: Ministry of Finance TransportSubsidies 8.89 There are three transport subsidies: a mass transportation subsidy to buses and microbuses, a transport voucher for students, and a subsidyto taxis. 176 8.90 Mass Transportation Subsidy. The public transportation tariff in Tegucigalpa is Lps. 3 for buses and microbuses. The last tariff adjustment took place inMarch 2002. The subsidy i s given to the owner of each bus and microbus in Tegucigalpa. The buses receive Lps. 575 per day, providedthey complete 4 laps intheir designatedroute; the microbuses receive Lps. 320 per day, provided that they complete 8 laps in their designated routes. The Ministry o f Transport employs 40 people to control the work of the buses and microbuses, including supervisors. The appropriation for this subsidy has remained at Lps. 114 million a years during the last years, so that the subsidy is suspendedwhenthe appropriation i s exhausted. 8.91 Studenttransport subsidy. This subsidy is directed at s high school students inthe major cities to helpthem cope with increasesinpublic transport costs. The subsidy is administered by the Ministry of Education. It cost Lps. 30 million in 2005 and Lps. 52.4 million in 2006. In 2006, the subsidy was given in 62 institutes (of an existing total of 263) with each student receiving a cheque for Lps. 400. According to the Ministry o f Education, these institutes are located inthe major cities of 10 Departments. 8.92 Subsidy to Taxis. This subsidy is being given to taxi drivers in seven cities and comprised a Lps. 1,118 one-time payment to eachdriver to compensatefor increase infuel costs. The Ministry of Transport estimates that there are 15,000 taxi drivers but as of the end of October only about 6,625 subsidies were granted at a cost of Lps. 7.4 million. It estimates that Lps. 8.4 million will be spent in2006 to benefit7,500 taxi drivers. Water subsidies 8.93 The state-owned water utility, SANAA, is responsible for providing water and sanitation in 33 cities, including Tegucigalpa. It covers 1.3 million people or 21 percent of the total Honduran population. SANAA operates under the Water and Sanitation Framework Law approved in August 2003, which created the National Commission on Water and Sanitation (CONASA) and a sector regulatory entity (ERSAPS). 8.94 Two types of subsidies have been identified in SANAA.'41One subsidy i s given by setting tariffs that do not reflect SANAA costs, therefore forcing the company to incur operating losses. The second involves cross-subsidies between large and smaller consumers which are implicit inthe tariff structure. 8.95 Tariff subsidy. SANAA's tariffs are highly subsidized. In2003, an attempt was made to raise tariffs, but the authorities were forced to roll back the adjustment after strong popular opposition. SANAA's tariffs have been frozen since then. Table 8.29 presents a comparison o f SANAA's tariffs in Tegucigalpa and the charged tariffs in other Central America cities. While SANAA's rates are only a fractionofthe average inthe other four cities for small consumers, but then catchup for largerconsumers. Overall, SANAAtariffs are about one-half the average level charged inthe other four cities. 8.96 In 2005, SANAA had total revenues of Lps. 381 million and total costs of Lps. 453 million, yielding a recorded loss o f Lps. 72 million. As 85 percent o f SANAA's current income 141. The Government does not get any return on the equity it holds on SANAA. This constitutes a substantial hiddensubsidythat is not consideredabove. The subsidy is estimatedat Lps. 555 millionin2005, basedon a stock of equity valuedat Lps. 5,555 million anda 10percentopportunitycost ofcapital. 177 goes toward paying personnel expenses, it has no resources left for other operating expenses, including its electricity bill owed to ENEE, which amounts to about Lps. 60 million a year. As of August 2006, its accumulated debt to ENNE was Lps. 275 million. On the other hand, according to SANAA, many public institutions (including the municipalities) do not pay their water bills. Public sector overdue payment to SANAA inTegucigalpa alone surpasses Lps. 140 million. (SANAA has indicated that public institutions have been paying their water bills more promptly inrecent months.) 8.97 SANAA'Soperating losses provide an approximation to the size of subsidy granted through its tariff structure. However, those losses also reflect the company's serious operational deficiencies. Currently, SANAA has about 103,000 registered clients. O fthis total, only 60,000 have meters and many o f those are reportedly not working. SANAA'Soperational accounts for 2005 indicate that 57 percent of the water revenues collected are from meters, 22 percent are based on estimates of consumption and 21 percent by established canon. Billing errors are equivalent to six percent of the amount billed. On the other hand, SANAA'Sstaff estimates its technical losses inTegucigalpa at 26 percent.142 Table 8.29 Average ResidentialWater Tariffs in SelectedCities(US$/M3) Consumption Tegucigalpa Average TegucigKA Managua San San Panama MYmonth (1) of the 4 Average Salvador Jose City CA cities (3)= (2) (1)/(2)*100 0-15 0.04 0.40 10.1 0.25 0.18 0.72 0.43 20 0.06 0.33 18.1 0.24 0.16 0.60 0.32 25 0.09 0.35 26.0 0.27 0.26 0.59 0.26 30 0.14 0.33 43.0 0.29 0.25 0.55 0.21 35 0.19 0.34 56.7 0.30 0.25 0.56 0.23 40 0.27 0.35 77.7 0.31 0.25 0.62 0.21 45 0.35 0.36 98.5 0.32 0.28 0.60 0.22 50 0.36 0.38 94.7 0.33 0.31 0.63 0.25 Source: SANAA and"Panama- Public ExpenditureReview", WorldBank, 2006. 8.98 Cross subsidies. SANAA tariff structure involves cross subsidies with higher tariffs for larger consumers. Industry, commerce and government in general pay higher tariff than residential consumers. Communities with public faucets and not for profit organizations (patronatos) pay lower tariffs thanresidential consumers; Table 8.30. Table 8.30 Water Tariffs, by Type of Consumer,2006 (LpsJm3) Consumption Residential Commercial Industrial Government Public Faucets m3/month andPatronatos 0-20 0.85-1.20 5.50 0-50 0.85-6.8 5.5-9.0 9.0 9.0 51-60 8.5-10.7 10.0 12.0 12.0 Moreof60 12.9 12.0 16.0 16.0 0-180 1.o More of 180 1.4 Note: Effectivetariffs since September 2003. A 25 percent charge for sanitationapplies to the water bill plus a fixedmonthly charge of Lps. 0.5 or Lps. 1.50 for the meter. Source: SANAA 142. Otherestimatesputthe losses at twice that level. 178 Agriculture Subsidy 8.99 In2006, the Government created an assistance program to increasethe food security of small-scale farming families called Technological and Productive Vouchers (Bono Tecnoldgico). The assistance includes improved seeds, fertilizer, and technical assistance delivered inkind to farmers and i s valued at US$74 (Lps. 1,400) per farmer. It targets farmers growing maize, beans, sorghum, and rice on less than 9 acres (5 manzanas). The program is expected to reach 70,000 farmers at a cost of Lps. 85 million in2006. This support was expectedto increase yields significantly (from 25 qq/mz to 40 qq/mz inthe case of maize) which would ensure an adequate production for self-consumption and generate a 30 percent marketable surplus. The impact of theprogram is currently being evaluated. Telephonysubsidy 8.100 HONDUTEL is a state owned enterprise that has the monopoly on fixed line telephony services inHonduras. Inthe mobile segment there are also two private operators. HONDUTEL has 449,500 subscribers of fixed lines and 5,000 of GSM. The two mobile operators have a combined numberof 2.1 million subscribers. 8.101 HONDUTEL's tariffs on local service do not cover costs and are subsidized by its international service. The high cost of HONDUTEL's international service has made it uncompetitive in relation to the mobile operators. The international rates o f the private mobile operators are less than one-third of HONDUTEL's; Table 8.3 1. VOIPhnternet international rates are much lower than those offered by the private companies and HONDUTEL, though it has disadvantages in both quality and accessibility compared to the fixed and mobile services. HONDUTEL's internet DSL (Digital Subscriber Line) connection costs US$ 15 and US$ 0.018 per minute, or similar to the cost of a regular local call. It has only about 1,500 subscribers. Cable services that offer internet and TV cost US$70 per month. Table 8.31 Price Structure of Telecommunication Services, 2006 LocalCalls OutgoingInternationalCalls (US$/minute) a/ (US$/minute) a/ HONDUTEL Fix lines 0.0 18 0.8411.42 Fix wireless (GSM) 0.039 0.8411.42 '' Public telephones (coins) 0.009 b' n/a Public telephone (card) 0.018 0.10 Mobile Operators 0.17-0.20 0.25 VOIP(Cafenet) " n/a 0.05 Note: a/Exchangerate: L 19/vS$. b'0.02613 minutes. Voice over internet protocol. 0.84 to the US " and 1.42 to the rest of the world. Source: HONDUTEL 8.102 To recover local costs, HONDUTEL estimates that it would need to increase local calls from US$0.018 to at least US$ 0.033 per minute (Lps. 0.62), where the latter cost reflects its average cost in local telephony under the operating standards o f an efficient company (i-e., without paying for excess staff and other inefficiences). HONDUTEL's connection fees are Lps. 179 40 for residential customers and Lps. 80 for commercial customers and include 200 minutes o f free local calls. Ifthese calls are valued at cost (Lps. 0.60 per minute) the connection fee should be at least Lps. 120to recover the price o fthe calls alone, let alone the cost o f the connection. 8.103 The cross-subsidy between local and international calls and the liberalization o f the telecom market are pricing HONDUTEL out o f the international service market. In 2005, HONDUTEL had revenues o f Lps. 4,700 million o f which about 60 percent was from its international traffic and 40 percent from local calls. InDecember 2005 the international traffic was liberated and now operators are not required to use HONDUTEL for incoming or outgoing traffic. Together with an uncompetitive price structure, this reduces HONDUTEL's revenues to about Lps. 3,500 million in 2006, with international and domestic traffic contributing similar shares to the total. Because o f declining profitability, HONDUTEL's transfers to the Government have been adversely affected and have declined further in 2006 by an estimated Lps. 343 million. This represents the indirect cost o f the implicit subsidies present in HONDUTEL's price structure andmore generally o fanadverse policy framework; Table 8.32. Table 8.32 HONDUTEL Transfersto the Government 2004-2006 (Lps. million) Actual 2004 Actual 2005 Estimated 2006 Transfer N/a 1292.0 528.9 Incometax andfee for the use of spectrum N/a 293.5 713.9 Total 1,500.0 1,585.5 1,242.8 Difference inrelation to previous years N/a +85.5 -342.7 Note: a/ cannonor fee for use of spectrum. Source: Ministry of Finance 8.104 Table 8.33 summarizes the subsidies analyzed. In2005 these subsides amounted to Lps. 4 mil million (or 2.6 percent o f GDP) and are estimated to have increased do Lps. 6 mil millions in2006 (3.5 percent ofGDP).'43 143. GDP in2005 is estimatedat 157,528 million and in2006, Lps. 174,668 million. 180 Subsidy Explanation Type Rationale Poverty Cost 2005 Estimated Targetin (Lps. Cost 2006 g millions) (LPS. millions) Electricity 3736.0 4739.0 300 kwh Reductioninthe electricity Direct Affordability Partial 275.0 275.0 bill for those that consume less than 300kwh Partialcost ENEE tariffs lagcost Hidden Affordability No 1,506.0 2,119.0 adjustment changes Article 46 Tariffs are higher for larger Cross Compensate ENEE Partial NIA NIA Subsidy consumer subsidy 1for costs incurred with smaller Bono 80 Reductioninthe electricity Direct Compensate for Partial 198.0 550.0 a/ bill for those that consume increasein less than 300 kwh transportationcosts Fueloil and Exonerationof taxes on fuel Hidden Lower cost of No 1,757.0 1,795.0 diesel oil usedto generate generation electricity Fuel 159.4 641.3 Price Pricesof gasoline, kerosene Indirect To compensate No 159.4 507.0 d eliminatedon consumers incurredwl smaller I I consumers Agriculture NIA 85.0 Technological/ Improvedseeds, fertilizer Direct Promotefood Yes NIA 85.0 Productive andTA to small farmers security of small Bono farmer families Telephony NIA 342.7 HONDUTEL Reducedtransfersfrom Cross Affordability of No NIA 342.7 HONDUTEL to the subsidy localcalls Government Total 4,039.4 6,054.8 181 E. Assessment ofthe Social Protection System 8.105 This assessment of Honduras' Social Protection system focuses on aspects related to the amount spent, relevance and scope, coverage, targeting, cost-efficiency, monitoring and evaluation, and institutional arrangements. It is based on the previous discussion and the comparison of the population at-risk andthe existing programs, as summarized inTable 8.34. Spending 8.106 International comparisons of spending on social sectors in general, and on social protection in particular, though not exempt of difficulties, provide a first approximation to the relative importance that countries attach to these sectors.144 Social Protection spending in Honduras (5.2 percent o f GDP) is slightly less than the average for the six LA countries (5.9 percent of GDP); Table 8.35 However, while Social Assistance spending in Honduras i s more thantwice the average for the LA countries, its Social Insurance spending is less than one-half. Overall, Honduras's social spending (14.5 percent of GDP) is similar to the LA average (14.7 percent of GDP), with relatively more spent on education and less on health. Table 8.35 InternationalComparison ofPublic SocialSpending S I SA SP Educatio Health Other Social Year Total n Social Total Honduras ` 2.2s/ 3.0b1 5.2 6.9`I 2.4 0.0 14.5 2005 LA Average 4.8 1.1 5.9 4.2 3.7 0.9 14.7 Argentina 8.3 1.4 9.7 4.1 4.3 1.1 19.2 2003 Chile 6.9 0.7 7.6 4.2 2.9 2.0 16.7 2000 Costa Rica 3.6 1.o 1.o 4.6 3.9 5.3 0.9 14.7 1999 Mexico 2.6 3.6 4.1 2.1 0.0 9.8 2002 Panama 5.0 1.o 1.7 6.7 4.0 6.1 0.0 16.8 2005 Venezuela 2.1 3.1 4.9 1.5 1.5 11.0 2000 Note: a/Includes all IHSS health spending estimated at 1.2 percent o f GDP. subsidies equivalent to 0.5 percent o f GDP. ``Adjusted to b' Includes energy and transport direct avoid double counting with education and health programs include in SA. Without adjustment, public spending on education is 7.2 percent o f GDP and health 3.2 percent o f GDP in2005. Average o f the six LA countries. ' Source: World Bank reports and for Honduras Table 35. 144. For a discussion o f some o f these difficulties see Marques, JosC SilvCrio "Central America, Cross-country Evaluation o f Social Safety Net Assessments (SSNAs)- Issues Paper", paper prepared for the World Bank, November 2002. 182 I I i -.. 5 I . n Lo Lo m v B n Lo i x Relevance andScope 8.107 Honduras' social protection programs seek to address the main risks affecting the poor and, therefore, are generally relevant. For all the risks identified there are inmost cases several programs that attempt to address them. On the other hand, the impact of the social assistance programs does not appear commensurate to the amount o f resources spent, as the number of poor in a situation of risk remains high. Indeed, the allocation of resources is biased in favor of subsidies that are not well targeted on the poor. Table 8.36 indicates that 16 percent of social assistance spending i s on subsidies on electricity and transport most of which do not benefit the poor (see targeting below). Children aged 0-5 years, who represent the most vulnerable group, have a very modest share of resources. Seniors and the indigenous and black population also appear to be at a disadvantage, considering the limited scope of social insurance programs (pension and medical insurance). Table 8.36 Distributionof SocialAssistanceResources,by Age Group, 2005 Age Group Annual Cost US$ 000 a/ Percentof Resources 0-5 3 1,640 12.8 6-17 64,622 26.1 18-64 8,033 3.2 65+ 1,134 0.5 Indigenousandblackpopulation 7,935 3.2 Householdsexcl. subsidies 94,2 11 38.0 Direct subsidiesb' 40,053 16.2 Total 247,628 100.0 Note: a/ Excludes social insurance.b' Electricitysubsidies (300kwh and bono SO) and Transport Subsides (buses andstudentvoucher) Source: Table 8.34. Coverage 8.108 The coverage of some programs that are important for the poor is limited, leaving vulnerable a large number of poor people. For instance, there are 200,000 poor children under 5 that suffer from chronic malnutrition, and the PRAF maternal infantile and IHNFA programs combined reach less than 100,000 children. One apparent exception is the Healthy School Program that reaches over one million children with school lunches and medical attention. Nonetheless, 500,000 childrenthat do not attend preprimary and the 200,000 children that do not attend primary school do not benefit from the program. The coverage of IHNFA programs are very limited, as the institution's limited resources are almost entirely absorbed by personnel costs. INFOP's vocational training courses have a limitedimpact, reaching less than 3 percent of the labor force. Housingprograms only reach 5,000 of the 200,000 households reporting over- crowded conditions. Finally, most of the subsidies on basic services do not reach the poor householdsbecause280,000 o fthem do not have safe water and 470,000 do not have electricity. 8.109 Honduras also has low social insurance coverage. The social security system covers about 18 percent of the population, and less than 6 percent of Honduras's seniors have a pension. Health insurance coverage is similarly low, and 13 percent of the population (900,000) does no have adequate access to health services. 185 Targeting 8.110 In the last several years, a significant amount of resources was allocated to social programs through the Poverty Reduction Strategy (PRS). However, these programs did no have the impact expected in part because of poor targeting, lack of coordination of the different intervention, dispersion of efforts, and lack of monitoring and evaluation. According to the last PRS Progress Report 2005 (UNAT June 2006), of the total amount of resources allocated to the PRS, 24.3 percent was spent on the poorest (first consumption quintile), while 45 percent benefitedthe n~n-poor.'~'The updated PRSP attributes the poor results obtained to "the lack of theprecise identiJication of thepoor and thefactors associated withpoverty; the lack of effective and comprehensive targetin of resources to combat poverty, and lack of coherence between objectives and program^".'^ B PRAF 8.111 The targeting of PRAF programs has relied on geographic targeting using INE's poverty map based on estimates of unsatisfied needs (mostly based on housing defi~iencies).'~~ Under regional targeting, all municipalities inthe selectedregion became eligible for the programs. The allocation of resources became a function of the poverty level and size of the population, with the weight given to population size declining over time. 8.112 In 2004, INE produced estimates of poverty at the municipal level by combining the informationfrom the 2001 National Population and Housing Census (CNPV) and the household survey. This targeting methodology involves two steps. First, the poverty map preparedby the INE is usedto identify the municipalities with the highest prevalence of households in extreme poverty. Secondly, the poorest villages within the selected municipalities are identified by variables that are closely correlated to poverty. 8.113 PRAF's main programs Table 8.37 EstimatedIncidenceof PRAFPrograms are well targeted on the poor. (Percent distributionof beneficiaries) Table 8.37 presents the Health (Maternal Education School Infantile) CCT CCT a/ Bag distribution of beneficiaries of Characterization Pro-poor Pro-poor Pro-poor three programs according to the Q1 Poorest 61.4 41.6 29.8 ENCOVI. As it can be observed, 42 22.1 34.2 21.2 the health (maternal infantile) and 43 11.9 16.8 29.7 education conditional cash 44 2.5 5.4 16.0 Q5 Richest 2.0 2.0 3.3 transfers (CCTs) are well targeted Total Nla Nla 100.0 on the poor. The school bag (a Urban 100.0 100.0 32.6 subsidy given to students to help Rural N/a N/a 67.4 purchase education materials for Note: a/ "Honduras Poverty Assessment". World Bank2006. Vol 11, the first, second and third grade Table 105, 106, 110-2, p. 185s. Source: ENCOVI2004. 145. "Estrategia paralaReduccibnde la Pobreza.Informede Avance, 2005". UNAT. Junio 2006. p.vi. 146. "Estrategia de Reduccibn de la Pobreza". Versi6n Actualizada a 2006, UNAT. Septiembre 2006. Borrador. (inthe SIERP's web page) p.28. 147. INE's poverty index takes into considerationthe access to potable water, sanitation, primary education, and state ofrepair ofthe housing, overcrowding, and education ofthe head ofhousehold. 186 o f primary schools) i s less well targeted but still one-half o f the beneficiaries are inthe first two quintile. The PRS Progress Report 2005 indicates that 51 percent o f PRAF spending benefits the poorest (Ql), 25 percent 42, and 13.6 percent Q3. 8.114 The other main targeting Table 8.38 EstimatedIncidenceof FHIS Program mechanism used in Honduras is Distribution Cumulative Beneficiaries the FHIS, which uses a formula 4 1 Poorest 23.3 23.3 that takes into account the (i) 4 2 lack 22.4 45.7 o f access to potable water (with a 43 21.2 66.9 weight o f 40 percent), (ii)child 44 19.2 86.1 malnutrition rate (20 percent), (iii) Total Q5 Richest 13.9 100.0 100.0 lack of access to sanitation (20 Source: "Honduras Poverty Assessment". World Bank2006. Vol 11, percent), and (iv) illiteracy rate (20 Table 101, p. 181 percent). Inaddition, FHIS tailors its menu o f projects to the needs o f the poor. Inrecent years, FHIS has also promoted participatory local planningto help match resources with the needs o f the communities and the poor. As indicated in the World Bank's 2006 Poverty Assessment, FHIS programs are pro-poor as 67 percent o f their beneficiaries are in the first three quintiles, and a disproportionately high number o f program beneficiaries are in the first quintile; Table 8.38. Education Programs 8.115 Data from the ENCOVI indicates that PROHECO, which seeks to extend preprimary and primary education coverage mostly inrural areas, is well targeted on the poor, with 83 percent o f the beneficiaries inthe first two quintiles and only 0.6 percent inthe upper quintile; Table 8.39. Incontrast, the scholarship program mostly benefits the middle and higher income groups, with 44 percent o f the beneficiaries in the top two quintiles and only 29 percent in the lowest two quintiles.The school lunch program mostly benefits the poor andthose inrural areas. The adult education programs -EDUCATODOS and PRALEBAH, benefit mostly the poor, with over two- thirds o fthe beneficiaries inthe first two consumption quintiles andvery few inthe fifth quintile. All in all, with the exception o f scholarships, the education programs are well targeted on the poor. 8.116 Table 8.40 presents the number o f students that benefited from scholarships and the amount awarded by type o f scholarship. The overall budget for scholarships is only US$ 3.3 million. (If the transportation subsidy for secondary students i s added, the total amount will increase to US$ 6 million in2006.) Support to poor students constitutes less than 14 percent o f this limited scholarship budget. 187 Table 8.39 Distributionof Beneficiariesof EducationPrograms(percent of beneficiaries) Group PROHECO Scholarships School luncha/ EDUCATODOS PRALEBAH Q1 Poorest 54.3 15.2 29.0 44.9 48.6 42 29.0 14.4 27.6 34.7 19.3 43 13.0 26.0 28.5 7.0 25.5 44 3.0 25.0 7.1 10.4 3.2 Q5 Richest 0.6 19.3 7.8 3.O 3.4 Total 100.0 100.0 100.0 100.0 100.0 Urban 4.5 42.5 NfA 15.0 20.6 Rural 95.5 57.5 N/A 85.0 79.4 No of observations 136 38 N/A 122 77 Expandedobservations 36,196 7,196 N/A 33,192 19,902 Note: a/ "Honduras Poverty Assessment".World Bank2006. Vol I1Tables 89, p. 174 Source: ENCOVI 2004 Table 8.40 Scholarships,2006 Number of students Cost (US$ 000) YO For poor students, Basic 3,784 240 7.3 For poor students, Secondary 2,795 147 4.5 For poor students, University 332 61 1.9 Sub-total 6,911 448 13.7 Excellencefor Basic 1,961 207 6.3 Excellencefor Secondary 7,056 560 17.1 Honors for Basic 3,447 273 8.3 Honors for Secondary 7,93 1 503 15.3 Honors, University 134 35 1.1 Graduation,preprimary rural 30,433 966 29.4 Graduation, Secondary 26,796 283 8.6 Teacher sonsfdaughters 128 7 0.2 Sub-total 77,886 2,834 86.3 Total 84,797 3,282 100.0 Source: Ministry of Education Housing 8.117 Most of the resources allocated to housing programs benefit the higher income quintiles; Table 8.41, According to the PRSP Progress Report 2005, 38.4 percent of the resources benefit the richest households (Q5). Table 8.41 State SubsidizedHousingPrograms,2004 (percentof Householdsthat Benefited) Public Housing Grants International Public Housing Combined Incidence Grants Counterpart Finance (YOof resources) Q1 Poorest 0.2 4.6 0.5 15.1 42 0.5 4.3 0.2 15.7 43 0.1 3.2 0.6 11.3 44 0.1 3.8 2.2 19.5 Q5 Richest 0.1 0.8 8.8 38.4 Total 0.2 3.2 2.9 Urban 0.2 2.0 6.0 Rural 0.1 4.2 0.4 Source:"Honduras Poverty Assessment". World Bank2006. Table 70, p. 159 188 Electricity subsidies 8.118 In 2006, the Government spent Lps. 4.7 billion in direct (Lps. 825 million) and hidden (Lps. 3.9 billion) electricity subsidies, equivalent to 2.7 percent o f GDP. The hidden subsidies are not targeted on the poor, while the direct subsides are only partially targeted on those that have lower consumption levels and are assumed to be poor. Tariffs are kept below cost to protect the budget o f those that have electricity service. As a result, ENEE incurs large losses and does not have sufficient revenues to make the required investment to increase coverage and improve service quality. Table 8.42 Incidenceof the 300Kwh Subsidy (% and numberof households) No of No of Distribut. Rate of Average Incidence Subsidy received HH HH HH with Subsidy Subsidy ("/I (Lps. million) without with electric. Coverage (Lps.) b/ electricity electric. (YO) a/ a/ Q 1Poorest 183,994 35,046 3.7 86.1 28.8 3.3 9.0 4 2 156,468 103,014 10.8 79.0 38.8 11.8 32.4 43 73,928 196,869 20.6 75.8 44.9 25.1 69.0 44 42,554 268,060 28.1 68.7 48.6 33.4 91.9 Q 5 Richest 13,141 351,423 36.8 39.9 50.6 26.4 72.6 Total 470,085 954,412 100.0 61.4 47.7 100.0 275.0 Note: a/ Estimates based on the ENCOVI and the structure of the subsidy fiom the ``Honduras Poverty Assessment". World Bank 2006. This takes into account that the rate o f subsidy drops sharply for the Q5 but the average subsidy increases with income/consumption. b/ Distribution of HHs with electricity times the rate of subsidytime the average subsidydividedby the total. Source: ENCOVI 2004 8.119 The direct subsidies include the 300 kwh subsidy (Lps. 275 million in 2006), and the Bono 80 (Lps. 550 million). Both subsidies are only partially targeted on the poor. Indeed, the two poorest quintiles receive only 15 percent o f the total value o f the subsidy (Lps. 275 million) while the richest quintile obtains 26 percent o f the subsidy. Furthermore, there are 470,085 households in Honduras that do not have electricity. Of this, two thirds are poor households (Q1 and Q2). Clearly this groups not only does not benefit from the electricity subsidies, but the likelihood that they will eventually receive electricity diminishes by the fact that these subsidies use up resources that could otherwise be utilized to expandcoverage o fthe system. 8.120 The Bono 80 also only partially targets the poor. First, to qualify for the subsidy one needs to be a client o f ENEE. Many poor families do not have electricity and therefore do not qualify for the subsidy. Secondly, Bono 80 gives a higher subsidy to those that consume more, and so is clearly regressive. O fthe total amount o f subsidy in2006, Lps. 550 million, about Lps. 200 million (37 percent) was received by families that consume less than 100kwh; the remaining 350 million (63 percent) was receivedby those with higher monthly consumption; Table 8.23. Fuel Subsidies 8.121 Fuel subsidies cost Lps. 641 million in 2006, both to cover the price differential and the reduction in taxes. The subsidies on gasoline do not benefit the poor because they do not have 189 cars. The subsidieson LPG (domestic) and Kerosene only partially benefitthe poor. A majority (88 percent) of the poor households still use firewood for cooking, followed by kerosene (4.5 percent) and LPG (3.3 percent); Table 8.43. Table 8.43 Fuelfor Cooking, 2004 (percent households) Firewood Electricity LPG gas Kerosene Other a/ Total ExtremePoor 97.2 0.7 0.4 1.3 0.4 100.0 Moderately Poor 81.6 6.2 5.4 6.8 0.0 100.0 Poor 88.2 3.8 3.3 4.5 0.2 100.0 Non-Poor 30.2 30.2 30.9 8.5 0.2 100.0 Total 56.0 18.5 18.6 6.7 0.2 100.0 Urban 26.1 31.7 29.9 12.2 0.1 100.0 Rural 87.2 4.7 6.8 1.1 0.2 100.0 Note: a/ Harvestresidueand other Source: ENCOVI2004 TransportSubsidies 8.122 In 2006, the government spent Lps. 175 million to subsidize public transportation (excluding Bono SO). An incidence analysis of the impact these subsidies in the 2006 World BankPoverty Assessment concludesthat they are pro non-poor, as the first quintile does not use publictransportation andtherefore does not benefitfrom the subsidies. Table 8.44 indicates that the use of mass transportation increases with the level of consumption and consequently richer quintiles appropriate more of the subsidy. While the subsidy per inhabitant is under Lps. 3 per year inthe second quintile, it is equivalent to Lps. 42 inthe fifth quintile. Table 8.44 MassTransportationSubsides,2004 PercentUse of mass Subsidy received Subsidy per inhabitant Lps. transportationa/ (Lps. millions) Q 1Poorest 0.0 0 0.0 4 2 3.3 3.8 2.7 43 16.8 19.2 13.5 44 27.8 31.7 22.4 Q 5 Richest 52.3 59.6 42.2 Total 100.0 114.0 16.1 Note: a/ From"HondurasPoverty Assessment". World Bank2006. Vol. 11, Table 98, p. 179. Source:ENCOVI2004 8.123 The transport subsidy given to high school studentsinurbancenters (Lps. 52.4 million in 2006) is not targeted on the poor as very few of them attendhigh school (less than 15 percent of the cohort). Indeed, 80 percent of the studentsthat receivedthe subsidy benefits are inthe fourth and fifth quintiles; Table 8.45. The authorities inthe Ministryof Education indicated that they seek to target the schools that are inthe poorest areas of cities, but there is apparently no explicit targeting criteria to allocate this subsidy withinschools. 190 Table 8.45 Student TransportationSubsidies Distribution Subsidy received (percent) (LPS. millions) Q1Poorest 0.0 0 4 2 2.9 1.5 43 19.9 10.5 4 4 32.4 17.0 45 Richest 44.7 23.4 Total 100.0 52.4 Urban 97.1 50.9 Rural 2.9 1.5 No of observations 73 Expandedobservations 8,224 Note: a/ Distributionof subsidy in2006 assuming the distribution of2004 (ENCOVI) Source: ENCOVI 2004 Water subsides 8.124 The subsidies to water and sanitation given through SANAA (estimated at Lps. 72 million in2006) do not benefit the 283,500 households (20 percent o f all households) that do not have access to safe water. They mostly benefit those in the highest quintiles who already have access to water from the public system. Table 8.46 indicates that only 10 percent of those that receive the subsidy are in the first two quintiles; while 70 percent are in the fourth and fifth quintiles. Table 8.46 Householdswith and Without Access to PipedWater, 2004 (YOof households) Without With Piped HHConnectto CoverageRate Distributionof Piped Water Public System of Public HHwithPublic Water Services System Q 1Poorest 91,887 127,153 10,383 4.7 2.3 4 2 86,278 173,204 38,492 14.8 8.6 43 50,720 220,077 78,53 1 29.0 17.5 4 4 35,721 274,893 129,619 41.7 28.8 Q 5 Richest 18,957 345,607 192,854 52.9 42.9 Total 283,551 1,140,947 449,899 31.6 100.0 Urban 45,089 682,15 1 426,890 58.7 94.9 Rural 238,462 458,795 23,009 3.3 5.1 Source: ENCOVI 2004 8.125 Table 8.47 summarizes the results of the preceding assessment by organizing the programs in descending order of progressiveness, defined by the participation o f the poorest consumption quintile in a given program. It identifies the PRAF maternal infantile program as the most progressive, with scholarships, housing programs, and subsidies being the least progressive. Infact, some o fthe subsidies do not even reachthe poorest quintiles. 191 Table 8.47 Progressivenessof the PrincipalAssistancePrograms Programs Poorest (Ql) Participation in program PRAFMaternal Infantile 61.4 PROHECO 54.3 PRALEBAH 48.6 EDUCATODOS 44.9 PRAF Education Voucher 41.6 PRAF school bag 29.8 School lunches 29.0 FHIS 23.3 Scholarships 15.2 Housing programs 15.1 Water subsidies 2.3 Electricity 300 kwh subsidy 3.3 Student transportation 0.0 Mass Transportation 0.0 Source:Tables 39-48 New TargetingMechanism 8.126 The updated PRS strategy14*and the Red Solidaria seek to establish new priorities and target mechanisms: i)The first prioritizedgroup isrepresentedbythe 80,000 poorestfamilies that live in Western and Easternregion of the country. ii)The second prioritized group is the next 120,000 families living under the most precarious conditions. iii)Thethirdprioritizedgroupistheremaining400,000 familieslivinginextremepoverty, which have not beenincluded inthe previous groups. iv) Finally, the fourthprioritizedgroup is the additional 300,000 families living inpoverty. 8.127 To identify the families inpoverty, a newpoverty map is being developed by combining the informationofthe 2001 Population andHousing Census andthe ENCOVI (2004). Ituses the per capita monthly expenditure from the ENCOVI to determine the poverty condition of the households. It also proposesto establisha newUnified Register o f Beneficiaries (PUB) of social assistance programs. The operation of the PUB will require the definition of eligibility criteria for eachprogram. 8.128 Improving the targeting of the poorest families is an urgent task given the large amount of resources spent insocial assistance programs that do not reachthe poorest families. However, it is also important that the programs target the most vulnerable groups within these poorest families, particularly women in child bearing age, mothers and their infants, and young children. As discussed earlier, it is crucial that young children have appropriate nutrition and early stimulation so that they can have a healthy development and avoid the intergenerational transmission of poverty. 148. Estrategia Para la Reduccion de la Probreza. Version Actualizada. Borrador. 2006. Gobierno de Honduras. http://www,sierp.hn/sierp. web/Biblioteca/erp-008.pdf 192 Monitoring and Evaluation (M&E) 8.129 The great majority of social assistance programs inHonduras do not have monitoring and evaluation systems in place. Their impact, therefore, cannot be determined. There major exceptions are the programs financed by multilateral banks including FHIS and PRAF.'49 In most other institutions, information and M&E systems are inadequate. Insome cases there i s not even the most basic information on the cost or coverage of the program. Inmost cases there is no informationthat permits an assessment o f program outcomes or impact. Very few programs have adequate controls or verification systems to ensure that the information generated is accurate and complete. 8.130 Progress has been made at macro level to monitor the impact of the PRS. UNAT has developedthe Poverty Reduction Information System (SIERP) that seeks to monitor the progress with the implementation of the strategy. The IDBhas recently approved a technical cooperation to support the development of strategic planning and monitoring and evaluation systems with emphasis on social expenditures. The objectives are: (i)to support and gradually implement strategic planning as a basic tool of government management; (ii) contribute to the integration to of the budgetary system andthe planning system; and (iii) to design a monitoring and evaluation system based on the Poverty Reduction Strategy indicators. These methodologies will be first developed in the Ministries of Health, Education, Public works, Transport and Housing, and Agriculture. Cost-Eficiency 8.131 The lack of information precludes any detailed analysis of the cost-efficiency of most social assistance programs. However, indirect evidence suggests that there must be substantial room to increase the cost efficiency of existing programs. For example, the evaluation of the PRAF program mentioned in paragraph 6.129 found that the program spent one-third of its budget on administrative costs, which is more than three times the less than 10percent spent ina similar program in Mexico (ProgresdOportunidades). The school lunch program is another area where there might be room to improve cost-efficiency. Currently, foodstuffs are sent to schools across the country where parents are organized to prepare the meals. This raises the need to store the foodstuffs at a central or regional location, transport the foodstuffs to the schools, and store them again locally. It is likely to be more cost effective to transfer finds directly to schools so that the school director with the parent association can buy the foodstuffs locally. This would save on logistical costs and yield apositive impact on the local economy. 149. The latest evaluation o f the FHIS was conducted in March 2005 by ESA Consultants; it found that (i)the FHIS made substantial progress intargeting its resources on the poor and ineliciting community participation in the selection and execution of projects; (ii)it had a positive impact on education, health, and sanitary conditions o f the population; and (iii) "had an effective but not excellent performance during the Mitch it emergency" as "the magnitude o f the disaster and the centralization o f its management", rested efficacy to its interventions. The evaluation conducted by IFPRI o f the IDB support PRAF program and published in March 2003, concluded that the project had less impact and higher administrative costs than expected and recommended its "re-engineering." 193 Institutional Arrangements 8.132 Honduras has a well developed institutional structure inplace, but needs to strengthenthe coordination and efficiency o f social assistance programs. There i s an active Social Cabinet in charge of formulating social policy, includingthat related to PRS and Social Protection, backed up by a Technical Support Unit, UNAT. The Social Cabinet also counts with a system of indicators to monitor de execution of the PRS (SIERP) and a virtual fund to finance the PRS (PRS Fund). To develop and implement the PRS, it can count on an Advisory Group (CCERP). For issues related to social protection, it has created a Social Safety Net Roundtable (Mesa de Protection Social), withparticipation of civil society and international donors. 8.133 The Government also recently created recently a Social Protection Network (Red Solidaria) with the objective of increasing the impact of PRS/social protection expenditures on poverty reduction through integrated actions in health, nutrition and education, as well as conditional cash transfers. The Red Solidaria is to be initially administrated through PRAF, and comprise the FamilyAllowance Program (PRAF), the Honduran Social InvestmentFund(FHIS), the National Institutefor the Family (IHNFA), andthe National Institutefor Women. F. MainConclusionsandRecommendations 8.134 Honduras has the institutions and instruments requiredfor developing an effective social protection system. It has an active Social Cabinet, a consultative body (Social Safety Net Roundtable) and a technical group, UNAT. It has many years of experience with conditional cash transfers, workfare programs, and other type of transfers. The challenge is to better focus the existing efforts andmake existing institutions, policies andprograms more effective. 8.135 With respect to institutional shortcomings, INHFA needs to be reorganized and better equipped to be able to play a role in social protection policy development and oversight. It major overhaul and more resource^.'^^ currently lacks the capabilities and resources to discharge its mandate effectively and requires a INHFA needs to gradually change its focus toward prevention and to decentralize its operations. IHNFA must work more closely with the Ministries of Education and Health, local governments, and services organizations to leverage its resources. 8.136 The social assistance programs inplace are biased towards subsidies, most of which are not well targeted on the poor. The direct subsides (electricity and transport subsides) amounted to US$ 40 million, or 16 percent of the total social assistance spending, estimated at US$ 248 million (Table 8.36). However, ifwe consider all subsidies-bothdirect, indirect, and hidden- the total amount spentwas US$212 million (2.6 percentofGDP) in2005 and an estimatedUS$ 320 million in 2006 (or 3.5 percent of GDP), far surpassing the amount spent on social assistance. 150. A recent evaluation o f its Center for Attention to Children concluded that ".,.there is neither education nor healthcare; 90 percent o fthe centers are deteriorated and lack teaching materials, kitchen apparatus, and some lack sanitary. The children caretakers have no training and salaries are very low...." ... Only 8 percent have higher education; incentives "InvestigaciQ Acerca del Estado Actual de la Primera Infancia y Su Atencion en las k e a s de Educacion y Salud en Honduras". Centro Intemacional de Educaci6n y Desarrollo, CINDE, Octubre de 2006 194 8.137 The government could generate substantial fiscal savings by targeting subsidies better, as followrS: e Electricity subsidies. The 300 kwh subsidy costs the Government Lps. 275 million a year. Those that consume less than 100 kwWmonth receive only 22 percent o f the subsidy and those that consume over 100 kwWmonth receive the remaining 78 percent; see Table 8.21. If the Government decides to focus the subsidy only on those that consume less than 100 kwh, it could save about Lps. 215 million a year. e Bono 80 is clearly regressive. It benefits only those that are clients o f ENEEand benefits most those that consume more. Ifthe Government decides to focus this subsidy only on those that consume less than 100 kwh, as was the original intent o f this subsidy, it could save Lps. 350 million a year. e Exonerations o f taxes on fuel oil and diesel used for electricitv generation represent a large subsidy. They cost the Treasury Lps. 1,795 million in 2006, o f which about 95 percent is exoneration on the private sector and the rest on ENEE. If the Government decides to reduce this exoneration gradually, say in equal installment during a five year period, it could collect Lps. 359 million during the first year.15' (Eliminating this subsidy will, o f course, put upward pressure on electricity generation costs and, thus, on the need for ENEE to raise electricity tariffs and operating efficiency. It is more efficient, however, to assist poor households that might be affected by such tariff increases with other, better targeted subsidies, such as a revised Bono 80.) e Fuel subsidies. As o f December 2006, the subsidies due to the price differential of fuel have declined substantially, What remains i s a small subsidy on LPG. Since the price o f gasoline has declined below the frozen price by more thanthe tax break, the Government could consider reinstatingthe tax and recovering Lps. 79 million intaxes. 152 e Telecom subsidy, HONDUTEL's transfers to the Government have declined because it maintains a tariff structure that cross subsidizes local calls with international calls, while facing increased competition from new operators. As HONDUTEL enters the mobile market and competition inthe sector intensifies, it should be allowed to set its prices and use its resources to invest and increase its competitiveness. On the other hand, HONDUTEL also should be required to pay taxes, fees for the use o f spectrum, and the return on equity to the Government. At this juncture, however, it is not possible to estimate how much HONDUTEL should be transferring to the Government under such conditions. a Partial Cost Adiustment o f Electricity and Water Tariffs. Similar to HONDUTEL, there are losses incurred by ENEE and SANAA for failing to set tariffs at levels that cover the full cost o f operation and capital depreciation. For 2006, those losses have been estimated at Lps. 2,119 million for the ENEE and Lps. 72 million for SANAA. One reason that the costs o f these enterprises are so high, however, is that they are inefficient. It is not clear that their customers alone should bear the full cost of those inefficiencies, 151. It assumes a recovery o f 20 percent on the amount foregone in2006. 152. The Government has announced that it would be procuring directly several types of fuels in the international market. It i s not clear how this decision will affect the distribution and pricing mechanisms, but one would expect that it will contribute to reduce fuel costs, therefore strengthening the case to eliminate existing subsidies. 195 so a tariff adjustment that eliminates losses entirely may neither be feasible nor advisable on economic grounds. As before, a more detailed study is required to identify the main sources o f ineffi~iency,'~~which could serve as the basis for estimating the subsidization losses due to inappropriate tariffs. These subsidization losses can range from 0 to the total observed losses observed inbothenterprises. 8.138 The potential fiscal savings that the Government could realize on the subsidies are summarized inTable 8.48 and are estimated to lie between US$53 million (0.6 percent o f GDP) and US$lS5 million (2.0 percent of GDP). Conversely, ifthese (or similar) policy changes are not adopted and subsidies remain unchanged, they will continue to drain the Government budget year after year. 300 khw Fuel Subsidies Exoneration o ftaxes on fuel oil and diesel ENEE Tariffsbelow Cost, SANAA Grand Total i 5,795 i i 1,003 3,537 - (as %2006 GDP) I (3.3%) I 1 (0.6% 2.0%) - 8.139 Which assistance programs should be strengthened? First, it i s important to keep in mind, that sustained reduction in poverty can only be achieved by tackling the root causes of 153. Recently the National Council for Water and Sanitation, CONASA, developed a plan to reform the sector with the support o f the World Bank and participation o f many sectors. The strategic plan (Strategic Plan for the Modernization ofthe Water and Sanitation Sector, PEMAPS) seek to promote sustainable development of the sector, including models for service delivery by the municipalities and other service providers. Also, the World Bank is in the process o f completing an Energy Strategy Study for Honduras, which also seeks to identifythe key sources o finefficiency inthe sector, and possible reforms to improve service. 196 poverty, which are often associated with lack of human capital formation and low education attainment. Improving education attainment of the poor takes time and requires the investment o f considerableamount of resourcesover long period of time. The programs that should receive priority are those that help to break the intergenerational transmission of poverty by building the humancapitalofpoor Hondurans early intheir lives. These programs are: a Maternal-infant programs that seek to ensure that children are born healthy and with adequate weight. Conditional cash transfers program have shown to be effective in promoting attendance at controls, but supply side issues (the capacity ofthe health system to respondto increasedemands) must also be addressed. a Programs directed at poor infants, particularly those focusing on nutrition and early stimulation. Young children that suffer from malnutrition or lack o f stimulation are bound to develop poorly and have low education attainment, thereby perpetuating poverty. To be cost effective, these programs should be community-based and receive the support from parents, NGOs and local governments. a Preprimary education (including bilingual multicultural education) and associated programs such as the Healthy School Program. a Programs that encourage children to enter and remain in school such as conditional cash transfers as well as scholarships and assistance for poor students so they can enter and progress inthe education systemtoward higher grades. a Scholarshipsand education assistance focusing on poor students. 8.140 None o f these are new programs andmost are beingconsidered inthe Red Solidaria. For the conditional cash transfers +rN>(l+n) =(1 (interest rate identity) yields the following difference equationthat governs the evolution ofthe domestic debt ratio: bN - bN [(l t.l +rN)/(l+g)]-v-o-p, Inthis equation, bt BJYt represents the debt-to-GDP ratio, ?e represents the inflation rate, rN represents the real interest rate on non-concessional debt, g represents the real rate of GDP growth, v [EtBCt EtBCt-l(l+ic)]/Yt, representsthe net concessionaldebt inflows as a share of - 213 Annex C GDP, o (Tt - Gt)/Yt,represents the primary fiscal balance, and p [(M,-Mt.l)/Yt represents the amount of seigniorageobtained by the government. C3. A country's debt level can be said to be sustainable ifthe ratio of that debt to GDP does not keep growing indefinitely under the current fiscal and monetary policy arameter settings. Expressed interms of the non-concessional domestic debt, this means that b must be equal to f t or less than bNt-l.Ifit turns out that the debt ratio i s growing, then it i s evident from equation (3) that the authorities can restore a sustainable framework by raising the primary fiscal balance, o, b a sufficient amount. It is useful inthis context to examine the steady-state case where bNt = bKt-l= bN. This allows usto rearrangeequation(3) as, (3 = bN[(rN g)/( I+g)] -v -p, - (4) which expresses the primary fiscal balance as a function of the steady-state domestic debt level, the levelof external concessionalinflows andthe amount of seigniorage. Theprimary balancein this equation represents the minimum surplus (or maximal deficit) needed to prevent the domestic debt level from following an explosive path. Note that the minimal primary balance is positively related to the real interest rate and negatively related to the rate o f GDP growth. I.e., if the economy grows more quickly, the authorities can afford a higher primary deficit without becoming fiscally insolvent. Also note that as the amounts of external concessionallending or of seigniorage decrease, the minimumsurplus requiredto keep the debt level steady increases. C4. The next step inthis Annex is to estimate the minimumprimary balance neededensure a sustainable fiscal account in Honduras. That will require some projections of the net concessional financing flows and seigniorage revenues available to Honduran policymakers in the longrun, as well as figures onthe domesticpublicdebt ratio andofthe real interest rate. The estimated miminum primary balance is then compared to the actual primary balance observed since 2000, in order to determine whether there i s any room for reducing the actual primary balance below the current level - thereby creating additional fiscal space for priority expenditures -without jeopardizing macroeconomic stability. C5. The nominal interest rate paid by Honduras on its external debt averaged out to 12.2 percent over 2001-06, which turns out to be close to 0 inreal terms, considering the evolution of the exchange rate and international inflation rates. If Honduras had unlimited access to such concessional financing, there is no limit to the sustainable size of the public sector deficit as long as real GDP growth remains positive. Clearly, however, access to concessional financing is supply-constrainedand not unlimited. Rather than attempting to model the decision processesof donors in allocating concessional resources, the remainder o f this Annex assumes that the average net flow of concessional financing provided during 2001-06 will continue indefinitely into the future.19* As shown inTable C.1, net inflows of concessionalresources averagedaround 0.9 percent of GDP over the last six years. This assumption is likely to constitute an upper bound on the amount of concessional funding available to Honduras. Honduras's per-capita income already has risen above the limit to be eligible for concessional lending from the World Bank (IDAthreshold), andwhile IDA is still providing concessional assistance, it is beingprovided on 198. This assumptionfollows a suggestion made inCuddington, John T. (1997), "Analyzing the Sustainability of FiscalDeficits inDevelopingCountries", World BankPolicy ResearchWorking PaperNo. 1784, June. 274 Annex C hardened terms as of 2007. As Honduras's income levels and creditworthiness indicators continue to improve, donors are likely to reduce the amount of concessional lending or harden their lending terms further. Table C.1. Honduras Key EconomicIndicatorsand FiscalParameters;2000-06 - Average 2000 2001 2002 2003 2004 2005 2006 2001-06 Real GDP growth rate (%) 5.7 2.6 2.7 3.5 4.7 4.1 5.5 3.9 Consumer price inflation (eop, %) 10.1 8.8 8.1 6.8 9.2 7.7 5.3 7.7 Net External Inflows (% of GDP) -0.9 1.1 -0.1 0.3 3.2 1.0 -0.2 0.9 - Net External External Financing 1.5 3.0 1.8 1.9 4.5 2.1 0.8 2.4 Net Interest 2.4 1.9 1.9 1.6 1.3 1.1 1.0 1.5 SeigniorageRevenues (% of GDP) Actual -- 0.9 3.5 0.6 3.9 4.9 3.3 2.9 Steady-stateestimate 1.4 1.2 1.3 1.0 1.6 1.5 1.1 1.3 memo: Monetarybase 14.3 13.7 16.0 15.0 17.1 19.8 21.2 17.1 % change 6.9 28.1 4.2 29.9 33.0 18.5 20.1 Primary Balance (% of GDP) 0.5 -2.7 -3.3 -4.6 -2.6 -1.7 -2.6 -2.9 Public Debt Indicators (% of GDP) Total public DebVGDP 77.4 75.2 75.5 81.1 79.5 60.6 39.4 68.6 Nominalexternal debtlGDP 70.4 70.7 70.8 70.6 68.2 50.1 28.9 59.9 Domestic public debVGDP 7.0 4.5 4.7 10.5 11.3 10.5 10.5 8.7 Real Interest rate (%) 3.5 4.7 2.7 5.5 2.8 4.0 5.4 4.2 Nominal rate (180day gov. paper) 14.0 13.9 11.0 12.7 12.3 12.0 11.0 12.2 Source: IMF, Article IV Consultations staff reports, various years. Note: The figures on net external inflows, primary balance and the public debt refer to the combined public sector. C6. The actual seigniorage revenues obtained by the public sector averaged almost 3 percent o f GDP during 2001-06. This high level o f seigniorage revenues, however, would not be consistent over time with the government's objective o f maintaining single-digit inflation rates. That is, the monetary base has been growing at an average rate o f 27 percent over the last three years, which is significantly higher than the average inflation rate duringthat period. Ifthat level o f monetary expansion were sustained, prices would eventually also increase at a similar rate. The steady-state estimate o f seigniorage revenues presented in Table C.l i s calculated by assuming a stable demand for money and a nominal monetar expansion rate of 8 percent, which coincides with the average inflation rate over 2001-06.1r9 this yields annual seigniorage revenues o f 1.3 percent o f GDP. C7. Inserting these estimated values for net concessional financing inflows and seigniorage revenues into equation (4), the minimum primary balance can then be calculated for different 199. The steady-state seigniorage revenue is calculated as [0/(1+0)]m, where 0 represents the rate of monetary expansion andm represents the monetary base as a share of GDP. 275 Annex C rates of economic growth, g, and the real interest rate on government debt, rN.2oo These resulting values for the minimumprimary surplus are presentedinTable C.2. Table C.2. Minimum primary fiscal balancefor debt sustainability Real GDP growth rate g =2% g=3% g = 4% g = 5% g = 6% Real interest rate rN= 4% -2.1% -2.2% -2.3% - - rN= 5% -2.0% -2.1% -2.2% -2.3% - rN= 6% -1.8% -2.0% -2.1% -2.2% -2.3% Note: The values in this table were calculated by using equation (4), and assuming net concessionalfinancing inflows of 1 percent of GDP per annum, seigniorage revenue of 1.3 percent of GDP per annum, and a domestic non-concessionaldebt level of 10percent of GDP C8. Note from Table C.2, that the minimum primary balance never falls below a deficit greater than 2.3 percent of GDP - which is the combined value of net concessional inflows and the sustainable seigniorage revenueswith single-digit inflation. This figure is slightly lower than the primary deficit that was assumed in the baseline projections of the joint World Bank/IMF Debt Sustainability Analysis prepared in 2006201,which exhibited a primary deficit that fluctuated between2.4 and 2.7 percent of GDP over the projection period o f 2007-2026. Inany case, it is important to note that the actual primary fiscal deficit reported by the combined public sector inHonduras during 2001-06 averaged2.9 percent of GDP. This figure i s bothhigher than the minimum sustainable deficit presentedin Table C.2 and also exceeds the baseline figures utilized inthejoint World Bank/IMFDSA 2006. C9. Based on the preceding comparison, there does not appear to be any significant room for increasing the primary fiscal deficit in Honduras as a means of financing an expansion on priority public expenditures, without jeopardizing macroeconomic stability or violating other public objectives, such as a single-digit inflation. It may be possible to increase the revenues generated through seigniorage, as indeed occurred during the first half of this decade, but that will eventually leadto a significant accelerationof inflation. I EmpiricalDebt SustainabilityThresholds C10. Research on debt in developing countries has shown that the circumstances under which countries encounter debt servicing difficulties and require exceptional financing are very diverse and appear to be driven by other factors in addition to the solvency analysis presentedabove. This has led to the investigation of empirical debt sustainability thresholds by the World Bank and IMF.202These analyses have indicated that developing countries typically exhibit signs of 200. As a point of reference, note fkom Table C.2 that the real interest rate on the domestic short term (180-days) governmentbills averaged around4.2 percent during2001-06. 201. That DSA, which is based on the joint Bank-Fund Low-Income Country Debt Sustainability Framework (LIC-DSF), had concluded that the total public debt i s on a sustainabletrajectory provided that sound fiscal policies are maintained. It also showedthe fiscal situation to be fairly robust under the most common stress tests, but pointedto a moderaterisk of debt distress inthe event ofmajor fiscal or external shocks. 202. See IMF and IDA, "Debt Sustainability inLow Income Countries - Proposal for an Operational Framework and Policy Implications", February2004. 276 Annex C debt-related distress when their total public debt levels exceed 43 percent of GDP. Moreover, since the capacity to service debt also depends on a country's overall macroeconomic and structural policy framework, these debt sustainability thresholds have been refined to allow for differentlevels of institutional capacity and quality of public policies. These thresholds, derived from an analysis of low income countries, are presented in Table C3, where the quality of economic management i s measured by the ratings obtained from the Country Policy and Institutional Assessments (CPIAs) that are carried out annually by the World Bank. C11. Honduras's current public debt and debt service ratios are well below the thresholds that signal potential debt-related difficulties. The only indicator inwhich Honduras comes close to a threshold level is in terms of the Net Present Value of Public Debt. Even here, however, Honduras is far from facing any immediate risk o f debt distress, considering that its CPIA ratings in2005-06 havebeenin"Strong" category. Table C3. Policy-Dependent Debt and Debt Service Thresholds ( percent) Quality of Policiesand Institutions Honduras Debt Indicators Poor Medium Strong 2006 NPV Debt as share of. GDP 30 40 50 29 Exports 100 150 200 73 GovernmentRevenues 200 250 300 152 Debt Service as share of Exports 15 20 25 4 GovernmentRevenues 25 30 35 12 Source: Debt threshold values are from IDA and IMF (2005), "Operational Framework for Debt Sustainability Assessments in Low-Income Countries - Further Considerations", Table 2, March 28. Figures for Honduras are from the Joint World Bank IMF,"Honduras: Debt Sustainability - Analsyis 2006". C12. On the other hand, it is also important to point out that Honduras's public and external debt indicators are currently so low mainly on account of the massive debt relief provided through the HIPC and MDRI initiatives since 2004, and not due to a particularly tight fiscal environment. It is important, therefore, to bring the primary fiscal balance within the limits discussed in the previous section to avoid any danger of raising the public debt once again toward the threshold levels indicated inTable C3. Summary C13. The preceding discussion showed that Honduras's public debt levels have fallen to very low levels, away from the empirical threshold levels that signal a significant risk of debt distress. This decline in the public debt, however, was mostly due to debt relief initiatives that favored Honduras, and not because of a particularly tight fiscal environment. Rather, as shown earlier, Honduras's primary fiscal deficit during the first half of this decade was very close, ifnot above, the maximum level beyond which the debt ratio (or the inflation rate) begins to increase again. This means that there is very little room, if any, for increasing the fiscal deficit further on a sustained basis to finance an increase in priority public expenditures. Such an increase in the deficit is possible for a limited period, but that would have to be compensated for by lower deficits infuture years. 277 Annex C 278 Annex D Annex D Measuringthe GrossDomesticProductofHonduras D1. This note concludes that the level o f GDP in Honduras may be about 15 percent higher thanthe figures recordedinthe nationalincome statistics. This estimate is basedonthe findings o f the 2004 Living Standards Measurement Survey, which yielded an estimate o f total private consumption inHonduras that was about 17 percent higher thanthe figure inthe official national income accounts. D2. Consider the national income accounting identity: Y = c +I+(X-M) where Y i s the Gross Domestic Product (GDP), C and I represent consumption and investment (both private and public) and (X - M) represents the trade balance (exports minus imports). Suppose that the national income accounts do not record all expenditures for the reasons described earlier. Generally, the expenditure categories that are the easiest to record are the foreign trade flows (X- M), public investment, f,and public consumption, 6;. It is therefore reasonable to assume that they are captured in their entirety in the national income accounting statistics. Private expenditure are more difficult to record in a comprehensive manner and, therefore, more likely to bypass official information gathering mechanisms. Accordingly, private investment, f can be separated into a recorded art, fR, an unrecorded part, fu. similar vein, private consumption can be written, 2 = PRande'. Given that the official Ina + national income accounts only include the recordedexpenditures, the official figure for GDP is: Y = (F+P) + (6 ; T+P)+(XT- MT); Le., it i s equal to the sum o f officially recorded private absorption, public absorption and net exports. The true level o f GDP, denoted Y*, equals the official estimate (Y) plus unrecorded private absorption. Hence, Y*= Y + (P+ P) D3. To estimate Y* relative to Y - i.e. the degree o f GDP sub-valuation - this note relies on the findings of the Living Standards Measurement Survey (LSMS) and an assumption concerning private consumption and investment behavior. The 2004 LSMS indicated that total private consumption is 17.03 percent larger than the officially recorded figures.204 This implies that 203. This methodology is identical to that developed by U.Ltichler for the Nicaragua Public Expenditure Review (2001). It had concluded that Nicaraguan GDP was about 70 percent higher than the official national income account figures. When the revised national income accounts (with base year 1994) were published in 2003, the revisedGDP figures average around67 percent higher thanthe earlier (base year 1980) figures. 204. According to the national accounts, private consumption in 2004 was Ls.103,219 million., while the 2004 LSMS indicatedthat private consumption was Ls.120,804 million. 279 Annex D f T = fRfu= *fR), + (1.1703 or fu= *fR) (0.1703 D4. It is assumed next that the same relationship betweenprivate consumption expenditures and private investment expenditures holds for unrecorded expenditures as for the recorded private expenditures. Inother words, (fR/fR) or = (fu/fu), (fU/fR) =(CpU/fR)=0.17 which yields that the total unrecorded private expenditures represent 17. percent of the recorded expenditures. Inserting this result into equation (1) yields Y'= Y + 0.1703[(fR+fR)/qx The national income accounts indicate that total private absorption to GDP, ( f R + f R ) / Y , averagedaround 90 percent during 1996-2005. This yields that, Y*= y[I+0.1703*0.903] = 1.15Y, suggestingthat the 'true' level of GDP is about 15 percent higher than the official figure. This estimate is substantially lower than those cited by some other observers, such as Agosin et a1 (2005), which mentions a figure inthe range o f 40 to 50 percent.205 D5. After this PER was completed, the Central Bank of Honduras published a revised series of national income accounts, based on the 1993 System of National Income Accounting and shifting the base year fiom 1978-2000. This new series revises upward the level of GDP by 19.3 percent 205. Agosin, Manuel., A. Barreix and R. Machado, eds. (2005), "Recaudar para crecer: Bases para la reforma tributaria enCentroamCrica," (Washington, DC, Inter-American Development Bank). 280 Annex E Annex E A Frameworkfor Analyzingthe Impact of FiscalReformon EconomicGrowth El. This Annex analyzes the impact of certain fiscal adjustments on economic growth, buildinguponthe findings from Gupta et a1(2005). Inthis study, the relationship between public expenditure composition, fiscal adjustment and growth i s estimated using three econometric models:206 Model A (Budget Composition): growth i s regressed on fiscal variables aimed at capturing the effect o f the composition o f the budget. These variables are measured in percent o f GDP and include: public sector wages and salaries, expenditures on other goods and services, transfers and subsidies, interest payments on government debt, capital expenditures, tax revenues, non-tax revenues, and grants. 0 Model B (Budget Balance and Composition o f Expenditure): growth i s regressed on fiscal variables aimed at capturing the effect o fthe budget balance andthe composition o f expenditures. The budget balance i s measured as a percentage o f GDP, while all expenditure items are measured as shares o f total public expenditures. The expenditure categories include: public wages and salaries, public transfers and subsidies, interest payments on government debt, public expenditures on other goods and services, and public capital expenditures. 0 Model C (Budget Financing and Composition o f Expenditure): growth is regressed on fiscal variables aimed at capturing the effect o f the deficit financing (both domestic and external financing inpercent o f GDP), and the composition o f expenditures as shares o f total public expenditures. E2. Gupta et a1 (2005) have estimated these models under several conditions and assumptions to check for consistency. The baseline regressions use a fixed-effects estimator and are performed using the whole sample and two sub-samples aimed to capture any nonlinear effects o f fiscal policy on growth @re- and post-stabilization countries). In addition a pooled mean- group estimator (Nested Model) i s also used to capture the effects o f both short-run and long-run dynamics and relax the assumption o f homogeneity o f short-run coefficients. Finally, robustness i s tested by running a Generalized Method o f Moments (GMM) estimator to address potential problems with endogeneity and serial correlation arising from the dynamic specification o f the models. E3. A summary o f the estimated coefficients associated to the fiscal variables in all the specifications reported by Gupta et a1 (2005) is reported in Table F1 and Table F2 for the long 206. All three models also include independent, non-fiscal variables inan effort to control for initial characteristics (initial level o f GDP per capita, private investment ratio, terms o f trade, labor force, and initial level o f primary and secondary enrollment rates). 281 Annex E run and short run models respectively. Except for the sample o f post-stabilization c0untries,2~~ the results from all specifications turnout to be consistent. E4. The estimation results inTables F1 and F2 yield the following basic findings: In general, fiscal adjustments do not appear to have harmed growth: a 1 percent improvement inthe fiscal balance raises growth by between 0.348 and 0.563 percentage points (between 0.46 and 0.8 percentage point under the short runspecification). Domestic financing tends to be more harmful for growth than external financing: a 1 percent increase in domestic finance reduces growth by roughly 1 percentage point (and more than 1 percentage point under the short run specification), while an increase o f 1 percent in external financing reduces growth by less than half o f a percentage point (roughly one-half o f one percentage point under the short runspecification). 0 Expenditure composition affects economic growth in interesting ways from a fiscal policy perspective. Inthe long run, spending on wages and salaries has the largest impact on growth: a one percentage point increase in the public payroll reduces growth by between 0.21 and 0.78 percentage points. In contrast, capital outlays exert a significant positive impact on economic growth (elasticities hover between 0.18 and 0.68 inthe long run).Finally, expenditures on other goods and services are also found to raise the growth rate, but only in the short term, while interest payments have a statistically insignificant impact on growth. Data E5. To analyze the long run consequences o f fiscal policies, a number o f simulations are performed next, based on the elasticities reported in Table F1,which measure the sensitivity o f economic growth to changes in the fiscal policy variables. These simulations seek to assess the impact o f adjusting the values o f various fiscal variables in Honduras to the average value o f these variables exhibitedby the countries ina regional group. Three groups o f countries are used as reference points for determining the changes in the fiscal variables used in the simulations. These are Central America (Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua and Panama), Latin America (Argentina, Chile, Colombia, Mexico and Peru) and Europe (Germany, Italy, Ireland, Netherland and Belgium). E6. Figure F1 presents the data on expenditures for Honduras and o f the other Central American countries, together with the averages exhibited by each reference group. As a percent o f GDP, Honduras exhibits (i) higher total spending thanthe other Latin American countries, but less thanthe Europeans, (ii) spending on wages and salaries than the Latin American and higher European averages, (iii) transfers and subsidies than any other group (only Guatemala has lower similar levels of transfers and subsidies at about 2.8 percent o f GDP), (iv) similar interest payments than the benchmark groups, and (v) higher spending in goods and services and capital than the other groups. Figure F2 presents the data on total and tax revenues. It indicates that Honduras ranks first among its Central American peers as a collector o f revenues, and also collects more than the LatinAmerican average, bothinterms o f total revenues andtax revenues. 207. A post-stabilization country is defined as one that has maintained an average fiscal deficit (after grants) below 2.5 percent of GDP during 1990-2000. Honduras qualifies as apre-stabilization country by this definition, 282 W & L m 03 N w 8 LI Y Annex E FigureEl.FiscalExpenditurt !004 (as percentage of GDP) 50 , TotalExpenditure Wages andsalaries 439 I 25 20 I5 13.0 10.0 I O 5 0 CRI SLV GTM HND NIC PAN CEN LAT EU CRI SLV GTM HND NIC PAN CEN LAT EU A M A M AM A M Transfern andsubsidies Interestpayments 25 1 21.5 1 6.2 5 0 CRI SLV GTM HND NIC PAN CEN LAT Eu CRI SLV GTM HND NIC PAN CEN LAT EU A M A M AM A M Other goods and services Capitalexpenditure I 5 I O 5 0 CRI SLV GTM HND NIC PAN CEN LAT EU CRI SLV GTM HND NIC PAN CEN LAT EU A M A M A M A M I Source: Author's computations based on IMF's Article IV Consultations staff reports. TotalRevenue Tax revenue 50 , 433 1 50 I 40 40 30 30 20 20 I O I O 0 0 1 I1 CRI SLV GTM HND NIC PAN CEN LAT EU CRI SLV GTM HND NIC PAN CEN LAT EU AM A M A M A M I / Source: Author's computations based on IMF's Article IV Consultations staffreports. 285 Annex E Simulations E7. Table F3 reports the results o f the simulations for the three models under consideration. For each model (A, B and C) we simulate the impact o f taking: (i) o f each the spending and revenue components to the three benchmark levels individually; (ii) the expenditure bundle simultaneously; (iii) the revenue bundle simultaneously; and (iv) the fiscal bundle simultaneously. It i s important to clarify that the results o f these simulations have to be taken more as an indication o f potential opportunities on the growth front (similarinspirit to the simulations performedinLoayza, Fajnzylber and Calderon, 2005), than as a forecasts that aim to predict the growth rate in response to a specific policy intervention. E8. Table F3 suggests that just by aligning expenditures on salaries and wages as a share o f GDP to the average level observed inother Central American countries, and not considering any other adjustment, Honduras could increase its growth rate by roughly 2.4 percentage points. Although this estimate is probably on the optimistic side, it i s noteworthy that public expenditures on wages and salaries are the source o f the largest potential gain. When all the expenditure variables are adjusted to the Central American benchmark levels simultaneously, Honduras's growth rate would register a more modest improvement o f about 0.5 percentage points. However, if we simulate the whole fiscal bundle, Honduras' growth would decline. E9. The potential gains are higher when Latin America i s taken as the benchmark. The results o f simulating the whole expenditure bundle indicate an average potential gain across models o f 0.8 percent. Inthis case also, the largest gain continues to come from reducing expenditures on wages and salaries to the benchmark level. A change in the whole fiscal bundle, on the other hand, yields additional growth o f 0.4 percentage points. When Europe is taken as the reference point, the gain in growth from simulating the entire fiscal bundle is greater than 1percentage point. Inthis case, however, most o f the gain comes from changes to the revenue side (andmore specifically from changes innon- tax revenues). E10. On the whole, these simulations yield two main messages First, it appears that fiscal reforms could result in significantly faster growth for Honduras over the long run. Secondly, among the different components o f fiscal spending, the main barrier to faster growth appears to be the level o f wages and salaries. 286 w -p! m f -m s - % % 8 8 - - - ? ? ? ? 9 9 z z a a 2 2 ; `9 4 n v) 0 . . o t ? ? t ? c\1 0 c c 7 4 0 9 7 4 9 4 w " 9 4 7 ? ? T o ? ? t o " 9 4 7 Nc?N? 4 4 7 4 -4w e ? 4 4 4 T N Y ? 4 4 4 4 r j w ? ? - 0 0 4 0 0 0 9 0 & et e & .-t2E' Po Y6t .-d3 E f c) d e Annex E Accountingfor measurement error El1. The previous analysis assumed that there is no measurementerror inthe datausedfor the simulations. One area of possible measurement error, however, is the level of GDP. Both Guatemala and Nicaragua have recently revised their national accounts (by updating the base year, which had previously been in the 1950~)~ yielding important changes. In the case of Guatemala, GDP was revised downward, but in the case of Nicaragua, the revised GDP turned out to be significantly higher. Clearly, a significant over or under-estimation of the GDP figures could have important implications for the preceding simulation findings, as that would exaggerate or reduce the differences between the fiscal ratios exhibited by Honduras and the benchmark averages. Absent any prior notion on whether the Honduran national accounts may be biased upward or downward, Table F4 reports the findings from simulating model A under three scenarios: (i) observed data (same as in Table F3) (ii)a scenario where GDP is the overestimatedby 30 percent, and (iii)scenariowhere GDP is underestimatedby 30 percent. a E12. The results inTable F4 indicate that even ifHonduras's GDP were underestimatedby 30 percent there would be important gains from bringing the level of wages and salaries to the averages observed in Latin and Central America. Moreover, this would be the only one among the fiscal adjustment alternatives that one could rely on for achieving faster growth. If, the on other hand, Honduras's GDP were currently overestimated, the potential growth gains fiom loweringthe public wage and salary bill would be evenhigher than under the baseline scenario. 288 W 'loo 4 0 0 ? Y O 'lev! 9 7 4 '"VI- 7 9 0 Y N N 0 0 0 c ? ? Y 0 0 0 u r m - c ? Y ? Annex E References Alesina, A. and R. Perotti (1995); "Fiscal expansions and fiscal adjustments in OECD countries." EconomicPolicy 21:205-48 Alesina, A. and R. Perotti (1996); "Fiscal adjustments in OECD countries: composition and macroeconomic effects." IMFWorking PaperNo. 96/70. Alesina, A., and S. Ardagna (1998); "Tales of fiscal adjustment." Economic Policy: a European Forum 27:487-546. Barro, R. J. (1989), "A Cross-country Study of Growth, Saving and Government", National Bureau of Economic Research Working Paper, No: 2855, Cambridge, MA, National Bureau of Economic Research. Barro, R.J. (1990); "Government spending ina simple model of endogenousgrowth." Journal of Political Economy 98(1):103-117. Barro, R.J. and X. Sala-i-Martin (1995); "Economic growth" McGraw-Hill, New York. Barro, R. J. (1997), Determinants of Economic Growth: A Cross-country Empirical Study, Cambridge, MA: MITPress. Bose, N., M.Haque and D. Osborn (2003) `Public Expenditure and Economic Growth: A Disaggregated Analysis for DevelopingCountries', University o f Manchester (Mimeo). Easterly, W. and S. Rebelo (1993), "Fiscal Policy and Economic Growth: An Empirical Investigation", Journal of Monetary Economics, 32(4), 417-458. Gemmel, N.(1983), "International Comparisons of the Effects ofNon market-Sector Growth", Journal of ComparativeEconomics, 7(4), 368-381. Grossman, P. J. (1990), "Government and Growth: Cross-Sectional Evidence", Public Choice, 65, 217- 227. Gupta, S., B. Clements, E. Baldacci and C. Mulas-Granados (2004); "The persistence of fiscal adjustments indeveloping countries" Applied EconomicLetters 11:209-212 Gupta, S., B. Clements, E. Baldacci and C. Mulas-Granados (2005); "Fiscal policy, expenditure composition, and growth in low-income countries" Journal of International Money and Finance 24:441-463 Haque, M., 2004, `The Composition of Public Expenditure and Economic Growth in Developing Countries', Global Journal of Finance and Economics 1(l), 97-117. Karras, G. (1993), "Employment and Output Effects o f Government Spending. I s Government Size Important?" Economic Inquiry 31(3), 354-369. Karras, G. (1996), "The Optimal Government Size: Further International Evidence on the Productivity of Government Services", EconomicInquiry 34(2), 193-203. Karras, G. (1997), "On the Optimal Government Size in Europe: Theory and Empirical Evidence", Manchester School, 65(3), 280-294. Landau, D.(1983), "Government Expenditure and Economic Growth: A Cross-country Study", Southern EconomicJournal, 49(3), 783-792. Landau, D. (1986), "Government and Economic Growth in the Less Developed Countries: An Empirical Study for 1960-1980", Economic Developmentand Cultural Change,35,35-75. 290 Annex E Levine, R. and D. Renelt (1991), "Cross Country Studies o f Growth and Policy: Some Methodological, Conceptual, and Statistical Problems", World Bank Working Paper, No: 608, Washington, D.C., The World Bank. Loayza, N., P. Fajnzylber and C. Calderon (2002); "Economic Growth in Latin America and the Caribbean: Stylized Facts, Explanations, and Forecasts", The World Bank. Lucas, R.E. Jr. (1988), "On the Mechanics of Economic Development", Journal of Monetary Economics, 22 (l), 3-42. Mendoza, E., G. Milesi-Ferretti and P. Asea (1997); "On the ineffectiveness of tax policy in altering longrun growth: Harberger's Superneutrality Conjecture." Journal of Public Economics 66:99-126. Perotti, R. (1999); "Fiscal policy in good times and bad." Quarterly Journal of Economics 114(4):1399- 1436. Ram, R. (1986), "Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data",American Economic Review, 7( l), 191-203. Romer, P. M.(1986), "Increasing Returns and Long Run Growth", Journal o f Political Economy, 94. 1002-37 Romer, P. M.(1989), "Human Capital and Growth: Theory and Evidence", National Bureau of Economic ResearchWorking Paper, No: 3173, Cambridge, MA, National Bureau o f Economic Research. Rubinson, R. (1977), "Dependency, Government Revenue, and Economic Growth, 1955-1970", Studies in ComparativeInternational Development, 12(4), 3-28. Solow, R. (1956), "A Contribution to the Theory of Economic Growth", Quarterly Journal of Economics, 71:65-94. Strauch, R and J. Von Hagen (2001); "Fiscal consolidations: quality, economic conditions, and success" Public Choice 109:327-46. 291 Annex E 292 Annex F Annex F The Maquettefor MDG Simulations(MAMS) F1. The Maquette for MDG Simulations i s a dynamic general equilibrium model that takes into account the links between public expenditures (both current and capital) on social services and infrastructure, and outcomes in terms o f MDG attainments and aggregate growth. The relationship between government spending and MDG outcomes, however, i s not a simple, invariable one for three essential reasons:2o8 a) The returns to scale o f government spending vary with the level o f service delivery. At low levels, increasing returns may prevail as network effects, learning effects and synergies are predominant. At high levels o f service delivery, government spending may suffer from decreasing returns to scale. Water supply, health care and education can be relatively easily provided in densely populated areas, but becomes increasingly expensive as coverage expands to remote areas. b) Effectiveness o f government spending depends on manyvariables. For example, spending on education becomes more effective if health conditions improve (reducing absenteeism at schools), if public infrastructure improves (facilitating access to schools), if income levels rise (and parents are less inclined to keep children at home) or if skill premiums increase (triggering a greater incentive to finish formal education). In general terms this means that spending on services becomes more effective if demand conditions for those services are more favorable. c) Costs o f service delivery change with macroeconomic conditions. The services are often skill intensive and in many cases also capital intensive. The more intense the MDG effort, the stronger the impact on costs as skilled labor becomes scarcer and financial conditions become tighter. From a general budgetary perspective the impacts on costs are even larger, because changes in macroeconomic conditions do not only affect MDG spending, but also other, non-MDG government spending (as well as the competitiveness of the private sector). F2. The first two aspects (changing returns to scale and impact o f demand variables) are captured in the 'MDG production functions' introduced in MAMS. The last aspect (macroeconomic interactions) is captured as the MDGproduction functions are incorporated ina dynamic economy-wide general equilibrium framework that also includes detailed fiscal accounts. F3. To keep the analysis simple and focused on the salient points, the model's MDG scope is constrained to goals which are likely to be most costly and have the largest effects on the rest o f the economy: universal primary school completion (MDG 2), reduced under-five and maternal mortality rates (MDGs 4 and 5), and increased access to improved water sources and sanitation (part o f MDG7). To the extent that a package o f interventions that curtails child mortality contributes to reducing the incidence o f major diseases other than HIV/AIDS, the model also implicitly tracks MDG 6. In addition, we address achievements in terms o f poverty reduction 208. The following four paragraphs are taken ftom Bourguignon, F. , C. Diaz-Bonilla and H. Lofgren, (2007) "Aid, Service Delivery and the MDGs in an Economy-wide Framework'' in Bourguignon, F. L. Pereira da Silva and M. Bussolo (eds.) "The Impact o f Economic Policies on Poverty and Income Distribution - Advanced Evaluation Techniques and Tools". 293 Annex F (MDG l), the model does not contain mechanisms for specific MDG 1-related although interventions. F4. The model disaggregation i s shown in Table G1. Given that the primary focus i s on government policies relatedto MDGachievement, the model maintains a simple structure for the rest o f the economy by aggregating all private activities into three sectors: agriculture, manufacturing, and services. Following standard practice, households maximize utility and producers minimize costs, and prices clearing behavior is assumed. Commodity prices and factor returns are determined in competitive markets and traded (exported or imported) and non-traded commodities are imperfect substitutes. In contrast to the private sector, the production structure in various MDG sectors is quite detailed: government commodities/activities include primary, secondary, and tertiary education, health, water and sanitation, and public infrastructure-all o f which are produced using a combination o f labor inputs o f varying skill levels and sector- specific capital stocks. The government finances its activities from domestic taxes, domestic borrowing, and foreign aid (borrowing and grants). Provision o f education, health, and water- sanitation services contribute directly to the MDGs. Table F1. The Sector Structure of the Model Activitieskommodities Private (3) Agriculture Manufacturing Services Public (7) Primary education Secondary education Tertiary education Health Water and sanitation Infiastructure Othergovernment Factors (11) Unskilled labor--workers who have not completed secondary school Skilled labor--workers who have completed secondary school Tertiary-skilled labor--workers who have completed tertiary education Capital @)--one stock for each model activity Institutions (3) Household Government Rest ofthe world F5. The MDG module i s a core component o f the model. For goals other than MDG 1, the level o f each MDG indicator is determined as the outcome o f a production function featuring inputs relevant for each particular MDG; this production function explicitly recognizes that marginal returns to MDG spending diminish as the economy approaches the target One 209. Mathematically, a two-level structure i s used in the MDG production block. At the bottom level (i), intermediate variables are defined in a set of constant-elasticity (CE) functions. At the top (ii), each 294 Annex F of the key features of our analysis i s the presence of synergies betweenMDGs, i.e. the fact that achievements in terms of one MDG can have an impact on other MDG outcomes. Thus, improvements in the water and sanitation targets, expressed as service delivery relative to the size of population, are a function of the growth in public per capita spending on water and sanitation, growth in government spending on public infrastructure, and growth of per capita consumption. Achievement of health MDGs depends on growth inpublic per capita spending on health, public expenditure on infrastructure, growth in per capita consumption, and improvements in water and sanitation service delivery. Although improvements in public infrastructure are not part of the MDGs, they serve as a key input in the MDG production function, and also contribute to overall growth by adding to the productivity of other production activities. F6. The treatment of the education sector slightly differs from other MDGs. The model explicitly tracks base-year stocks of students and new entrants through the three education cycles. Ineach year, students will successfully complete their grade, repeat it, or drop out of their cycle. Student performance depends on educational quality (growth in spending per student), improvements in household welfare (per-capita household consumption, which captures the demand side), growth in public infrastructure spending, wage incentives (the economy-wide wa e premium of completing the next cycle of education), and health status (proxied by MDG 4)!' The achievement o f MDG 2 requires that all students in the relevant age cohort enter the primary cycle and successfully complete each year within this cycle. In the model, this is translated into requiring that the rates o f lSt entry, graduation, and continuation to next grade grade within the primary cycle are at or very close to 100 percent for a six-year period prior to 2015. F7. The MAMS model does not explicitly track the progress of MDG 1 (poverty reduction). A simple approach to follow the progress on poverty reduction is to use a poverty elasticity of the growth inhousehold per capita consumption. This study takes a more sophisticated path by utilizing a macro-micro approach, where a set of aggregate results from MAMS are passedon to household survey data by means of a micro-simulation module. The data for our micro- simulations come from the 2001 Encuesta Permanente de Hogares de Propositos Msltiples (EPHPM) survey, which allows us to identify employment and wages by skill and sector. The simulations then apply changes inemployment, relative wages, and consumption per capita from MAMS to each individual inthe survey, producing a new distribution of income and translating the evolution of macro variables into poverty outcomes. Unlike the simpler poverty elasticity- based methods, the micro-simulation approach allows for four main avenues of escaping poverty: moving from agricultural employment to non-farm activities where the wages tend to be higher, upgrading individual skills (through schooling), changes inrelative wage changes, and an economy-wide growth component that equally benefits all households. F8. The production functions for education and the other MDGs have been calibrated to assure that, under base-year conditions, base-year performance is replicated and that, under a set intermediate variable is fed to a logistic function with diminishing marginal returns in the intermediate variable (and, indirectly to the arguments o fthe underlyingCE function). 210. A large micro-econometric literature exists on how to model the production and the demand for education. Our approach simply summarizes the most salient points of this literature in a simple reduced form equation. The main advantage o f our approach is its general equilibrium method but it i s not at all a substitute to more detailed sectoral analyses. 295 Annex F of other conditions identified by sector studies, the relevant targets are fully achieved. This approach also implies that the modeling o f MDG production does not mechanically link more inputs, i.e. increased production o f social services, to higher attainments, but reproduces, in reduced form, the functioning o f the sectors and the interactions o f demand and supply within them. Thus, the model consolidates knowledge from sector studies and incorporates various bottlenecks that are likely to hinder MDGachievement. F9. A major advantage o f MAMS is its ability to simultaneously determine MDG progress, supply and demand o f private goods and services, and factor markets equilibrium. Inthe model, this is accomplished through several links between the MDG module and the rest of the economy. Firstly, the increase in government service provision needed to reach the MDGs requires additional resources4apital and investment, labor, and intermediate inputs-that become unavailable to the rest o f the economy. As an example, this link allows the model to capture the wage hikes for skilled labor that may stem from the combination o f its small supply and rapid demand expansion (from MDG services that are intensive in skilled labor). These demand increases are likely to reduce the number o f skilled workers available to the private sector even as increased school enrollment reduces the overall labor force. On the other hand, by allocating graduates and dropouts from the education system to different segments o f the labor force (as a function o f the highest degree achieved), MAMS captures the supply side o f expanding education services, which works in the opposite direction by depressing relative wages o f labor categories with the most rapid growth. If government services rely heavily on these labor types, government costs and financing needs decline. Moreover, a higher average level o f education raises the productivity and the wages o f the labor force, with a positive feedback on growth, private incomes, government revenues, andMDGachievements. F10. The effects o f any program depend on its source o f financing, be it from foreign donors, domestic taxes, or from domestic borrowing. Other things being equal, higher taxes reduce private savings and consumption spending and have an immediate negative impact on efforts to reduce poverty and stimulate growth. On the other hand, increased domestic borrowing crowds out private investment with little direct impact on private consumption, but i s likely to spell pronounced negative long-run consequences for private capital accumulation and growth. Financing from foreign sources is also not free from perils, as increased foreign aid may lead to exchange rate appreciation with economy-wide repercussions, including a loss o f competitiveness in export sectors (`Dutch disease') and incentives for consumers to switch from domestic outputs to imports. At the same time, the pursuit o f the MDGs generates additional resources as it influences the composition o f the remaining labor force, raising the shares o f skilled and tertiary-skilled workers. The performance o f the rest o f the economy will also influence the ease with which different MDGs can be achieved. Higher private disposable incomes provide the additional resources that enable private households to draw more benefit from government health and education programs. More rapid growth raises government revenues, strengthening the ability o f governments to finance and operate efficient programs. 296 Annex G Annex G Modeling Continuation inthe School System G1. Continuation inthe school system was modeled using "survival function" techniques and household survey data for Hondurans between the ages of 7 and 65, most of whom have completed their schooling. A "failure" inthis treatment is definedas leaving the school system. This method allows the expected grade attainment to be conditioned on the level of schooling already attained. G2. Column 2 of Table G.l: Survivalin Schooland ConditionalExpectedGrade Table 2 shows Attainment inHonduras (3) Mean Expected estimated by Years (2) ProbabilityWill Attainment if Currently (4) Years of grade. The same data Completed Enrollin This Grade Enrolled School to go are the basis for Figure 0 100.00 5.6 5.6 2. notableThe two most 1 95.26 5.8 4.8 trends are the 2 87.51 6.3 4.3 sharp drop in 3 76.90 6.9 3.9 enrollment transition 4 65.42 7.5 3.5 57.52 8.0 3.O from grades 1 to 6 and 6 50.28 8.5 2.5 the fact that only halfof 7 26.60 10.6 3.6 sixth grade graduates 8 22.92 11.2 3.2 stay in school. The 9 19.41 11.8 2.8 pattern i s typical of 10 15.17 12.6 2.6 two-tiered, low-income 11 13.11 13.0 2.0 Latin American 12 10.77 13.4 1.4 countries with unequal 13 4.78 15.2 2.2 distribution of income. 14 3.49 16.1 2.1 2.74 16.6 1.6 After the 15 system is 16 2.08 17.1 1.1 "cleaned" of these 17 1.43 17.7 0.7 students the survival 18 0.54 18.8 0.8 function is quite flat, 19 0.3 1 19.3 0.3 indicating that a large 2o 0.09 20.1 0.1 proportion of the 21 0.01 21.8 0.8 remaining students will 22 0.00 23.O 0.0 be "going all the way", finishing secondaryschool. G3. A first pass at estimating expected grade attainment for the entire population that is still in school is average (mean) grade attainment for the immediately preceding cohort that has completed schooling. This would be fine for a developed economy where relations are relatively stable. Indeveloping countries, where relations change very rapidly it is unwise to ignore strong secular trends. G4. The link betweenthe survival function estimates and future educational demandis shown incolumns 3 and 4 of Table 2. The likelihoodthat a child will attain grade 12 is not the same if itis currently ingrade 1or already ingrade 11. Column 3 showsthis relationexplicitly: average expected grade completion, conditional on having attained a given grade. We can see, for instance, thejump inexpectedadditional years of schooling, once a child is enrolled in7* grade. 297 Annex G G5. More importantly, these conditional expected completion rates can be combined with estimated current grade-level enrollments to obtain a picture of overall future school attainment andtherefore also a profile of demand by grade level inthe near future. 298 Annex H Annex H The FiscalImpact of Early TeacherRetirement H1. The Estututo del Docente stipulates that seniority benefits can rise to 120% o f base salary. Under existing practice, and according to the PASCE agreement, there can be a maximum o f 7 seniority bonuses, each at 668.04 Lempiras, for a maximum o f 4,676.28 Lempiras. This raises the possibility o f reducing total salary expenditures by replacing more senior teachers with new ones through some sort o f early retirement inducement. H2. The retirement benefits that determine the incentive to retire and the national budgetary impact o f an early retirement program are laid out under the "Ley del instituto nacional de previsih del magisterio' (INPREMA)." H3. Contributions: Teachers contribute 7% of their total remuneration -including extra pay for seniority, education, hardship posts, and other positions occupied. The government contributes an additional 12% (private employers an additional 11%) o f total pay (Article 27). The base for determining these contributions i s the total remuneration from all teaching jobs and administrative positions held by the participant. Inaddition INPREMA generates revenues from interest payment on loans to participants and from other investments, including real estate transactions. H4. Contributions entitle the participants to a number o f benefits and financial services. The benefits consist of: e Retirement benefits e Voluntary severance pay e Life insurance e Handicap insurance The services include personal loans, mortgage loans, andthe automatic payroll deduction of loan repayments. H5. Retirement: Participants are entitledto a lifetime monthly retirement payments after age 50 ifthey have served for at least 10 years. The amount o f the monthly annuity they will be paid is a fraction o f the average salary they earned over their last 36 months o f service. What proportion they are to be paid inbenefits o f this 36-month average depends on years o f service. Participants receive 3% o f the 36-month average salary for each year o f service. This i s capped at 90% after 30 years o f service. H6. For the purpose o f calculating this percentage (but for calculating the 10 year minimumservice eligibility), service inrural areas counts for time-and-one-half. Thus two 50- year old teachers, with 10 years o f service, one rural the other non-rural, would both be eligible to retire, but the rural teacher would receive the annuity o f an urban teacher with 15 years o f service. The urban teacher with 10 years o f service would receive a 30% annuity, while the rural teacher with 10years o f (actual) service would receive a 45% one. Also, e Payments continue to a designated beneficiary for 60 months after the death o f a retired participant. 211. Decreto Ley Numero 1026. 299 Annex H e The "mandatory" retirement age is 60, but it can be waived on appeal up to age 70. H7. Voluntary Severance: Participants that choose to withdraw from the system (leave the teaching profession) before the minimum 10 years o f service or 50 years o f age may choose to leave their benefits with INPREMA, where they earn a statutory 6% per annum. They may also opt to take with them 70% o f the average monthly salary o fthe previous 36 months o f service for each year that they have served. Ifthey later choose to reenter the system they are given 5 years to repay this along with interest on that amount calculated at the average yield on INPREMA investments over the period. e Teachers who withdraw fkom the system after 10 or more years o f service but before they turn 50, are entitled to retire when they turn 50 unless they opted to take the cash severance Pay* Retirement Timing H8. The Honduran career in education is designed to begin after graduation from school and retirement at age 50. The lowest degree required i s "Maestro Titulado", meaning graduation from a normal school after a total o f 9 years o f formal education. Students that enrolled in first grade at the statutory age o f 7 can be as young as 17 when they assume their first teaching post. This would leadto 33 years o fteaching seniority at age 50. This is exactly the amount needed to receive the maximumpensionbenefits for the longest period o ftime. H9. The law governing teacher salaries contemplates automatic raises for time served according to Table H.1. Onthe other hand, the law governing INPREMA provides for TABLEH.lSeniority Bonus pensions rising at 3% per year o f service, Years of Service YOBonusSalary (21 years of service =loo) 5 15 66 capped at 30 years o f service, or a 90% 10 30 74 annuity. This 90% is calculated at the 15 45 83 average salary earned over the last 36 18 60 91 months before retirement. This system 21 75 100 appears to provide a very strong incentive 24 90 109 to retire at age 50 even before the full 30 27 105 117 years o f service andright to a 90% pension 30 120 126 are reached. H10. Table H.2 illustrates a hypothetical scenario o f a 44-year old man who is facing his 21St year o f service. For reference, his base salary is initially set at 100 in this 21" year o f service. Though he has more than the minimum 10 years o f service, he would not be eligible to retire until age 50. The amount (58) he could earn in pension benefits if he were 50 is shown in parentheses in column (5). But he is too young to retire. Since he does not lose these pension benefits ifhe continues to work his salary net o f the pension entitlement is the same as his yearly salary through age 49. H11. His salary continues to grow with seniority, reaching a maximumof 126 after 30 years of service at age 53. However, after age 50 he could have pension benefits ifhe stopped working. He forgoes the pension as long as he continues to work. This forgone pension entitlement now represents an opportunity cost o f working. His real compensation working i s thus the salary minus his pension entitlement. At age 52 his salary net o f pension entitlement is just 18% o f 300 Annex H what it was when he was 44. While his salary reaches its maximum at age 53 and the pension does so at 55, salary net o f pension i s highest at 49. TableH.2. SalaryNet ofPensionForgone u r n (Hypothetical44-year old teacher with 21years of service) la (4) fa Age Yearsof Salarv Pension Salarv Net of Current Pension Entitlement 44 21 100 (58) 100 45 22 100 (62) 100 46 23 100 (67) 100 47 24 109 (72) 109 48 25 109 (77) 109 49 26 109 (82) 109 50 27 117 88 29 51 28 117 94 24 52 29 117 99 18 53 30 126 105 20 54 31 126 112 14 55 32 126 113 13 56 33 126 113 13 57 34 126 113 13 58 35 126 113 13 59 36 126 113 13 60 37 126 113 13 61 38 126 113 13 H12. So ifteachers decide to continue working past age 50 this decision is not based on current salary net o f pension entitlement. It must either be motivated by the entire stream o f post- retirement benefits or by a cash flow problem. Table H.3 shows actual data from the Honduran Ministry of Education payroll database for August, 2006 for just over 58,000 public sector teachers on the payroll. About 12% (7,331) o f them have already reached 30 years o f service. Figure F.1shows the seniority/salary structure. H13. Table H.3 shows the number o f teachers according to years o f service. The mode occurs with 6% o f the teachers at 5 years o f service and then gradually tapers off. Despite the abrupt reduction in net earnings after 30 years o f service, a substantial number o f teachers (12%) continue working. FIGURE H1 Honduras2005 Monthly Salary by Years of Service (all Teachers) - Total MonthlySals 301 Annex H TableH.3 All PublicSchoolTeachers Years of Number of YOof Cumulative Years of Number of YOof Cumulative Service Teachers Total Y O Service Teachers Total - % 0 345 0.59 0.59 26 1007 1.73 83.37 1 1662 2.86 3.46 27 690 1.19 84.56 2 1920 3.31 6.77 28 970 1.67 86.23 3 2210 3.81 10.57 29 663 1.14 87.37 4 2537 4.37 14.94 30 629 1.08 88.45 5 3469 5.98 20.92 31 1028 1.77 90.22 6 2721 4.69 25.61 32 839 1.45 91.67 7 2263 3.90 29.51 33 593 1.02 92.69 8 2534 4.37 33.87 34 620 1.07 93.76 9 2328 4.01 37.88 35 574 0.99 94.75 10 2018 3.48 41.36 36 438 0.75 95.5 11 1763 3.04 44.40 37 416 0.72 96.22 12 1162 2.00 46.40 38 412 0.71 96.93 13 2259 3.89 50.29 39 349 0.6 97.53 14 1833 3.16 53.45 40 340 0.59 98.12 15 1662 2.86 56.3 1 41 273 0.47 98.59 16 1484 2.56 58.87 42 222 0.38 98.97 17 1367 2.36 61.22 43 147 0.25 99.22 18 1736 2.99 64.21 44 110 0.19 99.41 19 1426 2.46 66.67 45 76 0.13 99.54 20 1807 3.11 69.78 46 102 0.18 99.72 21 1352 2.33 72.11 47 37 0.06 99.78 22 1019 1.76 73.87 48 43 0.07 99.86 23 1141 1.97 75.83 49 23 0.04 99.9 24 1639 2.82 78.66 50 17 0.03 99.93 25 1728 2.98 81.63 >50 43 0.08 100 Source: INPREMA, 2006. H14. Figure H.2 shows a Kaplan-Meyer survival function estimate for teachers on the job. There is no clear breaking point in years o f service. The seniorityhetirement rate is relatively constant. FIGUREH2 Honduras(2005)TeacherSunrival Years inService Annex H H15. Theredo appear to be systematic differences across individualcharacteristics. Estimated hazard ratios are shown in Table H4. Ratios greater than one indicate variables likely to encourage quitting work as a teacher, while Table H 4 The Hazard of Teachers Quitting those less than one are associated with staying Variable Haz. RatioStd. Err.Z-score on. Estimates indicate that teachers are less # ofJobsHeld 1.22 0.02 14.75 likely to quit if they have more hours, are older, Salary 1.00 0.00 -6.2 are female. Administrators are also less likely Hours Worked 0.97 0.00 -10.34 to quit. Incentives to quit include holding more Age 0.85 0.00 -37.98 than one job and being one of the less qualified FemaleAdministrator 0.66 0.18 -1.56 0.73 0.03 -7.49 teachers ("empiricos JJ)212. So the incentive Maestro Empirico 8.54 1.68 10.92 System seems to be working, especially with MaestroTitulado 1.08 0.08 0.99 regard to teacher training: less educated teachers Profesor Empirico 5.06 0.53 15.36 are more likely to leave. The fact that ceteris SinTipoProf. 8.65 0.86 21.62 paribus women are also less likely to leave may TBcnicoUniv. 0.56 0.06 -5.75 reflect the fact they are probably saddled with most o f the childcare responsibilities and a teachingjob provides the ideal schedule for a working mother. 212. The comparison group (omitted) is "Profesor Titulado", middle and high school teachers with degrees in secondaryeducation. 303 3 3: m 2 00 2 3 3 2 0 3 41 I 4 3E \o 2 m c1 3 .3 2 2 N- I v,N 3 m -- I- 2 m 2 2 N" I \o m m 3 *- -2 3 2 N I m z 2 3 x 3 3 2 I 3 2 m 3 3 2 3 2 I N 2 N ", m c1 2 P m El 41 0 \o N N P 2 m m 2 -! N z v, d d m d 2 m 0 N N m 0m 00 m m d 00 N :- m 2 I- co Annex H H16. A teacher that continues working past age 50 e continues earning a salary e may increase current pay through 30 years seniority e may increase future retirementbenefits through 33 years seniority e loses pensionbenefits duringthe years worked beyond age 50 H17. The total impact o f these effects is illustrated in Table H.5 by continuing the previous example. It i s assumed that our 44-year old teacher will live until the age o f 80. H e then bequeaths the pension to a beneficiary for an additional 5 years after his death and discounts this bequest at the same 3% rate used for his own benefits. The third row shows salary rising from 100 to 126 after 30 years o f service, as provided in existing seniority legislation. H i s pension rises from 58 to a maximum o f 113 at age 55. H i s working salary net o f pension benefits forgone is shown inrow 5 to reach a maximumo f 109 at age 48 and then falling from age 50 and beyond. H18. If he continues to work, his pension benefits continue to change. Up to age 50 it is simply a gain intotal yearly earnings. After age 50 the relation is more complicatedbecause on the one handworking longer causes his yearly retirement benefits to increase, but on the other he is cutting into the number o f years over which he will earn those benefits. The discounted present value o f these two effects (PV) is shown inrow (6). Initially PV rises because seniority andthe yearly pension both rise untilthe 30* year o f service. Then even though the total benefit period declines, PV continues to rise because the moving average o f his last 3 years o f earnings is catching up to his maximum earnings. PV later falls since yearly benefits are constant at their maximum and continued work just reduces the total years o f benefits. H19. O f course if he continues to work he will earn a salary and lose a retirement year. The last row shows the total salary and P V effect o f delaying retirement for an additional year. The impact i s positive until age 53. Since it involves discounting, the result is sensitive to the rate o f discount used. The higher the discount rate, the less high future benefits will matter and the earlier it will pay to retire. For instance at the 23% rate o f interest INPREMA charges on loans to its members, the optimal retirement age falls to 50. H20. This analysis begs a question. Why do so many teachers continue working past 30 or even 33 years o f service? One financial reason could be cash flow. Notice that in the current structure, regardless o f how long they work, yearly pension benefits never exceed the yearly salary. Iftheir financial commitments -say a mortgage, children inschool, or an ailing relative - are higher than what they can pay out o f their pension entitlement, they will not be able to afford retirement. Another reason o f course is some people may like working. They would continue to work because they enjoy teaching more than anythingthey could do intheir leisure time. H21. Not much can be done to buy out the people who are essentially already nearly working for free. But ifthe problem i s cash flow, something may be done. INPREMA functions not only as a retirement fund, but also as a banker. In this capacity it provides mortgages and personal loans. It would not be difficult to examine INPREMA data and see if a significant percentage o f the more expensive, older teachers are saddled with debt. A cheap way o f inducing early retirement would thenbe to link it to debt restructuring. 305 Annex H Early Retirement and INPREMA H22. I s early retirement an option for reducing the teacher wage-bill? Any early retirement program should be preceded by a careful analysis o f its likely impact on INPREMA finances. INPREMA is a pay-as-you go system, much like the U.S. Social Security program. It i s not a vested retirement system. Contributions by current workers pay the benefits transferred to retirees. Early retirement reduces the contributions and increases the number retirees. It could have a deadly impact on the financial solvency o f the retirement system. Indeed, as shown in Table 2.14 o f Chapter 11, INPREMA's net worth is not projected to become negative until at least 2023 if teachers were to retire at the age o f 60, but it would already become bankrupt by 2015 ifthey were to retire at age 53, H23. Consider now the impact o f PASCE on the retirement decision. Two things determine retirement benefits: years of service and average salary over the 36 months o f service prior to retirement. This second provision is crucial to PASCE's impact. If a teacher with 30 years o f service were poised for retirement inDecember o f 2006, she would have earned the average o f 3 years o f pre-PASCE salary. N o w there i s an extremely strong incentive to wait at least 6 more years before retiring, and this is especially truefor the most expensive teachers, i.e. for the ones that would be targeted by an early retirement scheme. H24. Consider the example o f a primary school teacher withthe title o f "Maestra Titulada" and 33 years o f service. Her base salary would be 5,216.64, her bonus for the title would be 3,073.12, and her 7 seniority bonuses (at 668.04 each) would be 4,676.28, for a total o f 12,966.04 Lempiras per month. Since her seniority bonus reached its maximum after 30 years o f service, this would also be her average salary for the previous 36 months. She could have retired inDecember 2006 andearned 90% ofthis average, or 11,669.44 per month. H25. N o w PASCE i s in effect. If she works for the next three years, her salary will go up to 18,219.50 per month. But if she retires then, she faces the fact that her retirement benefits don't fully incorporate this raise, because they reflect the average o f the last 3 years. Ifshe retires in 2009, she earns retirement benefits o f 14,803 per months, but if she waits until 2011,they will go up to 16,398, almost 11% higher. PASCE is therefore going to have the effect o f delaying retirement by 3 to 5 years. This means that ifteachers are to be induced to retire earlier in order to reduce the total size o f the teacher wage bill, it will require a higher fiscal incentive now that PASCE i s ineffect thanit would have cost before. The OverallFiscalImpact of Early Retirement H26. An early retirement scheme may conceivably reduce the Education Ministry's fiscal outlays. This is conceivable if the oldest, most expensive teachers can be replaced with new teachers, who have not yet accumulated any benefits. Even if this were possible, however, it does not appear that there would be an overall fiscal savings for the public sector as a whole. Rather, the fiscal savings achieved by the Education ministry would be entirely offset by an teacher aged 50 who is poised to retire. As shown inTable F.5, her total earnings (base salary + increase in the liabilities o f the teacher retirement fund, INPREMA. To see this consider a bonuses) would amount to Lps.117 if she continues to work. Were she to retire, her pension receipts would be Lps. 88. Maintaining this experienced teacher on the payroll, therefore, has a marginal overall fiscal cost o f Lps. 29. The marginal fiscal cost of hiring a new teacher, on the other hand, is at least equal to Lps. 29. (To see this, consider that the maximal amount o f 306 Annex H bonuses that a teacher can receive i s around 300% of the base salary. This means that the salary ofthe 50-year old teacher, which i s normalized at Lps. 117 inTable F.5, can be decomposedinto a base salary that is at least 29.25, plusbonusesthat at most add up to 87.75. So, the newteacher would be earning at least the base salary of Lps. 29.25. In other words, replacing the older teacher with a young teacher does not yield any overall fiscal saving and could result ina decline inquality, assumingthat there is value inexperience. H27. Insum, there i s no clear advantage to be had from an early retirementscheme interms of reducing the overall fiscal cost of PASCE. Such a scheme would merely shift a fiscal burden from the Ministry of Education (or Finance Ministry) to INPREMA, thereby accelerating the latter's bankruptcy. Insofar as INPREMA is part of the public sector, the resources saved inthe Education budget would have to be spent on propping up INPREMA, without any net fiscal savings to behad for the public sector as a 307 Annex H 308 Annex I ANNEX I InstitutionalandGovernanceIssuesinthe HealthSector 1.1 After the MOF establishes the budget envelope, the M O H allocates its budget among three main recipients: the central level, the 20 Regional Health Departments (RHDs) 213 and hospitals, across six different programs and nine types o f expenses such as personnel services, non-personnel services, equipment and inputs andcapital goods. Figure 1.1 BudgetAllocation Stages Payment Mechanisms Source: Adapted fiom Ruano, E.2004. 1.2 In 2004 only 24 percent of the approved budget was allocated to the RHDs through programs 11 (ambulatory care and epidemiological control) and 12 (environmental and health promotion) -of which 24 percent o f this portion went to personnel expenditures (MOH, 2004). Budget allocation had previously followed historical trends and estimated average budget execution rates for 2001 to 2004 were less than 60 percent (Ruano, 2004). Since 2005, inorder to achieve a more equitable distribution o f resources, the M O H has been applying the following resource allocation criteria: 1.3 The MOH's reorganization o f RHDs and the new budgeting criteria have generated important challenges for the RHDsfor the following reasons: (a) the new allocation criteria reduced the weight given to population size. As a result, RHDs state that their resources have been seriously diminished and they are "resisting" the new allocation criteria ((Orstrom et. a12005). (b) although the MOH is trying to correct the serious imbalance in distribution o f personnel across the regions, the current Estatuto Medico accords physicians the privilege to move with their posts to another RHD. This hinders the MOH's control of personnel 213. In May 2004, the decentralization process moved forwarded with the establishment o f 20 Regional Health Administrations,: 18 based on the political divisiodclassification in the country and 2 that correspond to the Municipalities o f San Pedro Sula and Tegucigalpa. 309 Annex I deployment and to some extent, weakens the links between planning, implementation, andresults. (c) Budget allocations to the RHDs do not include funds from external aid, yet 39 percent of the resources from program 11 (ambulatory care and health promotion) come from this source o f funds. This could generate potential imbalances among the RHDs if there are important differences between the donors' and MOH allocation criteria. This external aid can also introduce imbalances among or fluctuations into the RHDs flow o f funds. For example, technical assistance provided to the departments or municipalities by international donors i s not quantified and not considered in the RHD budget. Sustainability issues could arise if this TA were required after the external aid is finished becausethe RHDswould have to finance it with their own funds. Table 1.1 ResourceAllocationCriteria applied to RegionalHealthDepartments Criteria Measure Weight (YO) total RHD's population Number o f inhabitants per RHD214 10 population with three or more unsatisfied basic Based on census data using access to 30 needs (NBI potable water, education, overcrowding, housing conditions Priorityhealth problems TB, AIDS,diarrhea andpneumonia, 20 hypertensionand diabetes Initial Consultation Annual number of first-time external 10 consultations ina health unit Production index total medical attention per year by age 5 group at the primary health care level) Health status life expectancy 10 Health status Chronic malnutrition inchildren at first 15 school grade level ( 6 to 9 years old) Source: Adapted from Ruano, 2004. Note: At a later stage, the MOH is considering introducing performance indicators into these allocation criteria. Decentralization, Budget Allocations, and Execution 1.4 The RHDs also faces some challenges inallocating and implementing their funds due to gaps or inconsistencies in the decentralization policy. For example, there is a lack o f regulation o f the decentralized funds. Health facilities can charge user fees or receive transfers from municipalities to complement M O H funding. These two sources o f funding have proven to be essential for health facilities to finance their personnel or to buy inputs (Fiedler 2000). However, these funding sources tend to be irregular because municipalities can either stop or cut their transfers and the revenues from user fees are subject to changes in demand. Thus, when these funds are unavailable, RHDs need to offset this cutback which can be difficult because o f their other commitments and because they are entitled to only a small amount o f discretional funds, i.e. 15-20 percent o f the total budget. The volatility o f municipal funding could be prevented if institutional arrangements are introduced that ensure a regular source o f funds for the health units.Inthe case o f user fees, although the RHDs are entitled to set the fee levels, they do not use clear and systematically applied criteria to ensure optimal fee levels based on the 214. The IHSS insuredpopulation is deducted for the municipalities o fTegucigalpa and SanPedro de Sula 310 Annex I particularities of eachhealth unit's situation (Fiedler et. a12000). Addressing this issue could not only improve the regularity of this source of funds, but also prevent potential inequities in the population's access to financial protection. 1.5 On the other hand, RHDs operational responsibilities are restricted to carry out supervision tasks, and they only have very limited discretional authority to reallocate funds thereby limiting the benefits of promotinga program-basedbudgetingapproach. Moreover, both M O H and RHDs also have limited institutional capacity to plan a realistic budget; resulting in their approved budgets being subject to many modifications during the fiscal year. Also, the RHDs supervision tasks are constrained due to central laws such as the Estatuto Me`dico. This Law regulates working conditions for physicians and even stipulates the process of hiring and firing them. Thus, as long as no other mechanisms can be implemented, the chances to reduce the high rate of absenteeism (73%) and the proportion of ghost workers (2.4%) found by the Public Expenditure Tracking Survey in 2000 are still low (World Bank 2003). Likewise, given the serious inefficiencies in the distribution of medicines by the central level as discussed in section E of this chapter, the RHDs and health units cannot be held fully accountable for not providingthe services programmed inthe operations plansor the quality of their services. 311 Annex I 312 Annex J AnnexJ Preparingthe Medium-termExpenditureFrameworkfor the HealthSector inHonduras J.1 This section describes the approach being usedto evaluate the potential cost and impact of Honduras' efforts to strengthen the contribution of health services in order to achieve health outcome goals in the context of the ongoing health sector reforms, the National Poverty Reduction Strategy or EW (Estrategia de Reduccion de Pobreza), and the Government's recently established results-based approach towards monitoring sector performance. In particular, it seeks to answer the following questions: o Howmuchadditional fundswould beneededto increasethe effective coverage of and quality of providing high impact health services interventions from the current level to meet the institutional goals (metas institucionales), and the 2015 MDGtargets? o What results, interms of the reduction in child and maternal mortality, could be expected (on the basis of the most recent knowledge available) from these extra funds? o If additional resources are not available or insufficient, what can be done to improvethe efficiency o f sectoral interventions? 5.2 To respond to these questions, the Government in collaboration with UNICEF and the WB has agreed to adapt the Marginal Budgeting for Bottlenecks (MBB) TooI2l5developed by the World Bank, UNICEF and WHO, and then validate the results through national and international expert opinion groups and comparethe findings with the outcomes o f other studies. 5.3 The MBB, an analytical costing and budgeting tool assisting policy makers to plan and managehealthprograms, will beusedto: Guidea country-based process ofidentificationof bottlenecks to the successful service delivery of high impact interventions; Calculate the incremental cost of removingsystemic and intervention-specific bottlenecks to achieve effective intervention coverage; and Estimate the potential impact of health interventions on health outcomes, such as rates of child andmaternal mortality, HIV/AIDS, nutrition, andmalaria- associated(which is an issue insome regional departments inHonduras) mortality. 215. While MBB focuses on marginal and incremental cost and impact, making it a particularly helpful tool to estimate the extra efforts and resources neededto reach the MDGs, it also demonstrates the results bought by current resources/investments in the health sector. It has been used in China, India (some priority states), Benin, Ethiopia, Ghana, India, Madagascar , Mali, Mauritania, Nigeria, Rwanda, and Sierra Leone. Four countries (Benin, Mali, Madagascar, Rwanda) have usedit to prepare their MTEFs. 3 13 Annex J Applying the MBB Tool2" in Honduras: Process and Methodology 5.4 The process has three mainphases: J.5 Phase I(March 2007). A joint UNICEF-WB team held working sessions with staff from key Ministry o f Health Department orient them to the approach, discuss the achievements and constraints inthe health sector, and to collect health intervention coverage baseline and cost data needed for the MBB analysis. Meetings were also held with the Vice Ministers o f Health and the Presidency, as well as representatives from the Ministry o f Finance to briefthem on the approach and determine their interest to be the first country in Latin America and the Caribbean to apply the tool. The Government agreed to adopt the approach and pilot it in the health sector in Honduras for the preparationo fthe 2008 planand budget which are due inJune. J.6 Phase I1(April 2007). A workshop will be held from April 17-19that will include policy makers, key health program directors, 20 regional health department representatives, and technical staff from the M O F and MOP to discuss and agree on the specific program strategies that would help define the basic packages o f interventions that will be used for the analysis. Agreements will also be reached regarding which priority policy options (for example whether conditional cash transfers will be used to improve utilization o f health and nutrition services, zinc supplementation, HIV/AIDS treatment, package, etc.) will be used for cost and impact simulations. 5.7 In generating the results, the MBB analysis will rely on existing information at the country level including survey data such as the DHS 2005/06 and ENCOVI 2004 and government documents (program reports, action planetc.), especially the National Health Policy Document 2006-2010 and the Health National Plan 2021, Additional information from studies undertaken at regional or sub-regional level will also used if Honduras-specific information will not be available. 5.8 Fivemainsteps will be involved: a) Analysis of the health system which includes: (i) the epidemiological profile of the country; (ii) the organization o f the health delivery system; (iii) standards that guide the the distribution o f facilities, personnel, and equipment; (iv) the health policies in place; (v) the health outcomes and coverage goals the system aims to attain; and (vi) the health care interventions already offered by the system. b)-Identijkation of high impact health interventions based on existing and planned service delivery strategies that would improve not only supply o f services, but also demand for the interventions; J.9 The selection o f effective interventions to improve child health will be based on the Lancet series o f child217and neonata1218survival (See table 7.32). These series offer a review o f the literature that identifies both preventive and curative interventions with proven effects in 216. MBB is comprised of five modules: (1) a health services delivery arrangements module; (2) a health service coverage module (sometimes also called bottleneck identification module) , (3) a costing module; (4) a budgetingmodule; and (5) animpact module. 217. Lancet, 2003, Child Survival: 361,362. 218. Lancet, 2005, Neonatal Survival. Published on-line, March 3,2005. 314 Annex J reducing neonatal and child mortality and which are feasible for delivery at high coverage in developing countries. The selection o f interventions to improve maternal health was based on a review o f the literature done by the World Bank, WHO and UNICEF on the basis o f a Cochrane analysis.219 Finally, the selection of interventions to effectively prevent and treat malaria and HIV/AIDS was based on evidence from the Roll Back Malaria Initiative and UNAIDS respectively. J.10 The interventions that will be used in the Honduras exercise will be based on the epidemiological profile o f the country, the effectiveness o f the interventions in accelerating progress in achieving the MDGs and sectoral goals, and on whether they are already offered by the system, or are plannedto be offered c) Identification of bottlenecks in the delivery of these interventions. To identify delivery constraints, an analysis was made o f bottlenecks that hinder increases inthe coverage o f the selected interventions. For this analysis, all the interventions were classified according to the mechanism by which they are delivered and thus according to the inputs and strategies needed intheir delivery (See Box 1). Box J.l. Classification of healthcare interventionsaccordingto their deliverymode. Family/ community based care: These interventions, which include some preventive measures and the management o f maternal and childhood illnesses, can be delivered by the households or the communities themselves under some guidance by health professionals. Insecticide treated bed nets (ITNs) for pregnant women and children under five, condom use, breastfeeding, and oral re-hydration therapy (ORT) are some examples o f interventions that are family or community based. Population oriented services: These services are delivered to all segments o f the population, regardless o f whether or not they are currently sick. They are usually delivered through periodic outreach or scheduled clinical services. This delivery mode includes the following preventive care interventions: immunizations, ante-natal care, family planning, etc. Clinical based individual care: These activities include all type o f individual curative care interventions that needto be delivered ina health facility and by a trained health care professional. They are offered ina continuous manner so that they can respond to unpredictable situations: a sudden illness, a delivery, etc. Source: Soucat et al. 2005. J.11 Interventions inside each service delivery group share similar constraints intheir delivery and thus the analysis o f bottlenecks was done for each o f these groups. This analysis will include an examination o f different indicators o f the availability, accessibility, utilization, continuity, and quality o f these interventions220: availability: these indicators assess the availability o f critical health system inputs such as drugs, vaccines, supplies, and/or humanresources. accessibility: these indicators examine physical access to health services, such as access to trained humanresources at community level, number o f communities/municipalities (depending on level o f data availability) are regularly reached by outreach services, andtime to reach facilities providing basic and emergency obstetric care services. 219. Cochrane reviews are based on the best available information about healthcare interventions. They explore the evidence for and against the effectiveness and appropriateness of treatments (medications, surgery, education, etc) inspecific circumstances. 220. Soucat et al. MBB Manual. 315 Annex J utilization: these indicators evaluate the use of health services, measuredby first contacts withhealth service providers, for example, the percentageof children receiving their first dose of DPT immunization, the percentageof women going for their first antenatal visit, etc.. adequate coverage: these indicators compare the actual contact with providers relative to optimal numberof contacts. For instance, inthe case of DPT this indicator will measurethe percentageof childrenof the requiredage that have receivedthe three doses; the percentage of women who go for at least four prenatal controls, etc.. quality: these indicators examinequality of care by assessing the skills o fhealthworkers, for exampleinterms oftheir ability to examinethe beneficiary, to diagnose, to provide the requisite interventions, to use the equipmentand to advice appropriately. One indicator of quality coverageinthe case of ante-natalcare would be the inclusion o fthe most important components of ante-natal care (e.g. bloodpressure, tetanus toxoid immunization, etc.). d) Formulation of a strategy to eliminate these constraints; and setting of coveragefrontiers or goals across time for these services; and e) Simulations of the cost and impact of the proposed strategies. These estimates will first be based on needs or the total resources needed to reach the MDGs and institutional goals assuming that there are no funding constraints. Finally, various policy scenarios will be estimated based on government priorities and allowing for various government's financing strategies (Le. different annual spending schemes dependingon resource availability). J.12 Phase I11(May 2007). Refining the model based on discussions with the Ministries of Health, Finance and Presidency, taking into account the country's macroeconomic situation and fiscal space, inorder to establish a mediumterm expenditureframework for the health sector. An annual budget for 2008 will then bepreparedand vetted, for final submission inJune. 3 16 Annex J Table J.1EffectiveNeonatal,child, and maternal healthinterventionsbasedon the Lancet Neonataland Child Survival Series and on a Cochrane review of the literature on effective maternal healthinterventions NeonatalInterventions Child Interventions MaternalHealth ~~ Pre-conception Preventive Interventions Oral contraceptives Folic Acid Supplementation Breastfeeding Condom promotion Antenatal Insecticidetreated materials Iron supplementation during pregnancy Tetanus Toxoid immunization Complementary feeding Folate supplementation during pregnancy Syphilis screening andtreatment Water, sanitation, hygiene ITNfor pregnant women Pre-eclampsia and eclampsia: Hibvaccine Antenatal steroids prevention (calcium supplementation) Intermittentpresumptive Zinc Clean delivery treatment o fmalaria Detection andtreatment o f VitaminA Assisted delivery asymptomatic bacteriuna Intrapartum Nevirapine andreplacement Access to Basic andComprehensive feeding Emergency Obstetric Care Antibiotics for preterm premature Measles vaccine pre-eclampsia detection and rupture o f membranes management Corticosteroids for preterm labor TreatmentInterventions Detection and management o f Oral rehydration therapy breech (cesarian section) Labor surveillance for early Antibiotics for pneumonia diagnosis o f complications Clean delivery practice Antimalarials Postnatal Antibiotics for dysentery Resuscitation o fnewborn baby Zinc Breastfeeding VitaminA Prevention and management of hypothermia Kangaroo mother care Community-based pneumonia case management Source: Lancet, 2003, Child Survival Series: 362 and Lancet 2005, Neonatal Survival Series, 2005 (Published on-line) and a Cochrane review o fmaternalhealth interventions (see Soucat et. al: MBB Manual). 317 Annex J 318 Annex K Annex K Recent and Projected Port Traffic in Honduras, 2000-2006 Source: ENP 319 Annex K ProjectedCargoTraffic in The PortsOfHonduras Unit of 2007 2008 2009 2010 Measure 1. MARKET SERVICES Metric Tons 9,106,669 9,540,527 9,995,433 10,472,423 Puerto Cortks 7,225,642 7,586,924 7,966,270 8,364,584 Tela 205,000 211,150 217,484 224,009 La Ceiba 3,450 3,622 3,803 3,994 Puerto Castilla 803,836 844,028 886,229 930,540 San Lorenzo 868,714 894,803 921,647 949,296 2. SERVICES TO SHIPS Ships 2,212 2,265 2,321 2,378 Puerto Cortks 1,800 1,845 1,89 1 1,938 Tela 27 27 28 29 La Ceiba 90 92 94 96 PuertoCastilla 161 165 169 173 San Lorenzo 134 136 139 142 3, CONTAINER SERVICES TEUs 596,960 626,807 658,148 691,055 PuertoCortks 509,588 536,067 561,820 589,911 PuertoCastilla 87,372 91,740 96,328 101,144 Source: ENP PlanningDirectorate. 320 Annex L Annex L SummaryRecommendationsfrom the 2001 and2007 PERsandthe 2003 PPIAF Report 1Sector Issue 2001 Actions Taken 2007 PossibleNext Recommendation Recommendation Steps Definea clear and Some incremental Recommendation Nextpriorities coherentstrategy: address achievements,e.g. remainsvalid as shouldbe: the large infrastructure gaps; the airport governmenthasnot developmentof a give greater weight to concession, a frst yet adopteda clear PPPpolicy for improvingaccess, quality roadsconcession andcoherentstrategy. ports andthe andefficiencyinservice inplace. Interest extensionof delivery ratherthan expressedinPPP concessions inthe maximizingshortterm for portsand roadsector. fiscal gains- roads. Mustbe groundedinclear Limitedprogress Recommendation Studyto be carried objectivesfor: promoting achieved. remainsvalid. out onthe competitionwhere Predominantrole Institutional institutional possible; improving stillplayedby requirementsfor options. Future operationalefficiency and SOPTRAVIwhich effectiveregulation roleof SOPTRAVI qualityofserviceto has limited needto be reviewed inmulti-modal customers. Merge capacityinthis andthe agency regulationshould existing agencies into one area. responsiblemustbe bereviewed. multi-sector regulatory Superintendency givenclearer and body for autonomy and for concessions strongerpowers. effectiveness existsbut its autonomyis weak andthere are I questionsabout the legalframework e.g.the legalityof I roadtolls. Subsidies Restructurewith a view No changes have Recommendation Decision required to: addressingissue of takenplace remains valid. inregardtothe improvedaccess to though, inregard future ofbus servicesto have an impact to urbantransport, subsidiesin onreducingpoverty, as the restructuringof Tegucigalpa(see opposedto improved subsidieshas been below) affordabilityfor those studied(see with existingnetwork below). access. Improve analytical basefor targeting subsidies Expenditurepatterns Generallyaccepted Implement EnsureSEFIN shouldplacepriority on andbeingapplied agreementsreached projections are maintenanceactivities inpractice,suchas on future funding of alignedwith those inroads. However roadmaintenance(see of sector ministries aggregate below) inregardto expenditureson maintenance maintenanceare requirements. still insufficient I (see below) 221. The thematic recommendationsappliedgenerally to infrastructure 321 Annex L Sector Issue 2001 Actions Taken 2007 PossibleNext Recommendation Recommendation Steps Sub-Sec Ports Infusiono f capital and Progress being Policy statement on Bankcould provide management expertise to made interms o f role o fPPP inports i s further support to expand capacity and port performance, needed, together with PPP training and to reduce enduser costs IMO/TSA targets any consequent the development metto facilitate adjustment inthe o f policy accessto U S legal framework. statement; key market. Port performance indicators. indicators and Roatan port associated facilities have been benchmarks to be concessioned developed. Capacity inPuerto Cortes has to be increased Airports 4 international Performance o fthe airports are being concession would be operated under a reviewed on a regular 20 year basis. The residual concession. The role o f SOPTRAVI contract has has to be reviewed however had to be andprobably re-negotiated and downsized. the delivery of capital improvements is delayed. Rail Limitedservices FNHshouldbeshut There is a are being provided down and its debts potential, limited at an annual loss consolidated future role for rail by FNH, a public that could be company. explored ifprivate Services are partners are unviable, even interested in subsequentto a investment and 2003 downsizing management. program. Bank SOPTRAVI should provided resources pursue. to study options, also financial audit andasset valuation. Roads - Invest inwidening Good progress Concerted effort is Master plan needs Investment sections o fthe "north- being made. needed to improve to be updated, south" highway. Investmentinthe rural access, inturnto focusing on Investigate economically corridor is going support rural strategic upgrade viable concessions for ahead as planned, competitiveness. o froad corridors. construction and supported by a Study o f The priorities for management. Low cost donor coordination concessioning o f SOPTRAVI are: projects to expand access instrument (mesa other sections on the more contracting inremoteareasby de cooperantes). north-south and o f out; strengthen improving gravel roads A frst concession other key corridors social is under operation. should go ahead. management; 322 Annex L Sector Issue 2001 PossibleNext Recommendation Steps strengthen has beenmade on reviewof SOPTRAVI transportservices improvingrural andimplement management. roads. reforms. Workable axle load control systemmust beput inplace. Roads- Priorityfor public Gradualprogress At least 85% ofall Mesa de Maintenance expenditure-meetingfull has beenmade. publicnetwork cooperantes will costs ofmaintainingthe Pavedroads maintenanceby 2010 monitorplan roadnetwork. Ensure nearing 100% inlinewiththe implementation. effective andefficient use coveredthough revisedplan, Developproposal of dedicatedfundingby unpavedroads buttressedby the for movingtowards FondoVial. Ensure maintenanceis agreements reached 100%maintenance transfers are inline with lagging. The with the donors(mesa ofthe "red basica what idmandatedinlaw. micro-enterprise de cooperantes). municipal". Keep Mostroutine maintenance schemehas been Reviewofthe Fondo the micro- activities will becarried successful. Vial law to be carried enterprises out by micro-enterprises Concerns are that: out-focus onfunds programgoing. provisions ofthe flow, management Selective law are still not andgovernance introductionof beingmet; arrangements. servicelevel governance of contracts. FondoVial. UrbanTransport Bussubsidiesdon't A schemeto Decisioninprinciple Poverty impact benefitmarginalurban replacethe current is neededto go analysisneedsto areas whichhave no subsidyhasbeen forwardonthe basis be carriedout in services. Allocation studied. However of least cost subsidy regardto any new mechanismdoes not this schemedoes andcarefulroute schemeto be encouragecompetition, not address the selectionandthen introduced. lowersoperators' concernabout allocationona dependenceonsales service levelsand competitivebasis. revenue. Leastcost will not encourage subsidyschememay be new competition. preferableway to go. 323 Annex L Recommendationsemanatingfrom the 2003 PPIAF Report Legal, Institutional andRegulatory The (1998) General framework law for concessioning (Decree 283-98) created a regulatory arm (Superintendency of Concessions and Licenses for this purpose). It is already working inthecase ofthe airports, buthas aweak structure with little autonomy. Thereport proposed a single transport regulatory agency with members appointed for 5 years (with overlapping terms). This agency would absorb, inter alia: enforcement of weight and axle load control and control of interurban cargo transport (SOPTRAVI's efforts here were regarded as inadequate), There is some ambiguity on whether tolls can be charged. They are specifically excluded in the case of municipalities through another law (Decree 134-90). However, San Pedro Sula and Puerto Cortes still do this. Inaddition the Constitution (Article 81) may be interpretedas outlawing tolls on national roads. The lack o f global vision in transportation planning, e.g. development of Trujillo/Puerta Castilla as an alternative Caribbeanport, implementation of the concept o f the "canal seco", i s mentioned as evidence of the inadequacy of the current structure for policy, planning and regulation for transport. Sector Issues Roads - Road Fund law is not being followed in regard to transfers. No specific funding is available for urbanroad maintenance. Better specifications and supervision of contracts are neededto improve works quality. Improvements inroad safety would come, inter alia, from enforcementof law and its regulation. Master Plan.needs updating. Ports -Hightariffs are damaging competitiveness. Restructuring could help increasereturns and improve efficiency. Priority should be given to increase competition and inflow of private capital and management expertise. There is an urgent need to expand container and dry bulk cargo capacity at Puerto Cortes. A detailed sub-sector strategy and supporting legal framework should be developed. Urban Transport - The recommendation is to abolish the Tegucigalpa subsidy scheme. Improvement of regulation of urban transport could be achieved following the San Pedro Sula model. 324 4 Y $ ? i $ 2 2 ? ? Y ? a m p * m N $m 2a 5r h ' C - u 3 y d d 2 L o Lon c! ? ? ? m a w 2 % 2 I v n n 00 T ? Y ? Y ? n m i O W b m m n n b o ' C Q ~ 8 m 2 :?%: ; 2; m p L o m r m n n 5 N 2 'm2? un P (9 p m a ? ? 1?'4 m N n n " 2 $ 2 ? ?-*La ? ? ? ? Y o -OL on -N n Loa 2 5 2 6 % " Y '-10m m p 4 ? V l Y N N ? ' C m m a r n am n 10 bc, 0N b m n o f O m @ M N r n (0 n W N 00 OIO m o nN m m n r ?'II f P " N W ? L o N N m N Y N Y ? N m n O $ N O W N L o W Lo N f W m N 'C W W m N N Z ' C P Lo < < $ 4 z z z i -- rmn ?r '4a r - n N N N r g z 8 m c 9 w b o b o o m b 9 I- N m r; fu 42 x c .e Y v1 .e 0o r r i i v ; d~ Wvo d) O ~W o - ? m m o w * o w m B m w 3 O ~ W - W N - 3 9 b 8 vi 0 VI N h h b N h O N - 0 8 0 u1Onoo-hvl N 8 N N 0 3 0 N 0 0 h i? m z 6i? vi m O O 1 O d N O O r n i? Bi? a 9 : : m , N N: : ., . . . . . . # a 0 m i:8- .. . . . . . . . . . . ,, .. .. .. p ! : : : ; : 8N. 0m 1 : : : 0 m : : : : : " : : r. 0 c ci i ". ' . . N 0 : : : : : 1 : : ; R ci I' 0 : : : : : 0 : : : m t- VI. N 0 : : : : : y : : : I' '9 * VI N 0m m ? a ? ? V I v l O W ~ v l v l m : q q o q : W ? ? q q : : : q q : : : 2 " : : q y y : : 9 9 9 .. . . . . . , . . . , . . I . I . . . N - 0 0 0 N N O 110 0 0 0 3 - 0 s p: y q o v : vms mQI0 : : : o q : : : : : : y q 1 ., . . . . . . . . . . . . . . . . I . - 3 0 0 - N $ 6 m o 0 0 0 - - ? VI 2 P S Y T Y Y y q y 1 0 q q q - 0 2 2 8 a : : : : : q : : : 8 r n 0 0 0 N 2 2 2 ~ t c b - o r n o o o oz q- '0?oy-oo' " o - 0 0 N d 0 0 N rn 8N N 0 N c0 *I c c Pa N 0 0 R P h 0 y poqqqq zz; q q q - o q q - y o q q q y o m y 5 : : : : :q : : : c ta 0 0 N 3 * o o o r n CQVIVI-omoooo * N 0 0 - 0 0 oC w - , 1 - - 0 0 G ba d 2 a d 2 k 2 d W 2 d VI 2 .-bB8 v) a QQ b N 0 8 y c": : : c " m: : : :' $ 2 2 ,. . . .. . . . . . . . . . . . . . . . . . . 0 xz3 2 2 : : ? Q,W 3 O N e m * * z o 2 2 ; 8 :: B N m 0 2 ' 9 m w q q y ON. 5 o o " o q " . q ~ o1 Q , q ? ? y \ q a : : : e - 0 - V I O N vic m e o ~ o o o o b - - e w N 3 y : : : : : N 0 N x * N N : N e 0 - .I I 0 0 0 N E41e 8 N QI 0, 2 m 00 2 k2 ;fl 2 m VI 2 w Y ., . . . . . . h,? : : : h a :: : .. .. .. .. 0 3 N * s o $ 3 h h v! 3 Y q l ? ? ? y : : : w o l b o l m c r o ~ ~ m m ro m. N N N. - f o m m m m o - ? ' 4 N Y ? " ? m m m m c . r o a 3 - m - N * - ' 9 : : : N 5 N * , : : : vi ? D ? ' 4 ? ? q : : : m m o l m v r**- r o - o b0 o - - f 3 3 q : : : 3 f I-olVIN ? N ? ? w m m m- is m m m 4 X - o o o - - ~ m - o o o o o - o o o o o o o o - ~ ~ " w 0 - - ~ 0 0 m N w " - ~ N " ~ - - N 0 " " - N 1 Y S " " - " " 9 " N " 9 - 4 " 0 w - " p ! 0 p ! N - m o o o o - ~ m o o o o o o - o - o o o o o - - ~ N "lNn" ~o "o "o" 9 - 9 - m 9 q - " N n q - - " 0 N " m o o w m - 0 0 0 0 0 - 0 - 0 0 0 0 0 V I - 0 d o 0 o o o v ; m0-0o o o o o N o - o o o o o0t - - -r v VI N N N N - m. t-, q- p! q 0 f - cy '? w , q N * d o 0 o 0m0Nm r i 0 o 0 o o oNN o o o o-,o0, o0 I N o m VI N t- 0 0 0 N w * - " w, w oN, Vp!m 0, N s'om N o P v1 w - - -8 o m m m q m, m \o m w q - - - e, e, -'m' 0' m' d' w vi vi vi-'r;N' 0' ICI 0 rj -' 0' 0' 0 0 d 0 e m - 4 .e Y .e Y v) m s