Madagascar - Education and training in Madagascar : towards a policy agenda for economic growth and poverty reduction (Vol. 2 of 2) : Main report

Report No. 22389-MAG Madagascar Education and Training in Madagascar Towards a Policy Agenda for Economic Growth and Poverty Reduction Volume 2: Main Report June 22, 2001 Human Development IV Africa Region Document of the Wotid Bank  CURRENCY EQUIVALENTS (Exchange Rate Effective June 21. 2001) Currency Unit = Malagasy Francs (MGF) MGF 1.00 = US$0.000150 US$1 = MGF 6,657 Fiscal Year: January 15'- December 31st Abbreviations and Acronyms AGENATE - Agence Nationale d'Evaluation ARIF - Association Regionale Inter-professionnelle pour le developpement de la Formation professionnelle BEPC - Brevet d'Enseignement du Premier Cycle CEPE - Certificat d'Etudes Primaires El6mentaires CERES - Centre de Ressources des Personnels des Etablissements d'Enseignement Technique et Professionnel CFP - Centre de Formation Professionnelle CiSco - Circonscription Scolaire (school district) CNFTP - Conseil National de la Formation Technique et Professionelle CNTEMAD - Centre National de Tele Enseignement de Madagascar CONFEMEN - Conference des Ministres de l'Education des Pays ayant le francais en partage DEA - Dipl6me d'Etudes Approfondies DETP - Direction de l'Enseignement Technique et Professionnel DEUG - Dipl6me d'Enseignement Universitaire G6neral DGFTP - Del6gation Generale du Gouvernement a l'Enseignement Technique et a la Formation Professionnelle DirESEB - Direction de l'Enseignement Secondaire et de l'Education de Base ENI - Ecole Nationale d'Informatique ENS - Ecole Normale Superieure ENSET - Ecole Normale Superieure pour l'Enseignement Technique ESP - Ecole Superieure Polytechnique ESSA - Ecole Superieure des Sciences Agronomiques FIFP - Fonds d'Intervention a la Formation Professionnelle FRAM - Fikambanan'ny Ray-Amandrenin'ny Mpianatra (school-based parents association). GIREFTP - Groupements Inter-Regionaux d'Etablissements de Formation Technique et Professionnelle HIPC - Highly Indebted Poor Countries IHSM - Institut Halieutique des Sciences Marines IOSTM -Institut d'Odontologie-Stomatologie Tropicale de Madagascar IST -Institut Superieur de Technologie LTP -Lycee Technique Professionnel MBDPA -Ministere charge du Budget et du Developpement des Provinces Autonomes METFP -Ministere de l'Enseignement Technique et Professionnel MinESEB -Ministere de l'Enseignement Secondaire et de l'Education de Base MinESup - Ministere de I'Enseignement Superieur ONCE -Observatoire National des Competences pour l'Emploi ONETFOPP - Office National de l'Enseignement Technique et de la Formation Professionnelle Prive PASEC - Programme d'Analyse des Systemes Educatifs de la CONFEMEN PAT - Personnel Administrif et Technique PDES - Plan Directeur de l'Enseignement Supdrieur PDI - Plan de Developpement Institutionnel PNAE - Programme National pour I'Amelioration de l'Education PRAGAP - Programme de Renforcement et d'Amelioration de la Gestion Administrative et Pedagogique TEFISO - Teknica Fiofanana Socialista Vice President : Callisto Madavo Country Director: Hafez Ghanem Sector Manager: Arvil Van Adams Task Team Leader: Jee-Peng Tan  Madagascar Education and Training in Madagascar: Towards a Policy Agenda for Economic Growth and Poverty Reduction Volume 2: Main Report This volume contains the analysis upon which the summary of policy challenges reported in volume 1 is based.   Table of Contents Acknowledgements ................................................ viii Executive Summary ............................................... ix Chapter 1: Demographic and Macroeconomic Context ................................................1 The demographic context ............................................................ I Economic growth and current conditions ............................................................3 Overall government finances ............................................................4 Overall public spending on education .............................................................6 Prospects for and implications of increased spending on education .............................................................8 Conclusion ............................................................ 11 Chapter 2: Coverage and Structure of the Education System ............................................... 12 Aggregate enrollment levels and trends ............................................................ 12 Public and private sector roles in education ............................................................ 13 Levels and trends in gross enrollment ratios ............................................................ 16 Grade-to-grade student flow profiles in primary and secondary education ............................................................ 18 Indices of overall student flow efficiency ............................................................ 25 Policy perspectives on grade repetition and dropping out in primary education ............................................. ....... 28 Co(nclusion ............................................................ 32 Chapter 3: Education Finance ................................................ 34 National spending on education ............................................................ 34 A close-up view of public spending on education in 1998 ............................................................ 40 Levels and patterns of public spending per student ............................................................ 50 Coonclusion ............................................................ 55 Chapter 4: Some Aspects of Equity in Education .............. ................................. 57 Overview of disparities in enrollments ............................................................ 57 Disparities in patterns of student flow ............................................................ 59 The incidence of public spending on education ............................................................ 63 Assessing the impact of alternative interventions to improve equity ............................................................ 66 Conclusion ............................................................ 68 Chapter 5: Primary and Secondary Education ................................................ 69 The demand for schooling ............................................................ 69 Overview of the supply of services ............................................................ 71 Teacher allocation across schools, redeployment and recruitment ............................................................ 77 Economies of scale in service delivery ............................................................ 89 Learning outcomes in primary and secondary education ............................................................ 96 Decentralization and local management of service delivery ............................................................ 106 Conclusion ............................................................ 110 Chapter 6: Vocational and Technical Education ............................................... 112 Historical and general context ............................................................ 112 The demand for technical and vocational education ............................................................. 113 The supply of technical and vocational education ............................................................ 117 The production of services in the public sector ............................................................ 120 The internal efficiency of vocational and technical education ............................................................ 127 The transition of technical education graduates to higher education .................... ........................................ 129 Conclusion ............................................................ 131 Chapter 7: Higher Education ............................................... 132 Historical context ............................................................ 132 Structure of the higher education system ............................................................ 134 Higher education staff ............................................................ 142 The internal efficiency of higher education ............................................................ 148 Unit costs in higher education ............................................................ 151 Conclusion ............................................................ 156  iii Chapter 8: Education and the Labor Market ..................................................................... 157 Structure of employment, income and returns to education ................................................................................. 157 Output of graduates and their absorption into the workforce ................................................................................ 161 Skills development in the workforce .................................................................................. 170 Conclusion .................................................................................. 176 Technical Note on Regression Models on Student Achievement .......................................... 178 Specifying the basic achievement model .................................................................................. 178 Correcting for selection bias .................................................................................. 178 Evaluating the impact of selection bias .................................................................................. 179 Evaluating the impact of differences in resource endowment and managerial efficiency ..................................... 179 References ..................................................................... 206 Boxes Box 1.1: Madagascar at a glance .................................................................................. 2 Box 3.1: Comparative perspectives on teacher salaries in Madagascar .................... .................. ........... 46 Box 5.1: What has been learned from effective reform strategies in education? .......................................... 109 Tables Table 1.1: Population size and incidence of poverty, Madagascar 1962-1997 ..............................................................2 Table 1.2: Madagascar's economic performance in comparative perspective, 1970-1999 ...........................................4 Table 1.3: Overall government revenue, Madagascar 1990-99 ...................................................................................5 Table 1.4: Overall government expenditure, Madagascar 1990-99 ...............................................................................6 Table 1.5: Overall public spending on education, Madagascar 1990-2000 ...................................................................7 Table 2.1: Trends in enrollments by level and type of education, Madagascar 1962-1999 ................... ................ 13 Table 2.2: Share of enrollments in the private sector, Madagascar 1986-1998 ........................................................... 14 Table 2.3: Gross enrollment ratios by level of education, Madagascar 1970-98 (%) .................................................. 17 Table 2.4: Madagascar's gross enrollment ratios in comparative perspective, circa 1996 .......................................... 18 Table 2.5: Cohort survival rates, Madagascar 1992 and 1998 .................................................................................. 20 Table 2.6: Grade-specific enrollment rates, Madagascar 1992 and 1998 .................................................................... 22 Table 2.7: Inter-cycle selection, Madagascar and selected world Regions, 1990s ...................................................... 23 Table 2.8: Repetition rates by grade and sector, Madagascar 1992 and 1998 ............................................................. 24 Table 2. 9: Efficiency of student flow in primary education, Madagascar 1998 ......................................................... 26 Table 2.10: Indices of student flow efficiency by cycle of schooling and sector Madagascar 1998 ........................... 27 Table 2.11: Correlation between grade repetition in primary education and survival rates and student achievement 29 Table 2.12: Access to complete cycles of instruction at the primary level, Madagascar 1998 .................................... 31 Table 3.1: Levels and trends in public spending on education, Madagascar 1990-2000 ............................................. 35 Table 3.2: Distribution of public spending on education by level, Madagascar 1990-2000 ....................................... 36 Table 3.3: Household spending on education, Madagascar 1997 ....... ...................................................................... 37 Table 3.4: Government and household spending on public and private primary education, Madagascar 1999 .......... 39 Table 3.5: Government employees in the education sector, 1998 ............................................................................... 41 Table 3.6: Teacher salaries and the distribution of teachers by salary grade, Madagascar 1998 ................................ 43 Table 3.7: Distribution of school-level non-teaching staff by salary grade, Madagascar 1998 ................................... 44 Table 3.8: Current public spending by function and level of education, Madagascar 1998 ....................................... 45 Table 3.9: Distribution of current public spending on education by business categories, Madagascar 1998 .............. 48 Table 3.10: Public spending per student by level of education in the public sector, Madagascar 1998 ...................... 51 Table 3.11: International comparisons of current public spending per student ........................................................... 52 Table 3.12: Decomposition of public spending per student in public schools, Madagascar 1998 .............................. 53 Table 3.13: Teacher salaries and pupil-teacher ratios in Madagascar and othere countries ........................................ 54 Table 4.1: Gross enrollment ratios by province, urban-rural locality and gender, Madagascar 1990-99 .................... 58 Table 4.2: Distribution of school-age population and enrollments by locality and gender, Madagascar circa 1999 ..59 Table 4.3: Disparities in entry rates to grade I, Madagascar 1997 .60  Iv Table 4.4: Disparities in rates of survival to grade 5, Madagascar 1997 ..................................................................... 61 Table 4.5: Disparities in grade 5 enrollment rates, Madagascar 1997 ......................................................................... 62 Table 4.6: Incidence of public spending on education among members of a cohort, Madagascar 1999 ..................... 64 Table 4.7: Comparative perspectives on the incidence of public spending on education ............................................ 65 Table 4.8: Share of public spending on education benefiting the poorest and richest population quintiles ................ 66 Table 4.9: Simulations of the equity impact of alternative structures of unit spending on education and student flow ................................................................................................................................................................ 67 Table 5.1: Number and distribution of primary and secondary schools, Madagascar 1998 ........................................ 71 Table 5.2: Provincial distribution of primary and secondary schools, Madagascar 1998 ........................................... 72 Table 5.3: Characteristics of public and private schools, Madagascar 1998 ............................................................... 73 Table 5.4: Characteristics of public and private primary schools by province, 1997-98 ............................................. 74 Table 5.5: Public subsidies to private schools, Madagascar 1993-99 .......................................................................... 75 Table 5.6: Subsidized private schools and teachers, Madagascar 1993-99 ................................................................. 76 Table 5.7: Pupil-teacher ratios in primary and secondary schools by sector and locality, Madagascar 1998 ............. 78 Table 5.8: Provincial regression estimates of the relation between number of teachers and pupils ............................ 80 Table 5.9: Distribution of public primary schools according to the adequacy of teacher complements, Madagascar 1998 ................................................................................ 81 Table 5.10: Regression estimates of the relation between number of teachers and students in public secondary schools, 1998 a ................................................................................ 83 Table 5.11: Distribution of public secondary schools according to staffing patterns .................................................. 84 Table 5.12: Number of staff redeployed and recruited 1997-2000 .............................................................................. 85 Table 5.13: Teacher surpluses, deficits, redeployment and recruitment in public primary schools, Madagascar circa late 1990s ................................................................................ 86 Table 5.14: Predicted probability of receiving at least one of the teachers recruited during 1997-2000 according to the adequacy of staffing in 1997-98 ............................................................................... 87 Table 5.15: Teacher surpluses, deficits, and redeployment in public lower secondary schools, Madagascar circa late 1990s ................................................................................ 88 Table 5.16: Size distribution of primary schools and teaching arrangements in public and private sectors ................ 91 Table 5.17: Size distribution of secondary schools, 1997-98 ................................................................................. 94 Table 5.18: Governm ent grants for public secondary education, 1995 and 1999 ....................................................... 95 Table 5.19: Teaching loads and subject-specialization among secondary school teachers ......................................... 96 Table 5.20: Trends in pass rates on national end-of cycle examinations, Madagascar 1989-99 ................................. 98 Table 5.21: Disparities in French and Mathematics test scores, Madagascar, 1997-99 .............................................. 99 Table 5.22: Correlates of second graders' combined scores in Mathematics and French . ........................................ 101 Table 5.23: Private school advantage over public schools in test scores, Madagascar 1997-8 l .............................. 103 Table 6.1: Enrollments in technical and vocational education and secondary general education a ........................... 114 Table 6.2: Access to general secondary education and technical and vocational education by household quintiles. 1 1 6 Table 6.3: Gender differences by field of study in the public sector ......................................................................... 117 Table 6.4: Characteristics of service delivery in technical and vocational education ................................................1 18 Table 6.5: Number of institutions and pupils enrolled in the public sector ............................................................... 119 Table 6.6: Diversity and specialization in course offerings in public sector institutions, 1998-99 .......................... 120 Table 6.7: Trends in enrollments in public technical and vocational education ........................................................ 121 Table 6.8: Access to public technical and vocational education by province ............................................................ 122 Table 6.9: Student-teacher ratios in vocational and technical education ................................................................... 122 Table 6.10: Regression estimates of the relation between number of teachers and size enrollments and diversity of course offerings a ........................................................................... 123 Table 6.11: Ratio of students to non teaching staff in public technical and vocational education, 1998-99 ............. 124 Table 6.12: Current unit cost of teaching activities in CFP and LTP, 1998 a .................................... ....................... 124 Table 6.13: Unit costs of vocational and technical education as a multiple of the unit costs of secondary general education a ........................................... 125 Table 6.14: Current unit cost by field of study a. ................................1....................2..................6........... . 126 Table 6.15: Regression estimates of the relation between total cost and size of enrollments a. .... ............................ 127 Table 6.16: Output of graduates from public technical and vocational education ..................................................... 128 Table 6.17: Internal efficiency of technical and vocational education ................................................................ 129 Table 6.18: Entry rate of LTP graduates into higher education (public sector only) ..... . .......................................... 130  v Table 6.19: Distribution of new entrants into higher education by type of programs ............................................... 131 Table 7.1: Entry rate to higher education and trends in enrollments by sector ................................. ......................... 135 Table 7.2: Trends in enrollments in the university faculty sector a....................................................................... 137 Table 7.3: Characteristics of course offerings in the university faculty sector .......................................................... 137 Table 7.4: Number of students in the &oles sup&rieures and instituts attached to the universities a ........................ 138 Table 7.5: Characteristics of the course offerings in &oles superieures and university instituts .............................. 139 Table 7.6: Enrollments in the ISTs ....................................................................... 140 Table 7.7: Characteristics of course offerings in the ISTs ...................................................................... 140 Table 7.8: Enrolments in distance education ...................................................................... 141 Table 7.9: Characteristics of private higher education ...................................................................... 141 Table 7.10: Number of permanent teaching staff in public higher education a......................................................... 142 Table 7.12: Distribution of permanent teaching staff by grade in the universities .................................................... 144 Table 7.13: Rate of remuneration for extra teaching hours ......................................................... 145 Table 7.14: Number of extra hours of teaching across the public universities a ........................................................ 145 Table 7.15: Ratio between students and faculty across the public universities taking into account the cost of overtime teaching .............................................. 146 Table 7.16: Number of administrative and technical staff in the public and private sectors a ,. 147 Table 7.17: Trends in the number of administrative staff in the universities ............................................................ 148 Table 7.18: Number of graduates from public higher education . .................................................................. 149 Table 7.19: Number of students and repeaters in the public and private sectors ....................................................... 150 Table 7.20: The schooling career of a cohort of entrants to public higher education ....................... ......................... 151 Table 7.21: Unit costs in public higher education .................................................................. 152 Table 7.22: Unit costs in distance education and traditional higher education ....... . ................................................. 153 Table 7.23: Unit costs in private institutions d,................................ ,..,.,....,.............................,.. ,,,..,,,............... 153 Table 7.24: Spending on student aid and the number of students receiving grants ................................................... 154 Table 7.25: Student grants in public higher education in relative terms .......................................................... 155 Table 7.26: Share of grant-holders students by field .......................................................... 156 Table 8.1: Trends in the economically-active population a,.....,,,,,,.,,,,. ,..,.,,,,.,,,,,,.,,,,.,,,,.,,.,. 158 a Table 8.3: Distribution of employment by type a ........................................ 160 Table 8.4: Salary gaps between wage eamers in the formal and informal sectors .................................................... 160 Table 8.5: Rates of return to education .......................................................................... 161 Table 8.6: Output of graduates and their absorption on the labor mnarket, Madagascar circa 1999 ........................... 163 Table 8.7: Percentage distribution of wage eamers by educational attainment relative to workers in the same jobs, Antananarivo 1996 a. .......................................................................... 166 Table 8.8: Unemployment and job search duration by level of education, Madagascar 1993 & 1999 ..................... 167 Table 8.9: Educational attainment of the workforce .......................................................................... 171 Table 8.10: Main business constraints encountered by Malagasy firms, 1998a ......................................................... 172 Table 8.11: Formal sector firms that organize in-service training, Madagascar and other countries a. . . 173 Table 8.12: Predicted probability of participating in in-service training a 17.4.,.................................................,,,,,,, , 174 Table 8.13: The supply of in-service training a,,..,,,,.,...,.,,,,,,,,,.,,,,,,,,,,,,..,,,,,,.,,.,,,. 175 Tablee8.14:Hourly costofytrainingbyrtype a,.,,,,,g.,,,....................................,...,,,,,,.,,,,,,.,,,,,,..,,.,.,,.,,,,,.,,,.,..,,,.,.................,,....,.,.,,117 Table .4:Horl 8.14:triin b Hourly........cost.......of....training...........by....type.......a... 176. 7 Table 8.15: Hourly costs of training for skills upgrading by econornic sector of training offered ............................ 176 Figures Figure 1.1: Trends in extemal debt service and public spending on education, Madagascar 1990-99 ..........................8 Figure 1. 2: Relation between external debt service and public spending on education, Madagascar 1990-99 ............8 Figure 1. 3: Cross-country relation between public spending on education and years of school life expectancy, circa 1993 .............................................................................. 10 Figure 2.1: Private schools enroll large shares of primary and secondary students in Madagascar, circa 1996 .......... 16 Figure 2.2:Student flow in primary schooling is relatively inefficient in Madagascar, circa 1993 ............................. 27  vi Figure 2.3: Impact of availability of comnplete cycle of instruction and reduced repetition rate in primary education, Madagascar 1998 .................................................................................... 32 Figure 3.1: Staff distribution by function at the various levels of education, Madagascar 1998-9 ............................. 42 Figure 3.2: Distribution of current public spending on education by function, Madagascar 1998 .............................. 49 Figure 4.1: Grade attainment profile of cohort ages 15-19, Madagascar 1997 ........................................................... 63 Figure 5.1: Relation between number of publicly-paid teachers and number of pupils .............................................. 79 Figure 5.2: Relation between number of teachers and pupils across public secondary schools, 1998 ........................ 82 Figure 5.3: Relation between spending per pupil and school size across public primary schools, 1997-98 .............. 90 Figure 5.4: Relation between spending per student and school size across public lower secondary schools, 1997-98 ............................................................................................................................................................... 93 Figure 5.5: Relation between spending per student and school size across upper secondary public schools, 1997-98 ................................................................................................................................................................. .93 Figure 5.6: Fourth Grader's Test Scores, Madagascar and other countries 1999 ........................................................ 97 Figure 5.7: Relation between per pupil spending and test scores across public schools, Madagascar, 1997-8 ......... 105 Figure 5.8: Relation between exarnination pass rates and spending per student, Madagascar 1998 ......................... 106 Figure 6. 1: Share of technical and vocational education in selected developing countries (circa 1990) ......... .......... 115 Figure 6.2: Relation between the number of teachers and the number of students ................................................... 123 Figure 6.3: Relation between the total current cost and the number of pupils enrolled ............................................ 126 Figure 8.1: Relation between per capita GNP and share of labor force in non-agriculture, 1995 ........... .................. 159 Figure 8.2: Relation between per capita GNP and coverage in higher education, circa 1995 ................................... 169 Appendix Tables Table Al.1: Overall governnent revenue and expenditure, Madagascar 1990-1999 ................................................ 181 Table A 1.2: Trends in total public spending on education, Madagascar 1990-2000 ................................................. 181 Table A3.1: Public spending by level of education, Madagascar 1990-2000 ..................................... ...................... 182 Table A3.2: Personnel in public primary and secondary schools, Madagascar 1990-2000 ...................................... 182 Table A3.3: Distribution of government employees attributed to various education sector budget votes, 1991-2000 ............................................................................................................................................................... .183 Table A3.4: Distribution of government employees under various education sector votes, 1991-2000 ................... 183 Table A3.5: GDP per capita and per worker, and take home pay of education sector government employees . ....... 184 Table A3.6: Salaries of teachers and other wage earners of similar qualification, Madagascar 1997 ....................... 184 Table A3.7: Distribution of staff by function and level of education, Madagascar, 1998-9 ...................................... 185 Table A5. 1: Provincial distribution of public and private sector primary and secondary schools, 1997-98 ............. 186 Table A5.2: Percentage of primary schools offering full- and part-day instruction, Madagascar 1997-98 ............... 186 Table A5.3: Regression estimates of the relation between the number of teachers and the number of pupils with provincial dummy variables across primary and secondary schools ..................................................... 187 Table A5.4: Regression estimates of the relation between teacher allocation and number of lower secondary students across public schools .................................................................................. 187 Table A5.5: Distribution of public primary schools according to adequacy of teaching complement, Madagascar 1998 .................................................................................. 188 Table A5.6: Distribution of public secondary schools according to adequacy of teaching complement, Madagascar 1998 .................................................................................. 189 Table A5.7: Correlates of the allocation of teacher recruited for public primary schools during 1997-2000 ........... 190 Table A5.8: Cost functions by type of public primary and secondary schools ......................................................... 190 Table A5.9: Fourth grader's scores on the Monitoring Learning Achievement (MLA) tests, 1999 ......................... 191 Table A5. 10: Second and fifth graders' year-end test scores in Mathematics and French in the PASEC study . ...... 191 Table A5.11: Sample characteristics of pupils and schools in the 1997-98 PASEC survey for Madagascar ............ 192 Table A5.12: Probit estimates of a child's probability of attending a private school, Madagascar 1997-8 .............. 193 Table A5.13: Correlates of fifth graders' combined scores in Mathematics and French, Madagascar 1997-8 ......... 194  vii Table AS.14: Correlates of second and fifth graders' combined scores in Mathematics and French in public schools, Madagascar, 1997-8 ................................................................................. 195 Table A5.15: Logit regression estimates of the relation between pass rates on national examinations and public spending per student, Madagascar 1997098 ................................................................................. 196 Table A6. 1: Organization of vocational and technical education ............................................................................. 197 Table A6.2: Number of institutions and number of students by field of study in the public sector, 1998-99 .......... 198 Table A8. 1: Employment in the formal industrial sector ................................................................................. 202 Table A8.2: Earning functions ................................................................................. 203 Table A8.3: Probability of participation in skills-upgrading ................................................................................. 204 Table A8.4: Earnings function including participation in skills-upgrading ......................... .................................... 205 Map: IBRD 20035R1  Acknowledgements This study is the product of close collaboration between the World Bank and the Government of Madagascar. The Bank team was led by Jee-Peng Tan (AFTHD), with Gerard Lassibille (Institut de Recherche sur I 'Economie de I 'Education, France) as the main team member; Lucia Navarro (Universidad de Malaga, Spain), Julia Lane (American University, USA) and Dandan Chen (YPP) as ad- hoc members providing specific inputs; Alain Mingat (AFTH2) as overall advisor; and Zafiris Tzannatos (MNSHD), Keith Hinchcliffe (SASED), and Jaap Bregman (AFTH4) as peer reviewers. The study benefited from overall management support from Arvil Van Adams, Hafez Ghanem, and Michael Sarris. The Malagasy technical working group consisted of senior policy and technical staff from the three education ministries, as well as representatives from other parts of government, as follows: Albert Rafalimanana, Dominique Fofa, and Joseph Sylvain, with overall management support from Andre Juckely Boto, and help with data compilation from Rabiary Andriaherisaina, Roger Andrianaivoson, Georges Gniarck, Jean-Claude Landry, Mariamo, Roland Rabeson, Honore Rajerison, Jose Rakotonaivojaoniriana, Jean Pierre Rakotovao, Jacqueline Ralisiarisoa, and Jules Razafimandimby (MinESEB); Daniel Robijaona, Emmanuel Rakotovao, and Ernest Randriamanampisoa, with overall management support from Hyacinthe Ralantonirina (METFP); Ying Vah Zafilahy, Chantal Rahalison, and Adolphe Randriantsoa, with overall management support from Lucien Rabenantoandro Rakotomanantsoa and help with data from Elisabeth Dama, Lala Rambeloson, and Daniel Razafimandimby (MinESup); Fleurette Nirinaharindrasana (MBDPA); Heri-zo Ralambomahay (PA1GEP/CTRSP/STA); and Harivelo Rajemison and Eric Ramilison (INSTAT/MADIO). Members of the technical working group wrote background papers (see Govt. of Madagascar 2000a, 2000b, 2000c, Nirinaharindrasana 2000; Rajemison 2000; Ralambomahay 2000; Ramilision 2000), as well as provided critical help in assembling the databases used for the anlysis presented in the report. The work has benefited from inputs of various kinds from others as well, including N. Andrianaivo, C. Andrianalimanana, F. Chenakarnkul, R. Kalidindi, C. Kostopoulos, E. Murray, A. Onivololona, S. Paternostro, T. Rabenandrasana, E. Rabeson, J. Rabetokotany, V. Rahelimanantsoa, M. Ranaivoarivelo, A. Ranarisoa, D. Randriamanampisoa, D. Ramiandrisoa, R. Raoilison, J. Razafindravonona, S. Sosale, A. Velompanahy, D. Viens and J. Zalduendo. Interactions with the following representatives of the donor community and NGOs have also been helpful: 0. Akpaka (UNICEF); P. G. Le Boul and G. Thuaudet (Cooperation Fran,aise); B. Coquelin (UNFPA); V. Hubert (EU); P. Lambiliotte (Aide et Action); Y. Polard (Association Franccaise des Volontaires du Progres); J. Randrianandrasana (UNESCO); and L. Ramanarivo (UNDP). The report has been financed by the World Bank with contributions from the governments of Madagascar, France and Norway.   EDUCATION IN MADAGASCAR: TOWARDS A POLICY AGENDA FOR ECONOMIC GROWTH AND POVERTY REDUCTION Executive Summary 1. Madagascar is a poor country with abundant and unique natural resources that can be mobilized to raise the standard of living of its population, nearly three-quarters of whom have subsisted below the poverty line for at least two decades now. That prospect has become more promising in the light of several recent developments. Most importantly the government has renewed its commitment to poverty reduction by articulating a publicly-discussed strategy with explicit goals for macroeconomic management and sectoral development in key areas. Signs of economic recovery are emerging, with GDP growth averaging 4.1 a year during 1997-99 (compared with only 1.9 percent a year during 1995-96) and projected to accelerate to 8 percent or more in the coming years. At the same time, the country's burden of external debt is being lightened under the World Bank-IMF Highly Indebted Poor Countries (HIPC) Initiative, with projected savings in debt service payment on the order of $50 million annually for at least the next 10 years. 2. Government spending on education has been rising over the past five years, and the increase can be expected to continue, in part because a large share of the HIPC savings are earmarked for the sector. Expectations are high for education to make a significant contribution to the government's poverty reduction agenda. For policy makers and managers in the sector the central issue therefore revolves around good stewardship of the available resources to advance that agenda. This report seeks to inform the discussion by documenting the current status of educational development in Madagascar and the key constraints on its performance. Scope of the report and its audience 3. The report deals with all the main education sectors (i.e. primary, secondary, vocational/technical and higher education), but the focus is necessarily limited by the complexity of the subject. In particular, non-formal education has been set aside for elaboration by others, even though it is an important element in any poverty reduction strategy. Similarly, neither early childhood nor such traditionally educational issues as issues as curriculum development, teacher education and pedagogical methods, are addressed in the report. Rather, the emphasis is on key aspects of education cost and finance and the link between spending options and educational outcomes. 4. The reports' primary audience are policy-makers in education in Madagascar, as well as those in other government ministries. However, given the subject matter and the context, it should also interest members of the multi- and bilateral donor community and non- governmental organizations active in education, as well as teachers, students, parents and others in the general public with a direct or indirect concern about the broader policy context of their involvement in education. 5. A key objective of the report is to contribute toward developing a widely-shared vision for educational development for Madagascar based on a common understanding of the challenges involved, and the part that each participant can play in realizing that vision. Open  x discussion of the report is an important part of the process for creating the common understanding. Such a process is of course only the first step toward policy development, not least because of the report's diagnostic focus and as well as its selective scope. As such, the report offers no policy prescription but leaves it development as a follow-up activity that should ideally involve the various actors in defining specific policy interventions to address key challenges in the sector as well as in defining the best way to implement them in order to achieve concrete progress over time. The economic development context 6. Madagascar is a predominantly agricultural country, with about 75 percent of the labor force in farming and other rural pursuits and 85 percent of the workers in informal sector jobs. Poverty is deep-seated and will require accelerated economic growth as a necessary, but obviously not sufficient, condition to overcome. Given the current structure of the economy, such growth in the coming decades will depend on two parallel developments: rising productivity in agriculture and the informal economy (which will continue to employ the bulk of the labor force), and expansion of the modem, and typically more productive, sector. Investments in education can help facilitate both developments, but the nature of the interventions needed are likely to differ by level of education. The worldwide historical experience suggests that no country has managed to develop its economy without putting in place a system of mass basic education to produce a literate and numerate population that can facilitate overall economic modernization. That experience also suggests that at post-basic levels of education, the production of educated labor needs to be linked closely to the pace of job creation in the modem sector. Thus, while a clear goal can be set to ensure that all children benefit from at least five years of primary schooling, the development of other levels in the system, especially higher education, must be calibrated, in qualitative as well as qualitative terms, to the labor market's absorptive capacity for educated labor. 7. Accepting this broad understanding of the direction for educational development provides a benchmark for assessing the various layers of management decisions needed to optimize the education system's performance as a whole. Initial conditions are important in this regard, for they delineate the system's current strengths and weaknesses, as well as the tradeoffs and policy challenges that must be managed to produce the best results within the inevitable budget constraints. Cross-country experience also brings a useful perspective by illustrating the scope of potentially viable choices, as well as the distance that Madagascar must travel to match other countries' achievements in education. In an increasingly globalized world where the basic skills and technical know-how and competencies of the whole population mark the difference between competitive and lagging economies, the education sector cannot afford to fall too far behind. Strengths of the education system 8. Madagascar starts out with some important assets in education. The country was one of the earliest to recognize the importance of mass basic education, and over the years the government has extended the network public primary schools to almost all of the country's 11,000 "fokotanys". The social demand for education appears to be generally strong, as suggested by the high entry rates to first grade (averaging 75 percent or more in all provinces except Toliara where the rate is still below 60 percent); the large private sector presence at all  xi levels of education; as well as the substantial private spending on education throughout the system (which make up an estimated 50 percent of the country's national investment in education in 1997). Unlike the situation in many low-income countries, girls in Madagascar are as likely to enroll as boys and the gap between them widens (although not dramatically so) only in certain fields in technical education and in higher education. Student learning in primary education compares favorably with outcomes in other low-income countries, with Malagasy children scoring above the sample averages in language, mathematics and life skills. 9. In terms of policy development some major achievements are already in place. In particular, Madagascar has during the past decade successfully managed the difficult and often politically-tense tradeoffs in resource allocation between higher education and the lower levels that many low-income countries continue to confront today. The share of spending on higher education now absorbs 13 percent of total public spending on education, instead of the 30 percent or so in the early 1990s. At the same time, there is growing recognition of the need for greater local participation in school management, and institutional mechanisms have already been put in place to formalize decentralization in education. 10. Perhaps the greatest asset of the system is the fact that it is poised for continued growth, given the highly favorable prospects for increased public investment in education. In such a context, the political economy of policy reform will be easier to manage in that some of the transitional costs of reforming the system's operational arrangements can be eased with strategic allocations of the incremental spending. Importantly, overall levels of teacher salaries in primary and secondary education do not appear out of line with market conditions and therefore do not present a major problem beyond the need for appropriate incentives to motivate improvements in service delivery. All in all, policy makers in Madagascar face an exciting situation in which the choices they make today can truly help propel the system toward an efficient and equitable path of expansion. Challenges for the next decade 11. Managing output from the education system. A major problem of the education system is that although entry rates to first grade are high, only about a third of the entrants successfully complete the full five-year primary cycle. Repetition rates are uniformly high throughout the primary cycle, averaging about 31 percent. These patterns of student flow have remained unchanged for most of the past decade, and they make Madagascar's system one of the least efficient in this regard. An estimated one-third of the country's spending on primary education effectively fails to produce school leavers with a complete primary education (which is generally accepted as the investment needed to achieve permanent literacy and numeracy); or typically takes longer than the nornal five years to produce the graduates. The latest data from household surveys show a decline in the share of those with no schooling in the labor force and a rise in the share of those with secondary schooling, but the improvement is largely an illusion since the same data also suggest that younger workers have far fewer years of schooling than their elders (4.2 years on average among those aged 20-24, for example, compared with 5.2 years among those aged 35-39). The clear deterioration in the educational attainment of the labor force raises questions about its compatibility with the skills needs in a modernizing rural economy. 12. At post-primary levels, the student flow indicators improve, although there remains room for improvement. Nonetheless, the most important issue here concerns the output  xii of graduates and their absorption into the labor market. In the aggregate, the volume of post primary graduates seems consistent with demand on the labor market, but at the top end of the educational ladder there is evidence of a substantial over-production, given that jobs for highly educated labor are currently expanding at perhaps 2,000 to 2,500 positions a year, while the system is producing some 3,700 higher education graduates annually. One consequence of the over-production (which has been sustained over many years) is that graduates find themselves in jobs for which they are over-qualified, and the adverse effect percolates down to school leavers lower down the educational ladder. Had enrollments not been brought under control during the 1990s, the problem would have been much worse. 13. The foregoing picture suggests that a major effort is needed to boost the performance of the primary school system. At the same time, as cohort survival rates improve, selection mechanisms would need to be put in place to ensure that enrollments at the post basic levels expand at a rate commensurate with the capacity of the labor market to absorb them into productive employment. The recent creation of the distance learning system in higher education, and the decision to let in private suppliers to the sector are important mechanisms to manage the excess demand for education at this level. 14. Improving equity in education. Madagascar has one of the most inequitable patterns of public spending on education among low-income countries. The share of the spending that benefits the top quintile is five times that which benefits the bottom quintile. Across members of each birth cohort, the best educated among them obtain an estimated 63 percent of the spending on the entire cohort. Making the pattern of spending more equitable will require a restructuring of the public spending per student in favor of primary education, as well as substantial progress in ensuring that children persist to the end of the primary cycle. Simulations suggest that by reducing public spending per student in higher education (currently 1.57 times the per capita GNP, compared to about 0.83 for countries in low-to middle-income countries in Latin America, Asia and the Middle East) by about half, while also raising the cohort survival rate in primary education from 33 percent to about 50 percent would bring Madagascar's system closer to the more equitable patterns found in countries of these regions. 15. Across provinces and income groups, there are substantial gaps both in the rate of entry to grade I and the survival patterns to the end of the cycle. More needs to be done, particularly in Toliara, as well as Fianarantsoa and Mahajanga, to raise entry rates. With regard to survival rates, the outcomes are poor in all provinces but the need to improve is especially great in Fianarantsoa and Toliara, where barely a fifth of the first graders currently complete their primary schooling. A combination of demand- and supply-side interventions will be needed to address the situation. As in most other settings, poverty and parental income play an important role on the demand side, but these factors are not directly within reach of policy instruments currently available to policy makers. Interventions on the supply side can have a direct impact, as well as work indirectly through demand by enhancing the characteristics of schools, and adapting the supply to make them more attractive (e.g. adapting the school calendar to reduce the opportunity cost of schooling, providing school feeding as a form of income transfer to poor families, and so on). We estimate that by ensuring that children have access to a complete cycle of instruction and by reducing repetition rates to about half its current level would together boost survival rates from 33 percent at present to more 56 percent.  xiii 16. Managing public spending on education. Bringing about the improvements in outcomes would require good stewardship of the public spending on education. At the aggregate level, the current spending on education is modest at 3.1 percent of the GDP, compared with 4 percent in most other low-income countries. A priority in overall public finance is to seek ways to increase the public resources available for education. Beyond the aggregate increase, it is also important to ensure that resources are channeled to primary schooling. During the past decade, as total spending on education was squeezed under the weight of a growing external debt, the government managed to its credit to redirect the falling aggregate spending toward basic education. The main beneficiary was secondary rather than primary education, however, so it will be important in the coming years to ensure that increases in spending do in fact reach all the way down to the primary level. 17. A closer examination of the pattern of aggregate spending suggests that more could be done to improve the distribution of public spending across economic functions. Ideally we would like to see a substantial share of spending going to support the core activities of teaching and learning, which after all is the main business in education. Especially at post- primary levels of schooling, too large a share appears to be allocated to administration and support services at the school level. In lower and upper secondary schools, for example, only half the government's total spending is spent on the core business of classroom teaching, while school administration and support services at the school level absorb around 20 percent of the spending. In vocational and technical education and higher education, the share of teaching services drops down to 34 and 29 percent respectively, while the share of administration rises to 22 and 28 percent respectively. Furthermore, within higher education, the system of remuneration for over-time teaching appears to be highly irregular, so much so that nearly 80 percent of teaching is paid on overtime. There also appears to be scope for rationalizing the current arrangements for student finance. Overall, the results suggest that although we have no benchmarks for identifying an ideal allocation of spending, the current patterns appear insufficiently focussed on the system's core business and strategic goals and could therefore benefit from a closer examination either to justify the pattern or to reform it. 18. Improving teacher allocation across schools. Beyond the system-wide problems in resource allocation, there is also room for improving the allocation of spending across schools. Because teachers represent the bulk of the public spending that schools receive to deliver services, their deployment across schools offers a good picture of the current patterns of allocation of spending in the system. The results suggest that there is substantial inconsistency in the allocation: primary schools with about 100 pupils, for example, could have anywhere up to 7 teachers, while some of them receive no allocation at all and must therefore rely on teachers financed entirely by parental contribution to the parent-teacher association. On average, Antsiranana and Toamasina appear to receive fewer primary school teachers than other provinces, but the differences across provinces pale in comparison to the inconsistency in allocation across schools within each province. In secondary education the pattern of inconsistency persists, although to a lesser extent. At all levels, the allocation favors Antananarivo (and especially the city of Antananarivo) as well as urban areas. 19. To its credit, the government took steps to improve the situation through two rounds of redeployment involving a total of nearly 5,000 primary school teachers and 1,000 secondary school teachers. However, given the large number of schools in Antananarivo with more teachers than would be consistent with system-wide norms based on enrollments, and the  xiv fact that only 104 teachers in the province were re-deployed in the two exercises, it is unclear how much of the inconsistency in staffing across schools has in fact been addressed. Since 1997-98 there has also been two rounds of teacher recruitment. The placement of the new recruits has tended to favor under-staffed schools: the probability that they have received a new teacher is estimated to be 13.5 percent (rising to 35.4 percent among those short by 3 teachers or more), compared with 8.9 percent for a school that is adequately staffed. There will be new rounds of teacher recruitment, and it is important to strengthen the focus on under-staffed schools, particularly those in rural areas. Given the current pattern of staffing, any new teacher allocation to schools in the city of Antananarivo would need careful and convincing justification. There are plans too to enhance the incentives for rural posting. The design of this instrument warrants thorough assessment to ensure that it effectively attracts and retains teachers in rural areas by offering sufficient inducements to compensate for the disadvantages of working in remote areas. 20. Taking advantage of scale economies and managing unit costs in small schools. Madagascar is a sparsely-populated country, so the government's policy of ensuring that schools, at least at the primary level, are physically close to the population they serve inevitably means that there will be a significant number of small schools. Indeed, more than 30 percent of the public primary schools enroll no more than 75 pupils each, far below the minimum economic size of about 250 pupils. In lower and upper secondary education, economies of scale are even larger, but small schools continue to be a prominent feature in the system. In vocational and technical education, and in higher education, the small size of the systems and its fragmentation across multiple locations combine to make it difficult to realize economies of scale in service delivery. In the Centres Formation Professionelle, for example, more than half the unit cost of service delivery is accounted for by the fixed costs, the high share reflecting the impact of the small enrollments in many of the schools. At these levels the problems are further compounded by the duplication of course offerings across institutions, so that the entire system is both costly and inadequately diversified. 21. The results carry two implications for cost management. The first is to examine carefully the spatial distribution of institutions at all levels to ensure that feasible options for school consolidation are not ignored. School consolidation can mean combining nearby schools offering the same level of instruction, or contiguous cycles. In Madagascar, many of the private schools are also small, but they invariably take advantage of economies of scope by combining different levels of schooling under one roof (e.g. primary and secondary, or lower and upper secondary), whereas most public school offer only a single cycle of instruction. Any plan for consolidation will inevitably increase the cost of schooling for some students who may have to travel farther, or even board away from home. Thus, all such plans would need to be combined with arrangements for student finance to ensure that deserving students from poor families are not adversely affected in the process. Doing so will call for a more strategic use of the resources currently allocated to provide student subsidies at all levels in the system. 22. The second set of implications pertain to cost management measures in the small schools. Multi-grade teaching will be important at the primary level, and apparently a large number of schools in Madagascar already practice it. Its use should in theory not affect children's access to full-day instruction, but the way it is implemented in Madagascar leaves large numbers of children in the public schools with less than three hours of instruction a day: for schools enrolling less than 50 pupils, for example, only-third of the children in the public system  xv receive full-day instruction, compared with more than two-thirds among private schools of similar size. At the secondary level, cost management in the small schools would require a closer look into teachers' workloads and subject specialization. At both the lower and upper secondary levels, many more public school teachers specialize in one subject area compared to their counterparts in the private sector. As a result, their teaching loads are significantly lighter: all else being the same, the difference implies an increase in unit cost in the public system by 28 percent in lower secondary education and by 45 percent at the upper secondary level. Clearly, any policy for cost management would need to review this situation and evaluate the implications for teacher training and development and perhaps also teacher incentives. 23. Enhancing student learning. As indicated earlier, Malagasy pupils perform well relative to the children in other low-income countries. Yet in absolute terms, the scores are low, with children correctly answering only half of the test items, and examination pass rates remain low as well. Comparison between the public and private schools suggest that while social selection does account for the superior perforrnance of the private sector, a significant source of the difference can be traced to differences in the effectiveness of school management. Looking within the public system itself, the results suggest that while there is a positive relation between school inputs and student learning, there is great diversity across schools, so that those with the same level of spending per student can achieve widely different outcomes. What this implies is that the pedagogical process in schools and classrooms probably lacks sufficient management oversight. A challenge in this area is therefore to equip teachers with the skills for more effective teaching, while also putting in place a system of accountability to motivate the application of those skills. Responding to the challenges 24. This report has attempted to offer a broad diagnosis of the main issues in the education sector in Madagascar. As indicated at the outset, its coverage is selective and the results must clearly be put in the country's broader context of human capital development, including policies pertaining to two areas specifically excluded here, viz. non-formal education and early childhood development. Even more important, as a diagnostic product, the report is only a first step toward policy development. That process must begin with a broad discussion among interested parties to agree on the assessment of the system's strengths and the challenges it faces, and beyond that, to explore the options for moving forward. The discussion is needed, not just because it has now become fashionable to talk about participatory approaches to poverty reduction and policy development, but because the business of education naturally involves a myriad of actors, each with a specific contribution to make. Everyone must pull in the same direction for the system to move forward, and the process of dialogue is needed to clarify and agree on roles. 25. The central ministries will obviously play key roles in leading and shaping the dialogue. There are policies over which they have direct control, such as staffing and resource allocation across schools and institutions, and here there is much scope for action to improve the management of the system. The ministries also recognize the importance of partnerships and local management of the educational process. The concept of "contrat-programmes" has been launched at the primary and secondary education levels, and may be extended elsewhere in the system. How much they will produce the expected progress in the schools is still unclear, but as with all innovations, including those that may be motivated by the findings of this report, there  xvi needs to be system of monitoring and evaluation to provide feedback for adjustment as plans are implemented and their impact materializes. The process of learning by doing is a critical ingredient in policy reform and needs to be put in place even as discussions get underway on the future direction for educational development in Madagascar.  Chapter 1: Demographic and Macroeconomic Context 1.1. Madagascar's abundant and unique natural resources hold great potential for improving the livelihood of her population (see box 1.1 for a brief description of the country). Since the mid-i980s various economic reforms have been launched to reverse the downward spiral of economy over the past forty years. Although the reforms have yet to produce a pattern of consistent and strong growth, the signs of an upturn are beginning to emerge. Given the anemic economy the state's investment in education has been modest at best, and even that modest investment has suffered cutbacks during periods of fiscal austerity, as exemplified by the pattern of spending during the 1990s. 1.2. In the context of the Highly Indebted Poor Countries (HIPC) Initiative and the government's explicit commitment to poverty reduction, the next decade promises better prospects for increased support for education.' What then is the scope for boosting the resources for education? How can the resources for the sector-including those currently allocated as well any extra that might become available-be used to maximum effect? This chapter focuses on the first question by documenting key features of Madagascar's demographic and macroeconomic context. It sets the stage for subsequent chapters of the report which take up the second question in more detail. The demographic context 1.3. Madagascar's population in 1999 was estimated at 14.6 million, up from 5.6 million at the country's independence in 1960. Like most other low-income countries, the rate of population increase has been and continues to be high, averaging 2.4 percent a year in the two decades following independence, and accelerating to just under 3.0 percent a year in the 1990s (table 1.1). This rate of growth is projected to continue into the next decade and a half, in contrast to the projected growth rate of 2.2 percent a year for Sub-Saharan Africa, and 1.3 percent a year in other low-income countries (World Bank 2000a). As in most high-growth populations, young people make up a sizable share of the population: those of primary school age (i.e. 6 to 10 year-olds), for example, account for 13 percent of the total. An estimated 75 percent of the population live in rural areas where agriculture is the main source of livelihood and employment. Among urban residents, about 40 percent live in Antananarivo, the capital city, where most people are engaged in petty trade in the informal sector.2 The HIPC Initiative was launched in September 1996 by the World Bank and the International Monetary Fund with the endorsement of some 180 govemments atound the world as a welcome approach to help poor, severely indebted countries reduce debt as part of an overall poverty reduction strategy. In September 1999, at the Annual Meetings of the World Bank and IMF, the HIPC Initiative was significantly expanded to provide more debt relief to more countries at a faster pace. This enhancement, and a redesigned strategy to link debt relief to poverty reduction, is intended to help eliminate debt as an obstacle to development and allow countries to invest more in their future. For more details on the Initiative and its application in specific countries, see the resources posted at htto://www.worldbank.ore./hipc/fa/ftao.html. For an elaboration of Madagascar's poverty reduction strategy, see Govt. of Madagascar 2000. 2 For more information on the labor market and its characteristics see chapter 8.  2 Box 1.1: Madagascar at a Glance Geography: Madagascar is an island nation located in the Indian Ocean off the southeast coast of the African continent. With a surface area of 592,00 km2, it is the fourth largest island in the world after Greenland, New Guinea and Borneo. The country has diverse soils and climates and many species of plants and animals found nowhere else in the world. Parts of the country are vulnerable to droughts and cyclones, and the environment is under increasing stress from massive deforestation as trees are cut for fuel-wood. People: Most of the Malagasy people trace their origin to ancestors from Africa and Indonesia. The two main ethnic groups, the Merina and the Betsileo, representing 26 and 12 percent, respectively, of the population, are mostly of direct Indonesian origin and live in the central highlands. The other major ethnic groups live mainly around the coastal areas and include the Betsimisaraka in the east (about 15 percent of the population), the Tsimehety in the north (7 percent), and the Antandroy in the south (5 percent). Economy: With a per capita GDP of US$235 in 1999, Madagascar is one of the poorest countries in the world. Agriculture is predominant, employing fourth-fifths of the workforce and yielding about one-third of the GDP. The creation of export processing zones in 1990 (which is part of a policy to attract foreign investment by offering favorable tax breaks to qualified companies) is helping to expand the country's modest manufacturing sector in such areas as textiles, clothing, essential oils and essences, woodcrafts, sisal and fisheries. Tourism and rnining are two potentially high-growth sectors. Political system: After decades of French rule, Madagascar became fully independent in June 1960. The country is now organized as a republic with a pluralist, democratic constitution approved by referendum in 1992. Political power is wielded at three levels: the presidency, the national assembly, and local and regional autiorities. Both the president and members of the national assembly are elected by universal suffrage. The constitutional changes approved in March 1998 are expected to strengthen the presidency as well as expand the role of local authorities. The political landscape is complex with shifting alliances across the many parties. Relations with the World Bank: Madagascar became a member of the World Bank in 1963 and received its first loan in 1966. Since then the Bank has provided policy advice, technical assistance as well as project and program financing to support the country's economic development. To date total cumulative lending amounts to more than US$1.6 billion of which US$650 million are for completed projects, US$550 nillion are for projects still being implemented, and US$400 million are for economic adjustment operations. Source: World bank 1999; and Economist Intelligence Unit 2000. Table 1.1: Population size and incidence of poverty, Madagascar 1962-1997 1962 1975 a' 1980 1993 a' 1997 Population growth rate (% p.a.) i 1960-1980 1990-99 Population Total size (millions) 5.7 7.6 8.7 12.2 13.7 2.4 2.9 Of primary school age (6-10) - - - 1.6 1.8 - 2.9 % in rural areas - 84 - 77 77 - - % living in poverty c/ Madagascar 43 - 47 70 75 - - Urban areas - - 65 68 - - Rural areas 72 78 - - Source: IMF 2000 for overall size of the population and MinESEB data files for size of the primary school age population (ages 6-10); and Govt. of Madagascar 2000d for the percentage of the population living in poverty (i.e. below the poverty line); and Govt. of Madagascar 1977 and 2000 for the shares of population in rural areas. a/ Refers to population census years; the dates shown in the other columnns coincide with the dates for which data are available on the incidence of poverty. b/ The growth rates are computed based on population sizes reported in IMF 2000 for the end years of the indicated periods; growth of the school-age population is for the period 1993-1999. c/ Defined as those unable to afford a food basket yielding the equivalent of 2,100 calories per day and to satisfy minimum non-food needs.  3 1.4. As the table above shows, poverty is rampant in Madagascar, and the situation has deteriorated rather than improved over time. Thus, whereas an estimated two-fifths of the population lived below the poverty line in 1962, the share has risen to three-quarters by 1997. With such a high overall rate of poverty, the share of poor people is high in both rural and urban populations. 1.5. Despite the country's poverty, most of Madagascar's health indicators show up favorably against those of other Sub-Saharan countries: the average life expectancy at birth was about 58 years in 1998, and the under-five mortality rate was about 162 per 1,000 live births in 1997, compared with the Sub-Saharan averages of 52 and 147 respectively (World Bank 2000c).3 HIV/AIDS prevalence is still relatively low, with an estimated infection rate of about 0.13 percent in the population aged 15 and above (World Bank 2000c). The current advantageous situation is no reason for complacency, however: the experience in neighboring countries, such as Tanzania, Uganda and Zambia, shows how just easily and rapidly the situation could get out of control in the absence of early preventive action. The government recognizes the risks, and has taken some of the most important precautions against an outbreak, including the establishment since 1988 of a national program to control sexually-transmitted diseases and HIV/AIDS, the creation of a sero-surveillance system to gather data and monitor the situation, and the use of high-level publicity to increase general awareness of the disease and encourage changes in personal behavior to minimize its spread. Economic growth and current conditions 1.6. The economy's dismal performance is the main reason for Madagascar's widespread poverty. Between 1970 and 1997 the GDP grew at only 0.5 percent a year, compared with the population growth rate of about 3 percent a year (World Bank 1999). As a result, the per capita GDP at 1997 prices fell from US$473 in 1970 to US$227 in 1997, a decline of more than 50 percent. The poor performance of the Malagasy economy comes into particularly sharp relief in cross-country comparison: among the 33 African countries for which the relevant data exist it ranks better than only Sierra Leone, Niger, and the Democratic Republic of Congo in the percentage change in per capita GDP between 1970 and 1997 (World Bank 2000b). Not surprisingly, when compared to countries which started out in the 1970s with the same or a smaller per capita GDP, Madagascar now lags far behind those countries (table 1.2). 3Still, half of all children under age five is estimated to be malnourished at present (World Bank 2000c).  4 Table 1.2: Madagascar's economic performance in comparative perspective, 1970-1999 COUntrY Per capita GDP as a multiple of Madagascar's Annual GDP growth rate, 1970-1999 (% 1970 1999 p.a.) Madagascar al1.0 1(0 0.5 (473) (227) Benin 0.9 1.7 3.3 Burkina Faso 0.5 1.1 3.6 Kenya 0.6 1.4 4.5 Lesotho 0.6 2.2 5.5 Zimbabwe 1.3 2.9 4.6 Indonesia 0.5 4.4 7.3 Pakistan 1.0 2.0 5.3 India 0.6 1.5 4.4 Source: World Bank 1999, updated with data for 1999 from the World Bank Statistical Infornation Management and Analysis (SIMA) database. a/ Figures in parentheses refer to the per capita GDP of Madagascar in 1997 US dollars. 1.7. Given Madagascar's many endowments and natural advantages and the fact that demographic pressures are no greater than those experienced by other low-income countries, the poor performance of the economy must be laid squarely at the feet of failed economic policies (World Bank 1999; IMF 2000). From the 1970s until the mid-1980s, the country pursued a development strategy based on economic nationalism, self-sufficiency and pervasive state involvement in and control of economic production. Excessive government regulation not only distorted incentives and production, but also discouraged the domestic and foreign investment needed to boost economic growth. The mid-1980s marked the beginning of new policies which were directed at fiscal and administrative reform, economic and financial liberalization, and disengagement of the state in direct economic production. Progress has been halting and was interrupted by political instability during 1991-93. Efforts in these directions have been renewed since the mid-1990s, and the signs of turnaround are just beginning to emerge. For the first time in many years, real GDP growth grew faster than the population, averaging 4.1 percent a year between 1997-99 (compared with only 1.9 percent a year in 1995-96); as a result, the per capita income grew at about 1 percent a year during this period. The upswing is good news indeed, but represents only the beginning of a trend that must obviously be sustained to produce tangible reductions in poverty. World Bank (1999) estimates that bringing the poverty rate down to 32 percent would be possible only with a GDP growth rate of about 7.5 percent a year over the next two decades4. Overall government finances 1.8. Better management of government finances is an important part of efforts to renew economic growth. Tables 1.3 and 1.4 document various indicators in this regard over the past 10 years. The government is dependent to a significant extent on grants from external In an undated article, Richard Jolly argues that a growth rate as high as 7.5 percent a year may not be needed if the development strategy puts more emphasis on policies that directly promote income-generating activities among the poor.  5 donors for its revenues, with this source representing about one-quarter of total revenues during most years of the decade. Including such grants, total revenues stood at 16.4 percent of GDP at the start of the 1990s, but has fallen to 15.1 percent by 1999; the latter represents a recovery since mid-decade, when the corresponding percentage was barely more than 11 percent. Excluding grants, current revenues average less than 10 percent of GDP over the decade, with mid-decade again showing the lowest percentages. Since 1997 the trend shows a strong upturn, reflecting increases in income tax revenues generated by renewed economic growth and improved tax collection. To improve the govermment's capacity to finance the country's many unmet development needs, World Bank 1999 suggests that by the late 2010s, it would be desirable to raise current revenues (i.e. income from taxes and other domestic sources) to around 15 percent of GDP. Table 1.3: Overall government revenue, Madagascar 1990-99 Total revenue Current revenue Grants as % of Year Billions of As % of Index As % of Index total revenue cuTrent FMG GDP 1990=100 GDP 1990=100 1990 752.8 16.4 100 11.8 100 26.9 1991 533.2 10.9 66 8.5 72 19.5 1992 752.6 13.5 82 9.8 83 26.0 1993 864.5 13.4 82 9.7 82 26.1 1994 1,035.9 11.3 69 8.3 70 26.5 1995 1,541.8 11.4 70 8.5 72 25.4 1996 2,090.7 12.9 79 8.7 73 32.7 1997 2,703.4 15.0 92 9.7 82 35.4 1998 2,864.3 14.1 86 10.6 90 24.4 1999 3,522.0 15.1 92 11.4 96 24.5 Source: computed from data on govemment finance from the World Bank Africa Live Database, and GDP data fiom the World Bank SIMA database; see also appendix table All.1. 1.9. Turning now to the expenditure side, table 1.4 shows that total expenditures have consistently exceeded total revenues, often by exceptionally large margins, for most of the decade; in 1994, for example, they were 1.74 times the government's total receipts. Current spending-that is total expenditure minus capital expenditure and debt service payment- represents between 53 and 65 percent of total spending throughout the 1990s; averaging around 10.8 percent of GDP over the decade, they match current revenues more closely than do total expenditures and revenues. The table highlights the role of debt interest as a source of fiscal distress during the 1990s: in the most difficult years at mid-decade, total debt interest payment accounted for nearly 30 percent of total government expenditure (or around 5 percent of the GDP), with most of the payment made to external creditors. The situation has improved considerably since 1997, with total interest payment declining to 2.1 percent of GDP by 1999.  6 Table 1.4: Overall government expenditure, Madagascar 1990-99 Total expenditure Current expenditure v Interest on all public debt Interest on extermal debt Year Billions of As % of GDP As ratio of As % of total current totaldevt.As %of total As%of As %of total As %of debtiers As %of FMGren, % Index1 ta revenue expenditure GDP expendituTe GDP t interest GDP FMG ~~~~1990=100 evnepayment 1990 784.1 17.0 100 1.04 53.4 9.1 8.9 1.5 68.3 1.0 1991 805.6 16.4 96 1.51 59.6 9.8 12.3 2.0 56.0 1.1 1992 1,120.9 20.0 118 1.49 58.8 11.8 17.1 3.4 40.3 1.4 1993 1,328.4 20.6 121 1.54 58.9 12.1 19.4 4.0 78.7 3.2 1994 1,803.6 19.8 116 1.74 64.8 12.8 27.3 5.4 94.0 5.1 1995 2,373.8 17.6 103 1.54 64.2 11.3 29.0 5.1 90.8 4.6 1996 2,883.4 17.8 104 1.38 59.1 10.5 26.3 4.7 82.8 3.9 1997 3,137.1 17.4 102 1.16 62.5 10.9 17.5 3.0 91.7 2.8 1998 3,819.2 18.8 110 1.33 56.2 10.5 14.5 2.7 84.2 2.3 1999 3,624.7 15.5 91 1.03 57.9 9.0 I 13.6 2.1 65.9 1.4 Source: computed from data on govemrnment finance from the World Bank Africa Live Database and GDP data fiom the World Bank SIMA database; see also appendix table Al.1. al Refers to the difference between total expenditure net of capital expenditure and debt service payment Overall public spending on education 1.10. In the context of overall government finances described above what has been the pattern of public investment in education? The relevant data appear in table 1.5. Total public spending on education averaged 2.5 percent of the GDP over the 1990s, but there has been substantial variation around that average, falling from a high of around 3.0 percent or more at the beginning of the decade, to a low of 1.8 percent in mid-decade before rising again to about 3.0 percent in the 2000 budget. As a share of total public spending, spending on education followed a similar U-shaped pattern, with highs at the beginning and end of decade and a trough in the middle years. Averaging around 14 percent over the decade, the share did not rise above 20 percent in any year. 1.11. As expected most of the public spending on education is accounted for by current spending (in which teacher salaries are the main item of expense). Its share rose to a high of 96 percent in 1995 just when overall spending on education was squeezed, and fell to a low of 71 percent just when spending rose again. The trend reflects the unsurprising fact that it has been easier to expand and contract capital rather than current spending on education in response to overall fiscal conditions. Thus, while total spending on education as a percentage of the GDP in 2000 has bounced back to the levels at the beginning of the 1990s, current spending as a percentage of the GDP is currently still only 86 percent as high as it was in 1990. Because the functioning of the education system depends on adequate levels of current spending, these trends suggest that the system's recovery from the fiscal austerity of the 1990s is probably still incomplete.  7 Table 1.5: Overall public spending on education, Madagascar 1990-2000 Current and capital spending on education a/ Current spending on education Year As % of GDP As % of total public As % of total public As % of GDP spendnntotal current spending spending on education interest oayment % Index % Index % Index % Index % Index 1990=100o 1990=100 1990=10 ______ 1990=100 1990=100 1990 3.0 100 17.6 100 85.9 100 2.6 100 28.3 100 1991 3.2 105 19.3 110 84.9 99 2.7 104 27.5 97 1992 2.8 93 13.9 79 85.0 99 2.4 92 20.2 71 1993 2.6 87 12.7 72 90.2 105 2.4 91 19.4 69 1994 2.1 71 10.8 62 91.5 107 2.0 76 15.3 54 1995 1.8 58 10.0 57 96.2 112 1.7 65 14.9 53 1996 2.0 66 11.1 63 85.9 100 1.7 66 16.1 57 1997 2.3 78 13.5 77 81.4 95 1.9 74 17.6 62 1998 2.6 85 13.7 78 81.8 95 2.1 81 19.9 70 1999 2.8 92 17.8 101 77.0 90 2.1 83 23.7 84 2000 3.1 103 - 71.1 83 2.2 86 - - Source: computed from data on public spending on education in Govt. of Madagascar (various years, a) for 1990-92; Govt. of Madagascar (various years, b) for 1993-1996; and Govt. of Madagascar (various years, c) for 1997-2000; and GDP data from the World Bank SIMA database; see also appendix table Al .2. a/ Includes curent and capital spending in MinESEB, METFP, and MinESup, as well as the salaries of ex-VSNs paid by local govenutnents, the salaries of physical education teachers paid by the Ministly of Youth and Sports, and central govemnment transfers to cormnunes for schools. 1.12. The pattern of public spending on education in Madagascar during the 1990s exemplifies the adverse impact that a heavy burden of external debt can have on the education sector. Figure 1.1 suggests that as interest on the country's external debt mounted in mid-decade, the share of education in total government spending declined. Plotting education spending relative to the burden of debt service brings out the negative relation even more clearly (figure 1.2). Not only was capital spending on education reduced, but spending to support the day to day functioning of the system also suffered the same fate. The evidence thus suggests that the education sector absorbed a disproportionate share of the fiscal squeeze associated with debt service payment during the 1990s. Each percentage point rise in debt service as a share of total government spending reduced current spending on education as a share of total non-interest current government spending by 0.54 percentage points; and reduced total spending on education as a share of the GDP by 0.05 percentage points.  8 Figure 1.1: Trends in external debt service and public spending on education, Madagascar 1990-99 3 0 2 5 &c 2 0 - U -Education spending