Report No. 28143-SE Senegal Policies and Strategies for Accelerated Growth and Poverty Reduction A Country Economic Memorandum April 3, 2003 Poverty Reduction and Economic Management 4 Africa Region Document of the World Bank ACRONYMSAND ABBREVIATIONS AfDB African DevelopmentBank AGETIP Agence d 'Exe'cutiondes Travauxd 'Inter&Public (Executing Agency for Works of Public Interest) AGOA Africa Growth OpportunityAct AOF Africque OccidentaleFranqaise APIX Agence de Promotion des Investissements et des Grands Travaux (InvestimentPromotion and Infrastructure Agency) BCEAO Banque Centraledes Etats de IYfrique de 1'Ouest (Central Bank of West African States) CDMTs Cadre de De'penses a Moyen Terme CEMAC Communaute'Economique et Mone'tairede 1;Ifrique Centrale (Economicand Monetary Communityof CentralAfrica) CET Common External Tariff CFA Communaute'FinanciBre Africaine (AfricanFinancial Community) CFAF CFA franc CIB Consolidated InvestmentBudget CIDA Canadian International Development Agency CNDCL ConseilNational de Dkveloppement des Collectivite'sLocales COF Contr8le des OpkrationsFinancieres CREA Centrede Recherches Economiques Applique'es(Center for Applied EconomicResearch) CWIQ Questionnaire UniJe'des lndicateurs de De'veloppementde 1'ESAM I1(Core Welfare Indicators Questionnaire) DCEF Direction de la Coope'ratione'conomiqueet$nanciBre DDI Direction de la Dette et de 1'Investissement DPS Direction de la Pre'vision et de la Statistique (Directorateof Forecasting and Statistics) DTAI Direction du TraitementAutomatise' de 1'Information DTC Droits ESAM-I EnquEte Se'ne'galaiseAuprBs des Me'nages (1994) (Household BudgetKonsumption Survey(1994)) ESAM-I1 EnquEte Se'nkgalaiseAupr2s des Me'nages (2001)(Household BudgetKonsumption Survey(2001)) FDD Fons de Dotation de la De'centralisation FECL Fonds d 'Equipementdes Collectivite'sLocates FIAS Service Conseilpour 1Investissement e'tranger FNR Foreign Investment Advisory Service FWA West African Federation GDP GrossDomestic Product GTZ Coope'rationtechniqueallemande HIPC Heavily Indebted Poor Country HIVIAIDS Human Immuno-deficiencyViruslAcquired ImmuneDeficiency Syndrome IFC International Finance Corporation 11 IMF International Monetary Fund IPRES Institution de Pre'voyanceRetraite du Se'ne'gal JICA Japanese International Cooperation Agency MDGs MilleniumDevelopment Goals MICS Multiple Indicator Cluster Survey MIGA Multilateral Investment Guarantee Agency NGO Non-GovernmentalOrganization NIP New Industrial Policy OECD Organization for Economic Cooperation and Development OHADA Organisationpour 1'Harmonisation du Droit des Affaires en Afrique (Organization to HarmonizeBusiness Law inAfrica) PAI-BAD Projet d 'Appui institutionnel- Banque africaine de De'veloppement PDEF Programme De'cennal de 1'Education et de la Formation (Ten-Year Education and Training Program PRSP Poverty Reduction Strategy Paper P T P Programme triennal d 'Investissementspublics QUID Questionnaire UniJie'sur les Indicateurs de De'veloppement RPED RegionalProgram on Enterprise Development RULC Relative Unit Labor Cost SMEs Small andMediumEnterprises SMIs Small andMedium Industries TOFE Tableau des Ope'rations FinanciBres de 1'Etat (State Financial Operations Table) UEMOA Union 6conomique et mon6taire ouest-africaine UNDP UnitedNations Development Program UNIDO UnitedNations ???/ USAID UnitedStatesAgency for InternationalDevelopment WAEMU West African Economic and Monetary Union WHO World Health Organization WTO World Trade Organization Vice President: Callisto E. Madavo Country Director: John McIntire Sector Director: PaulaDonovan Sector Manager: Robert Blake ACKNOWLEDGEMENTS This study was prepared during 2002 by a team o f economists from the World Bank and from Senegal. The latest data available before finalization of the report i s from Fall o f 2002. The document was sent to the Government in March 2003 and discussed with them inJuly 2003. The World Bank team was led by Nancy Benjamin (AFTP4) and includes Waly Wane (DECRG), Agnks Soucat (AFTHD), Linda English (MNSHD), Jan Walliser (AFTP4) and Ashutosh Dubey, David Lindauer, Anne Valenchik, Scott Jacobs (consultants). Logistical support was provided by Marta Berhane and Josette Percival (AFTP4). The Senegal team comprised a budget system reform working group led by Mamadou L. Deme, Conseiller Technique (MEF), NCnk Dieng, Directeur General des Finances, Daouda Diop, Directeur DCEF, Racine Kane, Expert de UPE, a poverty monitoring group led by SoguC Diarisso, Directeur de la Prkvision et de la Statistique (MEF), andProfessor Aly Mbaye and Frangois Diop (consultants). The project has benefited from the overall guidance o f Emmanuel Akpa (AFTP4), from the comments o f peer reviewers, Aristomene Varoudakis (MNSED), Jeffrey Hammer (DECRG), Philip English (WBPR), and many useful comments from the World Bank and IMF Senegal country teams. The team is also grateful to Alan Gelb (AFRCE) and to Abdoulaye Diop, Ministero f Economy and Finance, for their support. TABLE OF CONTENTS Overview .......................................................................................................................................... i Section I:Wealth Creation ........................................................................................................... 1 1. Growth. Poverty and the Labor Market.............................................................................. 1 A. Introduction.............................. ...... ................... ........................... 1 B. GDP Growth....... ............................. .........2 C. MeasuringPover ............................. ............................................................... 3 D. By How Much Should Poverty Fall?.............................................................................. 5 E. Poverty and the Rural Sector ......................................................................................... 7 F. WhatDo Labor Market Outcomes Tell Us?.......... G. I s the Labor Market to Blame?............................... H. Employment Policy in the PRSP............................ 2. Competitiveness. Manufacturingand Exports inSenegal ................................................ 33 A. Manufacturing in Senegal..................... ........................ 33 B. Explaining the Performance of Manufa 37 C. Policy Implications....................................................................................................... 42 D. Conclusions and Recommendations............................................................................. 44 3. Regulatory Reform. Market Performance. and Poverty Reductionin Senegal ................54 A. Introductio B. A Brief His C. Improving Regulatory Practices and Capacities in D. Regulatory Reform and Senegal's Informal Sector...................................................... 64 Section 11: Promotion of Human Capital and Basic Social Services ...................................... 71 4. Health. Nutrition and Population...................................................................................... 71 A. Childmortality B. Malnutrition ..... C. Fertility ............ D. Maternal mortality ....................................................................................................... 78 E. HIV/AIDS..................................................................................................................... 79 F. Water and sanitation ........ .... ..... ..............79 5. Education.......................................................................................................................... 99 A. Thestate of education in Senegal................................................................................. 99 B. Public expenditure management and Senegal's education sector ............................. 103 Section 111: Improving Conditions for Vulnerable Groups .................................................. 115 6. Commodity Tax Reformfor Poverty Alleviation........................................................... 115 A. Introduction................................................................................................................ 115 B. Aggregate Expenditureand Poverty Pro$les ............................................................. 117 C. Conclusions ................................................................................................................ 133 7. Pension Reformin Senegal............................................................................................. 135 SectionIV: Principles for Implementation of the Strategy .................................................... 140 8. Expenditure Tracking Survey inSenegal: The Health Sector ........................................ 140 A. Introduction............................................. B. The TrackingExercise: theflow of funds ................................................................... 144 C. Trajectory of theflows offunds .................................................................................. 145 11 .. 9. D. Conclusion.................................................................................................................. 155 MonitoringPRSP implementation.................................................................................. 159 A. Innovations in theMonitoring of the Poverty in Senegal........................................... 159 B. Support to the Statistical System ................................................................................ 160 10.C. Information Systems and Surveys to Support the PRSP............................................. 161 Improving the Efficiency o fPublic Expenditures...................................................... 167 A. Introduction..... ...... B. Outline of Possible Solutions ....... .................................. Listof Tables Table 1: Contributionto Growth .................................................................................................... ... ii Table 2: Senegal Economic Structure. 1980/2000.......................................................................... 111 Table 3: Real Output Average Annual Growth ............................................................................... v Table 4: Output and Employment Shares inSenegal, 1994 and 2001............................................ vi Table 5: Ownership Structure inthe Ten Largest Firms (1991)................................................... viii ... Table 6: Cohort Survival Rates, Grade EnrollmentRates, and Transition Rates Between Cycle, 2001............................................................................................................................................. xvii Table 7: Distribution o f per capita subsidyby education level and income (CFA), 2000..........xviii ... Table 1.1: GDP Growth inthe CFA Zone. 1994-2000 ................................................................ 23 Table 1-2: Alternative Estimates o f Poverty in Senegal............................................................... 24 Table 1-3: Trends inGroundnut Prices. Productionand Exports. 1994-2001.............................. 25 Table 1-4: Output and Employment Shares in Senegal ................................................................ 26 Table 1-5: "Hiring and FiringPractices are flexibly determined by employers" ......................... 27 Table 1-6: Firmclassification o f obstacles to firm operation and growth.................................... 28 Table 1-7: Firms' opinions about labor market reforms............................................................... 28 Table 2-1: Real OutputAverage Annual Growth......................................................................... 47 Table 2-2: Comparative ManufacturingPerformance for Senegal and SelectedOther A h c a n Countries........................................................................................................................................ 47 Table 2-3: Real Manufacturing Exports .Average Annual Real Growth Rate............................ 48 Table 2-4: Sector Shares inTotal Real ManufacturingExports................................................... 48 Table 2-5: Relative Unit Labor Costs .Senegal Against Comparison Countries ......................... 49 Table 2-6: Relative Producer Price .Senegal Against ComparisonRegions............................... 50 Table 2-7: Relative Wages in $ .Senegal Against Comparison Regions..................................... 51 Table 2-8: Relative Productivity .Senegal Against Comparison Regions................................... 52 Table 2-9: Estimation o f TFP and LP Levels and GrowthRates.................................................. 53 Table 3-1: Summary o f structural and regulatoryreforms in Senegal.......................................... 58 Table 4-1: HNP Indicators.Senegal in Sub-SaharianAfrica circa 1999(1) ................................ 72 Table 4-2: Senegal .Technical efficiency o f health services ...................................................... 91 Table 4-3: Senegal -Public healthexpenditue 1993-2000 .......................................................... 91 Table 5-1: Trends inEnrollment by Level and Type o f Education, 1997-2001......................... 100 Table 5-2: Cohort Survival Rates, Grade Enrollment Rates, and Transition Rates between Cycle. 2001 ............................................................................................................................................. 101 Table 5-3: Scores (grade 5) on standardized tests ina group o f countries ................................. 103 Table 5-4: Public Spending on Education (millions o fCFA and dollars) .................................. 104 Table 5-5: Distribution o fpublic recurrent expenditure on education by level selectedyears between 2000-2002...................................................................................................................... 104 111 Table 5-6: Distribution o fpublic education recurrent expenditure by levels and categories. 2000 Table 5-7: Distribution o fenrollments and recurrent spending by education level. 2000...........106 ..................................................................................................................................................... 106 Table 5-8: Unit cost per student by education level inthe public sector. 2000 ........................... 107 Table 5-9: Comparison of unit cost at different education levels ................................................ 108 Table 5-10: Disaggregation ofper student public spending. 2000 .............................................. 108 Table 5-11:Unit cost by education level (CFA). 2000................................................................ Table 5-12: Distribution o fper capita subsidy by education level and income (CFA). 2000 .....109 110 Table 5-13: Unit cost by gender and education level (CFA). 2000 ............................................. 112 Table 5-14: Incidence o fpublic education expenditure on the poorest and richest quartiles selected countries. % share.......................................................................................................... 112 Table 5-15: Unit cost by region and education level (CFA). 2000.............................................. 113 Table 5-16: Unit cost by gender. region and education level (CFA). 2000 ................................. 113 Table 6-1: Basic expenditure statistics for Senegal. 2001.......................................................... 118 Table 6-2: Expenditure groups size by region............................................................................ 120 Table 6-3: Regional poverty profiles .......................................................................................... 120 Table 6-4: Aggregate expenditure pattem o f income groups within regions ............................. 124 Table 6-5: Illustration o f aggregate gains inpoverty reduction................................................... 126 Table 6-6: Expenditurepattem for selected commodities .......................................................... 128 Table 6-7: Gini Expenditure Elasticities for selected commodities ............................................. 130 Table 8-1: Communes and districts inthe sample o f the survey ................................................ 143 Table 8-2: Flow o f Resources Trajectory ................................................................................... 145 Table 8-3: FDDtrends ................................................................................................................ 146 Table 8-4: the average year for the FDDcircuit .......................................................................... 148 Table 8-5: Steps for the trajectory o fFDDresources ................................................................. 149 Table 8-6: Health FDDreceived by the local Govemments and health providers ...................... 153 Table 8-7: Ratio o f amount received to amount disbursed.......................................................... 154 ListofFigures Figure 1: Senegal Trends o fUnder-Five Mortality Duringthe Past Decades.............................. Figure 2: Senegal Evolution o f Technical Health Personnel by Sector and Category..................xiv xvi Figure 1: Senegal Trends o fUnder-Five Mortality Duringthe Past Decades.............................. Figure 2: Senegal Evolution o f Technical HealthPersonnel by Sector and Category..................xiv xvi Figure 1-1:AnnualRate of Growth o f GDP inPercent.................................................................. 29 Figure 1-2: Activity ofpersons age 15 and up, inpercent:............................................................ 30 Figure 1-3: Labor Forceby Employment Status, inpercent5:...................................................... Figure 2.1: Compared Evolution o f Labor Productivity and TFP for Total Manufacturing: ........:32 Figure 1-4: Employment Status, Labor Force inSecondary and Tertiary Sectors, inpercent: ....:31 46 Figure 4-1: Senegal within global trends inGDP:per capita and under-five mortality circa 1999: ............................................................................................................................................. 73 Figure 4-2: Senegal - Trends o f under-five mortally duringthe past decades:............................. 74: Figure 4-3: Senegal - Inequalities inunder-five mortality duringthe past decades: .................... 75: Figure 4-15: Senegal - Inequalities inchild malnutrition: ............................................................. 76 Figure 4-5: Trends inFertility (age-specific fertility rates) in Senegal duringthe past two decades ......................................................................................................................................... 77 Figure 4-7: Senegal-Environmental ContaminationAccess to potable water and sanitation by region............................................................................................................................................. 80 iv Figure 4-8: Senegal .Level o f instruction of women o f childbearing ages. use o f condom among menwith occasional partners. assistance bytrainedhealthpersonnel at delivery. andchildren aged 12-23 months fully vaccinated by region (ESIS 1999) ......................................................... 82 Figure 4-9: Senegal .General population .Use o f health services and choice ofprovider among individuals who reported an illness duringthe four weeks preceding the survey by type o fplace o fresidence.................................................................................................................................... 85 Figure 4-10: : Senegal .General population .Use o f health services and choice o fprovider among individuals who reported an illness duringthe four weeks preceding the survey byregion ........................................................................................................................................... 86 Figure 4-1 1: : Senegal - Average household healthrelated expenditures (FCFA) per capita per year by region.......................................................................................................... 87 Figure 4-12: Senegal - Evolution o ftechnical health personnel by sector and category: ..............88 Figure 4-13: Senegal - Populationper technical healthpersonnel ratio by region and category in 2000 ............................................................................................................................................... 89 Figure 4-14: Senegal -Continuity o f maternal and childhealth services .................................... 91 Figure 4-15: Senegal -Public health spending (FCFA millions) by category: personnel, non- personnelrecurrent, and investment spending between 1993 and 2000........................................ 96 Figure 5:1: Distributionofpublic ................................................................................... 111 Figure 6-1:Expenditure Lorenz curves for the regions ............................................................... 119 Figure 6-2: Householdand Regional Expenditure Patters........................................................... 123 Figure 6-3: ConcentrationCurves............................................................................................ 125 Figure 6-4: Intersection Modified Concentration Curves............................................................ 127 Figure 6-5: Cumulative gain for households ............................................................................... 128 Figure 6-6: Expenditure Shares o f ProtectedCommodities ........................................................ 126 Figure 6-7: Concentration Curves for Meat, Sugar, Oil, and Rice .............................................. 130 Figure 6-8: Cumulative gain Curves for Rice vs Meat, Sugar, and oil........................................ 131 Figure 6-9: Cumulative gain curve for a reformwith many commodities .................................. 132 ListofBoxes Box 1-1: The De-Industrialization o f Senegal's Labor Force ....................................................... 20 Box 1-2: ChangingEmployment Status inthe Primary Sector ..................................................... 21 Box 1-3: Other evidence on the impact on labor regulations inAfrica ........................................ 22: Box 2-1: Senegal: Determinants o f Productivity Growth. 1974-99............................................... 45 Box 6-1:A Methodology for FindingRevenue.Neutra1. Poverty-Reducing Tax Changes.........134 PREFACE The work on this Country Economic Memorandum was launched as Senegal was finalizing its PRSP in early 2002. The work was brought to a conclusion as the implementation period o f the PRSP began. Thus it i s hoped that the C E M will increase understanding o f the main themes o f the PRSP and aid in the development o f implementation plans, as well as the reflections on the strategy undertaken inpreparation of the first annual progress report. The C E M was also undertaken in a context o f other significant investigations on PRSP themes, some still ongoing, the reports on which will convey analysis on questions that could not be taken up inthe current study. One o f the major studies recently completed i s the Diagnostic Trade Integration Study. This report examined all aspects o f barriers to increasing integration o f the Senegalese economy with the intemational economy. It examined Senegal's trade policy, trade institutions, customs services, as well as the comparative advantages and trade prospects for key trading sectors including agriculture, fisheries, textiles, and tourism. The CEM was substantially informed by the DTIS evaluation o f the adoption o f the Common Extemal Tariff and the impact o f regional trade agreements. Recommendations of the DTIS are currently beingprepared for implementation. One o f the most important studies related to the PRSP i s the examination o f poverty statistics, especially in light o f new data from the 2001-2002 household survey. Since the second round o f the survey was not completed until mid-2002, it was not possible to make use o f fully-processed results for the CEM. The data used inthe C E M represent preliminary results and are subject to change once all results are hlly processed. For the basic poverty rate, the C E M has relied on results reported in Senegal's PRSP, which, again, are preliminary. Due to data limitations at the time o f the preparationo f the PRSP, the poverty rate given in the document for 2001-02 is based on a type o f estimation. Specifically, a regression was prepared with the 1994-95 data to identify correlates o f adult-equivalent consumption using variables common to the 1994-95 survey and the 2001-02 survey. The coefficient estimates in the regression were then used to predict consumption using the values o fthe variables observed at the householdlevel inthe 2001-02 survey. A full poverty assessment, based on refined results o f the household survey and other data, i s planned for the near future. This poverty assessment will consider four main topics: trends in poverty and inequality; the microeconomic and macroeconomic determinants o f poverty; public services and poverty (education, infrastructure, and health); and the PRSP's development targets, related costs, and financing issues. One o f the main macroeconomic variables affecting growth and competitiveness o f the Senegalese economy i s the nominal exchange rate. As a member o f the West Africa Economic and Monetary Union, Senegal's currency i s the CFA franc whose parity 11 * . to the Euro is fixed by agreement among all WAEMU members, and whose issuance is governed for all members by a single central bank. Therefore, the nominal exchange rate i s not a policy variable under Senegal's direct control. The IMF monitors macro- economic conditions o f WAEMU members and o f WAEMU as a union, including the appropriateness o f the nominal exchange rate. The most recent review [IMF Board Document SM/03/75 o f February 031 has found the parity to be appropriate, however it remains under surveillance. For these reasons, the C E M does not take up issues concerning the nominal exchange rate, however, several measures o f the real exchange rate are examined. C E M chapters on human capital include treatment o f PRSP-related issues in health and education. The chapters do not present a comprehensive analysis o f all issues in either sector, but rather a portrait of constraints hindering progress toward PRSP targets and the main points o f interrelation first, between health and education, and second, between the public expenditure system and the delivery o f health and education services to the poorest citizens. Fortunately, a complete review o f health sector issues will be completed soon and will draw a full slate o fhealthpolicy recommendations. A Public ExpenditureReview o f the health sector i s also underway which will make further use o f results from the expenditure tracking survey, especially as concerns results on health facilities. The Education sector has also undertaken an expenditure tracking survey, and these results will be disseminated soon. Regarding social protection, the C E M focuses analyses on important topics intax incidence and pensions, but does not treat more widespread issues o f risk reduction affecting many vulnerable groups. However, an upcoming Rural Social Protection Strategy plans to expand the analysis o f formal and informal risk reduction and coping strategies by vulnerable groups, which should be complementary to the CEM. Two C E M chapters on tools for PRSP implementation were written by the Government o f Senegal - one on systems for monitoring PRSP implementation and one on budget system reforms. These government investigations will soon be reinforced by further diagnostics o f fiduciary standards undertaken jointly by the Bank and Government. Further, the Government i s now launching a study to develop means to improve their system o f fiscal decentralization and service delivery. Thus the CEM considers salient aspects o f the four PRSP pillars and presents an overview that synthesizes a collective sense o f recent policy performance and key constraints that inhibit reaching PRSP goals. The general picture is one o f significant economic growth mixed with disquieting inequities -- in income, health, education, and opportunities; macroeconomic stability matched with a still uncertain business climate, low investment and productivity growth; successful market liberalization paired with limited capacity to mobilize public resources infavor o f PRSP priorities. Insum, Senegal represents much unrealized potential as well as hope that the PRSP dynamic will provide ... 111 the motivation to remove constraints forming Senegal's greatest development challenges. The main conclusions andrecommendations are summarized below. CONCLUSIONSAND RECOMMENDATIONS Wealth Creation Senegal should foster the kind o f cooperation between unions, firms and government, that can shift all parties' focus away from dividing rents and toward the expansion o f employment and production. The most effective instrument inthis endeavor is to increase transparency, both in development and application o f economic policy and inall aspects o fpublic expenditure management. Transparency has powerful effects in the policy process - it encourages the development o f better policy options, and helps reduce the incidence and impact o f arbitrary decisions. Moreover, transparency helps consumers trust competition more because special interests have less power to manipulate government and markets. Of the various factors affecting competitiveness, productivity growth is the most sensitive to public policy. To raise productivity, Senegal will have to address a number of fundamental constraints relating to infrastructure, labor skills and other assistanceto small andmedium sized firms, financial markets, and labor relations. Improving the general primary, secondary and university education system, eradicating illiteracy and improving educational attainment o f girls and women are more likely to prove productive in improving standards o f living than are more narrow investments in vocational training. Creating an environment in which firms find it profitable to have a long term commitment to the marketplace and the labor force is more likely to stimulate appropriate human capital formation than are committees managing vocational training programs. Senegal can accelerate the transition of businesses from the informal to the formal sector by creating a new legal status for informal enterprises to govem a transition period toward formalization with specific incentives. Promotion of Human Capital and Basic Social Services Poverty clearly leads to poor health outcomes which can be overcome with properly targeted interventions ininfrastructure-especially water and sanitation-as well as selected health interventions. All these interventions must involve the target community in directing and monitoring the allocation of resources, and they must be geared towards improving health behavior which is the major determinant o fhealthoutcomes. The Senegalese education system is inefficient at the level o f both primary and secondary education because o f low rates o f inter-cycle transition and due to high dropout and repetition rates. If the government wants to achieve its goal of universal primary education, improvements in overall efficiency of the education system are critical. iv Spending on public education in Senegal i s very inequitable, especially inpost- primary education. The incidence o f govemment spending on education i s biased towards the richest segment o f the population, against girls and in favor o f urban areas. These biases should be removed by improving access to primary and secondary education. Improving conditionsfor vulnerablegroups 0 Barriers to market entry act like taxes, raising prices to consumers, and the impact o f such taxes on poor consumers needs to taken into consideration in judgingthese marketbarriers. 0 The credibility o f the pension system must be restored to ensure its usefulness as a source o f longterm savings for the economy. Principlesfor Implementation of the Strategy Systems of fiscal decentralization must be adequate to the delivery o f social services in all regions, so as to enable progress toward MDGs as targeted in the PRSP. Specifically the Government should: 0 Simplify the administrative and budgetary procedures for expenditure execution; 0 Gradually extend the use o f medium-term expenditure frameworks (CDMTs) inkeepingwith the budgetsby objectives previously established inthe education and health sectors, and with full participation from localities in developing and committing to objectives; 0 Introduce a high-performance information system on public expenditures; and 0 Ensure that this information system includes and profits from evaluation o f the impact o f public spending. OVERVIEW SENEGAL'S PRSP 1. Over the next several years Senegal's development policies will be guided by the strategy outlined in its Poverty Reduction Strategy Paper (PRSP). The PRSP declares that the national strategy rests on four axes: i)creation o f wealth; ii)promotion o f human capital and basic social services; iii)improving conditions for vulnerable groups; and iv) principles for implementation o f the strategy. Accordingly, the main themes o f this strategy concern the policies and institutions that can enhance economic growth and ensure that this growth contributes to poverty reduction and human development. The purpose o f the Senegal Country Economic Memorandum (CEM) i s to provide analytical work that can underpin a medium-term strategy for aiding the Government's implementation o f its PRSP. The analysis in the C E M supports the main themes o f the PRSP, growth and human development, poverty monitoring and the link to policy, and tools for improving the implementationo fpoverty-fighting programs. 2. These themes are compelling for the very reasons that made them the focus o f the PRSP. Economic growth over the last several years has been encouraging, but the impact on poverty reduction i s disappointing. Important progress has been achieved in some human development indicators, but other indicators have declined, or shown little impact for the amount o f resources spent. Attempts to improve performance o f poverty fighting programs have runinto constraints o f poor budget management and poor monitoring and evaluation. Thus while the PRSP establishes the broad objective o f the Government's strategy, a thorough analysis o f the issues and obstacles involved in meeting those objectives i s essential for developing a plan o f assistance adequate to reaching PRSP goals. 3. The study begins with an exploration o f the policy environment for implementation o f the PRSP. The remainder o f this chapter gives an overview o f this policy context by examining the record o f growth and reform since the CFA devaluation in 1994. The four sections o f the study are organized according to the themes of the PRSP: The first section explores various aspects o f the conditions for hture economic growth and wealth creation, including competitiveness, productivity, trade integration, regulatory reform, and labor markets. The next section evaluates the efficiency o f Government programs designed to enhance human development, including education and health. The following section treats some specific issues in protection of vulnerable groups. The final section investigates tools for PRSP implementation and improving the impact o f Government programs on poverty reduction such as budget reforms, monitoring and evaluation, and methods o f studying the distributional consequences o f public policies. .. 11 HOWDOES SENEGAL GROW? 4. The story o f Senegal's recent economic growth can help determine whether the seeds o f accelerated growth are well-planted in the current environment or whether current prospects are worrisome and call for new reforms. A natural point o f departure for measuring recent growth performance is the 1994 devaluation o f the CFA franc from 50 to 100 to the French franc, and now converted to the equivalent fixed value with the Euro. Before the devaluation, real economic growth was highly erratic, and net per capita income declined between independence and 1994. After the devaluation, real growth has been fairly steady, averaging 5 percent, and allowing for sustained increases in per capita income of 2 percent per year. Nevertheless, this episode needs to be examined in detail to discover what the devaluation did and did not do for Senegal's growth, and what remains to be tackled through other types o freform. 5. The devaluation certainly produced a one-shot improvement in intemational competitiveness, counteracting the effects o f over a decade o f substantial overvaluation. The currency realignment allowed the Government to signal that henceforth macroeconomic balance would be maintained on the basis o f an equilibrium exchange rate, where fiscal and credit policies would not be burdened with compensating for overvaluation. Such balance may have encouraged donors and investors, while investors and consumers were faced with a more efficient set o f prices guiding their import decisions. However, the supply response that could have been prompted, generating a surge inexports, largely failed to materialize. Table 1: Contributionto Growth (inpercent) 1996 1997 1998 1999 2000 2001 1996-2001 GDP 5.1 5.0 5.7 5.0 5.6 5.6 32.0 Private Consumption 2.4 5.8 2.3 0.9 3.6 3.2 18.3 Investment 2.5 0.3 2.7 2.8 -0.5 -0.6 7.3 Private 0.2 -0.1 1.6 1.3 1.3 -1.1 3.1 Public 2.3 0.4 1.2 1.5 -1.7 0.6 4.2 Government Consumption 0.3 -1.1 0.5 1.2 1.9 3.1 5.9 XGNFS -4.6 0.6 0.6 -0.6 0.3 -0.3 -4.1 MGNFS -4.4 0.5 0.3 -0.8 -0.2 -0.1 -4.7 Source :BCEAO Notes: (i) is exports o f goods and non-factor services; MGNFSi s imports o f goods and non-factor XGNFS services. (ii)Contribution to growth i s derived from real growth* share; it indicates what the overall growth o f the economy would be ifgrowth of the other sectors were zero. Source: World Bank calculations based on Senegalesedata. ... 111 6. Compared to other CFA countries, Senegal's exports suffered more from the CFA overvaluation and gained more from the devaluation, but the export response has nevertheless been quite small. Between 1995 and 2000 annual average export growth (in dollar terms) was 3 percent in Senegal and only 0.1 percent for the other CFA countries. Thus Senegal's exports grew more slowly than its GDP and more slowly than average world trade over the period. 7. Not only does the post-devaluation period fail to generate a surge o f export-led growth, there has been little structural change inthe economy as a whole, little structural change, indeed over the last twenty years. The post-devaluation episode actually saw more expansion and employment growth in nontraded sectors such as construction and transportationthaninmanufacturing, which lostjobs. Table 2: SenegalEconomic Structure, 1980/2000 Percent of GDP 1980182 1988190 1988100 Industrial origin Primary sector 19.7 20.7 17.9 of which agriculture (11.0) (10.6) (8.6) Secondary sector 15.4 18.6 20.5 (of whichindustry) (10.8) (13.3) (13.2) Tertiary sector 65.0 60.8 61.7 (of which government) (16.1) (9.9) (8.8) Use of resources Consumption 103.0 92.0 89.2 (of which public) (19.5) (14.2) (11.2) Investment -12.5 13.1 18.7 (of which private) (8.1) (9.1) (11.1) Gross domestic savings -3.0 8.0 10.8 Imports of goods and services 46.9 25.7 38.3 Exports of goods and services 31.7 21.8 30.3 Source: Senegaleseauthorities 8. The Government's GDP data indicate no systematic theme o f growth across sectors. Groundnuts output has dropped in importance compared to earlier decades and suffered especially during drought years inthe 1990s. Harvests rebounded recently with the return o f favorable weather, but global prices have not been robust, competition from Asian producers is strong, and Senegal's groundnuts basin is suffering from resource depletion. Fish and fish processing have also shown a surge o f output and exports with strong and growing demand abroad. However, the fish sector is also suffering from stock depletion and production has been dropping in recent years. The chemicals sector has definitely rebounded since the devaluation, though export growth has not fully recovered its earlier pace. The boom inconstruction and transport, while encouraging, i s difficult to explain except as a reflection o f growth in other sectors, where none have exhibited the kind o f sustained momentum that could lead the whole economy in the new decade. iv Table 3 shows output growth in the industrial sector and Table 4 shows changes in the structure o f employment. 9. The sources o f Senegal's growth are no more apparent when examining patterns of investment and productivity. The evidence examined in chapter 2 indicates that manufacturing productivity has been declining over the long term, which trend has not beenreversed by the devaluation. The chapter concludes that an important contributing factor i s the antiquated plant and equipment which result from the poor performance o f productive investment. Using cross-country regressions, O'Connell and Ndulu find that lack o fphysical capital accumulationi s the most important factor, explaining over halfo f Senegal's growth shortfall compared to the sample mean for the 1960-97 period. Low total factor productivity was the second most important factor for the whole period, however it becomes the most important factor during 1990-97. 10. A study of investment trends by the IFC finds that recent investment growth in Senegal i s respectable, but resembles that o f rich, capital abundant countries, rather than the high rates in emerging and rapidly growing developing countries. The study also finds that most bank credit i s for short-term loans (62 percent vs. 29 percent for medium term and 4 percent long term), that only two banks have sizable shares o f their portfolios inmedium and long term loans, and that foreign direct investment is negligible outside funds paid to acquire privatized utilities. The study includes an in-depth survey o f investments undertaken by a sample o f firms representing 75 percent o f firms in existence for three years. It concludes that most business investmentsare for maintenance andrehabilitationo f existing capacity, that few involve the expansion or creation o f new capacity, andthat those few are small insize. 11. Giventhe lack o f supply response to the devaluation, the fragile sustainability o f primary commodity exports, declining productivity inmanufactures, and low investment growth, the sources o f Senegal's future economic growth are not secure. The next section takes up Senegal's policy environment with the intent o f examining how the recent record o f growth can be sustained and even accelerated, as targeted in the Country's PRSP. MARKET REFORMSOF THE 1990s 12. Senegal began to dismantle its dirigiste and inward-looking policy regime in the late 1980s, starting with the New Industrial Policy (NPI) in 1986. These early reforms aimed at liberalizing international trade restrictions and domestic price controls. However, the policy change was not sustained and many tariff reductions were reversed and import quotas reinstated. A more consistent reform was launched inthe labor market with revisions o f the Labor Code that began in 1987. These reforms included ending the requirement for prior Government approval for layoffs and other measures designed to introduce more wage flexibility. Most significant were the new measures taken in 1989 allowing the use o ftemporary contract workers. V Table 3: RealOutputAverage Annual Growth (inpercent) Sectors 74-80 81-93 94-99 MiscellaneousFood Industries 10.01 -6.06 6.70 Fish-Processing 2.16 3.85 6.22 Oilseeds andproducts -9.55 4.12 -1.50 Grains andFlours -2.11 -1.60 12.57 Sugar and Candy 10.82 3.02 -4.84 Tobacco 0.42 1.73 3.32 Textiles 1.70 12.37 5.34 Apparel 1.09 8.18 5.21 Wood 14.33 7.13 14.23 Paper 2.32 -6.23 7.39 Printing 2.03 -1.50 11.87 Chemicals 5.35 1.58 12.46 ConstructionMaterials 6.18 -2.03 7.67 Metal-working industries 0.49 4.90 6.21 Energy 8.23 -2.95 1.47 Total Industry 0.01 1.79 5.05 8 8 0 W 8 0 ri 0 4 4 8 42e z C P 22 vii 13. In the 1990s, the same policy agenda was re-addressed with even stronger measures. In 1994, in addition to the devaluation, the Government decontrolled domestic prices for goods and services representing % o f sales for those remaining under control, about 12 percent o f GDP at that time. The Government also renewed its efforts at privatizing public enterprises which had peaked in number and value o f Government equity in 1987. Privatizations after 1994 reduced public holdings by about a third o f the value added inthe portfolio at that time. Plannedprivatizations o f Senelec, the national electricity monopoly, Sonacos, with a near monopoly in groundnuts processing, and the Dakar-Bamako railway, will remove most o f the rest o f parastatal value added from public control. 14. One o f the most significant reforms o f the late 1990s was the trade liberalization launched by the agreement in 1997 o f all WAEMU members to implement a Common External Tariff (CET) with streamlined and reduced external tariff rates and reduced nontariff barriers. For Senegal, the implementation of the CET reduced the number o f tariff categories from seven to four, peak tariffs from 70 percent to 42 percent, the average tariff from 36 to 14percent, tariff dispersion from 18 to 7 percent, and eliminated import quotas and licenses. 15. The implementation o f the CET inthe WAEMU region created a larger regional market in several different ways. First, the CET eliminated tariffs on most goods traded among WAEMU members, subject to rules o f origin. Second, the member countries adopted a common set o f business laws in an attempt to reduce the cost o f conducting business region-wide. Third, in order to manage the adjustments needed to maintain a monetary union and a customs union inthe same territory, WAEMU members adopted a set of macroeconomic convergence criteria and related surveillance mechanisms designed to stabilize fiscal and monetary policies throughout the region. Finally, the experience of managing these various means o f deeper regional integration expanded views o f the number o f possible areas where joint regional decisions could enhance the collective outcome. Discussions and agreements have ensued in transport, power, water and environment. 16. Inprinciple, the deepening ofregional integration andthe larger economic space it created should have been especially advantageous to Senegal. With its history as the colonial capital containing industry aimed at a regional scale, its port located favorably close to Europe and the U.S., and its extensive banking sector, Senegal would seem well- placed to take advantage o f a regional market. Nevertheless, these and the other reforms mentioned here have yet to yield any significant increase in investments, trade or traded goods production, or entry o fnew businesses. REFORMWITHOUT RESULTS? 17. In considering how the private sector may be responding to the current policy environment, it is worthwhile to note the observation made in Chapter 3 on regulatory reform and market performance: "The private sector has remained fairly constant insize, its energies focused less on development than on elaborating structures o f cross-holdings V...l l l and maintaining lines o f communication with government, which still controls the development o f key sectors,'' and "It i s widely perceived that important decisions are made inDakar by insiders who have the ear o f important politicians. Dialogue with the private sector was, and still is, often held behind closed doors, in the form o f discussing special favors andprivileges for businesses." 18. Ndiaye, in his study o f Senegal's economic growth performance, similarly describes a 40-year history o f highly concentrated productive resources operating by special concessions from government. This regime led to the dominance o f most major sectors by a small number of private firms, including foreign-owned firms, and public enterprises, during much o f the independent era. The main instruments o f providing favoritism and protection to certain firms were the conventions speciales which restricted entry and domestic competition. The conventions were implemented with various mechanisms including legal monopolies, licensing, guaranteed market shares and long- term concessions. Direct market measures included trade protection through tariffs and quantitative restrictions, domestic price controls, and various tax exemptions and subsidies implementedthrough the investment code. According to Ndiaye, by the mid- 1980s, firms under this regime produced 75 percent o f the value added in agro-industry (excluding fishing) 68 percent o f textiles and leather, and 90 percent of construction materials. Further, Ndiaye points out the low participation o f the Senegal private sector inthe largest firms: Table 5: OwnershipStructureinthe Ten LargestFirms(1991) (percent share-holding) Sector Private Private Public Senegalese Foreign Senegalese SAR Petrolrefining 0.0 90.0 10.0 SONACOS Vegetable Oil 0.0 12.6 87.4 SENELEC Electricity 0.0 0.0 100.0 CSPT Phosphates 0.0 49.0 51.0 ICs Fertilizers 8.2 68.5 23.3 css Sugar Manufacturing 6.0 94.0 0.0 AFRICAMER FishProcessing 50.0 50.0 0.0 SOCOCIM Cement 100.0 0.0 0.0 GMD Flour Mill 2.5 97.5 0.0 MTOA Tobacco 5.0 95.0 0.0 Source: Ndiaye 19. Sonacos the public groundnut monopoly listed in Table 5, i s a case in point. Though hndamentally a competitive sector, Sonacos was granted a monopoly on transport and processing o f groundnuts. While others are now allowed to compete, Sonacos uses its size and government backing to dominate the market, for example offering credit to fanners it does not require be repaid as long as farmer sell to Sonacos. Despite its dominant position in a major Senegalese export product, Sonacos perennially manages to lose money and a recent audit could not certify its accounts. Its one source o f i x profit i s processing imported food oil other than groundnut oil. This i s created by a low import tax on less than fully processed food oil and a high tariff on the final good. In principle any oil processor can exploit the margin created by this tax spread, but Sonacos i s able to use its size, its import buying power and processing scale, to dominate this market as well. 20. The conventions also had the effect o f raising key input prices, such as petroleum, as well as textiles and cement, to other manufacturers. In addition to the obvious inefficiencies and distorted market signals, the regime seems to have had a lasting impact inthe form of low investment-just enough to maintain monopoly profits, discouraged market entry and competition, and a persistent custom o f rent-seeking. The effects have also lingered in the labor market, where employers cite few problems with labor regulations, but substantial problems with labor unions. As Chapter 1 points out, an environment where monopoly or quasi-monopoly rents have been administratively protected i s most conducive to the development o f strong unions who will ensure that workers get their share o f the artificially highvalue added. 21. Given the market instruments used to protect the favored firms under the conventions speciales, the reforms described above o f the 1980s and 90s should have, and to some extent did dismantle the old system. Indeed, if the market liberalizations undertaken inrecent years had not been done, they would cry out to be done now. Still, we need to consider how such outward-oriented, trade and domestic market liberalizing reforms generated so little traded goods growth, so little trade growth, so little poverty reduction, and so little goodjob creation. THEAGENDA THAT REMAINS 22. One possibility i s that the reforms o f the 1990s will eventually have the intended effect; but after a colonial episode followed by 40 years o f preferences established through public policies protecting selected private interests, the private sector needs time to adjust. Hesitancy on the part o f investors may indeedbe a legacy o f the Government's initial approach to undoing the institutionalizedpreferences. As indicated above, reforms indomestic price controls, trade and tax preferences were initiated inthe 1980s and then largely reversed. As Ndiaye puts it: The regime ordained to protect the neocolonial economic order proved unable to sustain existing industrial structures or to create conditions that would have given rise to new ones. Ruling coalitions contain powerful and privileged interests vested in sustaining or undermining the coherence of state policy. This explains why state institutions become so fragmented and permeable in the first place. State power was not being used in any sustained, coherent or concerted way topromoteproductive investment. 23. However, tax and price liberalization in the 1990s, combined with labor market deregulation and exchange rate realignment, have been sustained by Government and X reinforcedby regional agreements among UEMOA members. Ifthe leaders o f emerging businesses begin to perceive that lucrative markets are now more contestable and tolerant o f competition, there may be more market entry, more investment, and greater economic efficiency in the near future. Nevertheless, A business survey in 2002 listed unfair competition as the number one complaint. A note o f caution is also struck inChapter 3 in noting that the private sector still invests much energy in seeking Government accommodation o f their business activities. They further note that the private sector faces and perceives a fair amount o f regulatory risk, i e a possible change o f policy setting that will adversely affect profitability, thus generating all the more incentive to focus business energies on communicationwith the Government, and all the more indication o f lingering government control over key sectors. 24. The other possibility is that the reforms, favorable as they may have been, didnot go far enough, and that further reforms are needed in order to reap the full benefits o f those undertaken so far. Various studies have investigated the remaining obstacles facing business and pointed toward the most urgent reforms needed now. Chapter 2 documents the widespread complaints about poor infrastructure, in particular, the unreliable energy supply. Senegal's energy shortages were well-forecast, and the Government spent some years trying to resolve them through investments by a private partner in the public electricity company, which efforts foundered most recently with the onset o f troubles in the global energy sector. Nevertheless, the energy problem i s well-known, and new supplieswill be online soon. Inadequate transport services have also posed an obstacle to business, and especially to trade, not only due to poor infrastructure, but also due to lack of competition on key routes. Telecom prices had been highuntilbrought down recently by cellular competitionto the fixed line monopoly. The highfixed line prices discouraged some local internet entrepreneurs and would-be services exporters. 25. The issue of regulatory risk is mentioned above, but is also compounded by weaknesses in the legal system. Legal procedures can often take years, and decisions in business cases do not bear a consistent relation to established principles. Thus contract enforcement i s not reliable, in particular the recovery o f collateral from non-paying debtors, except among multinational companies who can be sued in foreign courts. Also mentionedabove are the strong labor unions inthe formal sector, and the citation o f labor relations, rather than labor regulations, as a problem by business managers. What Chapter 2 documents i s the widespread complaint of business, especially inthe informal sector, o f the poor skills and training o f available workers. This concern crosses with those o fthe education sector, discussed below. 26. The administrative barriers facing investors in Senegal have been researched over time by FIAS. The most recent FIAS report notes substantial reductions in the time needed for registering and licensing new business investors. Transactions costs for paying taxes andpassing goods through customs have also beenreduced. However these improvements have been overwhelmed by the multiplying o f time required to acquire access to land. xi 27. The Integrated Framework study, focusing especially on barriers to trade, has recommended that UEMOA members negotiate an agreement that would eliminate remaining barriers to intra-regional trade. This would be an important step to creating a truly regional market, which could help some businesses to reach a large enough scale to become producers o f import competing goods, or o f nontraditional exports. At present, tariff-free crossing o f intra-UEMOA borders requires beingregistered on a positive list o f goods satisfying local rules o f origin. Peace i s maintained among the UEMOA countries by allowing each country to collect its own tariff revenues. Thus imports need to be targeted directly to the country o f destination, or tariffs will be charged anew in every country the imports pass through. Further, those who would like to transform imported inputsinone location and serve the regional market runinto resistance, even ifthey pass the rules o f origin. Downstream countries complain their consumers pay prices that cover import tariffs, while their Government does not benefit from the tariff revenue. Solutions for sharing tariff revenue that have been applied in other African customs unions could be adopted by UEMOA in order to allow for larger scale and more sophisticated operations. HOW DOESSENEGALADJUST? 28. Inconsidering the ways inwhich Senegalese investors may respond to incentives created by the recent reforms, it i s worthwhile to examine the cases o f two major reforms: labor market deregulation and trade liberalization. Essential elements o f the labor deregulation included removing restrictions on layoffs and on the use o f temporary contract workers inthe formal sector. After a long period with such restrictions inforce, the liberalization had the not-too-surprising result o f lowering formal sector employment. However, by many accounts, both formal and informal, the formal sector responded by extending its subcontracting relations with the informal sector. Thus much o f the formal sector work is actually done through subcontracting by the informal sector, where payroll and other taxes can be avoided, and where employment has increased. While the informal jobs pay less and lack security and social protection, the economy may actually be better off. Dependingon the expansion o f the informal sector, the job creation, work experience, absorption o f unskilledlabor, and development o f the informal sector, along with labor market flexibility, could be more valuable than the lost formal sector jobs. However, a thorough survey o fthe dimensions o fthe informal sector, its sectoral and size structure, would be needed inorder to have a complete picture o f the private sector, both in its inclinations to adjust to policy reforms and in its capacity for future economic growth. 29. The impact o f the trade reform, specifically on prices, has not yet beenthoroughly examined. One factor confounding the impact o f tariffreduction was the implementation of a Value Added Tax (VAT) reform inthe same year. The VAT reform was undertaken to conform with UEMOA directives and involved lowering the rates on most products from 20 to 18 percent and raising a few rates from zero or 10 percent to 18 percent. While it i s not obvious that such a reform should lead to price increases, this in fact occurred, and the Government came out to publicly condemn opportunistic price increases. Similarly with the tariff reductions, preliminary indications are that prices xii mostly did not come down on imported goods. Rather, the marginthat used to be taken by trade taxes is probably being absorbed inthe domestic economy at some point inthe distribution chain. Distributors who handle both imported and domestic goods say they have little incentive to lower the price o f imports relative to domestic products. The tendency o f prices to rise under what has so far turned out to be a revenue-neutral tax reform, and for import prices to hold when taxes are reduced, are both indications o f a lack o f competition in the market. Further, Chapter 3 also found prices to be sticky after domestic price decontrols. This suggests that future surveys o f the private sector should make a point o f examining sectoral concentration and segments o f the market that lack competition. THEINFORMAL SECTOR AND FUTUREECONOMICGROWTH 30. When estimating the prospects for sustaining or accelerating economic growth over the medium term, attention usually turns to Senegal's informal sector, and for good reason. The formal sector has proven to be relatively comfortable and rather un-dynamic. As shown above, it has not grown over recent years and its investments andjob creation are limited. The informal sector is the more diverse, dynamic, flexible, job creating, and holds the greatest potential for new growth. As Chapter 3 point out, the informal sector i s one o f the most positive aspects o f the economy and everything should be done to support it and help it expand. 31. Many have asked what the informal sector most needs, and the answers vary. Some say credit, though this does not distinguish Senegal's informal sector since all businesses say they lack credit. Further, an extensive system o f informal credit andmicro finance facilities serves the informal private sector. In this respect it helps to keep in mind the diversity o f the informal sector, as it includes some large scale, sophisticated operations, some engaged in extensive subcontracting with the formal sector. Others are small and less well-connected. The more successful and connected informal businesses find sources for financing. The challenge inthis credit market is like that o f the formal market - finding ways to establish credibility, capacity and willingness to repay loans, and willingness to surrender collateral inlieu o frepayment. 32. Others say the informal sector needs to formalize, or rather the economy needs them to formalize. There are some advantages to formalizing, such as greater access to different credit instruments, to contracts with foreign buyers, to government services, government contracts, and to VAT refunds. Chapter 3 proposes a program that would create a phase-in period for informal businesses to formalize. It would certainly improve equity in the corporate tax system to have more firms paying formal taxes, and would probably allow for a reduction intax rates, more so thanthe current system where only a minority of firms register formally and pay official taxes. It should be clarified that the transition program for informal enterprises i s not meant to apply to those large-scale enterprises, particularly in the importing sector, that evade taxes. These enterprises have no need for any special benefits and should be promptlybrought into the formal tax net. ... XI11 33. Still others say the informal sector needs the means to expand to sufficient scale to allow it to produce more tradable goods. The informal sector already finds and distributes to the local consumers o f a large number o f simple imported manufactures. It could, if sufficiently organized, produce some o f these goods as well. Similarly, foreign buyerso f labor-intensivemanufactures needreliable, large scale supplies, which could be provided through sufficient organization o f informal production. As mentioned above, a greater knowledge o f informal activities, at least in certain key sectors, will help to gage the potential contribution of the informal sector to new investments, competition, and foreign partnerships. It will also help estimate how far the new reform agenda needs to go to produce a greater response inprivate sector growth. GROWTH, POVERTY, AND HUMANDEVELOPMENT 34. Future economic growth is certainly essential for progress in reducing poverty. Nevertheless, Senegal's recent growth episode did not have a commensurate impact on poverty. After the 1994 devaluation, Senegal posted six years o f real economic growth averaging 5 percent. Over the same period, population growth averaged 2.8 percent, allowing for real per capita economic growth o f over 2 percent. Still, preliminary estimates from the 2001-02 household survey indicate a poverty rate o f 54 percent, down only four points from the 58 percent poverty rate measured in the 1994-95 survey, both using a poverty line of consumption at 2400 calories a day. The fact that this sustained episode o f growth generated so little impact on poverty probably indicates a substantial degree o f inequality. Such inequality could be due in part to unequal growth between urbanand rural areas, where the latter have the highest concentration o f the poor. Rural areas are overwhelmingly engaged in agriculture, which sector occasionally posts high growth rates, but usually in rebound from drought episodes which can devastate output. Agriculture narrowly defined, excluding fishing and processing, still engages over half the workforce inproducing only 10percent o f GDP. 35. Not only did rapid economic growth in the past six years fail to reduce poverty significantly, but improvements in health have been elusive. For example, chapter 4 reports an episode o f low economic growth inthe 1980s combined with a sustained effort to improve and redirect resources toward primary health care. This effort resulted in an impressive reduction in child mortality. However, during the 1990s, public health care efforts were fractured and less equally distributed. Thus in the high-growth 1 9 9 0 ~ ~ progress in child mortality rates was halted. The impact o f poorly targeted health services was probably compounded by lack o f economic growth and poverty reduction in the rural areas. The lesson from the health study is that poverty clearly leads to poor health outcomes which can be overcome with properly targeted health interventions, but these interventions must involve the target community in directing and monitoring the allocation o f resources, and they must help overcome the impact o f poverty by contributing to community education ina way that improves health behavior. xiv Figure 1 :Senegal Trends of Under-Five Mortality During the Past Decades USMRper 1000 country to reach the 2015 MDG 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Year 36. Health outcomes are measured in Senegal's PRSP in terms o f progress toward meeting the Millennium Development Goals (MDGs). These include targets on child mortality as well as on maternal mortality and communicable diseases. Yet, as the health study shows, child mortality is not susceptible to public health interventions inthe same way as the other two. Child mortality is largely determined by mother's education level and by water and sanitation standards. Health interventions largely affect child mortality by educating families on better sanitation andhealthbehavior. 37. The other two MDGs are more directly influenced through public health services. Chapter 4 shows the successful impact o f health interventions, even in overcoming low income levels, especially in the 1980s. However, the record o f the 1990s shows increasing problems, and increasing inequity, in delivering health services. The study points out difficulties in decentralization and deconcentration. The responsibilities for financial transfers and accountability were never fully reconciled between health districts established to respond to the Bamako Initiative, and local governments responding to Senegal's 1996 decentralization law. Further, health interventions inthe 1990s were less successfbl inaffecting health behavior. 38. The poor impact on behavior derives from two different dynamics. The first is the excessive proportion o f the health budget devoted to investment as opposed to recurrent spending. Donors, and government officials working closely with donors, have both contributed to the bias in favor o f investment, including construction o f new health xv posts and hospitals. The second part o f the problem i s the poor management o f recurrent spending. Health personnel complain o f poor working conditions, and households complain o f services that are low quality or too expensive. This derives inpart from the failure to maintain and support the involvement o f local communities in assuring the quality o fhealth services, as it hadbeen launched inthe 1980s. 39. Local communities had been organized as part o f the push on primary health care inthe 1980s. Theywere engaged inmonitoring bothpersonnel through absenteeism, and non-personnel spending, such as the availability o f medicine and supplies. Households paid what were initially token fees for medicine as part o f the means to allow users to ensure that medicine would be available. Once the ability to raise local health revenues by these means became apparent, the user fees rose far above the token levels. Reinforcedby the failures o f fiscal decentralization to provide needed health funds, local health facilities began to rely on these fees and to direct their efforts toward the curative activities that would earn them money, and away from the educational, behavior- influencing focus o f primary health care. Visits to health posts began to drop, with the chief complaint being the high cost o f care. Progress on child mortality began backsliding. Finally, the inequality o f the system increased as low-income, low- education regions no longer received adequate central-level transfers and community support that could raise health outcomes above the level predictedby their poverty. 40. This development brings out an unsettled relationship between the roles o f public andprivate health care services in Senegal. Chapter 4 shows that the clearest indicator o f unequal distribution o f health resources i s the distribution o f personnel. The study goes on to point out that most o f the increase inhealthpersonnel over the last decade has been absorbed into private practice, mostly in Dakar. Private practice is much more remunerative than public, and the productivity o f public health workers inDakar (patient visits per health worker) is far lower than elsewhere in the country. At the same time, user fees in public health centers increased 12 fold in the 1990s without any corresponding subsidy to households, while visits dropped due to high expense and health outcomes deteriorated. Numerous cases in other countries have demonstrated efficient and productive arrangements for combining public and private health care services. The results in Senegal indicate that this relationship has not been thoroughly thought through and decided. xvi Figure 2: Senegal Evolution of Technical Health Personnelby Sector and Category 1981 1994 2000 1981 1994 2000 I Nmrsts 1 Midrirts 1 Sources: MSPASKSDApril 2002, WorldBank, 1982. EDUCATION INEQUALITY AND 41 Education i s perhaps the most important factor in determining whether poor people can participate in economic growth. Education can also be a spur to economic growth. And, chapter 4 shows the education o f girls inparticular can have far-reaching effects on health and other dimensions o f human welfare. For these reasons, the Senegal PRSP gives particular emphasis to the fulfillment o f various education goals, the best- known o fwhich i s that o f universalprimary education by 2015. 42. Incomparing nationalindicators across countries, Senegal's progress ineducation is similar in certain respects to the results in health. Enrollment rates have shown impressive growth, but still education outcomes are disappointing for a country with Senegal's income. Test scores inmath and French are among the lowest inFrancophone Africa. Rates o f grade repetition are high and rates o f survival to higher grades i s low. The outcomes are all the more disappointing given the amount o f resources devoted to the sector. Education spending amounts to 32 percent o f recurrent spending, and Senegal's teachers are among the highest paid in Africa. The combination o f high spending and poor outcomes clearly indicates the inefficiency o f the system, both in terms o fthe allocation o f public resources and interms o f their use inbudgetedactivities. Chapter 5 estimates that the waste o f public resources due to highdrop out and repetition rates amounts to 40 percent o fthe entire recurrent education budget for 2000. xvii Table 6: Cohort Survival Rates, Grade Enrollment Rates, and Transition Rates Between Cycle, 2001 Survival Rates Grade Specific enrollment Transition rates Rates for a Cohort Grades Boys Girls Boys Girls Boys Girls 1 1.00 1.oo 0.85 0.84 2 0.87 0.85 0.74 0.72 3 0.71 0.69 0.60 0.58 Primary 4 0.57 0.55 0.49 0.46 5 0.47 0.44 0.40 0.37 6 0.34 0.30 0.29 0.25 7 0.12 0.09 0.10 0.08 Lower 8 0.10 0.08 0.09 0.07 35% 31% Secondary 9 0.08 0.06 0.07 0.05 10 0.06 0.05 0.05 0.04 11 0.03 0.02 0.02 0.02 Upper 12 0.02 0.02 0.02 0.01 49% 49% S&ndary 13 0.01 0.01 0.01 0.01 School Life Expectancy (implied) 3.7 3.5 Source: Based on data in Statistiques scolaires et universitaires, various issues, MOE, 2001, authors' calculation. 43. Senegal spends about 40 percent o f its public education budget on primary education. Of that, 95 percent i s spent on salaries; 89 percent for teachers and 6 percent for non-teaching personnel. Whether it is poor teacher performance, high administrative expenses, or lack o fresources for any teaching materialsbesides the teacher, the outcome from this spending i s abysmal. Primary completion rates are only 30 percent for boys and 25 percent for girls. Completion o f primary school i s considered essential for establishing sustainable literacy skills, and Senegal's literacy rates remain disappointingly low. A further concern arises because the official language o f the country is French, yet children only learn Frenchinelementary school. Thus government documents, legal system, business regulations, and export quality standards are accessible only to those who are literate inFrench, usually requiringprimary school completion. 44. In education, the active private segment of the sector seems to have a less disruptive effect on the public segment. Indeed it would be useful to have more expenditure data from private schools in order to compare unit costs with public schools, as well as the share o f wages versus non-wage expenses compared to public schools. This would helpidentifywhere inefficiencies are greatest inthe public system. 45. University education absorbs a hefty 26.5 percent o f total public education spending. This share covers both education expenses and scholarships for university students. The inefficiency o f spending so much public resources on university education where the benefits are overwhelmingly private, especially at the expense o f more effective primary education, is apparent even to the Government. Nevertheless, the xviii Government has yet to find a way around the obstacles presented by those with a vested interest inthe current system: teachers unions, administrators, and universitystudents. 46. Inequityin the distribution o f public education resources is apparent not only in the share allocated to Universities, but also inthe low allocations to rural areas, and inthe limited access to primary and secondary education in the rural areas. Table xx below shows the incidence o fpublic education subsidies across income groups, both within and across income levels. Table 7: Distributionofper capita subsidyby educationlevel and income(CFA), 2000 Income quartiles Primary Secondary1 Higher Total First quartile 1.0 1.0 1.0 1.0 Second quartile 1.5 2.8 1.7 1.8 Third quartile 1.6 3.3 4.3 2.4 Fourth quartile 1.7 4.0 7.3 3.1 Source: Ministryo f Educationand HouseholdSurvey data. `Includes lower and upper Secondary, and secondary technicalivocational. 47. It should be noted, however, that substantial inequity derives from the ineffectiveness o fresources allocated in generating education outcomes. Ifthe pattern o f survival rate and intake rate persists over time, future men and women would have, on average, only 3.7 and 3.5 years o f formal schooling as a part o f their human capital profile. These outcomes are lower than the average o f 5.6 years in Africa. This ineffectiveness diminishes the capacity of education to promote greater equity in the economy. 48. The impact o f poor education performance on the economy is substantial. The effects o f low education attainments on health i s noted above. Further, the system i s failing to build human capital at the rate required by the economy The informal sector in particular is hindered by the low skill level o f available workers. Senegal may not register as having especially highwages in the region, but the complaint o f employers i s that labor skills are too low for the amount o fwages they endup having to pay. SOCIAL PROTECTIONFORVULNERABLE GROUPS 49. Senegal's PRSP accords great importance to social sector programs inthe poverty fighting strategy. Nevertheless, the document also notes the slow progress inthe past and furthernotes the schedule o fyears that will likely beneeded to reachthe PRSP goals. As a result, the PRSP acknowledges the need to take measures that will help the poor and provide social protection for vulnerable groups inthe near term. xix 50. The government uses two main categories o f instruments for improving the living conditions o f vulnerable groups: direct subsidies or services and the structure o f taxes it uses to fund PRSP programs. Chapters 6 and 7 consider specific examples from both categories - indirect taxes and pension subsidies. 51. In poor countries with large informal sectors and poorly functioning tax administrations, indirect taxation, i.e. the taxation o f goods and services, is not only the main source o f revenue for the government but also the fiscal instrument with the greatest impact on individuals' welfare. Indirect taxation in developing countries accounts on average for two thirds o f the government's fiscal revenue. Inaddition to raising revenue, these taxes are also used to protect producers, whether in the form o f excise taxes or import tariffs. 52. One set o fproducts, mostly processed food items, has receivedvarying degrees o f protection through the variety o f recent fiscal regimes, including indirect taxes, price controls, tariffs and import controls. These products are: rice, meat, sugar, milk, wheat flour, refined oil, tomato paste and cigarettes. Evidence from the household survey on the structure o f consumption indicates that many o f these products are important in the budgets o f the poorest households, though none more so than rice, whereas meat is relatively more important for the non-poor. The analysis in chapter 6 shows that taxes could be redistributed across this set o f products in a way that substantially reduces the tax burden on the poor, while keepingthe total tax revenues constant. 53. Tuning to subsidies, the government has undertaken reforms o f the pension system to forestall pension deficits that could crowd out other PRSP programs, and is considering further reforms to reverse the erosion inpensionvalues due to low returns on investments. Pension reform i s o f great importance to workers, mainly, but not only for those in the formal sector. Pension deficits have become a growing liability on public resources as pension managers struggle to cope with demographic and employment changes. Restrictions on the growth o f government employment and reductions informal sector employment, combined with the maturation o f pension schemes set up in the 1960s, have lead to a rising ratio o f pensioners to contributors. The Government i s obliged to cover deficits o f the pension system with fiscal resources. A series o f parametric reforms inthe pension formulas have reduced this liability for the next several years, but it will eventually resurface. 54. To assess the role o f pension reforms inpoverty reduction, the Government needs to consider the contribution o f pensions to overall development goals. The first issue i s to avoid pension deficits that would compete with other social sector programs for public resources. The second issue i s to restore credibility o f the system which has eroded due to low or negative returns on pensions savings. The third issue concerns extendingsocial insurance to more workers and expanding the size o f the formal sector. The final issue is making better use o f pensions as a source o f long term savings for investing in the economy. xx TOOLS FOR PRSPIMPLEMENTATION: TRANSPARENCY AND INFORMATIONSHARING 55. Insum, discussion of the various issues above indicates that reachingPRSP goals could require a number o f politically difficult reforms. The historical culture o f rent seeking has seeped from production sectors into the social sector programs, and will prove difficult to root out o f the economy. Difficult reforms -- such as resource management in agriculture and fishing, changes intax and subsidy programs, changes in public expenditure management systems - will take time and require community involvement inorder to build sufficient ownership. However, one type o f reform should pose fewer difficulties and aid the process o f building ownership on the rest: Increasing transparency and information sharing. 56. The government has undertaken several initiatives in these areas. First, it launched expenditure tracking surveys in health and education, the former being the subject o f chapter 8. Expenditure tracking surveys help identify delays and leakages in the flow o f public funds, especially flows intended for the local level. Second, the government signaled in the PRSP the importance o f monitoring and evaluation, thus collecting and sharing data on progress toward PRSP goals. The government i s now working with a group o f donors to improve relevant data collection and network data sharing. Finally, in the PRSP, the government lists budget systems reform as key to PRSP implementation. The government has repeatedly emphasized that it must make clear to its citizens how public resources are beingusedand toward what objectives. The final chapter gives the government's views on the motivation and direction for budget system reforms. MONITORING IMPLEMENTATION PRSP 57. After extensive consultations with civil society and regional representatives, the Government has drafted and adopted documents laying out its approach to implementing and monitoring progress on the PRSP. The Cellule de Suivi du Programme de Lutte contre la Pauvrete (CSPLP), along with the Division de la Prevision et de la Statistique (DPS) and sectoral statistical services, i s charged with developing the poverty profile that will provide baseline poverty indicators, and these services will survey and maintain the statistics used to measure poverty outcomes. 58. Inkeepingwith its PRSP, the Governmentneeds to develop statistical systems for monitoring and evaluating progress in economic growth and poverty reduction. These systems should ensure timely, transparent reports, and consistent methodology. Previous assessments o f Senegal's statistical systems have noted lack o f coordination in data collection, poor survey methods, substantial delays in the publication o f some statistics, and lack o fresources. 59. Evenso, the Government will need to implement institutional reforms that go far beyond improvements in the statistical infrastructure. Given the increased emphasis on monitoring and evaluation cultivated in the context of budget reforms, the Statistics Directorate and the CSPLP will need to build bridges with sectoral units that will be xxi subject to statistical evaluation. Further, the Directorate and the Cellule will need to develop a plan to bring the results o f poverty monitoring to the attention o f decision makers so they can be used to improve the design or implementation o f ongoing programs. BUDGETSYSTEMREFORMS 60. Government staff inthe social sector ministries have been working with donors to define and implement sectoral strategies. These strategies, aimed at progress toward meeting the MDGs, are expressed in Senegal's PRSP. Nevertheless, an important segment o f problems blocking progress in human development cannot be solved within the social sectors alone. Inequitable allocations o f public resources, inefficient use o f budgetedallocations, poor tracking o f spending, no evaluation o f its impact, confused or dysfunctional systems for decentralized financial transfers and accountability are all obstacles to effective delivery o f social services. Reducing these obstacles requires collaboration with the Finance Ministry to reform basic budgeting systems, which reform will require cooperation with all the line ministries. 61. Reforms o f budget systems are typically categorized according to the episodes o f a typical budget: budget preparation, budget execution, and monitoring and evaluation. The last category includes both financial monitoring, such as audits, and qualitative or physical evaluation to assess the impact o f spending on intended outcomes. While Senegal has problems in all three areas, probably the greatest weaknesses are in budget execution and reporting, and the reform agenda should be lead by initiatives in these areas. A well-formulated budget will have limited impact if it does not connect to a system that will ensure that the written budget is in fact executed. Lengthy and cumbersome budget execution procedures are the source o f constant complaints from line ministries, local spending authorities, and donors. Budget execution procedures are probably also the stage most afflicted with governance problems. The heavy reliance on manual accounting procedures and highly fragmented computing systems allows for much private, individual control o f information. N o single information system connects the execution o f any budget line from beginning to end. Thus elaborate efforts are needed to produce budget reports o f a type that would allow the monitoring o f progress on a given program, as is done for health and education, but infrequently and only at the behest o f donors. 62. The lack o ftimely, within-year reportingon the execution o fthe budget is a major shortcoming o f the Government accounting and reporting systems. A select group inthe Finance Ministry discusses spending reports on a monthly, even weekly basis. However, the format o f these spending reports i s not clear because they are not more broadly distributed. Thus timely budget monitoring is conducted among insiders and experts ina way not conducive to any independent examination. The budget reconciliation law has not been validated by the Cour des Comptes since 1997because the draft laws were not presented in a form that met the required standards. Further, given the patchy systems o f record-keeping, it i s difficult to verify that final accounts really reflect actual spending. xxii 63. Since 1994 Senegal has consistently maintained fiscal balance as defined by targets set within IMF programs. Further surveillance on fiscal targets arises from UEMOA macroeconomic convergence criteria. Thus budget management in Senegal does not suffer from the inability to control total spending. Rather, Senegal's success at staying within fiscal targets despite some huge expenditure shocks indicates a tendency toward ad hoc adjustments in programmed spending. The details o f those adjustments may not be specified untilmonths or years later. InDecember 2001 the Government was able to spare 3 percent of GDP for unplanned transfers to the public electricity and groundnut companies and still produce fiscal a deficit that was smaller than targeted, such was the under-execution o f programmed spending. Investment projects are chronically under-executed, averaging a rate o f 68 percent from 1996-2000; however, with a combined recurrent budget share of 40 percent for health and education, these sectors are boundto bear some o fthe burdenfrom ad hoc adjustments. 64. So, statements above regarding the inequitable distribution o f resources in the health and education budgetsnotwithstanding, the problem o f poor budget allocations are probably second in priority behind making sure that written and voted budgets are executed and properly documented in a way that allows for timely, independent examination. The most stark issue in budget allocation, the share o f education resources devoted to university education, is largely political and does not result from lack o f understanding about the low social returnon higher education. Issues o f poor alignment betweenpublic resource allocation and public objectives need to be addressed, but best in tandem with efforts to improve the efficiency o f spending what has been allocated. 65. One immediate issue in budget preparation is the format in which the budget is presented. The current budgetary nomenclature presents the spending o f departments rather than programs. Each department has a separate chapter presenting lengthy and detailed administrative line items. The total package i s voluminous, and while copies are delivered to members o f parliament and other officials, the physical size o f the budget makes it difficult to distribute and to comprehend. Sectoral experts can largely follow their own chapters and can even infer the intended program or objective o f most o ftheir sector's line items. However, a working understanding o fthe whole budget, a sense o f its main objectives and the trade-offs across sectors, i s difficult to attain. The Government has committed to present Parliament with an orientation note that explains the main outlines o f the budget, but it had not done so in the previous two years. If the budget were publishedin electronic format, it would be easier to organize the raw information into different aggregations that make the purpose o fpublic spending clear. 66. The main reasons for presenting different formats o f the budget are to make it more executable, more monitorable, and better suited to commitments in fiscal decentralization. Efforts to make the objectives o f public spending clear need to be matched with efforts to monitor and evaluate the impact o f spending on achieving those objectives. Similarly, budget monitoring, especially at decentralized levels, cannot be accomplished without some specification o fwhat spending was supposed to achieve. xxiii 67. The social sector MDGs present a number o f national level objectives, such as literacy rates, that can only be implemented through locally delivered services. Senegal needs to resolve fiscal accountability issues at the local level; it needs to forge an agreement with local spending authorities on their role in meeting national level objectives, organize its budget according to programs that support those objectives, rationalize the share o f these program budgets that need to be transferred to local levels, and monitor the effectiveness o f local spending according to the agreed objectives. CONCLUSIONAND RECOMMENDATIONS 68. Since the 1994 devaluation, Senegal has completed its first sustained episode o f real economic growth, averaging over 5 percent. While economic growth is clearly essential for poverty reduction, the impact o f this six-year growth period on poverty has been disappointing. The limited effects on poverty are due partly to the nature o f the growth and partly to the nature o f institutions intended to ensure that this growth contributes to poverty reduction and human development. Senegal's PRSP is devoted to improving this recordover the next several years. 69. The recent growth episode received important impetus from two primary sectors that are already suffering from natural resource constraints - fishing and agriculture. Prospects for higher growth inmanufactures are clouded by the sector's limited response to substantial devaluation, trade liberalization and market liberalization. The poor performance o f traded goods inthe wake o f such outward oriented reforms is a clear sign of lack o f competitiveness, and the evidence examined in chapter 2 shows that declining competitiveness derives from low productivity, rather than high wages. The evidence further indicates problems with both labor productivity and total productivity, where the former suffers from low investment in physical capital and the latter suffers from poor human capital andpoor infrastructure. Poor infrastructure also limits growth in services, though services growth has surprisingly surpassed manufactures in the wake o f the devaluation. 70. There are manybright spots inthis growth record, and a great deal o fpotential for the future. Fish and horticulture show great promise ifnatural resource constraints can be managed; chemicals and textiles also show potential for export growth, as do telecommunications and other internet-based services. Annex xx includes several sectoral studies with specific recommendations for promoting growth in businesses in primary, secondary andtertiary activities. 71. Still, when posing the question o f how six years o f growth under stable macroeconomic conditions made so little impact on poverty, one must consider the unequal distribution o f gains from growth. The Government estimates the nationwide Gini coefficient to be 0.3, but reports a figure o f 0.5 for Dakar in 1996. Other estimates o f the country's Gini coefficient range from 0.5 to 0.6. This may be due inpart to higher economic growth inurban areas rather than rural, where most o f the poor live. If so, the Government needs to plan for more rural-urban migration and intensify efforts to provide primary education and literacy skills to rural populations. xxiv 72. Indeed, the inefficiency o f public education services, with low outcomes despite ample spending, is a major obstacle to improving growth and equity. The education system is not producing labor skills at the rate the economy needs, and it is not sufficiently spreading the benefits o f higher human capital, including being fit to earn higher wages, to disfavored groups o f rural and poor citizens. Neither education nor health institutions have succeeded in directing resources to primary services and to the poorest regions to the extent needed to make a bigger impact on poverty. This i s partly due to inadequate implementation o f programs as they have been designed, and partly due to central budgeting systems that hinder transfers to localities and that lack the kind of monitoring and evaluation system which would allow and require central and local authorities alike to account for funds used. 73. A group o f Senegalese researchers [CREA 20021 studying the factors affecting Senegal's economic growth has raised a similar list o f concerns. Using an econometric growth model estimated with data from 1960 to 1996, they find physical investment and human capital to have significant positive effects on growth, along with macroeconomic stability and, most importantly, institutional and structural reforms. Yet they concede that all the factors they include leave a large share o fthe variation ingrowth unexplained. They emphasize the importance o f several institutional factors, including the seeking and preservingo frents andmonopolization o fkey sectors: The consumer organizations also claim that reforms intended to promote competition through the rule of supply and demand that should govern the market, gave way, in fact, to a great opacity in business transactions. Further, as the culture of seeking and preserving rents is a fundamental aspect of Senegal's political economy, the dismantling of positions of monopoly and privilege is a precondition for the application of reforms needed to liberate the energies that growth requires. 74. Clearly this i s another piece o f the puzzle behind the unequal distribution o f gains from economic growth. Since this condition is apparent in rigid prices following major liberalizations, inownership concentration, in labor union behavior, inthe legacy o f past policy regimes, in the complaints o f consumer groups and business surveys, and in growth studies conducted by Senegalese researchers, the Government should take account o f the problem and consider ways to address it. 75. The Government can send a clear message to its citizens that the purpose o f structural reforms i s to make the market open to entry, new investment and competition. It can further enforce liberalization in those sectors where business leaders and consumers perceive the Government i s still protectingmonopoly interests. 76. The Government can launch a program to promote and aid the informal sector, taking account o f specific recommendations in chapter 3. It is worth noting that two xxv sectors widely perceived as having broad-based participation or open entry - fishing and agriculture - are facing natural resource constraints. The boom in nontraded goods - retail and construction services - i s encouraging and helps absorb marginal workers, but the country is unlikely to hit its growth targets if led only by nontraded services. Clearly a key area to focus support for the informal sector i s in production o f traded goods, specifically in business services that can help informal producers reach the scale needed to connect with better supply systems, foreign buyers, and to compete with existing operators. 77. The Government can work with its WAEMU partners to eliminate internal trade bamers and create a larger economic space, both for domestic and foreign investors. Such regional agreements can help increase competition and create opportunities to increase scale for a large number o fbusinesses. 78. Further, in the spirit of expanding opportunities for new businesses, the Government can be diligent in implementing an open and transparent procurement system for public contracts. Opportunities to qualify for government contracts need to be spread more broadly, with the added benefit that an open public procurement system may encourage more informal businesses to formalize. 79. The Government can help reduce the cost o f credit by promoting institutions that improve signaling in the credit market, that is, the certification o f ability and willingness to repay loans. Institutions of this type exist and are discussed inchapter 3. Inaddition, the Government should consider proposals to improve the legal enforcement o f contracts, including lending contracts. 80. Finally, the Government should undertake every effort to raise the efficiency o f education systems in order to increase the skill level o f the labor force. Absorbing one third o f current spending, education programs are a major part o f what the Government does to fight poverty. Education is a major factor driving improvements in health behavior and health outcomes. Education i s one o f the most potent tools the Government has to help equalize opportunities for the poor and redistribute gains from economic growth. Further, the labor skills are desperately needed by the private sector, especially the informal sector, which provides the greatest potential for new competition. 81. Nevertheless, given the legacy o f concentrating control o f key sectors, the traditions o f seeking and preserving rents are unlikely to disappear in short order. The Government needs to consider how these circumstances should affect policy approaches inseveral important areas. 82. The first area o f concern i s the privatization o f public monopolies and regulation of the newly privatized entities. Worldwide, such privatization has improved management, investment and customer services, but the allocation of rents between consumers and producers needs carefbl attention. The Government should absorb every lesson from the failed privatization o f the electricity company and use that guidance to improve the contribution o f private participation in electricity supply to the development xxvi o f the economy and the welfare o f its citizens. The privatization o f the telephone monopoly was successful, but in the absence o f an independent regulator, prices remained highuntilcompetition from new providers brought them down. 83. The privatization o f the public groundnut company i s now imminent, and the Government must decide on how to structure the assets it sells. Since the groundnut sector i s not a natural monopoly, Sonacos should be dissolved and the different oil processing plants sold separately to allow new competitors to operate inthe sector. Ifthe Government sells Sonacos as a single entity, it will create one private company that dominates the market the way the public company did. 84. Second, when addressing natural resource management issues such as fish stocks or the groundnut basin, the Government needs to assess the options available to the poorest participants inthese sectors. The openness o f the fishing sector to all comers and the low capital requirement for entry have been a source of some pride in Senegal. If access to natural resources is to be restricted temporarily as part o f a long term resource management plan, then greater opportunities need to be created for the most vulnerable workers, and perhaps public subsidies as well. 85. Finally, the Government needs to avoid letting public human development services become another province for rent-seeking. University students have already captured a fair portion o f the education budget. Those trained inmedicine are collecting highrents inprivate practice. Public health funds have driftedto construction crews and Dakar-based administrators, while local health posts in poor areas are left to fend for themselves with locally-generated user fees. 86. The underlying budget systems need to be revamped including both central and decentralized levels, while taking full account o f incentives influencing the behavior of civil servants in both the finance ministry and in social sector ministries. The Government should promote a new policy o f transparency and broadly distribute information on public spending and its outcomes as a means to reduce rent-seeking in public services. 87. The Government needs to renew and intensifyefforts incommunity monitoring o f local service delivery and quality. This monitoring should be integrated into budget reporting systems and PRSP implementation. Indeed the advent o f the PRSP and its implementation provide an ideal opportunity for the Government to gather and focus its efforts to speed progress on the goals chosen by its citizens to include in the national poverty reduction strategy. SECTIONI:WEALTH CREATION Between 1995 and 2001 Senegal's economy is estimated to have grown by over 5 percent per year. This represents strong growth performance both by Senegalese and regional standards. But despite this growth, Senegal has made only marginal progress inreducing poverty. This section poses the central question o f why six years real economic growth generated little progress on poverty reduction and examines the nature o f Senegal's production and economic reforms for insights on how to improve the future impact o f growth on poverty. 1. GROWTH, POVERTYAND THE LABOR MARKET A. INTRODUCTION 1.1 Senegal's 2002 PRSP reports that 57.9 percent o f Senegalese were poor in 1994 and this rate fell by only four percentage points, to 53.9 percent o f the population, in 2001. This growthlpoverty outcome i s both disappointing and surprising. It is disappointing because an increase in per capita incomes o f roughly twenty percent i s expected to reduce poverty by more than four percentage points. It is surprisingbecause devaluation and other market-based reforms are expected to reward a nation's abundant factor, which in Senegal includes unskilled labor. Senegal also i s reportednot to have an especially high level o f inequality - the PRSP reports a Gini o f only .30, other sources indicate a higher level o f .41. If such low-to-moderate income inequality has persisted, all income quintiles should have benefited appreciably from reported growth rates and poverty should have declined more rapidly. 1.2 There are three possible explanations for reported trends in Senegal's record o f strong growth andlimited poverty alleviation: (i) estimates o f GDP growth may be biased upward; (ii)estimates of poverty reduction may be biased downward; or (iii)the estimates are reasonably correct, but Senegal's growth path has not beensufficiently pro- poor. 1.3 This chapter examines these explanations. It attempts to confirm reported trends by investigating complementary data, especially on labor outcomes. We find that the low reduction inpoverty rates is consistent with labor market trends. Since 1994, Senegal has experienced an increase in reported unemployment. In addition, the labor force has not witnessed a significant movement o f workers from low productivity activity in agriculture to higher productivity work in industry and "modern" services. Instead, Senegal has continued to experience a "de-industrialization" o f its labor force with an increasing percentage o f workers engaged in more informal arrangements in commerce and other services. Even in agriculture the shift has been toward increasing numbers o f unpaid family workers. These are hardly signs o f declining rural or urban poverty. 2 1.4 In seeking to account for these trends, this paper focuses on labor policies. Despite long-held views to the contrary, we do not find evidence that labor regulations are a primary factor inthe slow growth o fmodem sector employment or, more generally, inthe disappointingpoverty outcomes reportedinthe PRSP. 1.5 The chapter proceeds as follows. The next two sections examine reported trends in GDP and poverty rates. Senegal presents a remarkably consistent trend in GDP growth that surpasses its record o f the preceding ten years and compares favorably with outcomes elsewhere in the CFA zone. A comparison o f the poverty rates reported in the PRSP with those o f earlier studies reveals many inconsistencies. This i s troubling for proper evaluation o f trends in poverty over time. But because the selection o f any poverty line i s somewhat arbitrary, what is critical i s less the comparison with other studies than a consistent approach applied to the 1994 and 2001 data used in the PRSP. Section IV evaluates Senegal's experience in poverty reduction by employing cross- country estimates o f the elasticity o f the poverty rate with respect to changes inper capita income. These crude benchmarks suggest that Senegal achieved, perhaps, as little as one- percent or more. third to one-half1o f the poverty reduction expected from six years o f growth o f five 1.6 Turning to explanations for reported outcomes, Section V considers "the stagnation o f agriculture" as a reason for the persistence o f rural poverty. While agriculture, especially the groundnut sector, showed little net gain between 1995 and 1999, since then it i s reported to have grown rapidly, but still was unable to raise many households above the poverty threshold. Section VI examines labor outcomes and identifies changes in unemployment, sector o f employment and employment status (whether a person is an employee, casual, unpaid or self-employed worker). As already noted, these changes are consistent with the persistence o f poverty despite strong GDP growth. Section VI1 argues that labor regulations are not to blame for the weakness in the industrial transformation o f the economy. These results are supported by evidence from firm surveys in Senegal and from similar surveys and academic research elsewhere in Africa and in other regions. Section VI11 concludes by evaluating the specific employment policies detailed in the PRSP. We conclude that the agenda o f active labor marketpolicies, with its emphasis on labor market information, will have little impact on the unemployment and employment transformation problems confronting Senegal. B. GDPGROWTH 1.7 Since 1995, the year following devaluation, estimates o f Senegal's GDP indicate that output has grown by 5.1 to 5.6 percent every year. This is a significant improvement compared to Senegal's recent past. It also indicates strong and consistent performance as compared to other economies bothinthe CFA zone and throughout Sub-Saharan Africa. These estimates are intended as a rough approximation only. Once detailed information from the 2001 ESAM I1becomes available, more refined estimates on expected poverty reduction should be computed. This will require data on the growth in mean household incomes or expenditures, the level of income inequality, and the poverty gap. A more refined estimate may reveal somewhat better performance in poverty reduction than reported here. 3 1.8 Figure 1 charts Senegal's annual increase in real GDP from 1985 through 2001. The period prior to 1994 indicates a high degree o f variability in growth rates. GDP contracted almost as often as it grew, and inno year did growth exceed the rates realized since 1995. In the post-devaluationperiod, output growth has been both higher than in the past and a picture o f consistency. As a result, real GDP by 2001 i s estimated at almost 50 percent larger than it was in 1993, the year prior to devaluation. 1.9 Relatively rapid and consistent growth also distinguishes Senegal from many o f its neighbors. Table 1 compares rates o f real growth for the 14 member nations o f the CFA zone. Despite pursuing the same exchange rate policy, some o f the CFA zone economies have fared better than others. Senegal has had one o f the better growth records and only a few o f these economies (Benin, Cameroon, and Mali) have managed this growth with relatively little variability from one year to the next. Incorporating estimated population growth reveals that about half o f the CFA group has seen per capita incomes stagnate or even contract. Senegal belongs to the other, better performing half, which has experienced per capita income growth o f at least 2 percent per annum. 1.10 At an estimated average annual growthrate inGDP o fper capita o f 2.24 percent, Senegal still lags some o f the strongest performers inSub-Saharan Africa. The discovery o f oil generated extraordinary growth in Equatorial Guinea, and Mozambique and Uganda both grew considerably faster. But since 1994 the number o f African economies that grew more slowly than did Senegal comprises a much longer list. Senegal begins this decade as one o fthe better growth performers inthe region. 1.11 Despite strong growth, poverty alleviation in Senegal has proved difficult to achieve. Employing a poverty line set at daily consumption o f 2400 calories per adult equivalent, Senegal's PRSP reports poverty rates falling only from 57.9 percent in 1994 to 53.9 percent in2001. Although per capita income has expanded by an estimated one- fifth since 1994, more than half o f the population still falls below the national poverty line. Even more telling i s that economic growth has only produced a four-percentage point decline in poverty. This raises questions about "the quality" o f Senegal's growth andthe distribution o fits benefits. C. MEASURING POVERTY 1.12 Before trying to understand why more rapid rates o f poverty reduction have not accompanied Senegal's recent economic growth, it is worth reviewing both available estimates o fpoverty and reasonable expectations for poverty reduction. The results from the PRSP are based on householdbudget/consumption surveys taken in 1994 and 2001. These estimates can be compared to those reported in the World Bank's Global Poverty Monitoring web site and in Senegal: An Assessment of Living Conditions, a detailed 4 World Bank poverty assessment released in 1995 but refemng to the poverty situation in 1992. 1.13 Table 1-2 assembles various estimates o f Senegal's poverty rates between 1991- 2001. The Headcount Index reveals wide variation in the incidence o f poverty, from a low o f 26.3% in 1994 to a higho f 67.8% in 1995, only one year later! Because o f a lack o f documentation o f how different estimates were derived, we can only try to infer why there is so muchvariance. 1.14 Three basic definitions o f poverty are employed by the various studies. The World Bank Global Poverty Monitoring website estimates use both a $l/day and $2/day poverty line. Since the dollar refers to a 1985 dollar, adjusting for inflation results inthe monthly poverty lines o f $32.74 and $65.48, respectively. The alternative approach i s to employ a measure o f 2400 calories per adult equivalent per day and to price the corresponding cost o f consuming the required food bundle. This approach yields significantly lower nominal values for the CFAF poverty line than does the $1 or $2 standard and, hence, lowers rates o fpoverty. 1.15 The estimates reported in the World Bank web site are so difficult to reconcile with one another that they may simply be misreported. The headcount index using $l/day ($2/day) falls from 45.4 (72.3) percent in 1991 to 26.3 (38.4) percent in 1994. But 1991-94 was a period of stagnation in the economy and it is hard to imagine why there would be dramatically fewer poor in 1994 than in 1991. The $l/day ($2/day) poverty lines in 1994, measured in nominal CFAF values, all have effectively doubled, consistent with the official devaluation. But this suggests there should have been an increase, not a decrease, inreported poverty as long as consumer prices rose by less than the change inthe exchange rate. The World Bank's 1995 poverty rate based on a $2/day measure i s equally anomalous. 1995 was a year o f relative prosperity and poverty should have fallen, not increased by almost 30 percentage points from one year earlier. Domestic prices confuse the picture even more as the corresponding CFAF poverty line in 1995 is lower, not higher, than in 1994. In sum, these reported poverty rates do not inspire confidence in capturing the true picture o f poverty in Senegal over the past decade. 1.16 Comparing poverty estimates based on calories also reveals considerable fluctuations over time. The higher CFAF price o f securing 2400 calories per day in 1994 and 2001, as compared to 1992 (compare lines 3,6 and 8 o f Table 1-2), could account for The web site addressis: ht~,:lJwww.worldbank.orcr/research/~ovmonitor/coun~details/Sene~al.htm ; "Senegal: An Assessment ofLivingConditions," WorldBankReport#12517-SE(May 1995). 5 some o f the higherpoverty rates reported inthe PRSP. Some o f this increase i s the result o f price inflation-the consumer price index increased by 30 percent from 1992 to 1994 and by 60 percent between 1992 and 2001. But the price index is based on urban consumptionbaskets and i s a biased deflator for most o f the poor, who live inrural areas. Giventhese urban price data, even if a better price index were available it is unlikely it could account for a three-fold increase in the CFAF value o f a fixed poverty line. If measurement o f the 2400-calorie poverty line has been held constant in real terms, then growth since 1992 was immiserizing, an unlikely outcome. 1.17 Alternatively, there may be some methodological differences in the calculations of the two calorie based poverty lines. One o f these differences might relate to the treatment o f "auto-consumption", that is, consumption o f own crops. The 1992 results are from harvest time, "the most prosperous time o f the year", and assume that a percentage o f necessary calories stem from auto-consumption. During "the hungry season", when such consumption can fall significantly, poverty rates, according to the report, could jump from 33 to 60 percent. The upper bound o f this range i s at least consistent with the rates reported for later years in the PRSP and it may be that similar auto-consumption adjustments were not made inthe PRSP calculations. 1.18 Because o f the importance o f understanding the impact o f economic growth on poverty, it would be helpful if a more consistent series o f poverty estimates were available. At a minimum, existing estimates should be sufficiently documented to permit analysts and policy-makers to compare them. Since the selection o f any poverty line i s somewhat arbitrary, consistency i s more important than the actual level chosen. Assumingthe PRSP employed a consistent approach inits calculation ofpoverty rates for 1994 and 2001, poverty did fall. The next question i s whether or not this decline was commensurate with the degree o f economic growth recorded duringthese years. D. BYHowMUCH SHOULD POVERTY FALL? 1.19 After seven consecutive years o f 5 percent or more real GDP growth, a four- percentage point decline in poverty is disappointing. But how disappointing is it? Should poverty have fallen by four or five times as much or only by fifty or twenty-five percent more? This counterfactual i s difficult to answer because the relationship between output growth and poverty reduction depends on many factors, including the initial level of GDP, the poverty line and the distribution o f income, and on assumptions about how growing incomes might otherwise have been distributed. 1.20 Estimates o f the elasticity o f changes inthe poverty rate with respect to growth in income are available based on cross-country studies. These elasticities pennit estimates of by how much poverty could have been expected to fall in light o f the extent o f income growth realized since 1994. 1.21 One of the frequently cited papers relating growth to poverty reduction is Ravallion and Chen (1997). The authors employ a cross-country data set covering 67 developing and transitional economies. The authors find that economic growth i s 6 associated, on average, with a decline inthe poverty rate. Furthermore, the magnitude o f the elasticity o f this relationship is dependent on the distance between the poverty line and mean household income (or consumption expenditures) per person. When the poverty line is at 100 percent o f mean income the elasticity for the non-ECA sample is - 0.64. (A ten-percent increase in mean incomes results in a 6.4 percent drop in the proportion o f individuals below the poverty line.) Whenthe poverty line i s at 50 percent of mean income the elasticity i s -1.57. 1.22 Inorder to apply the Ravallion and Chen elasticities to the Senegalese case it is necessary to know the ratio o f the poverty line to mean household income per household member, and the rate of growth o f the income measure. The PRSP does not report these data precisely, although they must be available from background work used to draft the PRSP, so it is necessary to approximate these values. Box 4 o f the PRSP (p. 25) reports an annual poverty threshold o f CFAF 143,080 per adult equivalent and annual mean household income per adult equivalent o f CFAF 204,118, for a ratio o f .70. But it is not clear ifthese values apply to 1991 or 1994. Annex 1o f the PRSP reports a 1994 poverty threshold o f CFAF 328/day, which i s equivalent to CFAF 119,720,a figure significantly lower than the Box 4 threshold. No amount i s reported in the Annex for 1994 mean household income so there i s no simple way to compute the required ratio. 1.23 Another way o f determining a range for the ratio o f the poverty line to mean income i s to start with the poverty rate. If the distribution o f income were normal and half the population were poor, then the poverty line would be at 100 percent o f mean income per person. If, as is usually assumed, income is distributed log-normally, the poverty line would be at a point below 100 percent o f mean income. In 1994, as indicated inthe PRSP, Senegal's headcount index stood at 57.9 percent. This suggests a poverty line that probably lies within 70 to100 percent o f mean income. Ravallion and Chen report elasticities o f the poverty rate to income growth when poverty lines are 75 (q= -0.95) and 100percent (q= -0.64) o fmean income. (Box 4 o fthe PRSP reports this same elasticity (q= -0.89), based on Senegal's experience, 1991-94, and where the poverty line i s 70 percent o f mean household income per adult equivalent. This value is slightlyoutside o fthe range reported byRavallion andChen.) 1.24 Inorder to apply the Ravallion andChen elasticities ina manner consistent with their derivation, it is necessary to know the rate o f growth o f mean household income per person. This figure i s not presented in the PRSP. A rough approximation is to use the growth rate o f per capita income, 20.5 percent from 1994-2001. According to national income accounts, private consumption in Senegal accounted for 75.4 percent o f GDP in 1994 and 77.7 percent in 2000. This result i s not inconsistent with the assumption that household incomes kept pace with national income. 1.25 If economic growth since the 1994 devaluation was distributionally neutral, if meanhousehold incomes per person grew at the same rate as GDP per capita, and ifthe poverty line in 1994 fell between 75 and 100 percent o f mean household incomes, then the poverty rate would have declined by between 13.1 and 19.5 percent. Interms o f the value o f the headcount index, it would have fallen from 57.9 percent in 1994 to between 46.6 and 50.3 percent in 2001. By only declining to 53.9 percent, the reduction in 7 poverty was only one third to one halfas large as what might have been anticipated. Stated differently, if growth had reduced poverty in keeping with cross-country norms, between 350 and 700 thousand more Senegalese would today be above the nation's 2400 calorie per day poverty line. 1.26 Inassessing the extent of poverty reduction in Senegal the results presented are subject to at least two important qualifications. First, there are many assumptions that cannot be fully justified without more data. Second, any poverty line i s a somewhat arbitrary standard and other dimensions o f poverty warrant attention. Perhaps many o f the severely poor witnessed an increase in caloric intake but not one sufficient to raise them above the poverty line. Such changes in the poverty gap would not be captured by the approach employed above, yet, if true, would represent significant gains in welfare for the hard-core poor. 1.27 The results on expected poverty reduction since 1994 imply that income inequality must have risen over time and raise the question o f why Senegal's post- devaluation experience was not more pro-poor. Examining the sectoral profile both o f output growth and o f employment outcomes offers some confirmation o f recent trends in poverty. These profiles also provide insight into why poverty reduction has been so limited despite Senegal's apparently good record o f aggregate growth. E. POVERTY AND THE RURAL SECTOR 1.28 As i s true throughout Sub-Saharan Africa, Senegalese poverty is disproportionately rural. The World Bank (1995) poverty study for Senegal for 1992 indicates that 80 percent of poor households were rural. A more recent IMF (2001) report indicates that rural poverty rates are more than 70 percent as compared to 30 percent in urban areas. In an economy where the population i s almost equally split between rural andurbanareas, poverty remains a heavily ruralproblem. 3 1.29 Given its rural nature, trends in poverty will depend upon growth in agriculture andor the movement of the labor force from largely agricultural activity to industrial and service sector pursuits. We begin by reviewing what has happened in the agricultural and, more generally, the primary sectors (agriculture, livestock, fishing and forestry) between 1994 and 2001. As the PRSP notes, "the stagnation o f agriculture" is a key determinant o f the modest reduction in poverty realized during the 1990s. This i s a satisfactory explanation until 1999. But the agricultural sector has done well the past two years - including 2001, the year o f the latest poverty survey - and recent trends require explanations beyond "stagnation" to explain the 2001poverty results. ~ Data for 1992 are reported in World Bank (1995), p. v. At that time Dakar represented 21 percent of the population but only 12 percent of poor households. Figures from the IMF (2001), p. 7 provide no reference to their source. According to World Bank (2002), Senegal's urban population has grown from 40 percent in 1990 to 47 percent in 2000. Slightly less than half of the urbanpopulation continues to live inDakar. 8 1.30 Trends ingroundnut production and revenues tell part o f the story. In 1992, poor rural households receivedmore of their cash income, about 30 percent, from groundnuts, than from any other single source. It was hoped that the devaluation in 1994, as well as other policy reforms affecting the groundnut sector, would improve rural incomes via higher farmgate prices, expanded production and growth inexports. Producer prices for groundnuts, which continue to be determined administratively, improved right after devaluation. Since devaluation, nominal producer prices have continued to increase, but at a rate probably well below the rate o f price inflation o f the bundle o f goods purchased by groundnut producers. Groundnut production from 1994 to 2001 has not consistently expanded despite generally favorable weather conditions, although much better harvests have been recorded recently. Export sales, at least until 2000, can only be considered disappointing (Table 1-3). Ifan expanding groundnut sector was expected to lift the rural poor out o fpoverty, these trends confirm its failure to do so. 1.31 The failure o f the groundnut sector to contribute more to the incomes o f the rural poor i s the result o f many factors. There continue to be issues about government intervention in the sector but there are also external constraints. Groundnut oil exports face inelastic demand and even when Senegal's groundnut production increases, as it did recently, falling world prices can offset the effects o f increasing production.4 1.32 Although the specific circumstances differ, the experience with groundnuts mirrors that o f much o f the primary sector. According to GDP estimates, the entire primarysector, and especially agriculture which accounts for about halfo fprimary sector output, showed little gain between 1994 and 1999. From 1994 to 1996, primary sector output grew slowly in real terms, the next two years it contracted, and in 1999 it recovered and stood at about 10 percent above the level o f 1994. The rural population expanded by about 5 percent over the same period, confirming the view o f "agricultural stagnation" reportedinthe PRSP. 1.33 Since 1999 estimates o f primary sector growth tell a different story. In constant prices, agriculture (primary sector) i s estimated to have grown by 21.3 (10.6) percent in 2000 and another 13.8 (7.6) percent in 2001. The years 1994 and 2001 correspond to the years o f the poverty surveys reported in the PRSP, and when taken as endpoints for growth inprimary sector output yield a 30 percent increase (again, mostly due to the last two years.) Between 1994 and2001, the ruralpopulation expanded by about 9 percent. 1.34 Despite what appears as a recent and significant growth spurt in primary sector output, both absolutely and relative to rural population growth, rural poverty appears not to have fallen accordingly. There are several potential explanations for this outcome. One is that the poverty gap in the rural sector was so large that even with growth o f 30 percent, most individuals continued to fall below the 2400 calorie poverty line. The Discussionof trends inthe groundnut sector i s basedon IMF(2001), pp. 6-11. Bycomparison,estimatesofthe contributionto GDP ofthe secondaryandtertiary sectors show annual growth each year since 1994. In 1999non-primary sector GDP was one thirdhigher than it was in1994. By 2001itwas almost 50percenthigher. 9 World Bank (1995) poverty study lends credence to this view, reporting a 1992 poverty gap of 16.4 percent for rural households based on a poverty line lower than the one used inthepoverty studies citedinthe PRSP. Alternatively, low rates ofpoverty reductionin rural areas may be the result o f how primary sector growth was distributed across households. If the distribution was fairly unequal, only a minority o f households may have benefited. Lastly, the GDP estimates for the primary sector in2000 and 2001 may be too high and there may have been less growth in the sector, an outcome in-linewith "the stagnation o f agriculture" and the less than expected fall inpoverty rates reported in the PRSP. 1.35 Even if2000 and 2001 were years o f exceptional growth inagriculture, there i s no guarantee o f continued rapid growth. The history o f Senegal's primary sector remains one o f cycles, and without a period o f prolonged and sustainable growth in agriculture, rural households will remain vulnerable to falling back into poverty. F. WHATDoLABOR MARKET OUTCOMES TELLUs? 1.36 Ifthe primary sector has been unable to lift rural households out ofpoverty, the altemative i s for individuals to find more productive work in industry and services, most often inurban areas. Since 1994 the non-primary sectors o f the economy have performed better than the primary sector. Combined, the output o f the secondary (construction, energy, industry and mining) and tertiary sectors (commerce, government, personal and professional services, and transportation) increased by 48 percent between 1994 and 2001, more than 50 percent faster than did the primary sector. Despite this more rapid growth, there is little evidence that a large number o f employment opportunities in Senegal were significantly transformed fi-om relatively low productivity work in agriculture to higherproductivity work inindustry and services. 1.37 Evidence on employment outcomes i s available from the same household surveys used to determine the incidence of poverty. These data indicate that despite strong growth inthe economy, unemployment increased betweenthe first round o f the ESAMin 1994/95 and the second round in2001 (Figure 2). Among the population o f working age (those over age 15) 6, 4.2 percent reported being unemployed in 1994/5 compared with 8.7 percent in2001. Since the share o f the working age population who were out o f the labor force actually rose a bit during this period, this represents a more than doubling o f the unemployment rate, from 6.7 to 13.8 percent o fthe labor force. Although the ESAM surveys gathered employment information for individuals ages 10 and above, a more standard definition of working age begins with age 15. This will cause some dramatic differences between these results and some reported elsewhere, since the population ages 10to 14representednearly 15 percent ofthe population in 1994. ' Calculations for 1994/95 were made using the ESAM Idata, incorporating the sampling weights. Calculations for 2001 are basedon the tables createdfi-omthe QUIDand from the raw data, bothincorporating samplingweights as well. 10 1.38 Approximately 1 million young people entered Senegal's working age group between 1994 and 2001 The increasing youthfulness o f the population probably accounts for the slight decrease in labor force participation, as a larger percentage o f the working age population remained in school or entered the peak child-bearing years. Total employment did rise, from an estimated 2.3 million workers in 1994/5 to 2.5 million in 2001, but this growth was not enough to absorb all the new entrants. Instead, there was an increase in the number o f unemployed persons over age 15 from about 168,000 at the start o f the period to about 435,000 by the end. Since devaluation the economy has created more unemployment than employment, and growing unemployment certainly i s consistent with Senegal's limitedprogress inreducing poverty. 1.39 The growing incidence o f unemployment i s particularly pronounced for women, 1.4 percent o f whom reported themselves as unemployed in 1994, compared with nearly 9 percent in 2001. This effect was not generated by an increase in labor force participation, since more women reportedthemselves out o f the labor force in2001 (52.5 percent) than in 1994 (50.6 percent). Since women constitute about half o f the working- age population, this decline in employment rates actually implies a small decline in the number of working women. Lower employment and incomes among women often are associated with poorer outcomes for children and families. 1.40 Table 1-4 presents evidence on employment trends. It compares changes in the share o f GDP from various sectors o f the economy with changes in the share o f employment in these sectors. In examining these data it i s important to emphasize that the +/- signs refer to changes inthe sector's share o f output or employment. Since 1994 all economic sectors identified inTable 1-4 experienced an increase in real output. But some sectors grew faster (slower) than others and claimed a rising (falling) share o f total GDP. 1.41 Primarysector output grew over time but contributed less to GDP in2001than it did in 1994. The decrease inthe primary sector's share of employment was even greater, at nearly 10 percent. Given the low productivity o f the Senegalese agricultural sector, in which about half the labor force produces less than one-fifth o f GDP, the relative movement o f labor out o f agriculture and into other sectors potentially is a vehicle for poverty reduction. But this depends upon the new economic activity that absorbs these workers. 1.42 Industry (including manufacturing, mining and energy), often the source o f expanding wage employment, commanded a larger percentage o f national output but a substantially smaller share o f employment, dropping from 12.3 to 8.6 percent. These results are consistent with a longer-term and troubling tendency toward the de- industrialization o f the labor force (Box l).Construction and transportation, both growing sectors o f the economy, saw output and employment expand by the same amounts. But the biggest increases in employment shares occurred inthe tertiary sector, especially incommerce and other services. 1.43 Commerce, mostly wholesale and retail trade, consists primarily o f informal activities; the self-employed constituted 87 percent o f that sector in 1994 and more than 11 90 percent by 2001. Without individual earnings data it i s impossible to make direct comparisons, but it seems unlikely that moving the labor force out o f agriculture into pettytrading activities will be a major force for poverty reduction. The same probably is true for the personal and professional services category, which experienced the largest increase inemployment share despite afall inits share o f GDP. 1.44 Personal and professional services cover a wide variety of activities, including employment inbankingand insurance, high-tech industries, the clergy, and tourism, or as civil servants, domestics and teachers. To the extent that the growth inthe importance o f this sector is associated with increases in employment in the more lucrative and stable occupations, this reallocation o f labor could be associated with poverty reduction. Employment in tourism i s an example, and there i s evidence o f some expansion in the number o f rooms, beds, and foreign arrivals, suggesting the potential for rising employment. But most employment growth in this sector probably is in its less remunerative areas.* 1.45 Self-employment i s much less common inpersonal andprofessional services than incommerce, butthere is evidence that the new employment inthis sector was also inthe informal sector. The share of service workers employed by individuals and households, rather than by private firms or government, rose from 55 to nearly 75 percent. Growth in the more informal, less skill-oriented branches o f commerce and the rest o f the tertiary sector, while obviously essential for the livelihoods o f those who practice them, i s not the kindo f higher wage, skill-based "modem" jobs on which the movement o f a population out o fpoverty must be based. 1.46 Another way o f thinking about what has happened inthe labor market over time is to look at the distribution o f employment status over time. The organization o f work takes a variety o f forms, and employment status i s a rough indicator o f the nature o f the relationship between employer and employee. It also can be used as an indicator o f the degree o f "formality" of employment. Figure 1.3 shows the distribution o f the entire labor force by employment status. "Employees" are workers who work "in continuous fashion" for their employers, even if that arrangement i s temporary or seasonal. Some changes in status, then, may not be obvious from the data if the temporary or seasonal share o f employment in this category expanded. "Casuals" i s a loose translation o f "Tticheron", referring to individuals whose relationship to an employer i s centered around performance o f an individual task or day's work. "Unpaid workers" includes two broad types o f workers: family workers and apprentices. The 1994 data allow us to separate out the two categories. In agriculture and commerce, family workers dominate the According to the Finance Ministrywebsite, public employment i s not the reason for the growth in the employment share of personal and professional services. Public sector employment actually shrank by 800 people during the period under consideration. Similarly, the Finance Ministryreportsno appreciable growth inthebanlungandfinancial sectors. Evidence from elsewhere inAfrica indicates that many apprentices are paid (Velenchik, 1995), butthe ESAMcategorizes them as unpaid. 12 category; in industry, transport and construction unpaid workers are almost all apprentices. Shifts instatus, therefore, have different meanings indifferent sectors. 1.47 Figure 4 presents the distribution o f employment status in the secondary and tertiary sectors combined. (Box 2 discusses trends in employment status in the primary sector.) The most notable feature is that the share o f employees in the sector has remained the same. Eventhe shares in self-employment show little change. The major difference between 1994 and 2001 i s the shift from unpaid workers to casuals. There are two possible interpretations of this change, and there i s probably some truth in both o f them. The decline in unpaid work in this sector implies a reduction in the number o f young people participating inapprenticeship programs, which can be an important source o f skill development among young people (Velenchik, 1995). Inthat event, a movement out o f apprenticeship into casual work could generate an overall decline inhuman capital investments. However, it can also be the case that enrollment in apprenticeship programs i s simply a way for unemployedyouths to fill time, without any necessary commitment to human capital formation on the part o f either the apprentice or the firm. In this case, a movement out o f unpaid apprenticeship and into casual work would indicate an improvement inthe options available to young workers. 1.48 Given the large size o fthe unpaid labor force in 1994, it seems likely that many o f them were using the apprenticeship to kill time, so that the expansion o f casual employment may well be making them better off. Without eamings data, it cannot be determined if casuals are paid substantially less than permanent employees are. Data from the Regional Program on Enterprise Development surveys (WED) elsewhere in Africa in the 1990's indicate that pay and benefits did not differ substantially between permanent and casual workers, although the latter are eligible for fewer protections. 1.49 The growth incasual employment may well have been causedby changes inlabor regulations (see below) and thus be o f some interest, but from a welfare perspective this i s not the most important thing the employment data show us. Despite considerable growth inoutput, and a significant departure o f labor from agriculture and the rest o f the primary sector, there has been little transformation in the employment status o f workers inthe secondary and tertiary sectors. Wage employment as a share of all employment has remained the same and "self-employed" or "unpaid", usually a sign o f low productivity and meager eamings, continues to describe almost two out o f every three workers. 1.50 Since 1994 the economy of Senegal has grown and some workers have certainly prospered. A constructionboom has expandedemployment inthis sector. Various sub- sectors, from fishing to tourism, have done well and, presumably, workers engaged in these activities have benefited. A smaller share o f Senegal's households was poor in 2001 than in 1994, and those that were far from the poverty line may have moved closer to it, indicating that the labor market has distributed the benefits o f growth to many people. But changes in the pattem o f employment in Senegal between 1994 and 2001 also are consistent with the nation's disappointing record o f poverty alleviation. This period was characterized by rising unemployment, a de-industrialization o f the labor force and shifts in employment status out o f more secure states in which human capital 13 formation would be encouraged into less secure self-employment activities. Given these outcomes, Senegal's poverty performance is not surprising. The next section will explore whether the labor market itself i s responsible for the lack o f growth in good employment opportunities. G. I S THE LABOR MARKETBLAME? TO 1.51 Highly interventionist labor market policies of the type found in much of Francophone Africa, including Senegal, have long been the subject o f criticism. The critique i s founded in the conventional wisdom that the distortions produced by government regulation undermine the labor market's ability to allocate resources efficiently, to provide incentives for the accumulation o f human capital, and to support broad-based growth and poverty reduction. Regulations on hiring, job security and the structure o f employment, and on minimum wages and other forms o f compensation, all are believed to have a negative impact on the labor market. 1.52 A number of studies o f Senegal's economy have blamed the labor market, at least inpart, for poor overallperformance. GuyPfefferman, writing inthe Forewordto Terrell and Svejnar (1989) describes his own 1968 book as concluding that "Labor market imperfections, mostly inherited from French colonial legislation and institutions.. .were so egregious as to stifle development.. ..Wage and other regulations intended to advance the workers' welfare failed to achieve their objectives." Twenty years later, Terrell and Svejnar, which i s still widely cited, argue that ... [as] current regulations contribute to " the underutilization o f labor, a relaxation o f the regulations concerning (paid) hours o f work, layoffs, and the use o f temporary (daily) workers would, along with transfer payments, improve the welfare o f all the parties."(p. 120). Most recently, the 1995 World Bank poverty study, Senegal: An Assessment of Living Conditions, posits that "...employers keep the number of regular employees hired to a minimumbecause they fear lack o f flexibility in set wages, benefits, and not being able to let employees go.. .Demand for the poor's most valuable asset, labor, thus was depressed." (p.17) If, indeed, labor regulations act as a significant constraint to job creation, then they might be expected to limit the dissemination o f growth to the poor. By the same token, deregulation o f the labor market, had those regulations actually been distortionary, should result inan expansion o f employment, especially ina growing economy. 1.53 Despite widely accepted arguments that distortionary regulation o f the labor market i s an impediment to development, evidence to support this argument is weak. In broad reviews o f evidence from aggregate data and the econometric and case study literatures, Freeman (1992) and Lindauer (1999) find little evidence from low andmiddle income economies to support the belief that labor policies reduce formal labor demand, crowd the labor force into the informal and rural sectors, or keep wages there low. For Senegal, even Terrell and Svejnar find relatively little direct evidence that labor regulations are among the more serious problems facing the industrial sector. Recent econometric work by Rama (1998) concludes ". ..wage misalignment cannot be traced down to the labor market policies and institutions in force in any obvious manner." Results from the WED surveys in a number o f countries, including Cameroon, CBte 14 d'Ivoire, Ghana and Zimbabwe, similarly find that employers do not rank labor regulations among the most intrusive o f government policies. 1.54 More recent survey data generally conform to this pattern. For the 2002 Africa Competitiveness Report, a sample o f firms in each o f twenty-four nations were asked to respond to the following statement on a scale from 1 to 10, with 1 being strongly disagree and 6 being strongly agree. "Hiring and firing practices are flexibly determined by employers." Senegal's average, at 3.88, ranked it 10 out of the 24 countries covered, and it was the francophone country where employers were least likely to complain about labor regulations. The ranking table from the Africa Competitiveness Report i s reproduced as Table 1-5. (See Box 1-3 for a summary o f survey results from other African nations). 1.55 Data from a World Bank sponsored Regulatory and Administrative Costs survey conducted on a sample o f firms in Senegal in 2002 provide comparable results. Respondents were asked to categorize a number o f labor regulations in terms o f their posing problems for firms. They were asked similar questions about other potential problem areas. Responses are reported inTable 1-6. There are two important features o f this table. First, among the labor market institutions and regulations, only unions rate as problems for most firms, and fewer than half o fthe firms rate them as moderate obstacles or worse. (The union issue is discussed in more detail below.) Second, all o f the other issues, some regulatory and some not, are substantially more problematic than any o f the labor market concerns. The exact same perspective is provided in a number o f web sites offering advice about the strengths and weaknesses o f different economies to international investors". This advice pays scant attention to Senegalese labor law, except to note the recent reforms, while giving great attention to infrastructure and capital market concerns. 1.56 Third, the most important labor market problem reported in Table 1-6 i s worker qualifications. Concerns about the quantity and quality o fhumancapital available among workers in Senegal i s frequently mentioned in employer surveys, from those done by Terrell and Svejnar in the late 80's to the work o f Golub and Mbaye in 2000. Contradictory results come from the Africa Competitiveness Report survey, in which Senegal ranks fifth based on employers' strong agreement with the statement, "The local labor market has a sufficient supply o f educated workers to meet the hiringneeds o f my firm." These same employers, however, characterized the primary, secondary and university systems as not providing rigorous training in language, math and science. It i s not clear from this evidence whether the perceived problem is in the availability o f workers with sufficient basic literacy and other skills to effectively perform factory work, or whether the problem lies with more sophisticated administrative, managerial and technical labor. loFor example, see www.tradeport.or8/ts/countries/sene~al/climate.h~1~ www.dkar.com/ccainvestclimate.htnx and www.unido-aaittx.com/unido-aaitpc/seneaal/basic- stance.htm1. 15 1.57 There are a number o f reasons why labor regulations might be expected to have much less impact than they would appear to on paper. First, the regulations generally cover only the formal sector, which constitutes a very small part o f the labor market. In Senegal, 75 percent o f the labor force are either inthe primary sector or self-employed in other sectors, and many o f the remaining workers are employed in the informal sector. Given that the labor force i s so much bigger than the formal labor market, one expects to find only a minimal direct impact o f labor market policies. That is not to say that these policies do not potentially have enormous indirect impact, insofar as they influence enterprise formation and expansion, and hence the creation o f employment, but their actual protective impact falls on an extremely small number o f workers. Whether this potential i s realized, however, remains questionable. 1.58 Second, enforcement of labor regulations is often weak. The 1997 Code du Travail o f Senegal states, inArticle L.3., that any employer having one or more workers i s subject to the labor code Inreality, no one argues that the government enforces these rules on firms o f only one or two employees, especially since the existence o f most such enterprises i s unknown to the government. Even among firms that are subject to enforcement o f labor legislation, many rules are only weekly enforced. Terrell and Svejnar found that most firms openly flaunted the rules that firms only hire workers sent through government employment agencies, preferring to find their own workers (usually at the factory gate) and "regularizing" their appointments through the labor ministry only after a probationary period o f employment. These hiringrestrictions are no longer part o f the Code du Travail. Weak enforcement, or enforcement focused only on the most visible firms (for example, multinationals), diminishes the impact o fthe legislation. 1.59 Third, not all components of labor legislation are bindingon firms. Inparticular, many industrial firms pay wages substantially above the legislated minimumwage. Inthe recent Regulatory and Administrative Cost survey, the lowest salary a firm reported paying was 10,000 CFAF per month. When the same firms were asked to report the lowest level o f the minimumwage structure applicable to them, the lowest response was 2105 CFAF per month, which i s certainly much less. Work with WED data for Zimbabwe and Ghana, (Velenchik, 1997, and Teal, 1996) also finds that larger firms pay wages far above those paid by smaller firms, and certainly far above the legal minimum. 1.60 Finally, and perhaps most important, the impact o f labor regulations has paled in comparison to the distortions generated by other policies inthe marketplace. One might expect that firms wishing to enter or expand could find the restrictions on firing and layoffs generate a marked disincentive to increasing employment. However, if the host o f other distortions that have traditionally characterized developing country markets render the environment too hostile to invite entry or expansion, then we would expect to l1 We believe this i s the most recent version of labor legislation. Article L.3 reads: "Toute personne physique ou morale, de droit public ou de droit privC employant un ou plusieurs travailleurs au sense de l'article L.2. est soumise aux disposition du present code visant les employeurset constituteune enterprise." 16 see very little impact of labor regulations. Labor market rigidities are second order problems relative to misguided macro policy, highly distorted trade environments, credit market constraints and corruption. 1.61 Inthe period leading up to devaluation, Senegal reformed its policy ina number of these areas, including the labor market. This period, then, provides an opportunity for examining the effect o f regulation by looking at firms' response to its removal or reduction, in an environment that has been, it is hoped, rendered much more suitable for investment, job creation and growth. 1.62 In addition to eliminating hiring regulations, the labor market reform had two particularly significant components. First, the rules requiring Labor Ministry approval to layoff permanent workers during difficult economic times were relaxed, so that firms no longer need explicit approval in order to conduct these layoffs. They are still required to meet with workers' representatives to explore all alternative options and send notes from that meeting to the ministry, and to wait two weeks before instituting the layoffs. The labor code also includes instructions on how to select workers for layoff, and how workers are to be informed. More important, employers are still required to pay significant severancepay. 1.63 The second reform relates to workers with fixed-term contracts, who are not permanent workers and thus not subject to the layoff protections described above. The reforms relaxed rules regarding the number o f fixed-term contracts a worker could have before automatically becoming permanent, by excluding seasonal workers, daily workers andworkers insectors where fixed term employment is "normal". Both o f these changes were intended to allow firms more flexibility in hiring in order to correct the disincentives described above. Ifthe reforms were effective, firms would be expected to use layoffs and more hiring o f fixed-term casual and temporary workers. Employers also would be expected to report that labor regulations are less burdensome. 1.64 Although data about layoffs, per se, are not available, information from the Regulation and Administrative Cost survey can address these same issues. Firms were asked to characterize the size o f their labor force. Among the 184 firms responding, 70 precent said their current firm size was just right, 10 percent felt it was too small, 15 percent said they didn't know, and only 5 percent felt that their current firm size was too large, This seems consistent with firms feeling able to release workers as they need to. The survey also included direct questions about the recent labor market reforms. The results are summarized inTable 1-7. The evidence i s mixed, with fewer than one third o f firms indicating that labor market reforms encouraged more hiring and only 15 percent indicating that they had responded to reforms by employing more unskilled labor. These surveyresponsesdo not paint apicture o f firms consciously making a strong employment responseto deregulation. 1.65 The final, and best, piece o f evidence in this regard comes from the pattern o f employment described in the previous section. The period from 1994 to 2001 saw a substantial increase in the share o f casual workers inthe labor force. It seems likely that this was a response to loosening o f the restrictions on this type o f employment. While 17 the rise in temporary and casual work has been described in the Senegalese press as troublesome, this i s exactly what would have been expected from the policy change, had the earlier regulations truly been binding. It i s true that casual workers have fewer protections than do regular employees, and are ineligible for some benefits. It i s also true, however, that these jobs offer more protection andbenefits than is likely to be found in self-employment or unemployment. When viewed in this light, some o f the conventional wisdom on the consequences o f labor market deregulation and the subsequent rise in the number o f casuals may be wrong. Rather than diminishing the welfare o f the workers affected, these developments may be welfare enhancing. 1.66 Conceming the labor market, Senegalese firms in2002 seem less concemed about labor regulations than they are about a lack o f qualified workers and problems with unions (Table 1-6). The PRSP recognizes weaknesses in human capital formation and advances steps to improve the quantity and quality o f education. For firms, the single most problematic area o f labor market institutions i s the relationship between employers and unions, which i s cited here, inthe 2002Africa CompetitivenessReport, inTerrell and Svejnar's earlier characterization o f industrial relations as "adversarial," and in the most recent case study work o f Golub and Mbaye (2000). The contentious relationship between unions and firms arises from Senegal's political and industrial history. While Senegalese industry was highly protected and uncompetitive, union-firm relations were centered around the division o f a reasonably fixed set o f economic rents. Cooperation between the parties was unlikely to generate a substantially larger pie, given the economic environment, so it was reasonable on both sides to engage in non-cooperative attempts to gain a larger share o f the pie. The unions had even less than usual incentive to focus on expanding employment rather than raising wages. The govemment, having created the monopoly rents through its own industrial and trade policy, then had an incentive to support union attempts to share at least part o fthese rents with the fraction o f the population able to get jobs in the sector, creating tight ties between unions and government that continue to exist. Intoday's more competitive industrial environment, in which firms are not as protected from intemational competition or supported by widespread input subsidies, there are fewer rents to share, and potentially much bigger gains to cooperation. The institutions, however, do not appear to have caught up with these changes, leaving a system of industrial relations that continues to be plagued by antagonism. Developing a better functioning industrial labor relations system should be a goal o f labor policy. 1.67 Overall, the evidence, from Senegal and elsewhere in Africa, indicates that the labor market i s generally not to blame for poor industrial performance and the failure o f these nations to attract significant foreign investment or grow sizable domestic industries. Given the extremely small size o f the covered sector in Senegal, it seems inconceivable that, even without regulation, the firms covered by labor laws would have grown enough to make a substantial dent inpoverty. We obviously do not know about the firms that did not enter because they were deterred by the legislation, but the responses o f existing firms to questions about regulatory importance make it unlikely that labor market policy i s the most important issue. Inour judgement, the failure o f Senegal's economy either to grow more quickly or for the growth that has occurred to be more pro-poor, probably, has had little to do with the operation o fthe labor market and the policies that govem it. 18 H. EMPLOYMENT POLICY INTHE PRSP 1.68 Senegal's PRSP i s a broad ranging document. Among its many policy recommendations are specific plans to improve performance in each sector o f the economy; proposals for enhancing macroeconomic stability; strategies for providing sanitation, drinking water, and health services to the poor; plans for expanding and improving primary education; policies to protect the most vulnerable people and the environment; and even proposals for improvements ingovemance. 1.69 All o f these proposals can, and should, be understood as ultimately promoting human resource development and employment. Inthat sense, the whole PRSP is about labor. It does, however, also contain some specific proposals for enhancing employment, as well as a number o fproposals for improving education and training. These concluding remarks will focus on those proposals. 1.70 The employment proposals contained in the PRSP can be characterized as "active" labor market policy on the part o f the government. For example, the proposal to "Create a national agency responsible for ensuringthe transparency and efficiency o f the labor market. Information program on the labor market,"(p. 59) implies more government involvement in the process o f matching workers and jobs. Such a plan seems to contradict the recent trend toward labor market deregulation, which included the removal o f restrictions on hiring practices. Although many firms complain about problems finding qualified workers, it seems more likely that this i s caused by shortages of highly skilled managerial and technical workers, rather than by information and transparency problems inthe labor market. In some developed economies - Germany i s a prime example - problems o f matching available jobs with workers who have the appropriate skills has proven problematic and contributed to frictional unemployment. But in Senegal and other developing countries the character o f unemployment is far different. Ifjobs remain vacant it more often i s due to an absolute shortage o f particular labor skills. Most unemployment results from individuals unable to find work because o f slow economic growth and a lack o f labor demand. 1.71 Similarly, a plan to "Establish systems to manage jobs, professions and skills" also implies government interference in labor market functioning that, even if it i s not detrimental to labor market performance, is probably not necessary either, and almost certainly isn't worth the expenditure o f limitedbudget resources. 1.72 On a more positive note, if the proposal to "Re-organize the labor market" means a revision o f the current system of industrial relations, this could be an important first step. Fostering cooperation between unions, firms and government, in order to shift all parties' focus away from dividing rents and toward the expansion o f employment and production, would be a positive change inthe labor market. 1.73 In terms of human capital formation, improving the general primary, secondary and university education system, eradicating illiteracy and improving educational attainment o f girls and women are more likely to prove productive in improving standards o f livingthan are more narrow investments invocational training. Government 19 i s probably less effective at determining which skills are needed and how they should be generated than firms are. Creating an environment in which firms find it profitable to have a long term commitment to the marketplace and the labor force is more likely to stimulate appropriate human capital formation than are committees managing vocational training programs. 1.74 Overall, the PRSP's relative inattention to labor market policy per se is appropriate. As this chapter has argued, labor market policies and institutions are not at the root o f Senegal's disappointingperformance inpoverty reduction. 20 21 22 0 C 0 N cn cn cn w co cn m H M N ' cn En Q cn cn H Ln cn cn w 8 E E 0 25 Table 1-3: Trends inGroundnutPrices,ProductionandExports,1994-2001 (1994/95 = 100) Year Producer Prices a Productionb Exports c 1994/95 100 100 100 1995/96 104 115 92 1996/97 109 87 74 1997/98 125 76 53 1998/99 133 81 56 1999/00 121d 114d 70 2000/01 121d 133d 107d Source: IMF (2001), "Senegal: Selected Issues" a. Producer prices refer to the price of oil groundnuts based on current CFAF prices per kilo. b. Productionrefers to estimates of groundnut output measured inthousands ofmetric tons. c. Exports refer to export values for calendar, not crop, years. d. Estimates. C C c c C 7 C c C r C C C C C 7 C C C 7 27 Table 1-5: "Hiring and FiringPractices are flexibly determinedby employers" (1 =strongly disagree; 6 =strongly agree) Rank Averaqe remonse 1 Uganda 5.20 2 Nigeria 5.14 3 Malawi 4.76 4 Ethiopia 4.68 5 Madagascar 4.50 6 Kenya 4.49 7 Ghana 4.48 8 Zambia 4.04 9 Lesotho 4.03 10 Senegal a 3.88 11 Botswana 3.85 12 EWPt 3.54 13 CBte dIvoire a 3.47 14 Cameroona 3.42 15 Tanzania 3.36 16 Namibia 3.31 17 Mozambique 2.87 18 Mauritius 2.86 19 Burkina Faso a 2.71 20 Swaziland 2.54 21 Zimbabwe 2.54 22 South Africa 2.41 23 Tunisia 2.39 24 Morocco 2.35 Source: World Economic Forum(2002), TheAfrica CompetitivenessReport, Table 5.10. a. Country inthe CFAF zone. 28 Table 1-6: Firmclassificationof obstacles to firm operationand growth (Inpercent) No obstacle Minor Moderate Major Very serious obstacle obstacle obstacle obstacle Labor code and 43.2 21.8 15.8 8.9 9.3 legislation Lack of qualified 29.3 8.7 23.9 26.1 12.8 workers MinimumWage 61.2 21.2 11.8 1.2 4.7 Social Security 41.0 26.0 14.0 12.0 7.0 PayrollTaxes Unions 40.0 17.5 13.8 10.0 18.8 Electricity 8.0 9.0 23.0 27.0 33.0 Tax rates and tax 6.0 6.0 27.0 29.0 31.0 admin. Corruption 4.0 23.0 22.0 21.0 30.0 Access to Land 8.0 15.0 27.0 22.0 28.0 Source: RegulationandAdministrativeCost Survey, 2002 Table 1-7: Firms' opinionsabout labor market reforms (Inpercent) Yes No Don't Know Have the structural 39.3 20.6 40.0 reforms currently underway had a positive effect onthe labor market? Have these structural 30.0 32.5 37.5 reforms encouraged you to hire more workers? Havethese structural 15.6 40.0 44.3 reforms encouraged you to hire more unskilled workers? Source: RegulationandAdministrative Cost Survey, 2002. \ \ 0 t e1 0 m I r t 0 0 0 0 0 0 0 F (0 In d- m N N m cl c, d si ed ! c, t 0 h ? I w 2. COMPETITIVENESS, MANUFACTURINGAND EXPORTSIN SENEGAL 2.1 The poor response o f traded goods output and exports to a series o f outward- oriented, liberalizing reforms i s a clear indication o f declining competitiveness. Before the 1994 devaluation, the overvaluation o f the CFA franc was detrimental to competitiveness. However, the devaluation, which produced a one-time improvement, did not reverse a long term decline in competitiveness. Was this due to too-high wages, or declining productivity growth? This chapter seeks to answer that question by first describing the structure and policy environment governing the manufacturing sector. It then defines and compares measures o f competitiveness in Senegal to other countries. Examining the factors driving competitiveness, it finds low productivity growth to be the key problem rather than highwage growth. It then discusses various factors affecting manufacturing productivity and explores their relative importance, as a guide to prioritizing policies to promote competitiveness. A. MANUFACTURING INSENEGAL 2.2 Before independence in 1960, Senegal had the most developed manufacturing sector o f the region. Dakar, then the capital o fAfrique Occidentale Francaise (the French West African Federationor FWA) was the focal point o f almost all o f the most important infrastructure in the FWA, including the port, the airport, and a relatively dense road network. Industries located in this area had as their natural markets all FWA member countries. After independence, however, manufacturing in Senegal started to experience some difficulties, as newly independent partners set up trade barriers and Senegal embarked on an inefficient inward-oriented industrial policy. Under various structural adjustment programs beginning in the 1980s and culminating in the 1994 currency devaluation, more market-oriented reforms were introduced without significantly entailing deep changes inmanufacturingpatterns. Sectoral Policy and Manufacturing: An Overview 2.3 Using industrial development as an impetus for sustained growth and economic development was the primary focus o f Senegal after independence. An important instrument the government relied on for this purpose was the investment code. First implemented in 1962 under Act 62-33, it granted generous tax and duty exemptions over 25 years to firms investing a minimum o f CFAF 1 billion. This minimum investment requirement was further reduced to CFAF 500 million in the 1972 code, which in addition generalized exemptions awarded to a larger bundle o f inputs. In 1977, the investment code was modified again as part o f Act 77-91 to put more focus on small and medium enterprises (SMEs) by setting the minimum required investment to the much lower level o f CFAF 20 million. 2.4 On the institutional side, the Ministry o f Industrial Development and Handicrafts (MDIA) as well as the Socie'te' Nationale de Promotion Industrielle (SONEPI) were 34 established to provide various types o f support to local businesses and to help boost industrial investment generally. Moreover, monopoly status coupled with huge tariff and non-tariff protection was granted to most firms. Prior to implementation o f the 1986 New Industrial Policy (NIP)more than 161 manufactured goods benefited from some type o f protection. The impact o f this incentive package on manufacturing development was quite disappointing, since output failed to significantly increase and average excess capacity in the industrial sector was estimated at 41% between 1981 and 1985. By the end o f the 1 9 8 0 ~Senegal was experiencing such serious balance o f payments and fiscal ~ crises that deep macroeconomic and structural reforms were necessary. 2.5 With the NIP reform in 1986, the government shifted the emphasis from import- substitution to export-oriented activities. Tariff and non-tariff protection were dramatically lowered, public enterprises were privatized, and various export promotion schemes, including drawback, temporary admission, export subsidy, and a preferential tariff regime, were applied to exporting firms. The outcome was disastrous for the Senegalese industrial sector (Mbaye 1998, Latreille and Varoudakis 1996). Most import substitution firms went bankrupt, while competitive exporting firms were yet to emerge; economic activities moved away from manufacturing to real estate and petty commerce, which were deemed less risky. In 1994 the currency was devalued, and in subsequent years very important trade and sectoral measures were undertaken to stimulate manufacturing activity. 2.6 Today, the Senegalese trade regime i s considered relatively open (World Bank 2002, Hinkle and Herrou-Aragon 2002). Since January 2000, Senegal has been under the Common External Tariff (CET) regime o f the West African Economic and Monetary Union (WAEMU ) countries, which has considerably reduced and streamlined tariff and non-tariff barriers. Tariff peaks are now set at 22% and the simple average tariff rate i s only 12%. These various measures have had a mitigated impact on manufacturing performance. Manufacturing Performance 2.7 Itis obvious that the 1994 devaluation andrecent trendsintrade policy havehad a sizable positive impact on aggregate as well as some sectoral indicators in Senegal. GDP growth has been above 5% since 199512,which compares favorably to other countries in the region. More to the point, Tables 2-1 to 2-3 depict genuinely favorable trends in Senegalesemanufacturing. The manufacturing share inGDP increased from 14%in 1994 to almost 18% in 2000, far above the levels achieved by Botswana (5%), Burkina Faso (12%) and Ghana (9%), but below Cote d'Ivoire (19%). In addition, real manufacturing exports grew at an average annual rate o f 12% in 1994-99, and real manufacturing output at an average annual rate o f 5% over the same period. 2.8 High performing sectors within the aggregate manufacturing sector are fish processing, chemicals, and building industries (table 2-1,2-3). The predominant actor in '*Except in2002 where GDP growth rate i s estimated at 2.4%. 35 this latter sector is SOCOCIM'3, a cement-producing firm that had until recently monopoly control and now faces increasing demand due to huge growth in housing construction following the 1994 devaluation. The firm exports very little o f its output and does so mostly to neighboring countries, particularly Mali. The building industry sector registered an annual growth rate o f 7.6% in its output in 1994-99. But its exports substantially decreased over the same period, because the domestic demand for building materials in Senegal increased dramatically after the 1994 devaluation. Another high- performing sector i s the chemicals industry, which is dominated by the Industries Chimiques du S&n&gaZ(ICs). ICs has as its main shareholders governments o f several developing countries (mainly India and Nigeria) in cooperation with Senegal, and specializes in the production o f fertilizer and sulfuric acid. Its output increased by 12% annually from 1994 to 1999, while it experienced a average annual growth rate o f 9.5 in exports over the same period. Other good performers include the miscellaneous food, grain and flour, wood, paper and printing, and mechanical industries, but their share in manufacturing is minuscule. 2.9 A feature o f Senegalese manufacturing is its high level o f concentration (table 2- 4). Three o f the fifteen sectors make up 52% o f total output and 93% o f total exports in 1999. These are, namely, fish processing, chemicals, and peanut oil. All three are export- oriented and process abundant natural resources found in the country: groundnuts, phosphates, and fish products. The other sectors are almost marginal within the total manufacturing sector, despite important advantages they benefit from to varying degrees. Inthis regard, the textilelapparel andsugar industries are worth singlingout. The latter is dominated by the Compagnie SucriBre S&n&galaise(CSS), a French company that grows and processes sugar cane in Richard Toll, located in the northern part o f the country. Since it was established, it has enjoyed various kinds o f protection, which seem to have survived the recent liberalization process. The prices it charges for its products are far above the world market; its mark-ups are the highest in the Senegalese manufacturing sector. They amount to 50% in 1999 after reaching a maximum o f 107% in 1993, although the company has yet to increase its share in total output to more than 7% since 1974 (Mbaye 2003). The textile/apparel sector i s generally considered to give an impetus to the industrialization process in developing economies, because it is highly unskilled- labor-intensive. In Senegal, it is among the lowest performing industries with less than 5% o f total manufacturing output and exports. Most firms in the textile industry have aged equipment far below international standards and are struggling to survive (Golub and Mbaye 2002), while in the apparel sector there are only very small firms on the borderline of the informal sector. Unless the textile and apparel sectors address the various bottlenecks that hamper their expansion it i s unlikely that Senegal will be able to significantly profit from the Africa Growth and Opportunity Act (AGOA), recently passedby the U S government.Export Performance 2.10 Senegalesemanufacturedexports have shown a general decline inthe long period, as illustrated inTable 2-3. In 1987 prices, manufactured exports rose from 162.2 billion CFA francs in 1974 to 254.3 billion in 1977 before steadily declining to 164.1 billion in l3In2002, we have another cement processing firm(Les Ciments du Suhel)that has startedoperating. 36 1999-an annual rate o f decline o f 1.79 percent in 1977-1999. A trough was reached in 1994, with some recovery after that. Table 2-3 and 2-4 shows the sectoral pattern over these periods. 2.11 We can distinguishtwo time periods: before and after the devaluation. Before the devaluation, many sectors experienced ups and downs, following the path o f aggregate real manufactured exports, which rose at an annual rate o f 3.1 percent, then fell 1.0 percent, respectively in 1974-83, and 84-93. This was the case for textile, wood, paper, construction materials, and metal-working industries. Prior to the devaluation, the Senegalese currency was clearly overvalued, and most sectors had difficulties exporting. In84-93, the vegetable oil sector suffered an annual decline of 13.55 percent, while the fish processing industries had feeble a growth o f 2.56 percent, followed in the following 1994-99 period by respective growth rates o f 4.55 and 39.39 percent. The Chemical industry grew 1.51percent annually in 1984-93. This sector is something o f an exception as regards output growth. Probably, because o f the presence o f the large firm Industries Chimiques du S&n&guE(ICs) which has almost guaranteed markets incountries like India and Nigeria, who are large shareholders inthe company. 2.12 Following the devaluation, the growth o f manufactured exports shot back up to 12.5 percent in 1994-99, compared to a decline o f 1.0 percent in the 1984-93 preceding period. In particular, the oilseed and chemical industries experienced an upsurge in exports. Even some industries not previously oriented towards exports, such as metal- working industries and printing, witnessed increased exports in the aftermath o f the devaluation. Some inward-oriented industries also saw exports decline (table 2-3). These include miscellaneous food products (-8.3 percent), construction materials (-34.4 percent) and grains and flours (-34.6 percent). A.1. THECOMPOSITIONOFEXPORTS 2.13 Manufactured exports, as total exports, are heavily concentrated in three sectors: fishing and fish-processing, the chemical industry (the ICs) and groundnuts (oil and oil cake). Table 2-4 provides information on the evolution o f these sectors from 1974-1999. These three sectors represented 81 percent o f manufactured exports in 1974. By 1999 this share increased to 93 percent. Within this subgroup, substantial change has taken place. In the 1970's, the groundnut oil sector dominated, with 56 percent o f manufactured exports in 1974. Inthe 1 9 8 0 with ~ ~ the creation o f the ICs, coupled with a decline in the groundnut harvests, the situation changed markedly. By 1999, the groundnut industry's share had plummeted to less than 12.6 percent while chemicals were up to 39.8 percent. The fishing industry also experienced a large expansion, and accounted for just over 40.6 percent o f manufactured exports in 1999. Table 2-4 hrther illustrates the changing composition o f exports. Since 1986, groundnuts' market share has declined. 2.14 An interesting further case is the apparel sector, which is highly labor intensive, and in which Senegal has a potential comparative advantage (Golub and Mbaye 2002). Yet its already meager share o f manufactured exports o f 2.4 percent in 1974 steadily 37 declined until the devaluation. Following the devaluation, there has been a partial recovery, but its share o f manufactured exports remains below 1percent. A.2. THEEVOLUTIONOFTHEWORLDDEMANDFORSENEGALESEEXPORTS 2.15 As noted previously, Senegal's exports are concentrated inchemicals, groundnuts and fishing. Unfortunately, most o f the latter have faced declining world demand-81 percent o f Senegalese export markets had growth rates below the global average export growth rate between 1997-2000. According to World Bank (2002), these adverse trends in demand are Senegal's mainproblem. Insome cases, Senegalese exports have grown more rapidly than world demand. For example, Senegalese canned tuna exports grew 300 percent between 1997 and 2000, even though world demand only expanded by 3.3 percent duringthis time (World Bank 2002). 2.16 Indeed, while declining world demand i s certainly a negative factor on the terms of trade, one should not make too much o f this factor. By itself, it can hardly explain the weak export performance, at least for the major export sectors. The domestic decline in output and productivity inSenegal inmany industries are surely more decisive. Tables 2- 1 and 2-3 provide information on output and exports growth by sector for various periods. Those sectors with declining exports also experienced declines in output. And, as argued by Golub and Mbaye (2002), the output declines are more closely linked to domestic problems than the world market. Also, for some Senegalese products such as chemicals, demand in developed countries actually increased substantially during much o f the time period in question and only recently slowed down. Furthermore, while it i s true that the demand for groundnut oil has declined, this did not prevent a number o f Asian and Latin American countries from gaining market share at Senegal's expense (Badiane 1993). EXPLAINING PERFORMANCEOFMANUFACTURING THE EXPORTS Indicators 2.17 The poor performance of Senegal's exports i s a clear sign o f declining competitiveness. This section examines competitiveness by comparing costs and efficiency between Senegal and its major competitors. It does so by constructing indices of relative unit labor costs o f labor. Unit labor costs measure the nominal wage divided by the marginal productivity o f labor, and thus increase with wages and decrease with labor productivity. 2.18 Labor productivity is defined as real value added per employee, and wages are labor compensations per employee. Real value added i s obtained as nominal value added deflated by a manufacturingvalue added deflator. To compare levels (as opposedto rates of change over time) o freal outputs and labor compensation across economies, they must be converted to a common currency. Real value added is converted at a PPP exchange rate and labor costs are converted using market exchange rates. See Mbaye and Golub 2002 and Mbaye 2003 for further details on data sources and methods. 38 2.19 Thus unit labor costs are influenced by three main factors: productivity, wages and the exchange rate. These indicators o f Senegal's competitiveness are compared to other countries and across sectors through bilateral relative unit labor costs (RULC) indices, relative to each o f Senegal's main foreign competitors for various manufacturing sectors. Then, such bilateral indices will be averaged to explain export trends for these industries. 2.20 The RULC indices specifically compare across countries relative productivity to the relative wage rate. Thus when the index is lower than one, Senegal's relative productivity is higher than its relative wage. Whenever it is higher than one, wages have outpaced productivity and Senegal will tend to lose market share to the benefit o f its foreign competitor. Comparison of Relative Unit Labor Costs 2.21 Table 2-5 depicts the evolution o f Senegalese RULC vis-&vis its major competitors in the aggregate and by regions, for the manufacturing sector. For these indexes, we compare Senegal to developing or transition countries in four other regions : Africa (Malawi, Zimbabwe, Mauritius), Asia (Korea, Singapore, Hong-Kong and India), South America (Mexico and Chile), Eastendcentral Europe (Hungary, Poland and Turkey). In the remainder o f the paper, regions refer to these country groups. At the beginning o f the 1970s, Senegal enjoyed a somewhat favorable competitive position relative to most o f the foreign developing countries in our comparison group, with a RULC inferior to 1 in most cases. The situation deteriorated very markedly at the beginning o f the 80's and kept worsening up to the 1994 devaluation. Immediately prior to the 94 devaluation, Senegalese competitive position was dismal since RULC was superior to 1for almost all countries and all sectors, and displayed a sharply rising trend. With the devaluation, RULC fell dramatically but soon started to appreciate again for many sectors against a number o f countries. 2.22 The same broad tendencies are clear when examining the relative producer price index, as shownby Table 2-6. Incomparison to all countries, Senegal's relative producer price dropped sharply in 1994, but then started rising again. Relative to Asian countries in 1999, the RELPReven surpassedits pre-devaluationlevel. Duringthe pre-devaluation period also, the general movements o f RELPR are quite similar to RULC. Thus, these two indicators o f the real exchange rate provide the same picture o f the time path o f Senegalese international competitiveness. 2.23 Turning to levels, we see that Senegalese absolute advantage has sharply deteriorated over time, in the sense that its RULC has moved from below the 1.0 threshold to well above 1.O over time. Senegal's initially favorable situation has several causes. First, until the early 1980s, Senegal's trade policy was typical o f many inward- oriented developing countries. Second, during this earlier period, Senegal had a relatively high capital-labor ratio due to the large role o f foreign firms associated with Senegal's past situation as capital o f colonial French West Africa (AOF). Inthe 1 9 8 0 ~ ~ all this i s changed, as most developing countries embarked on economic reform, with more favorable effects elsewhere than in Senegal. Senegal's capital stock-both public 39 and private4eteriorated steadily due to inadequate investment. Also, Senegal lost its captive markets inother newly-independent AOF countries as these countries established barriers to protect their own industries. This entailed reduced demand for Senegalese products and hence contraction o f output to inefficiently low scale, thus reducing observed productivity. The timidity o f NPIreforms and especially the inability to adjust the exchange rate renderedSenegal highly vulnerable to these changes. Thus, Senegalese RULC continued to rise. As noted previously, the devaluation reversed this trend only partially and temporarily. B.1. THEEVOLUTION COMPARATIVEADVANTAGE OF 2.24 Section 111.1discussed the overall trends inRULC. Here we focus on the sectoral distribution o f RULC-Le. Senegalese comparative advantage. 2.25 As noted earlier, SenegaleseRULC rose sharplyprior to the devaluation andthen improved temporarily. One exception to this trend i s the food sector which continued to suffer rising RULC even after the devaluation against almost all the other regions. Relative to Africa, Asia and Latin America, the average RULC for food increased from 1.47 to 2.26, 1.10 to 1.42, and 2.82 to 3.26 respectively, between 1986-93 and 1994-99. 2.26 The food sector includes such diverse products as beverages, fishing and fish- processing, vegetable oils, grains and flours, and sugar and candies. Some o f these are export-oriented like fishing, whereas others like sugar are inefficient and heavily protected, while still others such as beverages have quasi-monopolies. The U N D O data does allow some disaggregation by separating food and beverages. When we look at these two separately, we see that food products have considerably improved competitiveness after the devaluation, whereas beverages' RULC rose sharply, against all four comparisonregions. 2.27 All other sectors exhibited quite similar behavior over the sample period. In 1974, Senegalese RULC for most sectors, against all other countries, was uniformly around 0.6 to 0.9, except for the metal industry, a large outlier at 2.24. As noted earlier, RULC rose sharply for all sectors over the next 20 years. The textile and clothing sectors saw particularly sharp increases in RULC between 1974 and 1993 such that in the latter year, these two sectors had the highest RULC, along with the metal industry. Following the devaluation, RULCs again converged, with the more labor intensive sectors showing some o f the biggest improvements (clothing, textiles, wood). Still, RULC remained above 1.O even for these sectors inthe years following the devaluation. B.2. THEEVOLUTION THE COMPONENTSOFRULC OF 2.28 As noted earlier, RULC can be decomposed into exchange rates, relative wages, and relative productivity. Here we study the evolution o f these factors by industry and country. 40 B.2.1. RELATIVE WAGES DOLLARS IN 2.29 Senegalese nominal wage behavior has actually been quite moderate. Senegal's low rate o f nominal wage increase reflects the fact that wages are largely determined by two factors: inflation and labor productivity. Neither o f these elements increased on sustained basis in Senegal, but this i s not the case for many other countries. Asia experienced rapid productivity growth entailing large real wage increases whereas many Latin American countries hadbouts o fhigh inflation. 2.30 It is more meaningful to compare relative wages in dollars, which conflates the movements o f relative wages in local currency and the exchange rate into a single measure, as shown inTable 2-7. Usingthis indicator, Senegal's relative wages increased rather than decreased over the 1974-93 period. Senegalese wages in dollars rose by 70 percent against other African countries, 38 percent against Latin American countries, and 39 percent against European countries, but declined 62 percent against Asia between 1974 and 1993. 2.31 This rise in Senegaleserelative wages is mainly due to the CFA franc, which, has appreciated sharply in nominal terms against other countries over time, and the devaluation only temporarily attenuated this trend. After 1994, the CFA resumed appreciating vis-a-vis most currencies o f the comparison countries. For example, the CFA appreciated by 79 percent, 8 percent and 41 percent again the currencies o f Malawi, Mauritius, and Zimbabwe between 1994 and 1997. This i s because these countries also devalued their currencies during this time, as they had a number o f previous times inthe 1980s and 90s in response to internal crises. Malawi, for example, had a series o f devaluations in the 1 9 8 0 ~ ~ which resulted in 1987 in a 75 percent decline over the 1979 level, as the authorities targeted a real exchange rate o f25 percent the 1979 level (Mulaga and Weiss 1996). Other devaluations followed in 1992 and thereafter. More generally, from the 1980/85 base period, while the CFA franc depreciated by 38 percent in 1994-96, other African currencies depreciated by 65 percent (Adams et al. 2001). For the European countries, there is a similar story. Turkey has floated its currency since 1990, with a depreciation of 70 percent in 1994, following a financial crisis duringwhich it lost more than half its international reserves (Ozatay 2000). InEastern Europe, the transition o f the early 1990s gave rise to economic upheavals accompanied by large devaluations. Poland devalued by 81 percent in 1990 before opting for a crawling peg and further devaluations. Hungary also devalued in 1993 and moved to a crawling peg in 1995. 2.32 The Asian countries, by contrast, had stable exchange rates until the crisis o f the late 1990s. Singapore, for example, has pegged to the dollar since 1981 andhad only one period o f devaluation in the middle o f the 1980s, after which its currency has regularly appreciated innominal terms. The Asian crisis, however, entailed sharp depreciations o f many East Asian currencies, sharply worsening Senegalese competitiveness. 41 B.2.2. RELATIVEPRODUCTIVITY 2.33 So far we have examined the role o f relative wages in manufacturing and noted that while Senegalese nominal wages have been restrained, Senegalese competitiveness suffers from the constraint o f the fixed exchange rate o f the CFA franc. We turn now to relative productivity inthe same industriesand comparison countries (Table 2-8). 2.34 We can see that Senegal's productivity has declined relative to the bulk o f the comparison groups. For the African countries, while there are ups and downs, there i s a very clear downward trend in Senegalese relative productivity. While Senegal's relative productivity has exceeded 1 compared to the African average, it has almost halved between 1974 and 1999. This i s the result o f a sharp decline in Senegalese productivity and a concurrent positive growth o f the other African countries' productivity. 2.3 5 Relative to the European countries, Senegalese relative productivity has also trendeddownward and fallen below the 1.O benchmark, even though these countries have had serious difficulties o f their own with transition and structural adjustment in the 1990s. 2.36 But it is especially incomparisonto the Asian and Latin American countries that Senegal's relative productivity has dropped most dramatically. In the manufacturing sector as a whole, Senegalese productivity in 1998 fell to 28 percent o f the Latin American averageI4 and o f the Asian average, compared to 77 percent and 182 percent in 1974. As is well known, the Asian "dragons" have had spectacular growth o f manufacturing productivity, which increased by 2. 7 times in Hong Kong, 7.9 times in Korea and 3.4 times in Singapore." Chile's productivity more than doubled despite a fairly highinitial level in 1974. Inthese two regions, the weakest performances are those of India and Mexico. 2.37 Ifproductivity has so dramatically decreased relative to comparison countries, it i s mainly because it has sizably diminished for almost all manufacturing sectors. Indicators o f Labor productivity and TFP levels and variations have been computed for a broader level o f aggregation o f the Senegalese manufacturing sector: 15 sectors. Figure 2-1 displays the evolution o f labor productivity and TFP for the overall manufacturing sector. If we take labor productivity, the majority o f sectors display a decreasing trend across years from 1974 onwards. For total industry, there i s an average annual decrease rate o f 3.41% between 1974 and 1998. However, there are some outliers that show an increasing trend o f labor productivity duringthe sample period. These are: grain and flour (3.48%), sugar and candy (6.46%), and building materials (1.98%). The energy industry has a stagnant productivity over the same period (0.64%). Apart from those sectors, all the remaining display a decreasing trend in their productivity level with the worst cases being paper, printing, textile, and apparel. Ifwe now consider TFP evolution during the sample period, we still have a decreasing trend, with a negative annual growth rate o f l4 The drop i s especially marked vis-a-vis Chile. l5 The latter benefitedfrom more favorable initial conditions, with a higher level of productivity. 42 3.21% for total industry. But the number o f outliers narrows down to two: the building material and sugar and candy industries. And for total industry, labor productivity dropped in 1999 to only 68% o f its 1974 level, while TFP only amounted to 37% o f its 1974 level inthe same period. (Mbaye 2003). 2.38 In summary, if we consider the 3 components o f RULC, only wages in local currency have evolved in a relatively favorable manner for Senegal, reflecting relatively low inflation in the UEMOA region16. In any event, this variable i s not o f interest in itself since it does not consider the exchange rate. It i s rather relative wages in dollars along with relative productivity that determine RULC. For both o f the latter, Senegal has had a highly unfavorable experience. The productivity performance i s particularly dismal. Not only has it declined relative to other countries, but even in absolute terms, Senegalese productivity has dropped sharply, and has been almost halved between 1974 and 1999. If the country wants to improve competitiveness, clearly productivity i s the decisive factor to address. 2.39 Table 2-7 gives the results o f our calculations. It shows that the human capital variable has a positive and significant impact on productivity. Its coefficient i s significant at the 1% level both on the level and the growth rates o f TFP and o f labor productivity. In addition, this impact i s found to be robust to various specifications. The export variable i s also significant at the 1% level and robust, implying that sectors that are more open to trade are also the most competitive ones. Regarding our control variables, they have a mitigated impact on productivity. While the profitability variable i s found to be significant, robust, and positively signed, the financial variable is found to be very fragile with a level o f significance and a sign varying a lot with specifications. That, of course, does not mean that financial constraint has no effect on sector productivity; this result could also be due to our proxy variable not fully capturing such financial constraint. Another factor that i s highly significant on labor productivity i s not surprisingly the capital-labor ratio variable. It i s significant with the expected sign inall equations having labor productivity as a dependant variable. POLICYIMPLICATIONS 2.40 Human capital is certainly one o f the most important, if not the most important, factor determining output and productivity growth. Not only i s humancapital a feature in "growth accounting" but recent theories o f growth have emphasized knowledge accumulation and education as an impetus to gr~wth'~.Consequently, the first place to ~ l6 Senegal as other countries of the CFA zone, faces a dramatic trade-off betweenalow rate of inflation andthe lack of flexibility inexchangerate management. See DevarajanandRodrik (1992) for moredetails on this issue. l7 While some recent empirical work has questionedthe strongcausallinkbetweeneducationand growth (e.g., Sacerdotiet al. (1998), Pritchett (2001), Behabib and Spiegel(1994)), a consensus is emergingthat this i s due to poorly-measuredvariables (Krueger and Lindahl(2000), Mbaye (2002)). The linkbetweeneducationandgrowth andproductivity is strongandrobust(Cohen andSoto (2001), Jones (2001). 43 look for improving Senegal's productivity i s in its human capital base. One short- to medium-term policy would be increasing training o f its workers. 2.41 Training. The shortage o f intermediate-skilled technicians i s a key bottleneck in some sectors such as textiles/clothing and fishing. The public training institutes are under-funded and the private sector has not been able to provide viable alternatives. The government should work with the private sector to determine the most efficacious way o f remedying this situation, by restructuring the existing institutes and assisting trade associations informing private institutions. Other cross-cutting bottlenecks toproductivity growth 2.42 Infrastructure. Infrastructure i s essential for industrial development. In Senegal much o f the infrastructure is in poor condition and service i s often unreliable yet expensive. Most prominently, the frequent power outages are a grave problem. The privatization o f the electricity monopoly SENELEC should be pursued in such a way as to encourage greater investment and the improved quality o f service. The road network needs to be extended such that raw materials from the regions can be brought to Dakar at reasonable cost. Evenmore important, the existing road network should be maintained to stop the ongoing deterioration. Telephone service, provided by the privatized monopoly SONATEL i s adequate due to substantial investments. However, the cost o f local calls on the fixed lines remains well above Senegal's competitors and even many African countries inthe franc zone. Some competition i s inorder. 2.43 Antiquated Plant and Equipment. Senegal has had low levels o f productive investment". After some moderate growth in the 1970s, real investment in manufacturing declined in the 1980s and early 1990s. In 1993, real investment was at 35% its 1983 level. For some industries, the decline i s even more dramatic: for textiles real gross fixed capital formation in 1993 fell to below 20% its 1977 level, while for the vegetable oil industry the corresponding figure i s 31%. However, there has been an encouraging upsurge since the devaluation. This recent increase, though, has not offset the years o f low investment, and much o f the manufacturing capital stock i s antiquated. In the textile industry some equipment dates from the 1950s, Le., the time of French ownership! In the fishing industry, the trawlers have an average age of 23 years, whereas the normal useful life is 10-11years. 2.44 Financing. InSenegal, a number o f enterprises, especially the SMEs, lack access to credit. What credit i s available i s generally short-term. This i s clearly a constraint on investment and productivity growth. We suggest several concurrent paths to remedying this difficult problem. First, entry requirements and prudential regulations should be revisited so as to increase competition. Second, the legal and judicial system should be '*O'Connell andNdulu(2000) find capital formation as playing a central role inpoor economic performances of African countries. 44 changed to facilitate action against delinquent debtors. Doing so would surely increase the supply o f credit. Third, decentralized financial institutions should be encouraged to expand their horizons from micro-credit towards greater mobilization o f saving and loans to SMEs. One way to do this might be to situate some o f the lines o f credit provided by donors, such as the World Bank's APEX and the African Development Bank's development funds, inthese institutions. 2.45 Institutional Problems. While the tariff structure has been reduced and simplified, firms still complain about delays in obtaining duty reimbursements or exemptions that they are entitled to. Maximum time limits should be established for customs clearance. It might also be advisable to shift the decision about preferential tariff treatment to Customs, which i s in close contact with the firms in question, from the Ministry o f Finance, to speed the process. 2.46 Access to foreign markets. Senegalese firms often lack knowledge o f quality norms in U.S. and European markets, and the ability to satisfy these norms. An enhanced export promotion agency could help. Such an agency's mandate would cover the supervision and possible improvement o f export incentives, provision o f information about foreign norms and regulations to local firms, and assistance with making contacts with foreign firms. 2.47 Labor Relations. While some o f the highly restrictive labor laws have been eased, there i s still often an adversarial relation between unions and management that contributes to reduced productivity and lowered employment. Labor laws should be further reviewed with the aim o f increasing the incentive to hire rather than to protect existing workers. The govemment should also spur a spirit o f cooperation by both firms and workers. One way to do this is to involve workers more in the operation o f the enterprise andto provide profit-sharing bonuses. CONCLUSIONSAND RECOMMENDATIONS 2.48 In this study, we have sought to understand the obstacles to international competitiveness o f Senegalese manufacturing. Relative unit labor costs and relative producer prices have been calculated for Senegal relative to other developing countries in Africa, Asia, Latin America and Europe. These real effective exchange rate calculations show a steady deterioration o f Senegalese competitiveness from independence up to the 1994 devaluation. The devaluation entailed a temporary improvement and consequent boost to exports, but the deterioration resumed shortly thereafter. Even though wages rose relatively little indomestic currency terms, relative wages indollars increased much more inSenegal than its competitors. Concurrently, Senegaleseproductivity fell sharply, not only relative to other developing countries, but even inabsolute terms. 2.49 Raising productivity i s the best and only sustainable way to improve competitiveness. The membership in the CFA zone precludes use o f the exchange rate, and lowering wages is difficult and painful. Raising productivity i s also desirable in itself as it i s the only way to improve the standard o f living in the long run. To raise 45 productivity and improve competitiveness, Senegal will have to address a number of fimdamental constraints relating to infrastructure, training and other assistance to small and medium sized firms, financial markets, and labor relations. 2 Y 3 Table 2-1: Real Output Average Annual Growth (inpercent) Sectors 74-80 81-93 94-99 Miscellaneous Food Industries 10.01 -6.06 6.70 Fish-Processing 2.16 3.85 6.22 Oilseeds and products -9.55 4.12 -1.50 Grains and Flours -2.11 -1.60 12.57 Sugar and Candy 10.82 3.02 -4.84 Tobacco 0.42 1.73 3.32 Textiles 1.70 12.37 5.34 Apparel 1.09 8.18 5.21 Wood 14.33 7.13 14.23 Paper 2.32 -6.23 7.39 Printing 2.03 -1.50 11.87 Chemicals 5.35 1.58 12.46 ConstructionMaterials 6.18 -2.03 7.67 Metal-working industries 0.49 4.90 6.21 Energy 8.23 -2.95 1.47 TOTAL INDUSTRY 0.01 1.79 5.05 Table 2-2: Comparative Manufacturing Performance for Senegal and Selected Other African Countries (inpercent) 1980 1987 1994 1998 1999 2000 Manufacturing, value added (%of GDP) Senegal 10.58 12.84 14.37 15.90 16.87 17.77 Botswana 5.27 6.18 4.62 4.85 4.97 4.97 Burkina Faso 15.85 17.35 20.74 15.91 13.18 12.21 Cote d'Ivoire 12.81 18.78 16.47 19.55 20.48 19.11 Ghana 7.81 9.88 9.11 8.99 9.02 9.02 Manufacturing, value added (annual % growth) Senegal -4.49 8.93 -4.70 7.81 4.89 4.80 Botswana -11.39 24.36 1.48 5.90 6.00 6.00 Burkina Faso 9.24 6.67 1.66 1.72 8.28 5.02 Cote d'Ivoire 1.44 0.35 -1.10 0.82 -7.23 Ghana -1.37 10.01 1.65 4.08 4.61 3.74 Source: WDI 2002 48 Table 2-3: Real Manufacturing Exports Average Annual Real Growth Rate - (inpercent) Sectors 197483 198493 199499 Miscellaneous Food Industries -1.10 4.51 -8.34 Fish-Processing 16.98 2.56 39.39 Oilseeds and products -0.80 -13.55 4.55 Grains and Flours -31.82 23.50 -34.66 Sugar and Candy -0.63 -0.03 -3.33 Tobacco -11.06 -1.02 -1.56 Textiles 27.29 -5.32 4.02 Apparel -6.02 -15.84 6.93 Wood 4.97 6.30 -2.61 Paper -7.10 5.07 -0.79 Printing -4.98 0.01 54.95 Chemicals 18.63 1.51 9.55 Construction Materials -3.14 19.85 -34.45 Metal-working industries -3.61 2.78 3.12 Energy na N a na TOTAL INDUSTRY 3.15 -1.04 12.47 Table 2-4: Sector Shares in Total Real Manufacturing Exports (inpercent) Sectors 1974 1980 1993 1994 1998 1999 Miscellaneous Food Industries 0.96 1.07 2.20 1.41 0.32 0.09 Fishingand Fish-Processing 11.78 28.50 39.13 14.30 34.69 40.61 Oilseeds and products 56.28 24.23 8.23 18.74 16.51 12.63 Grains and Flours 2.24 0.17 0.46 1.68 0.37 0.11 Sugar and Candy 0.76 0.94 0.51 0.77 0.15 0.35 Tobacco 0.56 0.51 0.30 0.24 0.09 0.12 Textiles 2.42 3.66 5.50 7.21 4.80 3.17 Apparel 2.39 2.08 0.41 1.01 0.86 0.76 Wood 0.42 0.35 0.35 0.12 0.07 0.06 Paper 6.36 5.30 2.93 2.67 1.28 1.39 Printing 0.02 0.02 0.01 0.00 0.00 0.00 Chemicals 13.02 31.68 34.30 46.70 40.10 39.76 ConstructionMaterials 1.39 0.10 4.62 4.06 0.15 0.26 Metal-working industries 1.41 1.39 1.04 1.09 0.62 0.69 Energy na na na na na na TOTAL INDUSTRY 100 100 100 100 100 100 49 Table 2-5: Relative Unit Labor Costs - Senegal Against Comparison Countries 1974/80 1981/85 1986193 1994 1994199 Senegal/ African Countries Total 0.49 0.65 1.20 1.16 1.45 Food 0.56 0.65 1.47 1.45 2.26 Textile 0.60 0.86 3.77 2.27 2.58 Clothing 0.56 0.77 1.55 0.99 1.11 Wood 0.62 0.63 1.31 1.06 1.30 Paper 0.60 1.12 1.51 1.26 1.31 Printing 0.37 0.59 1.22 1.13 1.22 Chemical 0.50 0.96 1.28 1.04 1.13 Metalindustry 0.49 0.77 1.67 1.50 1.42 Senegal/ Asian Countries Total 0.67 0.69 1.16 0.90 1.25 Food 0.65 0.64 1.10 0.92 1.42 Textile 1.28 0.87 3.50 1.61 1.49 Clothing 0.81 1.22 2.87 1.48 1.47 Wood 0.76 0.57 1.30 0.73 1.05 Paper 0.38 0.73 1.17 0.97 1.41 Printing 1.06 0.94 2.09 1.82 1.83 Chemical 0.61 0.97 1.38 0.80 1.17 Metalindustry 4.59 3.84 6.98 4.95 4.65 Senegal/Latin Countries Total 1.34 1.32 2.74 1.49 2.44 Food 1.15 1.28 2.82 1.75 3.26 Textile 0.76 0.93 4.09 1.87 2.04 Clothing 1.30 1.37 3.39 1.36 1.52 Wood 1.34 1.41 3.26 1.24 1.87 Paper 0.93 2.13 4.28 2.08 3.51 Printing 0.94 0.78 2.10 1.48 1.74 Chemical 0.69 0.98 1.84 0.94 1.47 Metalindustry 1.05 1.10 2.72 1.76 2.37 Senegal/Transition EuropeanCountries Total 0.81 0.99 1.76 1.02 1.37 Food 0.70 0.89 1.88 1.25 1.81 Textsle 0.74 0.95 3.98 1.73 1.51 Clothing 0.76 1.06 2.10 1.05 1.34 Wood 1.01 1.02 2.45 0.81 1.20 Paper 0.73 1.24 1.94 0.99 1.38 Printing 0.81 0.82 1.91 2.02 1.83 Chemical 0.89 1.33 1.67 0.78 1.00 Metalindustry 0.95 1.17 2.27 1.31 1.41 ~~ 50 Table 2-6: Relative Producer Price Senegal Against Comparison Regions - 1975180 1981185 1986193 1994 1994199 Senegal/ AfricanCountries Total 1.16 1.12 1.94 1.87 1.87 Food 0.87 0.80 1.49 1.00 1.22 Tobacco 0.82 0.68 1.73 1.42 1.47 Textile 0.58 0.43 0.95 1.02 1.16 Clothing 0.75 0.64 1.25 1.27 1.37 Wood 2.11 0.70 0.62 0.82 1.50 Paper 1.72 1.17 2.98 4.09 5.04 Printing 1.54 2.28 7.19 3.33 7.37 Chemical 1.22 1.78 2.83 1.60 2.27 Metalindustry 0.89 0.74 0.97 0.77 0.97 Senegal/Asian Countries Total 1.22 1.15 2.40 2.17 2.99 Food 1.04 1.04 2.65 1.91 2.56 Tobacco 0.87 1.01 5.91 7.38 3.83 Textile 0.79 0.54 0.80 0.67 0.87 Clothing 0.92 0.72 1.28 1.12 1.50 Wood 1.96 0.47 0.54 0.56 0.96 Paper 1.30 0.79 1.51 1.77 2.98 Printing 1.17 1.01 3.24 1.77 3.24 Chemical 1.43 2.25 4.44 2.56 3.96 Metalindustry 0.92 0.62 1.16 0.86 1.19 Senegal/LatinCountries Total 1.56 1.26 2.10 1.53 1.66 Food 1.03 0.85 1.30 0.68 0.94 Tobacco 1.13 0.85 2.22 1.44 2.01 Textile 0.98 0.72 0.73 0.60 0.99 Clothing 1.15 0.88 1.14 0.88 1.19 Wood 1.83 0.46 0.31 0.27 0.57 Paper 1.50 0.87 1.30 1.60 2.65 Printing 1.13 1.26 3.34 1.91 3.17 Chemical 1.35 1.63 1.78 0.66 1.02 Total 0.90 0.66 0.65 0.41 0.47 Senegal/Transition EuropeanCountries Total 0.95 0.86 1.01 0.80 0.87 Food 0.86 0.97 1.33 0.79 0.96 Tobacco 0.88 1.03 2.26 2.51 3.20 Textile 0.64 0.53 0.68 0.46 0.62 Clothing 0.73 0.63 0.78 0.57 0.70 Wood 1.80 0.77 0.55 0.49 0.76 Paper 1.04 0.90 1.32 1.66 2.60 Printing 0.99 1.37 1.62 0.38 0.60 Chemical 1.07 1.89 2.43 0.99 1.59 Metalindustry 0.79 0.57 0.61 0.35 0.41 51 Table 2-7: Relative Wages in $ - Senegal Against Comparison Regions 74/80 81I85 86/93 94 94/99 Senegal/ AfricanCountries Total 2.23 2.33 3.37 2.03 2.27 Food 2.26 2.09 2.94 1.61 2.16 Textile 2.99 2.69 2.74 0.93 2.33 Clothing 2.76 2.58 2.95 1.26 2.31 Wood 2.49 1.90 2.58 1.32 1.67 Paper 3.01 2.47 3.00 1.78 1.54 Printing 2.01 1.48 1.45 1.18 0.98 Chemical 1.64 1.48 2.01 1.30 1.39 Metalindustry 2.12 2.00 3.31 1.75 1.64 Senegal/ Asian Countries Total 1.40 0.82 0.69 0.25 0.24 Food 1.16 0.62 0.55 0.17 0.19 Textile 3.68 1.49 0.80 0.12 0.20 Clothing 2.38 2.06 2.58 0.62 1.96 Wood 1.34 0.54 0.43 0.13 0.15 Paper 1.37 0.76 0.72 0.24 0.23 Printing 2.80 1.15 1.09 0.53 0.46 Chemical 1.07 0.71 0.60 0.23 0.23 Metalindustry 7.46 4.22 3.56 0.91 0.79 Senegal/Latin Countries Total 1.20 0.70 1.26 0.49 0.61 Food 1.19 0.90 1.63 0.50 0.73 Textile 1.30 1.12 1.35 0.24 0.55 Clothing 1.27 1.00 1.33 0.31 0.57 Wood 1.38 0.84 1.36 0.39 0.55 Paper 1.02 0.74 1.36 0.42 0.55 Printing 1.19 0.56 1.15 0.55 0.62 Chemical 0.87 0.67 1.05 0.40 0.50 Metalindustry 0.98 0.83 1.41 0.40 0.47 Senegal/Transition EuropeanCountries Total 1.73 1.96 2.47 1.03 1.15 Food 2.00 2.06 2.65 0.86 1.07 Textile 1.94 2.26 2.18 0.51 1.16 Clothing 1.88 2.17 2.27 0.66 1.15 Wood 1.68 1.51 1.89 0.61 0.71 Paper 2.13 2.03 2.81 0.83 0.81 Printing 2.23 1.86 2.30 1.37 1.27 Chemical 1.75 2.08 2.45 0.95 1.07 Metalindustry 1.77 2.19 2.69 0.82 0.83 52 Table 2-8: Relative Productivity - Senegal Against Comparison Regions 74/80 81/85 86/93 1994 94/99 Senegal/ AfricanCountries Total 2.80 2.13 1.99 2.06 1.73 Food 1.82 1.65 1.88 1.44 1.05 Textile 2.39 2.59 0.88 0.17 0.18 Clothing 3.76 1.81 1.04 0.63 0.82 Wood 3.06 3.52 1.44 3.18 2.89 Paper 5.45 4.25 6.81 3.76 1.91 Printing 1.95 2.70 1.54 0.66 1.21 Chemical 5.90 2.14 1.20 4.01 1.95 Metal industry 0.79 1.05 0.65 0.55 0.62 Senegal/ Asian Countries Total 2.69 1.60 1.04 0.99 0.76 Food 3.52 1.92 1.45 1.01 0.80 Textile 3.21 4.52 0.66 1.43 1.20 Clothing 3.34 2.07 0.68 0.28 0.42 Wood 2.37 2.40 0.68 0.95 0.79 Paper 6.30 3.84 2.95 2.81 1.28 Printing 2.76 2.84 1.24 0.49 0.79 Chemical 6.16 2.28 1.30 3.14 1.80 Metalindustry 1.06 0.97 0.56 0.37 0.41 Senegal/ LatinCountries Total 0.71 0.41 0.33 0.37 0.27 Food 0.58 0.39 0.39 0.31 0.24 Textile 1.16 0.60 0.27 0.21 0.21 Clothing 1.36 0.69 0.25 0.14 0.18 Wood 0.73 0.71 0.31 0.43 0.31 Paper 2.21 1.05 1.15 1.08 0.52 Printing 0.58 0.38 0.35 0.17 0.24 Chemical 1.61 0.60 0.56 1.67 0.87 Metalindustry 0.49 0.50 0.25 0.18 0.22 Senegal/Transition EuropeanCountries Total 1.54 1.25 1.09 1.38 1.11 Food 1.53 2.08 1.39 0.92 0.75 Textile 0.99 1.11 0.19 0.38 0.32 Clothing 2.06 1.41 0.69 0.54 0.96 Wood 1.60 1.93 0.90 1.87 1.47 Paper 2.81 2.38 2.18 2.72 1.15 Printing 1.43 1.95 1.13 0.62 0.86 Chemical 3.14 2.02 1.28 4.16 2.07 Metalindustry 0.67 0.99 0.72 0.67 0.91 53 Table 2-9: Estimationof TFP andLPLevels and GrowthRates Dependant PROF FFIN OPEN ACAPLA HK N R2 Variable -ATFP 0.008 0.003 0.014 368 0.67 (2.80)* (3.33)* (4.47)* -ATFP 0.010 -0.001 0.004 368 0.67 (3.06)* (-7.14)* (4.47)* -ALP 0.97 -0.08 2.65 0.34 368 0.13 (2.81)* (-0.96) (11.37)" (2.58)* -ALP 0.89 0.29 1.075 2.65 368 0.13 (2.72)* (1.98)** (2.58)* (11.37)* -TFP 0.31 -0.05 0.105 432 0.67 (5.53)* (-6.87)* (5.03)* -TFP 0.293 0.057 0.329 432 0.67 (5.08)* (2.48)** (5.03)* -LP 7.271 1.628 5.569 1.206 416 0.65 (4.62)* (9.06)" (10.88)" (7.00)" -LP 7.661 -0.320 1.206 1.782 416 0.65 (4.91)* (-3.32)* (7.00)* (10.88)* t-stat inparentheses, significant at 1% level, *at 5% level. Variables are taken inlog. A is a variation operator. * 3. REGULATORY REFORM, MARKET PERFORMANCE, AND POVERTY REDUCTIONINSENEGAL A. INTRODUCTION 3.1 Senegal's major challenges over the next few years will be to systematically unwindthe extensive state involvement inthe economy, discourage entrenched habits o f rent-seeking, build new capacities in the public administration, and create the market-based regulatory regimes and institutions that will support investment, innovation, and vigorous competition. In future, the strengths o f Senegal's entrepreneurs - clearly demonstrated by the thriving informal sector - could become the driving force o f the modem Senegalese economy. Already, risk-taking and innovation in the private sector are gradually improving, illustrated by the growth o f the informal sector and the increasing readiness of the best performing informal businessesto enter the formal sector. 3.2 Senegal began to liberalize its domestic markets almost a decade ago, following an economic crisis in 1994, and since then has expanded regulatory reform, both economy-wide and inkey sectors. Like other countries inthe region, Senegal has gone furthest with the deregulation and privatization components of its regulatory reform agenda. The monopolies, quantitative restrictions, entry barriers, exclusive import licenses (such as the state monopoly on rice imports), and other schemes that characterized production prior to 1994 have largely been eliminated. Electricity and telecommunications have been liberalized and other sectors are to follow. A competition law and an investment code outline a basic market framework. Regulatory reforms include a substantial part o f the larger structural program -market opening, financial sector restructuring, education and health reforms, regulation o f network industries, competition policy, commercial law, labor law, and corporate governance -- that i s establishing the scope and framework for market competition in goods and services. 3.3 Yet more remains to be done to reform regulatory processes, and even the progress made is vulnerable to reversal. Commitment to reform has not been robust. In part due to the historical political events of 2000, Senegal's first democratic transition from one party to another, followed by legislative elections inMarch 2001, political uncertainty has delayed decisions and implementation often faltered. A deeper underlying problem i s that the appropriate roles o f the state and the market have never been clearly distinguished, and patterns o f formal and informal intervention are similar to those before market liberalization. An IMF study in 1999 found that, although Senegal i s characterized by relative political stability, "this did not translate into consistent or sustained set o f economic reform policies, but rather into a gradual blurring o f the line between party and state, which has facilitated the intrusion of the state into the e~onomy."'~ This can be seen in important sectors such as groundnuts and textiles. 3.4 The regulatory and institutional environment i s still inadequate for private sector development. Despite average growth rates o f over 5 percent for the past l 9 Senegal: 2001 Article IV Consultation.IMF Country Report no.01/186. 55 several years, Senegal suffers from lack of investment and job creation inthe private sector, particularly the formal sector. It i s striking that little or no employment has been created in the formal sector. The formal private sector has remained constant in size, its energies focused less on development than on elaborating structures o f cross- holdings and maintaining lines o f communication with government, which still controls the development o f key sectors. Foreign investment i s below the African average, and production costs are high due to costs of inputs. The country's infrastructure deficits in electricity, rail, air, seaport and road have acted as a brake on the whole economy. The informal sector, where the country's entrepreneurial energies are emerging, faces daunting constraints on productivity growth for both capital and labor. Much o f the economy i s simply unaffected by market liberalization. Poverty rates remain stagnant at over 50 percent, and Senegal still ranks among the poorest countries in the world, with GDP per capita around US$600. The country's poverty reduction strategy o f February 2002 concluded that "growth remains fragile.. .the situation o f the primary sector i s precarious due to stagnation in the second sector, which i s insufficiently competitive, and this i s an important constraint onjob creation and revenue growth." This report corroborates that conclusion. 3.5 There are two contexts for assessing the state o f market competition in Senegal. State firms that are not privatized and the heirs to the previous state monopolies still dominate the large-scale picture, in sectors such as vegetable oil, sugar, and utilities. Here, the long history o f central control of the economy appears to support a tendency to maintain protection under new guises, despite the formal steps taken toward liberalization. In distribution and other industries, the potential for competition appears more promising. Regulatory barriers are falling, and the economy i s evolving toward higher-value activities as firms make the transition from informal trade to the formal sector, and from importation to assembly and then manufacturing. As domestic markets deepen, competition will become more important, as will the incentive o f the private sector to try to control markets. Already, there are persistent, though anecdotal, reports o f conditions and phenomena in distribution and manufacturing that imply the presence o f privately-imposed constraints on pricing or other aspects o f competition. It i s likely that such private arrangements perpetuate some of the old public policies o f control. Because there have been no serious investigations or actions by the competition authority, there i s little reliable evidence about these problems. The lack o f visible gains for the poor o f market liberalization, discussed below, could result partly from market power that captures the gains from trade. 3.6 What contribution have past regulatory reforms made to the economy o f Senegal? This question cannot be fully answered yet, since reforms are recent and effects are still working their way through the economy. Several years are needed for these reforms to significantly restructure the economy and change the relationship between investment and output. Regulatory and institutional change o f this magnitude prompts resistance from incumbent interests, makes heavy demands on administrative systems, and engenders short-run costs, such as increases in unemployment and prices, before longer-term benefits appear. Senegal's reform program has encountered each of these problems. Even if the program advances on schedule, it will require several years for Senegal to adapt to the demands o f a modem economy. Adopting 56 reform strategies i s easier than actually changing the laws, and changing the laws i s easier than changing institutional, behavioral, and cultural traditions. 3.7 More importantly, reform has been tentative and delayed, which has limited the impacts of regulatory reform. The major causes o f the solid economic growth in recent years are not structural reform, but other factors, including: 0 The CFA franc devaluation: In 1994, the externalvalue of the CFA franc was halved through a change inthe fixed exchange rate with the Frenchfranc. This gave a competitiveness boost not only to Senegalbut to the other countries o f the CFA franc zone. 0 Senegal's continuedlow inflationhelpedto maintainthat competitiveness and to improve business confidence. 0 Senegal, becauseo f its geographical position, benefited from growth inthe regionas a whole, including by acting as a distribution point for other countries inWest Africa. 0 Buoyant informal sector activity boosted growth through improved consumer spending. 0 Deterioration inpublic finances as a result of loss-making public enterprises was contained through improved revenue intake and external assistance. 3.8 IfSenegalistosustaintheselevelsofgrowthinthemedium-term, anew commitment i s needed to create the conditions for domestic competitive markets. Good regulation can play a key role inbringingabout a more dynamic and innovative private sector that can contribute to poverty reduction through new business opportunities, higher incomes, provision o f essential services, lower costs, and increased employment opportunities. Better and cheaper health care, transport, water and telecommunications can all contribute directly to living standards and the quality o f life. 3.9 Deregulation and privatization should be completed, but these reforms are not sufficient to drive growth. The more difficult aspects o f regulatory reform include creation o f an enabling environment for private enterprises, protecting vigorous and fair competition that rewards entrepreneurs, drives innovation and yields benefits for the consuming poor, re-regulating carefully where markets do not work, creating new market disciplines and regulatory regimes, and building institutions and public sector reforms based on values o f simplification, transparency, and neutrality. B. A BRIEFHISTORY REGULATORYREFORMINSENEGAL OF 3.10 Measures already taken in Senegal to improve the efficiency o f the economy have included economy-wide measures, sectoral measures, and institutional change. Table 7 presents an overview, while more details o f the changes and issues in key sectors are givenlater inthis report. 3.11 From 1989, there were changes inpolicy, but the turning point came in 1994 with the devaluation of the CFA franc. Devaluation stimulated import substitution, and drew attention to the need to create a more favorable investment climate to 57 complement the new incentives to investment. At that point, market liberalization began ineamest as part o f a general structural adjustment policy. 3.12 Progress has been greatest in reforming price controls, one o f the most damaging forms o f regulation. The good performance of the consumer price index has been noted, and in fact, as competition has increased, some prices dropped inrecent years, including for some food products and printing. Price deregulation has not unleashed inflation, because it was accompanied by more competition in many sectors. About 13 products continue to have price controls, eleven o f which are categorized as "essential products." Price rigidities are noted where govemment controls are still in place. Petrol price increases, advised by a commission that includes private sector membership, have pushed up costs in other sectors such as transport, and the experience i s that the prices are flexible upwards only, and that they do not reduce when the petrol price itself i s lowered. This intum indicates some anti- competitive activity in this sector. In other price-related developments, there have been, partly due to legislative elections, delays in implementingthe single rate VAT and reintroducing the pass-through mechanism for retail prices o f petroleum products. 3.13 In general, though, progress has been uncertain. In 2000, "the new administration had inherited the legacy o f weak implementation o f structural reforms, notably inthe groundnut and energy sector".20Progress onprivatization i s slow due to a tendency to agree on programs o f privatization, but to delay their implementation. MIGA reports that in June 2001, out of 20 companies that initially constituted the privatization portfolio defined in the 1994 privatization program, 7 were still not privatized, among them SONACOS (groundnut oil refinery) and SODEFITEX (cotton). In2000, the only privatization transaction that was recorded was the Sociktk Nationale des TransportsABriens (SONATRA, international air transport), which was bought by Royal Air Maroc. MIGA lists 11major companies as still to be privatized. 3.14 Other problems are evident in the post-privatization phase. Inprinciple there are no subventions to industryin Senegal. Subventions are forbidden by the UMEOA treaty. In practice, considerable supports have been given to two state owned companies through effective guarantees by the govemment o f their substantial debts. One o f these i s the heavily indebted groundnut company SONACOS. Its total debt amounted by May 2001 to 3.5 percent o f GDP. Attempts to privatize SONACOS in the past were unsuccessful, because investors were not interested in the conditions attached to the privatization, which were that the company should continue to provide seeds and fertilizers to producers on credit basis and maintain the integratedJiliBre structure. Privatization of SONACOS i s now planned for "after the 2002/2003 groundnut campaign". 20Senegal:2001Article IV Consultation.IMF Country Reportno.01/186 21Senegal:2001Article IV Consultation.IMF Country Reportno.01/186, pp.12-13 58 Table 3-1: Summary of structural and regulatoryreformsin Senegal (March 2002) Conditions for good market performance Actions taken, underwav, or to be taken inSenegal Market scope. The scope for free market 1.In1994, legalreformsformalizedtheprivatesector, and mtry and private sector competition onthe opened the economy to internationaltrade. Imports boomed 2asis of price and quality should be broad Senegaljoined WTO, MIGA, and OPIC. The UEMOA Treaty indeconomy-wide. State monopolies reduced obstacles to cross-border trade and restricted state jhould be restrictedto those very few areas aids. OHADA provided a secure business framework. The Nhere a clear publicinterest would be Investment code of 1998, the Investmentpromotion agency served. (APIX), and the Guichet unique for foreign investors have speeded upinvestmentapprovals. 2. Privatizationis underway ininfrastructureservices and other areas. Air Senegal was sold. Progress ina range of othc privatizations has beenuneven. Privatizations have been proposed, butnot completed, for the HBtel Mkridien- President, the urban transport company (the former SOTRAC), the groundnut company (SONACOS), the cotton company (SODEFITEX), the Dakar industrial estate (SODIDA), an industrial promotionandresearchcompany (SONEPI), and a real estate company (SICAP). The Dakar- Bamako internationalrailway was placed under private management beginning inDecember 1999 with the establishment of a new private company (SETI). Studies are underway for privatization of Senegal's airports. At the end of the privatization process, the planis that only utilities, industrial promotion, or social service companies will be publicly owned. 3. In1994, the prices of 24 goods and services, or about 12 percent of GDP, were subject to controls. Reformfreed the prices for about half of these products, accounting for 3/4 of the sales subject to control. But controls remain on 11 products that were deemed essential. 4. The scope of concession monopolies outside the utility sectors has been sharply reduced since the mid-1990s. 5. Some problems still exist incompetition inpublic procurement. 59 I Conditions for good market performance Actions taken, underway, or to be taken inSenegal services. Rules are needed to 1.Poor electricity supply has inhibited growth inthe establish rights and correct market failures economy. The electricity company (SENELEC) is being re- privatized. An independent regulator was established in 2001. A joint venture with Mali and Mauritania for hydro- electric power inOMVSwill go on-stream later in2002. An independent power generator is starting. 2. Intelecommunications, the principal operator (Sonatel) was privatized in1997. A regulator was established 2002. 3. Water and equipment remain state-owned, but supply services were privatised in1996.A regulator has not yet bee established for the sector. The regulator for water and sanitation will be structured inaccordance with the framework utilities law. 4. Transport may be regulated along the lines of the framework utilities law, or may be left to the competition authority. There is little or no economic or technical regulation at present, andsafety problems are evident. 1.Protectionof propertyrightsunderSenegal's civillaws appears to be secure. 2. The judicial process for resolvingdisputes about property parties. issues may be inefficient, however. Factor inputs. Businessesrequire inputs - ~ 1.Laborlawcontinues tobehighlyrigidandunsuitablefor capital, labor, and land.Functioning market economy markets inallocating these resources will 2. Financial sector reformis underway, butsignificant areas stimulate private sector development. suchas SME lending are undeveloped. 3. The system of legalinterests inland appears cumbersome anduncertain, makingitharder to use those interests as security. Market behaviour. Rules for market 1.A generalcompetitionlawaboutrestrictivepracticessuck behaviour should be providedby economy- as price fixing was an important part of the 1994 reforms. Yc wide frameworks such as competition its provisions on protecting competition have not been applied seriously. An initiative at the UEMOAlevelwill shi decision-making responsibility about restraints o n competition to a new regional regime. Competitive market entry is discouraged by tax favors for incumbents, bureaucracy, and insider influence. 2. Senegal's company law provides an adequate framework for organizing formal enterprises. 3. Senegal should consider adopting as formal guidelines f c allregistered public corporations the OECD Principles of Corporate Governance. This could be done quickly and easily, and would provide larger corporations with a benchmark for their behavior. Capable public sector. Reform of state 1.Planningis underway to reformthe state administration. organs is necessary to firmly establishthe 2. Decentralization of governing responsibilities could rule of law through which rights are made worsen the business environment inthe short to medium operational, andto improve governance term, andits implications and the regulatory competences c capacities so that the state can operate localgovernments should be carefully assessed. consistently with a market environment. Source: Jacobs andAssociates, 2002, va )ussources 60 C. IMPROVING AND CAPACITIES IN PUBLIC SECTOR REGULATORY PRACTICES INSTITUTIONS 3.15 The importance o f capable governance for economic development is now widely recognized. Senegal has taken steps ina few areas to buildnew capacities for regulatory governance, such as the creation a decade ago o f the cellule de reflexion sur la competitivitb, part of the Ministry of Economy and Finance, which i s now disbanded; creation and expansion o f the Guichet unique for investors under APIX; and the Centre de Formalite's des Entreprises under the Chambre de Commerce, de 1'Industrie et de 1'Agriculture deDakar (CCIARD), which are intended to cut through thejungle o f administrative barriers. 3.16 However, Senegal falls far behindgood international practices with respect to public sector capacities for market-oriented, reliable, and transparent regulation. In general, market-oriented policy reforms are undermined by poor govemment capacities: 0 Market-oriented policy and legal reforms have generally not been supported by new capacities, practices, and cultures in the public sector, despite a proliferation o f commissions, committees, research bodies, and other institutions (many resulting from uncoordinated donor recommendations). Numerous new programs and institutions have been set up, but they are fragmented, isolated, and unable to change the practices o f the bureaucracy. Senegal suffers simultaneously from institutional overload, and from lack o f capacity to carry out a sustained and effective reformo f policy reform. 0 Except for the utility sectors, there i s little attention inside the public sector to the quality and appropriate use o f regulations. The role o f the state in a free market i s not yet defined clearly enough to guide public officials in day-to-day decisions, and this is one reason why accountability for results continues to be poor. Indeed, the behavior o f Senegal's politicians in freely intervening at the expense o f market performance and incentives provides a negative example to civil servants intheir relationshipwith private enterprises. 0 Commitment to the market i s inconsistent. Problems are often addressed through state-led approaches rather than the market, for example, in the desire in Senegal for an export promotion agency to increase exports, and the replacement o f oil price controls with coordination among distributors and consumers. 3.17 In principle, the governance agenda is recognized as important. Senegal's document on poverty reduction describes la bonne gouvernance, including transparency, decentralization, and the fight against corruption, as a basis o f poverty reduction. It notes that better govemment capacities are needed if its strategies and plans are to be credible. Likewise, the 10th Plan de De'veloppement Economique et Social for 2002-2007 includes among its strategic orientations the promotion o f good governance and the rule o f law, including good governance at local levels. However, most of these governance measures relate to the need to fight corruption and improve public resource management and procurement, that is, the transparency, 61 accountability and efficiency o f public spending and the justice system. There i s little ineither strategy about improving the regulatory capacities ofthe public sector or its ability to serve the needs of a growing domestic market. A wider governance debate on public services was launched in 1999 on an action plan to improve public service organization and operation, and to match the new missions assigned to the administration with the required human and financial resources. In a reform environment with a tradition o f creating institutions that subsequently do not function well, this is an important debate. 3.18 The blame for too much intervention, lies onboth sides: the public sector has a long tradition o f examining and approving individual investment projects as its primary regulatory tool, while the private sector, reasonably, prefers to have government backing and support for each project to reduce regulatory uncertainty. Excessive formalities in Senegal stem in part from a culture o f governance in which businesses face highregulatory risks inthe implementation stage that can be managed only through extensive negotiations and commitments directly with responsible government officials. Investments requiredlongreview and careful negotiation by the public sector. This approach, contrary to market principles, involvedthe public sector deeply ininvestment decisions, increased opportunities for corruption, and slowed the entire investmentprocess. 3.19 Because o f the uncertain administrative environment, simple deregulation could actually slow private sector development in Senegal. Private enterprises would face higher regulatory risks if reduced govemment intervention upfront i s not followed by simplification o f procedures, transparency o f rules, and accountability in regulatory implementation. Face-to-face negotiations can be replaced only with reliable and transparent market rules. 3.20 Another trap must be avoided in trying to increase regulatory certainty. Some governments have adopted detailed procedures to increase certainty and avoid delegationo f broad discretionary powers to regulators. These rules tend to concentrate on a mass o f procedural details ("rights and obligations") rather than on setting down substantive criteria for decisions (policy results). But an accumulation o f procedures tends to increase the arbitrary nature o f administration, because it is difficult to know or comply with all requirements, leaving administrators to decide which rules to enforce, and how. The pursuit o f certainty in regulations can produce so much complexity and detail that it reduces the performance o f the whole system. Again, simplicity, transparency, and accountability are the most important characteristics o f a market-friendly regulatory environment. Enhancing the transparency of laws and regulations through better consultation with theprivate sector and a securelegal registry. 3.21 Transparency is key to regulatory quality.22Inaddition to democratic values o f openness, transparency inregulatory decisions and applications helps to cure many of 22This paragraph is adapted from Jacobs, Scott, "The Second Generation ofRegulatory Reforms," Paperpresentedat IMF Conferenceon Second GenerationReforms,8-9 November 1999, Washington, D.C. The paper can be found at www.imf.orgiextemal/pubs/ft/seminar/1999/reforms/jacobs.htm 62 the reasons for regulatory failures - capture and bias toward concentrated benefits, inadequate information in the public sector, rigidity, market uncertainty and inability to understand policy risk, and lack o f accountability. Transparency at any stage has powerful upstream and downstream effects in the policy process - it encourages the development o f better policy options, and helps reduce the incidence and impact o f arbitrary decisions inregulatory implementation. Moreover, transparency helps create a virtuous circle - consumers trust competition more because special interests have less power to manipulate govemment and markets. Transparency i s also rightfully considered to be the sharpest sword inthe war against corruption. 3.22 Transparency has democratic as well as economic implications. An increase in the activity o f civil society - such as non-governmental organizations (NGOs) -- in many countries has put a higher value on govemment transparency. In Senegal, too, the participation o f civil society has become more important due to an increase inthe number of NGOs. Govemments are seeking to accommodate these changes by developing improved models and approaches for better informing and involving citizens inthe policy-making process. 3.23 Domestic trends toward openness are reinforced by a widening set o f international trade-related disciplines on regulatory transparency, such as GATS requirements. Foreign firms, individuals, and investors seeking access to a market must have adequate information on new or revised regulations so they can base decisions on accurate assessments o f potential costs, risks, and market opportunities, but have greater difficulties than domestic market players in obtaining information. Regulatory transparency has also been improved by the growing use o f intemational standards, which reduce search costs and increase certainty for consumers and market players. 3.24 Regulatory transparency i s increasing slightly in Senegal, due to the activities o f APIX and the guichet unique when rules are applied. Yet regulatory transparency i s still far from satisfactory. 3.25 Two major reforms could place Senegal within the mainstream o f good international practice. The first reform i s establishment o f a permanent Business Advisory Group to consult with the private sector on new projets de loi, dbcrets, arre"tbs, and dkisions. Public consultation (the active seeking o f the opinions o f affected groups) when developing new laws and regulations i s used in all industrialized countries to improve transparency and quality. As Senegal moves toward the rule o f law, in which legal instruments play a more important role in driving social and economic behavior, consultation will be increasingly used as a bridgebetweenlaw and the diverse interests o f a complex society. 3.26 Consultation methods used in Senegal are traditionally based on formal hand- picked advisory groups who provide general feedback on policy issues. For example, one important groupe de consultation technique i s chaired by the Minister o f Finance and Economy and meets once a month. It reports to another consultation group chaired by the Prime Minister that meets every three months. The President, too, has hisownconsultationgroup. 63 3.27 These groups are, however, not well-designed nor well-placed to react to a continuous stream o f detailed technical proposals, and they are too vulnerable to capture. Senegal does not now have an effective mechanism for the private sector to engage with the state on new laws and regulations. Processesfor discussingprojets de loi, ddcrets, arre^tds,and ddcisions with the private sector are unevenly applied and discretionary, and do not ensure a high-quality response. It i s often unclear who has been consulted, or how they were selected. Concepts o f open access to government legal drafts are not widely accepted inthe Senegaleseadministration. 3.28 Senegal's government already recognizes the need to buildbetter consultation mechanisms that evolve with the needs o f society. The 2002 Document de strategie de reduction de la pauvrdte` calls for the creation o f mechanisms of consultation between the public administration and the private sector. More extensive and pro-active forms of consultation are illustrated by the development o f the strategie de reduction de la pauvrdtd, which involved extensive participation at national and local levels by actors in the public and private sectors, development partners, and NGOs. This form of consultation i s not practical, however, for individual regulatory decisions. 3.29 Some forms o f consultation used in Senegal are actively harmful. It i s widely perceived that important decisions are made inDakar by insiders who have the ear o f important politicians. Dialogue with the private sector was, and still is, often held behind closed doors, in the form discussing special favors and privileges for businesses. This kind o f private dialogue with business interests is widespread in transition countries with a tradition o f close relations between businesses and the state, but is undesirable and dangerous to economic development in a market economy. It supports corruption, protectionism, and competition abuses that harm the most vulnerable members o f society. It discourages investment by outsiders who do not have a special channel to the officials inside the government. Governments indulginginprivate dialogues with well-placed business interests pay a heavy cost in terms o f slower investmentand growth. The mechanisms recommendedbelow should replace this kind o f closed-door consultation. Another kind o f harmful consultation i s to permit special interest groups too much access and influence over "their" policies. Reform o f the Labor Code is a case in point, where trade unions have exerted powerful influence to maintain an extremely damaging set o f regulations that, because they hold back growth, also hurt trade unions in the long-run. Consultation cannot become captured by special interests who do not share the interests o f the larger society. 3.30 Another problem with consultation in Senegal i s that it i s often used to plan, but with little follow-up or impact. Many conseils interministeriales are appointed to review problems and come up with recommendations, but often nothing happens afterward. This wastes resources and engenders cynicism. Another complaint i s that the consultation council is sometimes over-weighted with public sector officials who would rather see their plans adopted than to listen to other ideas. 3.31 Lack o f consultation i s not entirely the fault o f the public sector. Senegal's business sector was traditionally more in favor o f protections and subsidies than o f competition, and is not regarded as well-organized compared to other countries. Participation o f enterprises in private sector organizations i s at a low level, perhaps becauseentrepreneurs rely more on their personal interventions with government than 64 on the influence of the business organizations. The lack of influence o f business organizations means that new entrepreneurs find it hard to break into the circles o f influence that operate inSenegal. Butthe business organizations seem to be reforming today, and hence it i s easier and more useful now to construct an open consultation process. Business organizations such as the Employer National Confederation o f Senegalese Employers (CNES), the Conseil National du Patronat du Sknkgal (CNP), and the Chambre de Commerce are moving to provide services for business services rather than to protect individual members, though this transition has not been completed. They are adopting a wider view o f broad-based market development in driving economic growth in Senegal. Thus, they are more responsible consultation partners. 3.32 Adapting laws to evolving market needs will require a systematic public consultation strategy that i s clear about i)the design of public consultation and the involvement o f major affected interests, ii)the time period o f consultation, and iii)the treatment o f comments from the public. 3.33 Systematic consultation procedures provide guidance and training for civil servants and provide benchmarks to all parties as to their roles. This enhances confidence inthe consultationprocess, andmeans that it i s likely to bebetter balanced among participating interests, and less prone to capture by small, organized groups with major interests in the outcome. Further, adopting a consistent process across ministries permits better co-ordination for regulatory quality initiatives across a wide range o f policy areas. 3.34 Senegal needs to move forward with legal reforms without delay and a few hundred regulatory instruments each year will require consultation with the private sector. Ministries need a consultation system that collects as much relevant information as possible, as quickly as possible. An approach often used in European countries, and by the European Commission, i s a permanent Business Advisory Group whose members represent a range o f interests. InSenegal, such a Group should contain large and small enterprises, domestic and foreign investors, and formal and informal enterprises. All meetings o f the Group and its recommendations should be public. D. REGULATORYREFORMAND SENEGAL'S INFORMAL SECTOR 3.35 This report has noted that the most important anti-poverty mechanism in Senegal today i s the flourishing informal sector, where most new jobs are created. Senegal has a dual economy, in which 200,000 employees are in the formal sector, while millions o f others, up to 90 percent o f the work force, work ina regulation-free environment. One estimate attributes 85 per cent o f GDP to the private sector, with 60 per cent coming from the informal sector and 25 per cent from the formal.23 The informal sector seems to have been stable or increasing in size over the past decade due to stagnation in the formal sector coupled with new economic opportunities for the informal sector. Liberalization o f the rice market, for example, expanded the informal sector inthat area. Some have estimated that the informalsector evades VAT 23Systeme des nations unies. Evaluation commune de la situation du Senegal. Septembre 2001, p.108 65 payments o f 10billion CFA annually. The boundary betweenthe formal and informal sectors i s blurred,since there i s wide-spread tax evasion even inthe formal sector, and the informal sector i s used as a cost-avoidance strategy for the formal sector by hiring informal enterprises as suppliers. 3.36 Senegal will have a large and perhaps dominant informal sector for decades to come. No policies should be adopted that inhibitthe capacity o f the informal sector to create jobs, but the large informal sector should not be a matter o f laissez faire. Senegal would grow faster and more equitably if the most successful elements o f the informal sector could be brought into the formal sector. Constraints on growth and productivity in the informal sector are severe, as are the human consequences for its employees due to lack o f access to social protections. There i s good reason why most exporters are in the formal sector, since international markets demand a level o f professionalism and quality that the informal sector cannot meet due to difficulties in recruiting quality staff. Formalization would also diversify the tax base (perhaps allowing a reduction in overall rates) and the financing o f social programs. Bringing the informal sector progressively into the mainstream- through incentives rather than coercion -- would expand the size o f internal and external markets, increase opportunities for entrepreneurship and employment, raise incomes intowns and rural areas, and boost consumer demand. 3.37 Rather than focusing on the informal sector as an "unfair competitor" to the formal sector, it will be more useful to see the informal sector as an incubator to build skills and assets before an enterprise enters the formal sector, with its opportunities such as larger markets and costs such as higher production and labor costs. The informal sector has traditionally been an entry point into the formal sector. For example, o f the 60 members in the Societk de domaine industriale (SODIDA), some 57 had their origin inthe informal sector. There i s evidence that, as opportunities for market entry expand, the informal sectors are increasingly playing exactly this kindo f useful role in the development o f Senegal's business sector. The most successful o f Senegal's informal firms are evolving from simple retailing and marketing to investment in distribution channels to investment in production facilities. They are moving from importers to domestic producers. They are investing incapital and labor, and are adding more domestic value. This trend i s almost entirely undocumented, but anecdotes suggest that there are a significant number o f such firms who are not very different from their formal counterparts, have deepened their activities through the supply chain and would find it advantageous to move to the formal sector if a manageablemultiyear transition program were inplace. 3.38 The capacity o f the best performing informal enterprises to enter formal markets is also suggested by the increasing organization o f the sector by the Union Nationale des commercants et industrielle de Senegal (UNACOIS), which represents the interests o f SMEs, distributors, and informal enterprises in policy reforms. UNACOIS was created in 1989 and today has 100,000 members and holds a vice- presidency position in the Chambres de Commerce. UNACOIS sees a small but visible trend inenterprises who wish to expand their opportunities by formalizing and listing with APIX, but who are hstrated by the complexities, inefficiencies, and corruption o fthe public sector, such as customs. 66 3.39 There i s an emerging strategy in Senegal to encourage the progressive movement o f informal' enterprises into the formal sector. The Centre de Gestion Agre`e (CGA) is a mechanism to formalize existing businesses. It has assisted some movement from the informal to the formal sector, but i s not seen as sufficiently ambitious in coordinating between ministries, and as ineffective in the absence o f a tax amnesty. Due to fiscal and regulatory barriers, the transition is difficult for all except the most successful o f informal enterprises. What i s needed in Senegal i s designation o f a special legal status, such as "Enterprises inTransition", accompanied by a much more attractive package o f incentives and aids for such enterprises. The following menuo f incentives i s not meant to be exhaustive, but illustrative o f the kind of incentive-based program that could be designed to formalize the Senegal economy over time. 3.40 Phased levels o f simplifiedtaxation over time, and access to special tax credits for investments in training and physical capital, with a possible tax rate o f zero. The level o f taxation i s a key issue. Informal enterprises have scale constraints and employees in informal sectors have lower skills and less training than those in the formal sector. For these reasons, the capacity o f informal enterprises to pay taxes i s very limited. To raise productivity inthe transition phase and to provide incentives to enter the program, training and investmentcredits should be provided, at a level that permits Enterprises in Transition to channel all their tax liabilities into human and physical capital investments for at least the first 3-4 years o f existence. 3.41 Tax simplification will be essential for enterprises entering the formal sector. Indeed, introduction of the TVA complicated life for SMEs, since a single tax d'egalisation for SMEs was suppressed as part o f the TVA reform. Given the level o f accounting o f most informal enterprises, this program should be as simple as possible to administer, and accounting help and training may be necessary, perhaps from the one-stop shop recommended below. The TVA might be too difficult to apply in the first few years for these enterprises, and hence an altemate tax might be necessary. The total tax load should be managed. The goal i s a phased-in tax that permits the enterprise several years to raise productivity and profits before entering fully into the tax system. 3.42 Independently, the Ministry o f Finance has set as a top priority the reduction of tax evasion by informal sector operators by developing a system o f presumptive taxes that reduces accounting burdens, in effect retuming to the pre-VAT solution. Another solution might be continued development o f the VAT system, with selective thresholds designed to encourage significant parts o f the informal sector to participate. An interesting problem is that most informal operators are not fluent in French, which, as the official language, is used in the promulgation o f laws and regulations. Thus many operators face significant difficulties incomplying with the tax laws, even iftheywantedtocomply.Thelanguagedifficultiesmayneedparticularattentionin the first years o f formalization. 3.43 Preferential access to SME or microfinance credit windows. The informal sector needs access to credit to improve the quality of products with new machines. Financial supports (such as access to small loans) and other supports (for instance, access to workshop areas) could be made conditional on registrationo fthe firm. 67 3.44 Phased-in payments under the social security, medical insurance and pension systems, with phased-in benefits. Senegal does not have an integrated social security system, butthree differentinstitutions: the CSS (Caisse de Sicurite' Sociale) for the family benefits, sick or maternity leave and workplace accidents. For the CSS, the payments are 7% o f salaries for family benefits and sick and maternity leave, plus 1% to 5% o f salaries for workplace accidents (the percentage depends o f the category which depends on the risk). The total percentage o f CSS i s between 8 and 12% o f salaries, which i s entirely paid by the enterprise, with a ceiling o f 60.000 CFA per month per worker. the IPM (Institution de Pre'voyance Maladie) for medical insurance. Companies having less than 100 employees contribute to an I P M to cover the worker and family. The percentage o f salary i s 6%, which i s equally shared by the enterprise and the worker. The ceiling i s 60.000 CFA permonth. the IPRES (Institution de Pre'voyance Retraite du Sdnigal) provides pensions. For the IPRES, the percentage i s 14% for low salaries, with a ceiling o f 200.000 CFA per month (this i s called the Rdgime general de retraites) and 6% for high salaries (this i s called the Rigime complimentaire des cadres). These payments are co-supported by the enterprise (60%) and the worker (40%). Informal enterprises claim that they already pay the equivalent o f social charges by contributions on the occasions o f births and marriages for their employees. The social security system does not take such payments into account. UNACOIS argues that ifthe Caisse obligation is applied in addition to these traditional social arrangements, then the traditional payments should be credited as part o f their Caisse obligations. There is also concern about the quality o f services. 3.45 With respect to Enterprises in Transition, the several kinds o f payments seem very high, well over 20 percent o f labour costs, not including the employee costs. One way to phase inthese costs and services i s to place the various payments and services on a parallel schedule, keeping inmindthat coverage would be zero ifthe firm stayed in the informal sector. The important point is to get the firms registered, and on a payment schedule that, in the long-term, progressively extends coverage o f the country's social safety net. 3.46 Training servicesfor managers. Managers o f informal enterprises will have to acquire many skills to operate and grow in the formal sector. Senegal's Chamber o f Commerce already has implemented a training program to helpentrepreneurs develop business plans and build their businesses. The learning curve will be shorter and growth faster with a training program dedicated specifically to Enterprises in Transition in areas such as accounting and billing. This program could be self- financing, ultimately, with reimbursements from Enterprises in Transition that have successfully made the transition. In export sectors, specific training in quality standards and quality processes will be useful, as suggested above for the business sector as whole. 3.47 Set-aside programs in public procurement for Enterprises in Transition. A method to stimulate new businesses in many developed countries i s the set-aside 68 program for public procurement. The set-aside means that a pre-determined proportion o f purchases, contracts and subcontracts for property, commodities and services for the public sector are given to eligible businesses through competitive tendering. Such programs usually includeprotections for the public interest, such as if acceptance o f the best bid or proposal will result in the payment o f an unreasonable price, the public sector can withdraw the designation o f the set-aside. Set-asides for Enterprise in Transition should be allowed in those industries and procurement areas in which the government determines that Enterprises in Transition could provide some portion o f the goods or services. The ministries should also encourage participation in such tenders by keeping lists of eligible Enterprises inTransition, and informingthem of new opportunities. 3.48 Efficient justice through special priority procedures for contract and dispute resolution in the commercial courts. Informal enterprises have very limited access to the formaljustice system. Since enforceable contracts are generally not made between informal enterprises, they use means o f informal arbitration that preserves reputation and relationships, and that seem to work well. If they wish to use the court system, they are required to regularize themselves and pay their taxes and social security, which bars the informal sector from the justice system. An Enterprise in Transition will be able to enter judicially enforceable contracts. On the other hand, cases in the court system and the new arbitration system can be costly and lengthy. 3.49 Exemption from minimum wages for a period of time (perhaps five years), with a possibility of extension. Minimumwages in Senegal are set at 36,000 CFA per month (around US$50). However, labor qualifications and productivity in the informal sector are far lower than inthe formal sector, and time i s needed to upgrade the quality and value-added o f labor. Exemption from minimumwages for a period o f 4-5 years would also encourage more rapid job creation in the initial years o f the transition. 3.50 Other exemptions from the rigid Labor Code, which is too rigid evenfor the formal sector. Senegal still maintains a labor law based on French colonial laws that i s seen as an essential component o f socialprotection for workers inthe formal sector. Strong trade unions have prevented any serious reform o f the Labor Code, but such reform is essential to enable Senegalese businesses to adjust to market opportunities and create new jobs in the formal sector. Fortunately for the Senegalese economy, practices in the labor market are more flexible than what would appear from reading the present Labor Code. Even a Labor Code that i s substantially liberalized, however, i s likely to require significant changes in how Enterprises in Transition use labor. These changes should be made over a phase-in period to avoid sudden shocks and costs that discourage formalization, investment, and new job creation. Hence, Enterprises in Transition should be exempted from compliance in the first 2-3 years with any restrictions in the new Labor Code that reduces flexibility in signingpart- time, time-limited, weekend, or seasonal contracts, or inworking more than 40 hours per week (forbidden under formal contracts) or indismissingworkers, or inproviding paid leave (5 weeks a year, or 7 weeks for a mother with a child less than 14 years old) or other benefits. These requirements should be carefully phased inover a period of 3-5 years, untilthe EnterpriseinTransition becomes fully normalized. 69 3.51 Create an easily accessible one-stop shop for business formalities for Enterprises in Transition. Progressive transition would be supported by flexible and simple registration procedures. Everyone agrees that, despite years o f effort, formalities are still too burdensome for enterprises. One o f the reasons why Senegal's businesses suffer high costs from red tape i s that ministries do not communicate. There is little coordination among multiple ministries in imposing reporting requirements on businesses. The secure regulatory registry recommended above will be a useful mechanism in coordinating red tape, but Enterprises in Transition need more direct assistance particularly an easily accessible one-stop shop for permits, licenses, and approvals. To the extent that special transition requirements are established, a one-stop shop i s necessary to avoid confusion with businesses already inthe informal sector. The one-stop shops should be located around the country, and could be located in APIX, the Chamber o f Commerce, or preferably a joint APIX- Chamber o f Commerce operation, or even contracted out to the private sector. 3.52 To become formal, an enterprise must have a legal existence. This requires a NINEA (numero identijkation national des enterprise et associations) issued by the Ministry of Finance, and registration in the Registre de commerce kept by the Tribunal de Commerce in the Ministere de Justice. Registration is carried out by the Chambres de Commerce, but the Tribunal keeps the register. In addition, another number, the NITI, is needed for importers, and is administered by the customs service. Unification of these various numbers into a uniquenumber would be a useful step. 3.53 A less ambitious reform would allow Enterprises in Transition to accomplish all these registration requirements through a single contact point. While a simple version o f the one-stop shop merely provides information in the form o f application forms and contact details, Senegal should establish a one-stop shop that coordinates inter-governmental co-operation and issues a complete registration. This kind o f one- stop shop "makes the documents walk, not the client" and reduces the costs and delays o f the transition. 3.54 One-stop shops are difficult to set up, since they require extensive upfront coordination and planning among participating ministries, but where they have been well designed, they deliver good results. In Italy, the one-stop shop operated by municipalities replaced 43 authorizations needed to start up a new business, and reduced the time required from 2-5 years to 3 months on average, and 11 months maximum for big investments. 3.55 The Senegal government should commit to establishing a one-stop shop for registration o f the Enterprises in Transition, perhaps building on the A P E guichet unique, but preferably a joint operation between the Chambres de Commerce and APIX to reduce costs and expand national coverage. Already, as noted, discussions are underway on a Centre desformalities des Entreprises, and this Centre could as a first step focus on the needs o f Enterprises inTransition. As a first step, such a Centre should compile a comprehensive electronic list o f all approvals, permits, and inspections o f businesses needed for formalization. The unit should identify those requirements that could be coordinated or consolidated, and should initiate such actions among ministries where possible. A plan should be developed for establishing 70 one-stop shops offering a clearing-house for applications from informal enterprises and subsequent communications between applicants and ministries. 3.56 Itshouldbe clarified that the transitionprogram for informal enterprises is not meant to apply to those large-scale enterprises, particularly in the importing sector, that evade taxes. These enterprises have no need for any special benefits. They exist only with the tacit agreement o f powerful supporters in the govemment, and their existence should be considered a costly form o f corruption rather than a special case for a developing economy. To remedy this problem, the govemment needs to launch a program to enforce the tax and other laws against these enterprises. Increasing the gains of growthfor thepoor 3.57 Regulatory reform has an important role to play as par o f a coherent antipoverty strategy. Regulatory reform will benefit the poor mainly by increasing the value o f entrepreneurship and investment, which should create newjobs. The most important anti-poverty mechanism in Senegal today i s a flourishing informal sector. The discussion above examines how regulatory reforms can aid the informal sector in contributing to job creation and investment,not only through indirect effects, but also through strategies such as accelerating the transition of the best-performing informal enterprises into the formal sector and improving linkages in export sectors to small producers. Since new jobs are most likely to occur in services in urban areas, the urban poor are more likely to benefit from the recommended reforms. In contrast, increased quality inexport sectors should expand opportunities for small producers in rural areas. 3.58 Other anti-poverty benefits can be far-reaching and indirect. For example, general growth increases tax receipts that can finance the social safety net. Competition in sectors such as transport, electricity, telecommunications and health can lower prices and improve services, increasing economic activity in many other sectors, even if the poor are unlikely to benefit much from direct consumption o f these services in the market. Universal access programs in utility sectors will allow groups now deprived o f such services to consume more o f them inthe future. Senegal should monitor these programs carefully to ensure that they promote both efficiency and access for target groups. 3.59 Another positive indirect effect o f regulatory reforms that increase market size will be on key sectors in household consumption (food, clothing) or that impact indirectly on the costs o f these consumption items, such as wholesaling and transport. More generally, measures to increase the size o f the relevant market will increase the over-all intensity o f competition and increase the purchasingpower o fhouseholds. 3.60 There i s no universal model for the right regulatory system, since solutions must be designed to fit within the specific circumstances o f Senegal's values and institutions, and its development needs. However, since Senegal i s competing in Afiican, European and global economies for capital and markets, intemational expectations and experiences for high-quality regulatory regimes can provide valuable benchmarks for action. SECTION 11: PROMOTIONOF HUMAN CAPITALAND BASIC SOCIAL SERVICES Future economic growth is certainly essential for progress in reducing poverty. Nevertheless, Senegal's recent growth episode did not have a commensurate impact on poverty. After the 1994 devaluation, Senegal posted six years o f real economic growth averaging 5 percent. Over the same period, population growth averaged 2.8 percent, allowing for real per capita economic growth o f over 2 percent. Still, preliminary estimates from the 2001-02 household survey indicate a poverty rate o f 54 percent, down only four points from the 58 percent poverty rate measured in the 1994-95 survey, both using a poverty line o f consumption at $2.00 a day. The fact that this sustained episode of growth generated so little impact on poverty probably indicates a substantial degree o f inequality. Such inequality could be due in part to unequal growth between urban and rural areas, where the latter have the highest concentration o f the poor. Rural areas are overwhelmingly engaged in agriculture, which sector occasionally posts high growth rates, but usually in rebound from drought episodes which can devastate output. Agriculture narrowly defined, excluding fishing and processing, still engages over half the workforce in producing only 10 percent o f GDP. To enable its poor people to participate inthe h i t s o f growth, Senegal will have to bolster their human capital. The PRSP's second axis is, appropriately, the promotion o f human capital and basic social services. This part o f the report focuses on the two most important aspects o f human capital development-health and education. 4. HEALTH, NUTRITIONAND POPULATION 4.1 Not only did rapid economic growth in the past six years fail to reduce poverty significantly, but improvements in health have been elusive. An episode o f low economic growth in the 1980s combined with a sustained effort to improve and redirect resources toward primary health care resulted in an impressive reduction in Senegal's child mortality. However, during the high-growth 1990s, with public health care efforts fractured and less equally distributed,, progress in child mortality rates was halted. The impact o f poorly targeted health services was compounded by lack o f economic growth and poverty reduction inthe rural areas. The lesson i s that poverty clearly leads to poor health outcomes which can be overcome with properly targeted interventions in infrastructure-especially water and sanitation-as well as selected health interventions. All these interventions mustinvolve the target community indirecting andmonitoring the allocation o f resources, and they must be geared towards improving health behavior which i s the major determinant o fhealth outcomes. 72 Senegal's Health, Nutrition and Population Outcomes A. CHILD MORTALITY 4.2 Health outcomes in Senegal's PRSP are measured in terms o f progress towards the Millennium Development Goals (MDGs). By Sub-Saharan Africa standards, Senegal is a good performer in terms o f most o f these health, nutrition and population (HNP) outcomes. Senegal's life expectancy at birth, 54 years, i s well above the regional average o f 49 years (see table 1). This i s a consequence o f lower infant mortality (70%0) and under-five mortality (145%0) rates, both o f which are associated with levels o f infant malnutrition (18%) that are well below the regional averages. However, the level o f matemal mortality remains at an unacceptable level inthe country (510 per 100,000 live- births). In addition, compared to the levels observed in Sahelian countries, such as Burkina Faso, Mali andNiger, fertility has started declining in Senegal. These results are not surprising, however, since Senegal combines one o f the lowest level o f prevalence o f HIV/AIDSinAfrica, higher GDPper capita and a higher rate o furbanization than most o f its regional counterparts. Table 4-1: HNP Indicators- Senegalin Sub-Saharian Africa circa 1999(1) Country Life expectancy Infant Under 5 Maternal Total Prevalence Infant at birth (yrs) Mortality Mortality Mortality Fertility of H I V Malnutrition (per1000live- (per 1000 (per 100000 Rate (per 100) (weightfor age: births) live-births) live-births) (number of 90 99 (1) per 100) childrenper woman) Senegal (2) 54 70 145 510 5.7 1% 18 SS Africa 49 107 173 822 5.4 8% 31 Burkina Faso 45 106 199 810 6.4 7% 46 Guinea 47 115 181 670 5.3 2% 24 Madagascar 58 95 156 490 5.2 0.5% 40 Cote d'Ivoire 47 102 171 600 4.9 10% 24 Ghana 61 63 101 210 5.0 25 Uganda 42 83 131 510 7.0 - 26 Mali 54 123 237 577 6.7 1.5% 4% 40 Niger 49 136 274 700 7.5 2% 50 Sources: (1) UNICEF <(Lasituation des enfants dans le monde 2001D. For Sub-SaharanAfrica, average values for the period 1990-1996.(2) Directionde laPrtvisionet de la Statistique, ((SituationEconomiquedu Stntgal 1999))(life expectancyat birth); MICS 2000 (infant and childmortality) andDHS surveys 1986, 1992et 1997(fertility). 4.3 Nevertheless, when compared to other, equally poor, developing countries, Senegal's outcomes are well below the median. With a level o f under-5 mortality o f 145%0in 1999, Senegal i s among the 25% poorest performers among countries with GDP per capita between $300 and $600 in 1999 in the world. Senegal's levels o f under-five mortality are higher than Ghana (101%o) and The Gambia (11O%O), countries with lower 73 levels o f GDP per capita in West Africa. They are much higher compared to countries with lower levels o f income in other developing regions such as Nicaragua (43%0), Bangladesh (89%0)andVietnam (42%0)(see Figure4-1). Figure 4-1: Senegalwithin globaltrends in GDP per capita and under-five mortality circa 1999 300 N 250 y =4648.3~-''~~~' R' =0.7664 200 0 4: i 25 '150 100 Bangladesh d 50 0 10 100 1000 10000 100000 per capita GDP, US$ 4.4 Furthermore, the accelerated decline in child mortality that Senegal experienced in the 1980s has been lost in the 1990s. Child mortality declined slowly in Senegal between 1960 and 1980: during the first twenty years o f independence, under-five mortality declined only from 300%0to 250%0as a consequence o f poorly targeted health interventions and continued high prevalence o f malnutrition and fertility. A primary health care strategy based on a package o f priority health programs targeted a childhood diseases and increased social mobilization and community participation contributed to an acceleration o f the rate o f decline of under-five mortality during the 1980s: under-five mortality declined at an annual average rate o f 3.8% resulting in absolute child survival gains inten years comparable to gains achieved inthe preceding twenty years (Figure 2). Inthe 199Os, however, the levels of under-five mortality stagnated between 140%0and 150%0as evidenced by successive DHS surveys (DHS 1992-93, DHS 1997, ESIS 1999). With infant mortality stagnating around 70%0,childmortalityhas increased to 85%0at the national level in the late 90s. Had Senegal's progress in infant and child mortality continued at the pace o f 1980s, the country would have reached the 2015 MDG (see figure 4-2). With the 1990s having been a lost decade interms o f child survival gains, this target can only be reached only with an accelerated rate o f decline over the next decade. 74 Figure4-2 :Senegal- Trendsof under-fivemortalityduringthe pastdecades country to reach the 2015 VDG \ I 4.5 Most o f the major causes o f under-five mortality in the country could be averted based on available health technologies and sustained health and demographic behavioral changes. Neonatal mortality, where changes have been slow over the past decades, continues to be associated with prematurity and low birth weight, neonatal tetanus, pneumonia and other acute respiratory infections, and birth trauma; in addition, higher levels o f neonatal mortality continue to be associated with shorter birth intervals exemplifying the behavioral interrelations between fertility and infant mortality in the Senegal context. The high levels o f child mortality, however, continue to be associated with acute respiratory infections, diarrhea, malaria, measles and tetanus in combination with malnutrition. 4.6 Better targeting o f child health interventions to the poor will be necessary, for the heaviest burdeno f avoidable deaths before age five i s borne by the poorest segments o f the population. Inequalities inunder-five mortality in Senegal are among the worst inthe West Afkican region. The major changes inunder-five mortality which occurred since the 1980s have not significantly altered child health inequalities in the country. The poorest 20% o f the population continue to experience under-five mortality rates which are twice higher compared to the richest 20% of the population (see figure 4.3). Furthermore, health inequalities are among the highest inthe West Africa region: countries with under- five mortality rates inthe neighborhoods o f Senegal's levels such as Ghana, Cote d'Ivoire and Togo experience lower child health inequalities; in addition, Senegal's child health 75 inequalities are much higher than its Sahelian neighbors with much higher levels o f under-five mortality such as Mali, Niger and Burkina Faso. Figure 4-3: Senegal - Inequalitiesinunder-five mortality during the past decades USMRper 300 1000 250 200 150 100 50 0 I Poorest Second Middle Fourth Richest B. MALNUTRITION 4.7 Malnutrition i s also concentrated among the poorest segments o f the population. Although Senegal experiences among the lowest levels of child malnutrition in West Africa, socioeconomic inequalities in child malnutrition are among the worst observed in the sub-region. Underweight and stunting prevalence displayed similar socioeconomic patterns in the 1980s; in the early 199Os, however, both underweight and stunting prevalence deteriorated consistently among the poorest 40% o f the population, while no consistent changes were observed among the better-off. As a result, while underweight prevalence among the poorest 20% stands at twice the level o f the richest 20%, stunting prevalence among the poorest 20% has deteriorated to three times the level o f the richest 20%.. Accordingly, in addition to targeting child health interventions to the poor, increased efforts would be necessary to foster synergies between health and nutrition interventions to support child survival and development gains among the poor. 76 Figure4-4 :Senegal Inequalitiesin childmalnutrition. - Underweight 40 35 30 25 20 EiDHS1986 IDHS1992 15 10 5 0 Poorest Second Middle Fourth &chest Stunting 40 I r I I I Poorest Second Middle Fourth Richest 4.8 The poorer the region in Senegal, the higher the levels o f infant and child mortality and child malnutrition. The poorest regions o f Kolda and Kaolack experience infant and child mortality levels twice as high as the better-off regions o f Thies and Dakar; similar disparities are observed in the prevalence o f malnutrition. The region o f Ziguinchor i s an exception, because o f the status o f women in the region, relatively high level female education, and patterns o f health behavior. The large gaps in infant and child mortality and child malnutrition betweenregions suggest that targeting by region o f interventions in the health sector and health-connected sectors (water, food security and education) can be beneficial. The institutional framework o f the ongoing decentralization reforms can support this strategy. 77 Figure 4-5: Senegal - Regional disparities in infant and childmortality and child malnutrition (USMR: ESIS 1999-Malnutrition:MICS2000) IProportionoI EU5MR ElMalnutrition 25 Kolda Kaolack Ziguinchor Tambacounda Diourbel St Louis Fatick Louga Thies Dakar C. FERTILITY 4.9 The acceleration o fthe decline o f infant and childmortality over the 80s has been followed by the onset o f fertility decline in Senegal. Total fertility rates have fallen from 6.6 in the middle o f the 80s to 6.0 at the beginning o f the 90s and 5.2 at the end o f the 90s. Most o f the change in fertility, however, has occurred inthe urban areas and among educated women-mostly due to delayed entry into first union; and to a lesser degree, increased use o f modem contraception. The segments o f the population that could have benefited the most from a decline in fertility are still experiencing relatively high levels of fertility : rural and illiterate women who live in the poorest regions o f the country. While total fertility rates have declined to 3.9 among urban women, they still stand at a relatively high level o f 6.1 in rural areas. Similarly, they have dropped at 2.7 among women with a secondary level o f education, 4.1 among women with a primary level o f education, compared to 5.9 among women who have never been to school. The relatively low levels o f fertility in Dakar (3.9, Ziguinchor (4.6), Thies (5.0) contrast with total fertility rates above 6 inthe poorest regions o f Fatick (6.0), Tambacounda (6.4), Kaolack (6.4) andKolda (6.3). 78 Figure 4-6: Senegal- Trends inFertility (age-specific fertility rates) in Senegal during the past two decades Per 1000 300 -I 250 200 150 100 50 0 1 I I I I I I 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Age D. MATERNAL MORTALITY 4.10 Unacceptable levels o f maternal mortality continue to underscore the highrisks o f givingbirth, poor access to life saving technologies and the low status o f women inthe country. The levels o f maternal mortality are still very highin Senegal: over the 90s, the maternity mortality rate was estimated at 510 per 100,000 live-births. Demographic surveillance sites in the rural areas suggest that this national average hides very large disparities between rural and urban areas and between regions. Most maternal deaths are preventable with appropriate health care. But in Senegal, matemal health service coverage and quality continue to suffer. Although antenatal care coverage expanded in the rural areas during the 80s and the 90s, the majority o fbirthdeliveries continue to take place at home. The country's health delivery system has failed to maintain a network o f qualified midwives inthe rural areas and the poorest regions. It i s inthese same segments of the populationwhere reproductive factors o f maternal mortality, such as early maternal age and highparity, continue to be sustained by female illiteracy, low status o f women and other socioeconomic and cultural factors. Without sustained efforts for extending quality maternal health services inthe rural areas and the poorest regions o f the country, it would be very difficult for Senegal to reach the specific MDG relative to maternal mortality. 79 E. HIVtAIDS 4.11 Senegal's prevalence o f HIV/AIDS i s among the lowest in sub-saharan Africa. Estimated at 1% in 1999, the prevalence o f HIV/AIDS among adults was evaluated at 1.4% in2000. Although 14,000 of the 24,000 alive adults infected in2001 were women; prevalence rates among women aged 15-24 years range from 1.12 to 2.97% compared to a range o f 0.39 to 1.02 among man of the same age group suggestingthat the epidemic may be progressing faster among women24. Given the relatively low level o f prevalence among pregnant women (0.7% in 1996), mother-to-child transmission i s relatively low: among the 100,000 cases o f infection, only 3000 are children. However, the number o f surviving orphans aged under 15 years o f age due to HIV/AIDS was estimated at 15,000 at the end o f 2001. Improving Senegal's HNP Outcomes 4.12 Improving HNP outcomes, especially for Senegal's poor, will require a systematic effort to relax the constraints on better health. These constraints are, in rough order o f importance: (i) environmental factors, such as access to potable water and sanitation; (ii) behavioural factors, such as household health practices, the ability o f women to support these practices; and (iii) availability, cost and quality o fhealth services. the F. WATER ANDSANITATION 4.13 Child mortality and morbidity i s strongly associated with water and sanitation conditions at the community and household levels; in addition, higher risks o f acute respiratory infections are associated with indoor pollution. In Senegal, only 55 % households have access to potable water either through pipedwater at the householdlevel (32%) or through public fountains (23%) in 2001. For the majority o f Senegalese communities, wells andother sources continue to be the main sources o fwater. 24UNAIDS, 2002. "Senegal - Epidemiological Fact Sheets on HIV/AIDSand Sexually Transmitted Infections: 2002 Update". Unicef and World Health Organization. 80 Figure 4-7: Senegal -Environmental Contamination Access to potable water and sanitation by region Dakar Thies Louga Fatick st LOUIS Diourbel Tamba W F l w h T o i l e t Ziguinchor Kaolack EPublicFountains Kolda Source : Questionnaire Unifie des Indicateurs de Dkveloppement (QUID)2001. 4.14 The poorer the community, the lower the access to potable sources o f water (see figure 7). It is only in the richest region of Dakar where the proportion o f households with pipedwater is above 70%; inthe next region o f Thies, this proportion falls to 27%, it levels off at less than 3% in the poorest region o f Kolda. Outside o f Dakar, the main source o f water o f the majority o f households are wells, public fountains ranking as the second source: at the extreme end, 95% and 80% o f households inKolda and Ziguinchor respectively depend on wells as their main source o f water. Access to good or moderate sanitation conditions displays similar levels and patterns inthe country. Thus, one o f the first constraints to better health for the poor i s the contamination o f their community physical environment resulting from very low access to potable water, poor sanitation conditions and other environmental factors contributing to higher exposure to disease vectors. Yet, it i s not clear how these considerations are incorporated in public policy decision-making in Senegal. For instance, decisions on water and sanitation are the responsibility o f the infrastructure sectors, whereas the main beneficiary will be Senegal's health sector. Behavior 4.15 Not only do the poor live in communities where exposure to communicable diseases is relatively high, but also cultural and social factors sustain behaviors in many of these communities that are detrimental to better health. In many rural communities, belief systems and representations o f illness constrain preventive measures' initiatives and/or delay the substitution o f traditional health practices or remedies for more effective health practices. For example, beliefs that amulets may provide protection against communicable diseases, such as sexually transmitted diseases or other common diseases, are pervasive in many Senegalese communities. Such beliefs prevail also relative to the 81 protection o f the health o f mothers during pregnancy and child health, and continue to delay the adoption o f more effective measures o f reproductive and child health. 4.16 High social demand for children sustains early marriage customs, levirate and excision practices, childbearing at early and older ages and short birth intervals which continue to compound health risks for mothers and their children. At the same time, such demands continue to divert women's investmentaway from formal education, acquisition o f knowledge and income generating activities or other health enhancing strategies. As a result, while health is highly valued inthe community, its maintenance in the family and the community remains the role o f women who are mostly illiterate, with very low social status and limitedaccess to household and community resources. 4.17 Inaddition to the levels of income, the level of education ofhouseholdmembers, especially the mothers, remains one o f the main factors o f health behavior change at the household level, and consequently o f child health and reproductive health outcomes. Most o f the decline in fertility which has been observed in Senegal over the past two decades occurred mainly among women with at least a primary level o f education. In addition, multivariate analyses o f infant and child mortality based on data gathered over the 80s and the early 90s, suggest that female education continues to be one o f the main contributors to health improvements in Senegal, although its effects have been diluted with the reversal o fchild survival gains inthe late 90s. These local patterns are consistent with the effects o f female education on HNP outcomes observed extensively in sub- Saharan Africa and other developing areas. Hence, the evidence suggests that not only will the low level o f instruction and literacy o f women, resultingfrom poor performance inthe education sector inthe past decades, continue to puta downwardpressure onHNP performance inthe country, but also disparities in female education between the poor and the better off, and between poor and relatively rich regions, will continue to challenge efforts to reduce health inequalities in the country (see figure 8). The importance o f female education to health outcomes reinforces the point about the multisectoral nature o f the problem-and the need for a multisectoral approach to solving it. The budgetary reforms associated with implementingthe PRSP could play a major role here. 82 Figure 4-8: Senegal Level of instruction of women of childbearing ages, use of condom - among men with occasional partners, assistanceby trained health personnel at delivery, and childrenaged 12-23 months fully vaccinated by region(ESIS 1999) Proportion (??) ELIMan: use of condomwith occasionalpartners 100-I___ 0Assistedbytrainedpersonnelat delivery I 0Children12-23monthsfullyvaccinated 90 80 I iH 70 , n .. 50 40 30 20 10 0 Kolda Kaolack Ziguinchor T a d a Diourbel St Louis Fatick Louga ?hies Dakar 4.18 Adoption o f health care practices and household behaviors which are supportive of better health i s slow inthe country. That education and literacy o f household members contribute to better HNP outcomes inthe Senegal context results from the slow adoption of health care practices and household behaviors which are supportive to better health in the country. For instance, in the midst o f the HIV/AIDS epidemic, 32% o f adult male continue to maintain occasional sexual relations; during these occasional encounters, however, only 48% o f men use a condom even though more than 90% o f adult male report adequate knowledge o fways o fprevention against the disease (figure 8). 4.19 Inefficiencies in the consumption o f reproductive health services continue to be sustained by the slow adoption o f household behaviors which support better health for women. The gap between knowledge and practice continue to be observed in the adoption o f family planning andmodem contraceptive methods: although 87% o f women of childbearing ages report knowledge o f any modem contraceptive, only 17% have ever used any modem method and 7% were currently usingany modem method in 1999.The use o fprenatal care has steadily increased since the middle o f the SOs, and nearly 82% o f pregnant women receive at least one prenatal visit by the end o f the 90s, but around 50% of deliveries continue to take place at home and only 48% o f deliveries are assisted by trained healthpersonnel (figure 8). 4.20 Such inefficiencies are also observed in child care practices and the consumption of child health services. While breastfeeding is widely practiced inmost segments o f the population, the adoption o f practices o f exclusive breastfeeding before the age o f four 83 months still fail to take hold in all categories o f Senegalese households: less than 2% o f children 5-59 months o f age have been exclusively breastfed in the country in 1999. In addition, only 11% o f children are regularly weighted within formal growth monitoring programs. While in neighboring Gambia, the use o f insecticide-treated bednets reaches 35%, it stands at 11% in Senegal. Finally, although 61% o f children between 12-23 months o f age were vaccinated against measles in 1999, only 42% were fully vaccinated (figure 8). 4.21 In most regions, the community environment is changing towards a more supportive environment as a consequence o f the explosion o f the associative movement, the emergence o f the rural commune and community involvement in health. In many communities, rural communes and community-based organizations (CBO), mostly migrant associations, women groups and youth associations, have started embracing a greater role in health. From the more punctual and visible interventions which had prevailed in the OS, such as buildingrural maternity clinics and health posts and social mobilization events, CBOs are extending their role to hygiene and environmental actions, greater involvement in health prevention and promotion in collaboration with health personnel, including HIVIAIDS prevention activities, to emerging community-based health insurance schemes. 4.22 While such supportive changes in the community environment may be sustained by ongoing decentralization and democratization movements, they still have to be mainstreamed into deliberate health strategies to provide for more sustainable community-based support systems for health in general, and for supportive household behavior changes inparticular. Indeed, how health policies and strategies would buildon community resources to develop and to widen the bases o f internal response capacities to increasing health needs inthe country remains one o f the main challenges to sustainable health gains inSenegal. Health interventions 4.23 Finally, while specific health interventions, as well as the health system in general, can have a profound effect on HNP outcomes, Senegal's health system suffers from problems o f access, affordability and quality-all o f which need to be addressed if health interventions are to have a favorable effect on HNP outcomes. 4.24 Access. Senegal's health infrastructures still reflect the imprints o f successive health development strategies implemented since independence. The top two tiers o f the national health delivery system are constituted by the national hospitals all located in Dakar (Dantec, Fann, Principal, Grand Yoff, Thiaroye), and the network o f ten regional and departmental hospitals which was developed mostly during the first two decades of independence under central government initiatives. 4.25 The basic health infrastructure which includes the lower tiers o f the health delivery system, however, results from central efforts as well as local initiatives. From 34 health centers in 1960, the number o f health centers stagnated duringthe 60s and 70s and increased steadily in the 80s and 90s to reach 47 in 1988 and 53 in 2001: these health 84 centers constitute the first referral facility at the basic health infrastructure level. They supervise extended networks o f primary health care facilities where the multiplicity o f actors under the implementation o f the primary health care strategy during the 80s and the 90s i s translated in the diversity o f structures: health posts and dispensaries, rural matemity clinics and health huts, most o f which were built under local initiatives, including efforts from urban municipalities, rural communes and community-based organizations. From 200 at independence, the number o f health posts increased to 376 in 1980, 659 in 1988 and 828 in 2001; the number o f rural matemity clinics grew ffom 189 in 1980 to 502 in 1988, and declined to 411 in2001 as a result o f the transformation o f rural matemities into health posts over the 90s. Finally, health huts numbered 1,160 in 2001. 4.26 Despite this growth, the geographical accessibility o f health services i s still very limitedinthe rural areas. While 80% o fthe urbanpopulation livewithin30mn of a health facility; in the rural areas, only 42% o f the population live within 30mn from health facilities and 43% live beyond 1 hour from health facilities. Large disparities between regions continue to prevail. In the two poorest regions o f Kaolack and Kolda, the geographical accessibility o f health services i s twice as low as the richer region o fDakar. Between these two extremes, the disparities ingeographical accessibility are explainedby and large by the relative importance o f local initiatives. These pattems suggest that under the institutional environment resulting from the 1996 decentralization reforms, such disparities will continue unless effective central initiatives support the equitable improvement o f geographical accessibility throughout the country. 85 Figure 4-9: Senegal Generalpopulation Use of health services and choice of provider - - among individualswho reported an illness during the four weeks preceding the survey by type of place of residence. 90 80 70 60 50 40 30 20 10 0 Urban Urban Rural Rural Senegal Senegal 1992 2001 1992 2001 1992 2001 Source: Questionnaire Unifik des Indicateurs de Dkveloppement (QUID)2001, EnquCte Sur les Prioritts (ESP) 1992. 4.27 Affoordability. Furthermore, even when accessible, the facilities are not being used. The levels o f utilization o f health care summarized in figure 9 suggest that the use of modem health care i s very low in the country. Only 16% o f individuals who reported an illness during the four weeks preceding the survey sought care either with modem or traditional providers (figure 9); 20%, among children under 5 years o f age. Traditional healers continue to be an important source o f health care; however, for most households, public health posts and centres remain the main sources o f curative care, followed by public hospitals. In a few regions, Dakar and Thies in particular, a relatively important proportion o f the sick who enter the modem health delivery system choose private providers. 4.28 The QUID2001 data suggest that a large share o f the population is not satisfied with the health services. Nearly 40% o f the sick who sought care reported not being satisfied with care received: 46% among the sick who chose a public hospital, and 47% among the sick who chose a public health center and or a public health post. For all sources o f care, costs were cited by 61% o f the sick as being the main raison of their discontent with care received; 35% reported long waiting time among the reasons o f their dissatisfaction, compared to 18% and 12% who mentioned lack o f drugs and quality of the treatment respectively. Such attitudes were reported among urban as well rural areas, inall regions andamong all socioeconomic groups. 86 Figure 4-10: Senegal General population Use of health services - - and choice of provider amongindividualswho reportedan illness during the four weeks precedingthe survey by region Proportion(YO) 35 30 13PrivateHospital 0PublicHospital 25 PublicHealth Center/Post W Traditional Healer 20 ~l 15 10 5 0 Kolda Kaolack Ziguinchor Tamba Diourbel St Louis Fatick Louga Thies Dakar Source : Questionnaire UnifiC des Indicateurs de DCveloppement (QUID)2001 4.29 The results o f the QUID 2001 are consistent with the ESIS 1999 results on women identification o f the main constraints to seeking health care inthe country. About 25% o f the ESIS 1999 sample o f women reported that the cost o f drugs was the first constraint for seeking care, while 23% indicated that the availability o f services was the major constraint. The costs o f consultations, the geographical accessibility o f services, and the availability o f drugs were reported among the secondary constraints for seeking care. 4.30 The primacy o f costs among the reasons o f dissatisfaction with health care is consistent with the levels o f household health-related expenditures reported inthe ESAM I1survey. Indeed, Senegalese households spend an average o f FCFA 6,066 per capita per year, corresponding to 3.2% o f total household expenditures (figure 10). Household health-related expenditures per capita average up to FCFA 10,300 per year in the Dakar region, while in the remaining regions, they range from FCFA 4,000 to FCFA 6,000. Given the low levels o f utilization o f modem health providers, such levels o f health- related expenditures are relatively high; moreover, since they are concentrated among households who needed care only, their levels suggest that they may contribute not only to the exclusion from needed health care, but also to sources o f indebtedness or impoverishment among low income families. 87 Figure 4-11: Senegal - Average household health related expenditures (FCFA) per capita per year by region 12,000 10,000 8,000 6,000 4,000 2,000 0 Kolda Kaolack Ziguinchor Tamba Diourbel StLouis Fatick Louga Thies Dakar ource : ESAMI1- 2001 4.31 Quality. Another reason for the underutilization o f facilities i s the absence o f qualified medical or health personnel in them. Disparities between regions in the availability and quality o f health personnel are even wider than disparities observed relative to the geographical accessibility o f health services: Senegal has not been able so far to sustain networks o f physicians and qualified midwivesinthe poorest regions o f the country and rural areas. Accordingly, solving the human resource problem o f local health services inthe poorest regions o f the country and regaining the thrust o f communities and the populations are one o f the main challenges Senegal will have to face in the coming decades for reproductive health and child health improvements to regain a trajectory towards meeting the HNP MDGs. The technical health personnel employed inthe health sector has continued to grow duringthe last two decades; however, the patterns o f growth of different categories and the distribution between sectors have varied significantly during the period (figure 12). The number o f health professionals (medical doctors, dentists and pharmacists) have increased steadily from 458 inthe year 1981to 961 inthe year 1994 and 1221 inthe year 2000: inthe year 2000, 58% o f health professionals were physicians, and 30% pharmacists. Most o f the growth o f the health professionals in the health sector, however, was absorbed in the private sector, including pharmacists as well as physicians: in 2000, 354 physicians were employed in the private sector compared to 306 pharmacists. Incontrast, while the number o f nurses have increased steadily between 1981 and 2000, most o f the growth was absorbed in the public sector. The number o f midwives employed inthe health sector stagnated between 1981 and 2000, growing only from 458 to 607 agents. 88 Figure 4-12 :Senegal Evolution of technical health personnel by sector and - category 1981 1994 2000 1981 1994 2000 Nrrsts Midrircs Sources: MSPAS/SSDApril 2002, World Bank, 1982. 4.32 The benefits o f the development o f the human resource base o f the health sector are distributed unevenly. Figure 13 suggests that profound disparities between regions continue to prevail inthe availability o f human resources. The ratio o f population to the number o f physicians is around 5000 in the richest region o f Dakar: while the region's share o f the population stands at 24%, 65% o f the country's physicians are located inthe region. In contrast, in the intermediate regions o f Thies, St Louis, Diourbel and Tambacounda, the population to physicianratios range from 17,000 to 30,000 inhabitants per physician. In the extreme region o f Kolda, the population- physician ratio reaches 85,000. Relative to the norm of 10,000 inhabitants per physician, very wide gaps in the availability o f physicians are observed in all regions, except Dakar. The distribution o f midwives betweenregions i s very similar to the distribution o f physicians; but relative to the norm o f 5000 inhabitants per midwife, the gaps in the availability o f midwives are even wider in all regions, except Kolda, Kaolack, Ziguinchor and Fatick. Only the distribution o f nurses relative to the size o f the population displays an even pattern across regions; inaddition, all regions, except Louga, present a population to nurse ratio close to the norm o f 5000 inhabitants per nurse. 89 Figure 4-13: Senegal Population per technical health personnelratio by region - and category in 2000 86476 50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 blda b l a c k Ziguincbr Tamba Dimubel StLwL Fatick L q a This Dakar ,..33 The patterns o f growth o f technical health personnel and their distribution between the private and the public sectors and betweenregions reflect not only the slow growth o f recruitments in the public sector under stabilization and structural adjustment reforms starting in the 80s, but also the growing solvency o f the demand for health care inthe urban areas, in Dakar in particular, which have sustained the growth of private practices in the country. In addition, disparities in the distribution between regions o f technical health personnel employed in the public sector, physicians and midwives in particular, are aggravated by the economic and social disparities between Dakar and other regions, the centralized management o f human resources and the weaknesses o f deliberate incentive measures to serve inthe poorest areas o fthe country. 4.34 Hence, while additional efforts from local communities have supported the growth o fbasic health infrastructures throughout the country under the implementation o f the primary health care strategy, and continue to support recurrent costs o f local health services since the health financing reforms o f the 90s, the public health sector has failed to support these local efforts by the substitution o f community health workers with better qualified healthpersonnel ingeneral, the substitution o f matrons with qualified midwives in particular. Indeed, the implementation of the primary health care strategy and the 90 decentralization o f the health delivery system during the past two decades have been based on a human resource base composed mainly o f nurses and community health workers. 4.35 Finally, Senegal has been unable so far to extend a network o f physicians and qualified midwives inthe rural areas andthe poorest regions o fthe country. Such failures inthe management of human resources will continue to hamper efforts to improve the quality o f local health services in general, reproductive health services in particular, outside the region o f Dakar. It i s no surprise that the population has grown impatient and discontent with local health services as evidenced by response on satisfaction with health services. As the prime stakeholders o f health services, health personnel are neither content with current working conditions: before 2001, health personnel unions have maintained an enduring strike over more than three years demanding the improvement o f their working conditions. Indeed, solving the human resource problem o f local health services in the poorest regions o f the country and regaining the thrust o f local communities are one o f the main challenges Senegal will have to face in the coming decades for reproductive health and child health improvements to regain a trajectory towards meetingthe HNP MDGs. 4.36 A symptom o f this variability o f qualified medical personnel is the gross inefficiency in health services. Much can be achieved through better use o f available resources. For instance, there i s sizeable inefficiency inthe production o f services inthe Dakar region. The number o f children immunized by DTC3 per nurse is 57 in Dakar, while it levels up to 260 in Louga, 176 in Thies and 121 in Fatick. As for maternal health, the same phenomena i s observed. While a midwife ensures 111prenatal visits in the public sector in Dakar, a midwife in the region o f Kolda produces 2086 visits. The same i s observed for deliveries with the personnel in the poor regions having overall a higher level o f productivity than the personnel in richer regions. The same can be observed for inpatient care. InDakar, the inpatient public sector activity is also very slow when examined in relation to the large quantity o f personnel available. In Dakar, 60 inpatient days are registered per medical doctor, while in Kolda, this number climbs to 1692. Finally, the data suggest that there i s room for large efficiency gains inthe delivery o f health services if supported by effective measures o f relocation o f health personnel from the region o fDakar towards the poorest regions o f the country. 91 Table 4-2: Senegal Technical efficiency of health services - Kolda Kaolack Ziguinchor Tamba Diourbel StLouis Fatick Louga T h i h Dakar # of DTF3 per nurse 116.5 214.2 166.0 103.7 121.5 93.6 121.7 260.6 176.1 57.3 # of antenatal care visit per 2085..5 817.7 441.8 969.6 613.8 355.8 591.4 801.4 424.0 111.2 midwife # of assisted delivery per 666.6 473.6 1377.0 353.0 289.7 132.0 277.9 362.7 260.05 84.9 midwife Occupancy rate of 43.9 38.0 42.8 42.2 83.1 72.0 27.9 56.5 85.5 60.0 hospitals'beds # of inpatient days per 1622.8 198.1 742.1 559.0 741 461.1 1587..4 982.5 486.7 61.0 physician # of inpatient days per nurse 80.3 29.5 103.8 83.9 61.0 112.6 129..6 204.7 166.5 38.2 Source :Senegal H N P CSR o n Poverty andHealth (forthcoming). 4.37 The quality and impact o fhealth services couldbe improved also ifthe continuity o fmaternal and childhealth services was increased. It i s only inthe Thies andZiguinchor regions that the continuity o f maternal health services, measured by the progression ratio between the first antenatal visit and the fourth antenatal visit, reaches 50%. In most regions, it ranges between 20% and 40%, and levels down to 11%inthe region o f Dakar. Similarly, the continuity o f child health services, measured by the progression ratio between the first dose o f DTC (DTC1) and the third dose o f DTC (DTRC3), i s very low. While the continuity o f child health services is above 60% in the Ziguinchor, Thies, Dakar, Fatick and Louga regions, its falls below 40% in the regions o f Kaolack and Tambacounda. Figure 4-14: Senegal - Continuity of maternal and child health services 100 60 I j 40 20 i 0 Kolda Kaolack Ziguinchor Tamba DiourbeV St Louis Fatick Lauga 'Thihies DakX Source: MOH and World Bank computations 92 What is to be done? 4.38 The two most important instruments for improving health outcomes in Senegal will be community participation, andbetter public-expendituremanagement. Communityparticipation 4.39 Community participation has been a hallmark o f Senegal's primary health care strategy since the 70s. From the emergence o f community initiatives which have shaped the primaryhealth delivery system inthe urban residential areas o f Pikine since the 70s, through the rural health experiments o f the Sine Saloum (current regions o f Kaolack and Fatick) during the 80s, and the interventions o f organizations o f migrants in the Senegal river valley, communities have always been in the forefront o f the primary health care strategy inSenegal. Through village pharmacies, healthhutsto rural maternity clinics and health posts, community initiatives have supported the development o f the basic health infrastructure inthe country, as discussed insubsection 4.2. 4.40 Buildingfrom the punctual and visible interventionsinbasic health infrastructure development, community participationinthe health sector has evolved since the 70s with the emergence of health committees involved inthe management o f local health services. Health committees have been strengthened by the ministry o f health as a deliberate strategy under the Bamako Initiative since the early 90s to provide for a vehicle for the participation o f the population in health promotion, social mobilization and the management o f health services. Everyhealth post, health center and health district inthe country has its health committee composed o f elected representatives o f its respective catchment's area population. 4.41 While they have supported the decentralization o f health services and have contributed to increased flexibility o f local health services in the country, health committees have not fully delivered yet relative to expectations. Health committee membership has been dominated by adult male as other elective positions inthe country, contributing therefore to the exclusion o f women and the youth whose respective community-based organizations could have served as bridges between essential health interventions and communities to reach their target populations. Consequently, while health committees are highly involved in the financial management o f local health services, their involvement in health promotion and social mobilization activities have beenrelatively limited inthe 90s. 4.42 Community participation, however, i s undergoing important changes in many communities beyond the confines o f the Bamako Initiative health committees. Community-based organizations, including women groups and youth associations are serving as the main local resources inthe fight against sexual mutilation and HIV/AIDS. Inaddition, community-based healthinsurance systems are emerginginmanyparts of the country since the early 90s buildingon the local community participation culture, mutual aid and solidarity values in villages and urban neighborhood and the organizational capacities o f the associative movement. These changes are being supported by the environment resulting from the 1996 decentralization reforms as community-based 93 organizations are beingpromoted as the main partners o f local government units - urban municipalities and rural communes - with increased responsibilities inmany development areas, including health. These changes are wideningthe bases o f local response capacities which, if harnessed through effective local partnerships for health, might prove as the most valuable resources towards meeting the HNPMDGs inthe country. Public expenditure management 4.43 The priority in the health sector should be better use o f public resources. Since the 80s, the main sources of finance o f health services in Senegal have been central government funding, private households, and international assistance2'. Despite the translation o f the high priority on human development in the health sector through the government's efforts to increase the share o f the public health sector in the national budget, central government funding inthe health sector has continued to fluctuate around 0.7% and 0.9% o f the country's gross domestic product (GDP) over the 90s. Public spending inthe health sector, however, has continued to benefit from increasinginflows o f international assistancethroughout the 90s. These additional efforts have raisedpublic spending in the health sector fi-om 1.21% o f GDP in 1993 to 1.71% in 2000 (table 4). From around US$8.4 in 1993, per capita public health expenditures have dropped to US$5.0 in 1994, as a consequence o f the devaluation o f the FCFA; despite increased internal efforts and external assistance, per capita public health expenditures were able to catch up with the early 90s' levels at the end o f the decade only: they have reached US$8.0 in 1999 and US$8.2 in 2000 (see table 4-3). Hence, while Senegal's levels o f public health spending compare favorably with other sub-saharan African countries, the country shares inthe regional constraints to raise real public spendinginthe health sector as a consequence o f the level o f development and structure o f the economy and rapid population growth. 25 Senegal has not developed yet a system of national health accounts, accordingly the contribution o f private social insurance inthe financing o f health services i s not well documented. Enacted since 1975, the coverage o f the social insurance system, however, is limited to employees o f the formal sector and their families which represent a small proportion ofthe Senegalesepopulation. 94 Table 4-3: Senegal -Public health expenditures 1993 Public Health Expenditures FCFAbillions Share (%) of total public expenditures Investment Share (%) Per capita Year Recurrent Treasury External Total Recurrent Investment Total Of GDP (US%) 1993 11.8 0.6 6.2 18.6 4.7 10.5 5.9 1.21 8.4 1994 15.8 0.6 6.3 22.7 5.O 6.3 1 5.3 1.12 , 5.0 ~ 1995 16.7 0.5 12.1 29.3 5.3 8.2 6.2 1.31 7.1 , 1996 18.8 0.9 12.5 32.2 6.0 7.3 6.5 1.36 7.3 ~ 1997 19.9 2.1 11.0 , 33.0 6.3 8.0 6.9 1.29 6.4 19981 23.7 3.6 4.1 31.4 7.6 3.9 6.2 1.15 5.9 19991 34.4 6.0 5.4 45.8 9.8 4.7 7.7 1.57 8.0 2000; 40.5 5.0 8.7 54.2 9.9 7.1 9.0 1.74 8.2 20011 43.6 8.6 11.0 , 64.1 9.1 7.3 8.4 1.90 9.4 4.44 In addition to central government funding, local governments have always contributed to public funding o f local health services since independence. But except in the urban communes o f Dakar, St Louis and Kaolack, local government funding has traditionally been relatively low and sporadic as a consequence o f limited fiscal and administrative capacities: only the commune o f Dakar has been able to develop its health administration capacities and to sustain significant investment and recurrent spending in basic health services. Between 1998 and 2000, local government sources o f funding for health services fluctuated around FCFA 1.5 and 1.8 billions per year, accounting therefore for a relatively low share o f total health spending in the country. Nonetheless, although a limited source o f public funding inthe health sector, local governments have started to play an important role in the institutional arrangements o f health financing in the country as a consequence o f ongoing decentralization reforms, as it will be discussed below. 4.45 Finally, private households have always been an important source o f financing in the health sector: in 1990 and 1997 for instance, it was estimated that households contributed up to 40% and 44% respectively of total health spending in the count$6. Recent estimates o f private household spending from the ESAM I1 survey, however, support that household health spending have reached between US$9 and US$10 per capita in 2002, suggesting that private households have evolved as the first source o f funding o fhealth services inthe country. But the public health sector continues to fail to capture and to recycle a significant share o f this internal source o f funding; however, 26Source: Banque Mondiale, 1994.Pour une meilleure santd en Afiique :les lecons de I'expereince. WashingtonDC; World HealthOrganization, 2000. The WorldHealth Report 2000 -Health Systems : Improving Performance. Geneva. 95 resources mobilized by health committees through user-fees at public health facilities ranged from FCFA 5.8 billions in 1998 to FCFA 7.8 billions in 2000, and have become an important source of recurrent funding for health services in general, basic health services inparticular. 4.46 The allocation o f public resources in the health sector has taken a path which i s biased towards investment and non-personnel recurrent spending and does not pay yet due attention to the motivation of healthpersonnel. The allocation o f public spending in the health sector has experienced significant changes over the 90s as a consequence o f efforts to rationalize public spending in the country, including a higher priority on investment and human development, and efforts to control recurrent spending. Public investment in the health sector has witnessed a relatively rapid growth throughout the decade compared to the growth o f health personnel spending. From FCFA 6.8 billions in 1993, public investment inthe health sector has grown steadily inthe first half o f the 90s to reach FCFA 13.1 billions in 1997; it had regained its pace under the health sector integrated development program (PDIS) to reach FCFA 13.7 billions in 2000 (see figure 4.15). Between 1995 and 1997, public investment accounted nearly 40% o f total public spendinginthe health sector; between 1998 and 2000, it share dropped to 25%; in2001, however, it increased to 32% of total spending. 4.47 Most o f the growth in public recurrent spending resulted from non-personnel recurrent spending. While personnel spending fluctuated around FCFA 9 billions between 1993 and 1995, and increased to FCFA 15.5 billions in 2001, non-personnel spending almost doubled between 1993 and 1994, from FCFA 3.8 billions to FCFA 7 billions, as a response to the devaluation of the FCFA, and increased to FCFA 22.3 billions in 1999 and FCFA 26.1 billions in 2000. It i s only from 1999 that personnel spending started growing steadily. Hence, throughout most o f the 9Os, while public investment andnon-personnel spending have all increased significantly, the relative share of personnel spending has deteriorated from 43% of total spending in 1993 to 26% in 2000: such patterns in the allocation o f resources have been met with a growing frustration among health personnel as evidenced by their enduring strike over the second half o f the decade. 4.48 The allocation o f public resources in the health sector has taken a path which is biased towards investment and does not pay due attention to the motivation o f health personnel. The allocation o f public spending in the health sector has experienced significant changes over the 90s as a consequence o f efforts to rationalize public spending in the country, including a higher priority on investment and human development, and efforts to control recurrent spending. Public investment in the health sector has witnessed a relatively rapid growth throughout the decade compared to the growth o f recurrent spending. From FCFA 6.8 billions in 1993, public investment in the health sector has grown steadily throughthe 90s to reach FCFA 27.4 billions in2000 (see figure 4.15). Before 1995, public investment accounted for less than 40% o f total public spending in the health sector, by the end o f the 9Os, however, public investment's share o f total spending increased to 50% o f total spending. Hence, although public recurrent 96 spending inthe health sector grew at a sustained pace, the balance o fpublic recurrent and investmentspending has deteriorated duringthe 90s. 4.49 Most o f the growth in public recurrent spending resulted from non-personnel recurrent spending. While personnel spending fluctuated around FCFA 9 billions between 1993 and 1998, non-personnel spending almost doubled between 1993 and 1994, from FCFA 3.8 billions to FCFA 7 billions, as a response to the devaluation o f the FCFA, and increased to FCFA 10.4 billions in 1999 and FCFA 12 billions in2000. It is only from 1999 that personnel spendingstarted growing steadily. Hence, throughout most o f the 90s, while public investment and non-personnel spending have all increased significantly, the relative share o f personnel spending has deteriorated: such patterns in the allocation o f resources have been met with a growing hstration among health personnel as evidenced by their enduring strike over the second half o fthe decade. 4.50 The distribution o f health facilities and health personnel confirm that the largest share o f public spending continue to be concentrated in the richest region, Dakar. Moreover, the technical efficiency o f service delivery in the Dakar region is very low compared to other regions in the country. In addition, recent trends in internal public investmentinthe health sector has favored central administration services, secondary and tertiary hospitals over the primary health care level as a local response eventually to the concentration o f external support on basic health infrastructure development and priority public health programs. Hence, as external support moves from project support to budgetary support in the health sector over the next decade, more attention should be focused on the efficiency and equity o f public expenditures. More attention should be focused also on the incentive regime associated with institutional arrangements o f health financing mechanisms inthe country. Figure4-15: Senegal-Public healthspending(FCFA millions)by category:personnel,non- personnelrecurrent,andinvestmentspendingbetween 1993 and2000 30,000 25,000 20,000 15,000 10,000 5,000 0 1 I 8 I I I I I I 1993 1994 1995 1996 1997 1998 1999 2000 2001 97 Note: See note under table 4 for the important changes innon-personnelrecurrent spending starting in 1998. 4.5 1 Institutional arrangements o f health financing mechanisms in the country are not oriented towards the production o f results. The institutional arrangements o f Senegal healthfinancing mechanisms have undergone important changes duringthe last decade as a consequence o f health financing reforms o f the early 90s and the devolution reforms o f 1996. First, the health financing reforms o f the early 90s have moved health financing arrangements in the country from a regime based on public financing principles, which have prevailed since the country's independence, towards health financing arrangements based on the combination o f central public funding and user-fees in public service delivery organizations. To support the decentralization reforms within the health sector and the emerging district-based health system in the early 90s, public financing o f basic health services has been strengthenedthrough the introduction o f the health district as a budgetary unit in the ministry o f health budget, significant increase o f the levels o f central funding o f non-personnel recurrent costs, and the delegation o f administrative authority over local budgets to district medical officers. Mechanisms o f allocation o f resources prevailing inthe hospital sub-sector - line-item budgets elaborated at the central level o f the ministryo f health to purchase local health service inputs - served as the main mechanisms for allocating resources to health districts with no demonstrable relationwith the size o f the target population, objectives and results o f health district operations. The main incentive change resulting from these reforms were (i)the position o f district medical officer became a highly valued and motivating position in the health administration hierarchy during the first half o f the decade, and (ii) district medical officers were encouraged to consume the line-item credits in order to claim higher budgetsinthe following years. 4.52 Inaddition, these changes inthe levels andmechanisms ofpublic fundingofbasic health services were followed by the adoption and implementation o f the Bamako Initiative (BI) in the country. Symbolic consultation fees have been applied in the basic heath services since the early SOs, but starting in the early 9Os, however, user-fees became a deliberate resource mobilization strategy in the health sector based on the BI model. As a result, resource mobilized through out-of-pocket spending by users in a average public health post and a typical health center were multiplied respectively by 6 and 12 between 1991 and 199927.Hence, while public personnel spending inhealth posts and health centers remained nominally unchanged, health personnel were faced with in increasing private source o f funding generated mainly by their curative activities and efforts, that they controlled through their influence over local health committees. Accordingly, while the BIwas implemented to revitalize basic health services, incentives associated with local resource mobilization and management strategies have been biased towards curative care activities. 2'Diop, F. et als 2001. Decentralisation et systeme de santk au Senegal :une synthese. Bethesda, MD. Partnerships for Health ReformProject, Abt Associates Inc. 98 4.53 Finally, the increasing resources o f public health facilities have grown as an unaffordable burden to households and users. Users o f public health facilities are more and more discontent with the health services mainly as a consequence o f the costs o f consultations and drugs. Inaddition, a large share o f the populationidentifies costs as one o f the main reasons for not using modern health services. The poorest segments o f the population are likely to suffer disproportionately from partial and total exclusion, and the burden resulting from the increasing importance o f private financing in public health facilities. 4.54 Second, the devolution reforms o f 1996 have altered mainly the public sector component o f the institutional arrangements resulting from the early 90s financing reforms as a result o f the fiscal decentralization measures accompanying the devolution o f health to local government units. Starting in 1997, regional hospitals and health districts' budgetshave been transferred to local government units, regions and communes where health centers are located. The devolution reforms did not take into account the reality and the organizational structures o f the district health system, and have dispersed the lines o f accountability within basic health services in the country. District medical officers and local health personnel have to negotiate their health budgets and to be accountable to local elected officials o f many local government units covered by the health district and health committees. Nevertheless, they continue to receive directives and instructions from, and are expected to remain accountable towards, the ministry o f health bureaucracy and multiple vertical programs, while most o f the resources which support their operations and activities are controlled by the local fund holders (local government units and health committees). 4.55 It could be expected that as local government fiscal capacities are increased, as a consequence o f fiscal reforms and increased transfers from the central level, and elected local officials' sense o f ownership and responsibility o f local health services i s institutionalized, that basic health services may benefit from increased internal public funding. For these additional resources to be translated into positive changes in the performance o f basic health services, however, synergies between central and local public funding and other sources o f funding have to be harnessed within performance- based, transparent and accountable organizational structures. Such a process o f adaptation of local health services to the devolved health system may take time, while Senegal needs to realign its health financing strategies over the next decade to meet the 2015 HNP MDGs. 5. EDUCATION 5.1 Education i s perhaps the most important factor in determining whether poor people can participate in economic growth. Education can also be a spur to economic growth. And, as previous chapters have shown, the education o f girls in particular can have far-reaching effects on health and other dimensions o f human welfare. For these reasons, the Senegal PRSP gives particular emphasis to the fulfillment o f various education goals, the best-knowno fwhich i s that o funiversal primary education by 2015. 5.2 The importance o f education to poverty reduction in general, and to Senegal in particular, stands in sharp contrast to the disappointing performance o f Senegal's education system. Although the gross primary enrolment rate stood at a reasonable 71.6 percent in 2001, the inefficiencies inthe system, high dropout rates, and misallocation o f public resources all make the 2015 goal unreachable. At the same time, the identification of the problems in the system points to directions for reform, so that significant improvements are possible. After describing the state o f education in Senegal, this chapter examines changes to the system, in particular the management o f public resources, that could accelerate progress towards the PRSP education goals. A. THESTATE OFEDUCATIONINSENEGAL Enrollment 5.3 About 1.4million studentswere enrolled at various levels o f education inSenegal during2001, o fwhich 2.2 percent attended pre-primary, 77 percent were inprimary, 13.5 and 4.3 percent respectively attended lower and upper secondary general school. Technical/vocational education sub-sector i s very small with less than 1 percent o f students while higher education currently absorbs 2.1 percent o f all students. Pre-primary education is not widespread with 31650 students (Table 6) attending preschool in 2001; this implies a gross enrollment rate o f less than 15 percent suggestingthat a majority o f children do not have access to early childhood education. 5.4 Over the past five years the trend in primary enrollments have exceeded population trends o f school-age children resulting in a gross enrollment rate o f 71.6 percent in 2001. This pattern i s expected to persist over the next few years as the government tries to reach its goal o f universal primary enrolment by 2010. To reach that goal it i s essential that the intake rate in grade 1 should reach 100 percent by 2005; this looks unlikely given the current intake rate o f 85 percent and the level o f inefficiency in the system. 100 Table 5-1: Trends in Enrollment by Leveland Type of Education, 1997-2001 Year Pre-Primary Primary GeneralSecondary Pre- Higher Lower Secondary Upper Secondary Vocational khcation 1997 18513 875661 151735 59511 25846 1998 19880 954758 155790 59162 25233 1999 23625 1026570 172469 64985 26113 2000 25392 1034065 186138 58983 8972 27000 2001 31650 1107712 195012 61274 8101 30788 Source: Statistiques Scolaires et Universitaires, various issues, MOE 5.5 The above goal o f universalprimary education i s closely linkedto the MDG goals o f 100 percent completion rate (PCR) in primary education by 2015. PCR in Senegal is currently around 30 percent due to a high degree o f inefficiency brought on by large dropout and repetition rates. It should be emphasized that achievement o f universal primary enrollment would not by itself lead to 100 percent PCR, which depends on lowering dropout rates to less than 2 percent and repetitionrate to five percent. 5.6 Access to secondary education i s severely limited, mainly because o f insufficient number o f public schools resulting in a gross enrollment rate o f only 16.4 percent in 2001. Even in existing schools access to full schooling i s not guaranteed - less than 50 percent o f public schools and only 22 percent o f private schools offer schooling up to Grade 13 (Terminal). Demand for secondary schooling i s projected to rise rapidly, both because o f higher primary enrollments and also due to the expected increase in the population o f children between the ages o f eleven and seventeen. This segment o f population is projectedto increase to 2.1 million in2015 from 1.5 million in2000. 5.7 Prevocational education has relatively few students and its share in overall enrollments has not changed significantly during the past five years. Higher education enrollments are also very small, but as discussed later, the sub-sector absorbs a much larger portion o ftotal education expenditurez8. 5.8 The overall gross enrollment rate provides an aggregate indicator for the level o f access to the education systembut it i s not very useful for policy purposes. Its usefulness i s constrained because it does not incorporateinformation on dropout and repetition rates. This hinders it in two ways: first it is a weighted average across all relevant grades but each grade has the same weight. Thus its estimation in an education system with significant number o f dropouts would result in an overestimate because o f larger number of students in lower grades2'. Second, high repetition rates imply the presence o f a significant number o f overage children in total enrollments; this inflates the value o f the '*Lack o f data on higher education is a major constraint o n a meaningful discussion o f the sub-sector. Iti s expected that it would be covered ingreater detail ina later document. 29 Grade specific enrollment shares are used as weights; ifdropouts are significant they result inlower number o f students inupper grades. The estimated value o f gross enrollment rate is, therefore, biased upwards towards lower grades. 101 gross enrollment rates since the numerator inits definition includes all children ina cycle irrespective o f age, while the denominator includes only the number o f school-age children in the cycle. This is a crucial problem in primary and secondary education cycles inSenegal where boththe dropout and repetitionrates are high. Survivalrates 5.9 In the absence of historical information related to a specific cohort, a methodology that utilizes cross-sectional data on non-repeaters (i.e., non-repeaters o f various cohort inthe same year) i s used to construct survival rates. The limitation o f this methodology is that it ignores differences in cohort sizes, intake rates and continuation rates to subsequent grades across actual cohorts. Table 5-2: Cohort Survival Rates, Grade Enrollment Rates, and Transition Rates between Cycle, 2001 Survival Rates Grade Specific enrollment Transition rates Rates for a Cohort Grades Boys Girls Boys Girls Boys Girls 1 1.00 1.00 0.85 0.84 2 0.87 0.85 0.74 0.72 3 0.71 0.69 0.60 0.58 Primary 4 0.57 0.55 0.49 0.46 5 0.47 0.44 0.40 0.37 6 0.34 0.30 0.29 0.25 7 0.12 0.09 0.10 0.08 Lower 8 0.10 0.08 0.09 0.07 35% 31% Secondary 9 0.08 0.06 0.07 0.05 10 0.06 0.05 0.05 0.04 11 0.03 0.02 0.02 0.02 Upper 12 0.02 0.02 0.02 0.01 49% 49% Secondary 13 0.01 0.01 0.01 0.01 School Life Expectancy (implied) 3.7 3.5 Source: Based on data in Statistiques scolaires et universitaires, various issues, MOE, 2001, authors' calculation. 5.10 Application of this methodology to Senegal data indicates (Table 8) that out o f every 100 boys and 100 girls in grade 1, 34 boys and 30 girls reach grade 6 (the end o f primary cycle); only 6 boys and 5 girls reach grade 10; andjust 1boy and 1girl are able to reach the final year o f secondary cycle. Results for primary are not very surprising as there are very few students who repeat or repeat or drop in this cycle. While repetition and dropout rates for private schools were not available, it is not expected that they would be large enough to change the overall pattem o f survival. 102 5.11 Grade-speciJic enrollment rates: The above survival rate analysis pertains only to children in the education system but does not provide information on the proportion o f students ineach cohort who actually attend school. Multiplyingthe gross intake rate3' by the survival rate in each subsequent grade provides the grade-specific enrollment rates that are reported in Table 8. Summing up these rates across all grades provides the school life expectancy o f only 3.7 years for boys and 3.5 years for girls in Senegal. This suggests that if the pattern o f survival rate and intake rate persists over time, future men and women would have, on average, only 3.7 and 3.5 years o f formal schooling as a part o f their human capital profile. These outcomes are lower than the average o f 5.6 years in Africa and comparable to 4.0 and 2.6 years for boys and girls in Guinea. It implies that if Senegal wants to use human capital as a factor for economic growth, it should improve the quality and accessto education, especially with regards to secondary education. 5.12 The transition o f student between primary to lower secondary and from lower secondary to upper secondary in Senegal takes place in grades 6 and 10 respectively. In Senegal the rates o f transition form primary to secondary are very low at 35 percent and 31 percent for boys and girls respectively. To raise the efficiency o f student flow, these transition rates would have to improve significantly if the aim i s to make six years o f education as part o f human capital profile for all students. Transition rate31are relatively low at 49 percent for boys and 49 percent for girls between lower and upper secondary school cycles. 5.13 It is important to comprehend at which point students are lost to the education system. Since few children reach the last year o f primary cycle, and the inter-cycle transition rates are low, it i s clear that the selection process inthe school system operates through a process of dropping out both within and between school cycles. In other words, o f the 1boy who reaches the last year o f upper secondary out o f 100 first graders, 66 (67 percent o f a total o f 99 dropouts) drop out within the primary school cycle, while 22 boys (22.5 percent) drop out during transition betweenprimary and secondary school cycles. About 6 (6 percent) drop out within secondary cycle, while the rest (3 boys, 3 percent) leave the system at the end o f grade Thus a total o f 25 boys drop between different cycles, while 73 do within the three school cycles. 5.14 Inconclusion, ifSenegal wishes to improve the humancapitalprofile ofits future labor force, it has to improve its rate o f transition between education cycles and improve the performance and internal efficiency o f primary and secondary schools. At present, a substantial number of students are leaving the school system with incomplete schooling. This results in a reduced pool o f qualified students for admission into post-secondary 30 It i sa ratio o f all non-repeating children ingrade 1to the population inthe corresponding age group o f the non-repeating grade 1students. 31 These transition rates are for a specific cohort andhence are net o frepeaters. 32 One student drops out within the upper secondary school cycle. 103 education and impacts the quantity and quality o f human capital available for improving the long-term prospects o fthe economy. Quality 5.15 The disturbing trends in completion rates are probably related to the quality o f schooling. Independentassessments o f that quality confirms that education in Senegal i s not delivering on its promise, even when compared to other countries in the sub-region. Senegal participates in the evaluation o f education outcomes that are conducted by PASEC. These evaluations are based on a series o f standardized test (in French and Mathematics) given to students in a cross-section o f Francophone African countries. These results place Senegal at the bottom o f the group o f countries that participated in these evaluations. This i s surprising because the group contains countries whose per- student spending in primary and per-capita GDP is lower than Senegal. InFrench, for example, the average score i s 16.6 in Senegal (Table 10) and i s significantly different than the average score o f 24.3 in Cameroon. Similarly, the score in Mathematics is barely above that o f Cote D'Ivoire but much lower than Burkina Faso and Cameroon, countries that are at a lower level o f economic development when compared with Senegal. Gender disparity (Table 11) inattainment i s low inFrench and infavor o f girls (by 10 percent) in Mathematics. More significant are the differences in educational attainment between rural and urban areas, with the latter leading the former by 26 percent and 15 percent inFrench and Mathematics scores respectively. The same study indicated higherrates o frepetitioninurban areas while gender differences were small. Table 5-3: Scores (grade 5) on standardized tests in a group of countries Countries French Mathematics BurkinaFaso 20.9 20.6 CBte d'Ivoire 21.2 17.2 Cameroon 24.3 23 Senegal 16.6 18.3 Source: Pasec, 1998 B. PUBLICEXPENDITUREMANAGEMENTAND SENEGAL'S EDUCATIONSECTOR 5.16 The disappointing perfonnance o f Senegal's education sector is particularly troubling because Senegal spends a higher proportion o f its GDP on education than most low-income countries. Clearly, there is a problem with how these resources are deployed. A detailed analysis shows that public expenditures in education are skewed in favor o f higher levels o f education, in favor o f salaries over other recurrent costs, and in favor o fthe non-poor. 104 Total spending 5.17 In Senegal, both the government and households play an important part in the financing o f educational services. Most private schools in Senegal receive some financial resources from the government; like many other countries private provision o f education i s higher in secondary education incomparison to primary. Households contribute intwo ways: they pay tuition inprivate schools; they also provide for some pedagogic inputs for their children. Richer households also pay for after-school tuition for their children. Table 5-4: Public Spendingon Education(millionsof CFA and dollars) Year Total Current Capital CFA Dollars Index CFA Dollars Index CFA Dollars Index 1990 46157 107 100 41857 97 100 4300 10 100 1995 82525 165 179 76852 154 184 5673 11 132 1996 91976 180 199 84437 165 202 7539 15 175 1997 98419 169 213 87185 149 208 11234 19 261 1998 107700 183 233 88146 149 211 19554 33 455 1999 119536 194 259 92313 150 221 27223 44 633 2000 131955 185 286 98690 139 236 33265 47 774 2001 137715 188 298 108596 148 259 29119 40 677 Source: Statistiques Scolaires et Universitaires,various issues, MOE 5.18 Public spending. Government spending (Table 5-4) on education has gone up by about 198 percent since 1990 (in nominal terms), and this increase has occurred disproportionately in favor o f capital expenditure, which has grown up by 577 percent compared to 159 percent for recurrent expenditure. Capital expenditure has been steadily increasing duringthis period except for a slight decline in2001. 5.19 An examination o f distribution o f recurrent expenditure between different sub- sectors (Table 5-5) indicates a shift away from secondary andtowards primary education. The share o f administrative expenditure has declined over the same period. While recurrent expenditure inprimary and secondary has risen consistently since 2000, the rate of growth for primary i s higher and for secondary i s lower than that o f total recurrent education expenditure. Data on sub-sectoral capital expenditures i s not available but it is reasonable to expect that it would exhibit volatility similar to that o f total capital expenditure shown inTable 5-4. Table 5-5: Distributionof public recurrentexpenditureon educationby level selected years between2000-2002 Year Total Pre-Primary Primary Secondary Technical/Vocational Post-Secondary Other' US$ % US$ % US$ % US$ % US$ % US$ % million million million million million million 2000 99 0.7 0.7 37.9 38.4 20.3 20.6 23 2.3 25.8 26.1 11.7 11.9 2001 2 109 0.7 0.7 41.7 38.4 22.4 20.6 25 2.3 28.4 26.1 12.9 11.9 2002 3 115 0.8 0.7 48.7 42.2 20.9 18.1 39 3.3 30.6 26.5 10.4 9.0 Source: Statistiques scolaires et universitaires, various issues, MOE 'Includes administrative staff 223Budgetnumbers were available for 2001 and 2002 105 5.20 Primary education expenditure has increased over the past few years and there is evidence that most o f the resources are being devoted to paying the salaries o f additional teachers. It i s expected that the share o fprimary would have to go up further to reach the goal o f universal primary enrollments by 2010. This is necessary because student enrollments are projected to rise at a faster rate than expenditure on primary education. Capital expenditure i s also projectedto grow with a rise inclassroom construction inboth primary and secondary education. 5.21 Private spending. Households contribute a moderate share o ftheir income to total education spending in Senegal. Most o f the information on private education spending comes from recent survey household survey estimates the amount spent on pedagogic materials. More than half o f the households reported monthly expenditures on education with an average spendingo f around CFA 1500 (US$ 3) per month. Summingup across all households, private spending on education amounted to CFA 2.5 billion. Duringthe same year, total public spendingon education was equivalent to CFA 138 billion. Taken together, public andprivate spending in 1999 amounted to about 4.2 percent o f GDP with private spending equivalent to less than 0.1 percent o f GDP. 5.22 In terms of distribution of private spending between levels of education, rough estimate suggest that primary education received 22 percent, secondary 50 percent, followed by pre-primary at 11percent. Households account for about 2.2 percent o f total national spending on education, a proportion that i s equivalent to spendingby India33(2 percent), Jordan (2 percent), and Malaysia (4 percent) but much lower than Madagascar (40 percent) and Kenya (38 percent)34. Composition of spending 5.23 Total public spending by different expenditure categories for each level o f education35i s shown in Table 20. Salaries constitute the major portion o f total recurrent expenditure in education, especially in primary and secondary (general and vocational/technical) while their share inhigher education is not as high. Public spending on non-salary inputs constitute a very small portion o f the education spending while transfers absorb a significant part o ftotal recurrent spending. 5.24 Total salaries have been further divided between teaching personnel and non- teaching personnel (administrators in schools, administrators in regions as well as in central ministry, and support staff such as cleaners, janitors etc). This breakdown shows that inprimary education, salaries constitute 95 percent o f recurrent expenditures; salaries o f teachers and non-teaching personnel account for 89 percent and 6 percent respectively o f primary education recurrent expenditures. Similarly, in secondary general, 94 percent 33Total private expenditure onprimaryand secondary education for these countries is used for comparison. 34Data for these countries is available at EDSTATS. 35Subsidies to private schools inprimaryand secondary education levels are included. 106 of total recurrent spending is on salaries, divided between the salaries of teachers (76 percent of recurrent) and non-teaching personnel (18 percent). It is significant that while the share of teachers' salaries declines between primary and secondary education, the share of non-teaching personnel has risen. Thus administrative functions constitute a larger fraction of per student expenditures at higher levels of e d ~ c a t i o n ~ ~ . Table 5-6: Distributionof public education recurrentexpenditureby levels and categories, 2000 EducationLevel Primary Secondary Secondary Higher Total I General1 Technicai Education Salaries 95.0 94.4 89.0 49.7 84.6 Teachers 89.0 76.4 Non-teaching 6.0 18.0 personnel Goods & services 5.0 5.6 11.0 5.1 3.9 Transfers 0.0 0.04 0.0 45.0 11.5 Total 100 100 100 100 100 Source: Estimatedfromdata providedby MOE 1Includelower andupper secondary 5.25 Operational expenditures (goods and services) rise across education levels37and before falling to 5.1 percent for higher education (Table 20). The share for higher education would rise to 9 percent in the absence of scholarships that account for 45 percent of all expenditures in that sub-sector. The large share of scholarships in higher education in Senegal i s disturbing in the context where the demand for higher education i s strong as evidenced by highrates of repetition inthe terminal grade of upper secondary education. Moreover, as indicated in the later section on equity, the bulk of the benefits of undergraduate education are captured privately and the majority of students come from families whose income i s more thanthe average income inthe country. Table 5-7: Distribution of enrollments and recurrent spendingby educationlevel, 2000 EducationLevel Enrollments Percentage Recurrent Percentage spending pre-Primary 25392 1.9 669 0.8 Primary 1034065 77.0 37857 43.5 Lower Secondary 186138 13.9 11927 13.7 UpperSecondary 58983 4.4 8394 9.7 Secondary Technical 10667 0.8 2310 2.7 Higher Education 27000 2.0 25802 29.7 Total 1342245 100 86959 100 Source: Statistiquesscolaires et universitaires,various issues, MOE 36 The same would be true for higher education expenditure but transfers (mainly scholarships) constitute the major share o f expenditure thereby reducing the share o fnon-teacher salary. 37 These expenditures do not contain resources that are made available from other sources to the education sector. These include the "Depense commune" from the Ministry o f Finance that pays for electricity, water and telephones services for schools. It also does not contain resources given directly to local councils for the maintenance o f schools (fonds de dotation), nor does it include funds providedbydonors for books andother pedagogical inputsundervarious projects. 107 5.26 InTable 21, overall student population is distributedbetween different levels o f education. Not surprisingly, primary education has the largest share with 77 percent; lower secondary has 14 percent while upper secondary and higher education have 4.4 percent and 2 percent respectively. The distribution o f expenditures in the same table, however, reveals the disparity in funding between different levels o f education. Primary education's (with 77 percent o f students) share in total expenditures i s only 43.5 percent while higher education's (with 2 percent o f students) share is 30 percent. Since salaries constitute the major share o f expenditures, these differences in per-student expenditure are traceable to disparities in staffing between different levels o f education. It is demonstrated in later sections, that significant differences in student-teacher and student- non teaching staff ratios across education levels result in disparities in per student expenditures. The situation is more complicated in higher education because o f significant subsidies (scholarships) to students that further ratchet up the cost. Giventhis disparity inper student expenditures, the increasing trend inthe share o f higher education over time is indeed a worrisome sign. 5.27 Expenditures per student at different levels. Total spending on education i s disaggregated to generate estimates o f per student spending (unit cost) in order to view the pattern o f spending across education levels, compare it with international experience and understand the causes for differences in unit spending across these levels. The estimates in Table 22 suggest that unit cost vary widely between education levels with CFA 36610 for primary education and CFA 955637 for higher education. Unit cost rises as one moves through the education levels with spending in lower secondary and upper secondary 1.8 and 3.9 times more per student than in primary, spending in secondary vocational/technical 5.9 times more, and finally, in higher education it rises dramatically to a multiple 26.1 times more than inprimary. As shown in later sections, most o f these differences can be traced to variation instaffing within each sub-sector. Table 5-8: Unitcost per student by educationlevelinthe publicsector, 2000 EducationLevel Enrollments Total Public Unitcost Spending (mn) CFA Dollars As a multiple of primary Primary 1034065 37857 36610 50 1.0 Lower Secondary 186138 11927 64076 87 1.8 Upper Secondary 58983 8394 142320 194 3.9 Vocational/Technical 10667 2310 216556 295 5.9 Higher Education 27000 25802 955637 1304 26,l Source: Estimatedfrom data provided byMOE 5.28 With the availability o f unit costs, it is now possible to estimate losses due to inefficiencies caused by dropout and repetition rates in each sub-sector. As noted in a previous section, the efficiency was 0.35 in the primary cycle and 0.27 in the secondary cycle. These inefficiencies result in a waste o f resources amounting to CFA 24.6 billion (US$ 35 million) in primary and CFA 14.9 billion (US$ 21 million) in secondary, for a grand total o f CFA 39.5 billion (US$ 55 million) in the two cycles. This total loss is equivalent to 40 percent o fthe entire recurrent education budget for 2000. 108 Table5-9: Comparisonofunitcost at different educationlevels Multiple of primary unitcost Primary Secondary Secondary Higher education General' technical/vocational Senegal 1 2.3 5.9 26.1 Burkina Faso 1 1.9 n/a 12.8 Guinea 1 1.9 16.6 32.0 Mali 1 9.4 14.9 14.2 OECD 1 1.4 n/a 2.9 Source: Estimatedwith data from Statistiquesscolaireset universitatires,MOE; World Bank 1Includes lower and upper secondary 5.29 Senegal's pattern o f education spending i s similar to other francophone countries and different from the average for OECD countries. While most countries inthe region have nearly the same multiple for secondary general education, spending per student in Senegal is much lower for secondary technicalhocational education. Higher education unit cost as amultiple ofprimaryis muchhigher inSenegal; only the multiplefor higher education in Guinea i s larger. It i s clear from discussion in previous sections that the multiple for higher education can be reduced by a considerable margin if the proportion o fhigher education spending on scholarships is reduced. 5.30 The overall unit cost for any sub-sector is an aggregate measure than can be decomposed into different components, mainly salary and non-salary spending. Since unit cost is defined as a ratio o f total recurrent spending and the number of students at each education level, dividing total spending between teachers, non-teachers and other school inputs makes it possible to delineate those components o f unit cost that are responsible for its variation between education levels. Total spending on teachers' salaries at any education level, for example, i s affected by their average annual salary and the student-teacher ratio at that education level. Higher average salary will increase while a larger student-teacher ratio will decrease total salary spending. Table 5-10: Disaggregationof per studentpublicspending, 2000 Spendingper student Multiple of primary education EducationLevel Primary Secondary Primary Secondary General General1 Teacher salaries 32583 63346 1.0 1.9 Annual average 14393 95058 1.0 6.6 Student-teacher ratio 44 37 1.0 0.8 Non-teachingstaff salary 2197 14923 1.0 6.8 Annual average 13714 53377 1.0 3.9 Student-non-teacher ratio 624 67 1.0 0.1 Non-salary inputs 1831 4634 1.0 2.5 Source: Estimatedfrom dataprovidedby MOE 1Includeslower andupper secondary 5.31 This methodology has been applied to primary and secondary (general) education levels and the results are illustrated in Table 24. It i s evident from the table that overall 109 spending per student in secondary general education is about 2.3 times that o f primary. An examination o fthe different components o f unit cost reveals that the main differences arise from disparities in(i) teachers' salaries, which are 1.9 times higher insecondary, (ii) non-teaching staffs salaries, which are 6.8 times larger than inprimary. Both differences are a consequence of higher salaries for teaching and non-teaching staff in secondary education (which are larger by a multiple o f 6.6 and 3.9 times respectively than corresponding salaries for primaryeducation), and, a smaller student-teacher and student- non-teaching staff ratio insecondary (80 percent and 10percent respectively o fprimary). Spending on non-salary inputs, at 2.5 times that o f primary, also contributes to higher overall unit cost insecondary. 5.32 To summarize the above results, most variations in the proportion o f unit costs attributed to teachers' salaries are caused by differences in average salaries and the student-teacher ratio. Higher salaries and lower student-teacher ratio result in higher level o f per student spending on teacher salaries in secondary education. This is a consequence of the fact that interms o f overall expenditure in each sub-sector, salaries o f teachers amount to 76 percent in secondary and 89 percent in primary. This compensates, to a significant degree, the large differences inaverage salaries and student- teacher ratio. But non-teacher salaries are 18 percent o f total expenditures in secondary and only 6 percent o f total expenditures inprimary. It i s here, that the combined effect o f higher average salaries and lower student-non-teaching staff ratio makes its impact on the unit cost, ,with the result that non-teaching staff salaries are 6.8 times larger than in primary. Incidence ofpublic expenditures 5.33 The inherent biases in public spending in Senegal-in favor o f higher levels o f education, in favor of salaries-have an impact on the incidence o f public spending, which i s decidedly skewed towards the non-poor andboys. Table 5-11: Unit cost by education level (CFA), 2000 EducationIndicator Primary Secondary1 Higher Subsidy per student 36610 82903 955637 Multiple of primaryunit cost 1.0 2.3 26.1 Subsidy per capita 3972 2132 2707 Source: Ministry of Education andHousehold Survey data `Includes lower andupper secondary and technicaYvocationa1 5.34 Inthis analysis unit costs are considered equivalent to subsidy per student. The subsidy per student rises across education level. The per capita subsidy on the other hand, declines across education levels as the number of students decline across successive education levels. The total education subsidy per capita i s CFA 8812 in the year 2000, o f which primary subsidy per capita was CFA 3972. Secondary and higher education per capita subsidy was CFA 2132 andCFA 2707 respectively. 110 5.35 Inequalities inthe distribution o f public education spending can occur for various reasons. The main cause, usually, is the discrepancy in access to education, especially secondary and higher education. If80 percent o f students inthe universitycome from the richest quintile and only 5 percent come from the poorest, then a major portion o f subsidies to higher education are captured by the rich segment o f the population. Thus disparity in access to education translates into unequal distribution o f public education spending. If these differences persist for a significant period of time they would have tremendous negative repercussions on efforts to reduce poverty and decrease inequality inthe population. 5.36 The analysis o f primary per capita subsidies by income groups indicates that the poorest quartile receives a much smaller subsidy than the richest quartile. The per capita subsidyinprimary education for the richest was 1.7 times larger than for the poorest. But the disparity increases at the secondary level where the per capita subsidy for the richest quartile i s 4.0 times that of the poorest. The gap i s even larger for higher education where the richest group is able to capture 7.3 times more subsidy than the poorest quintile. 3.37 The combination o f primary, secondary and higher education subsidies yields all education subsidies per capita. The poorest quartile received CFA 4227 per capita in education subsidies, o f which 64 percent were received through primary education and only 18 percent through higher education. Incontrast, the richest quartile received CFA 13202 per capita in education subsidies, o f which 35 percent was received through primary education and42 percent through higher education. Table 5-12: Distributionof per capita subsidyby educationlevel and income(CFA), 2000 Incomequartiles Primary Secondary1 Higher Total First quartile 1.0 1.0 1.0 1.0 Secondary quartile 1.5 2.8 1.7 1.8 Thirdquartile 1.6 3.3 4.3 2.4 Fourth quartile 1.7 4.0 7.3 3.1 Source: Ministry o fEducation andHouseholdSurvey data 'Includes lower and upper secondary andtechnicallvocational 5.38 Interms of absolute expenditures by education levels, the poorest group gets 17 percent o f total spending on primary education while the richest get 29 percent. But, the richest group captures three times the overall subsidy than the poorest group. The share of the richest group in total education expenditures was 37.5 percent compared to 12 percent for the poorest. 111 Figure5-1: Distributionof publiceducationspendingby level and Lorenz distribution,2000 100 75 50 25 0 0 25 50 75 100 CumulativeShare of Population Sources: IntegratedHousehold Survey (2000), Annuaire Statistique, MOE (1999/2000),MOF (2000) 5.39 In order to better illustrate disparities across income groups, the distribution of public spending is shownjointly with the Lorenz distribution o f household expenditures for the year 2000. The diagonal line i s a locus o f all those points where the cumulative share o f the population equals the cumulative share o f recurrent public education spending. Public spending on primary and secondary education was weakly progressive in 2000 because their distribution lied in between the diagonal line and the Lorenz distribution. Spending on higher education, incontrast, was more regressively distributed than the Lorenz distribution. 5.40 Girls received about 72 percent of the subsidy received by boys through all education spending. In other words, for every CFA spent on girls' education, boys received a subsidy equal to CFA 1.4. 112 Table 5-13: Unit cost by gender and education level (CFA), 2000 EducationIndicator Primary Secondary1 Higher Subsidy per capita 3958 1908 2707 Subsidy per capita for boys 4240 2401 3349 Subsidy per capita for girls 3677 1419 2070 Source: Ministry of Education andHouseholdSurveydata 1Includeslower and upper secondary andtechnicalivocational 5.41 Gender disparities are lowest at the primary level with girls receiving 88 percent o f the subsidy received by boys. But the disparities rise across education levels. For example, at secondary level, girls receive only 59 percent o f subsidyreceivedby the boys while inhigher education the percentage i s at 62 percent. The gender gap, moreover, i s widest for the poorest quintile at every educational level. Inprimary education, girls in the poorest quartile receive 83 percent o f the subsidy received by boys; at the secondary level they receive only 74 percent while at the higher education level they receive 63 percent3'. 5.42 A comparison across Sub Saharan African countries indicates that in 2001 Senegal and Guinea have the most unevendistribution o f education subsidies (Table 28). Most o fthese countries have higher primary enrollment rates than Senegal. This suggests that there are two effective ways to increase equity in public education spending: (1) increase the access to primary education, and (2) raise the share o f primary education in total education expenditure. These are also two o f the major goals o f the sector expenditure program and attaining them would go a long way inincreasing equity intotal public education expenditure. Table 5-14: Incidence of public education expenditure on the poorest and richest quartiles selected countries, % share Country Year Education spending benefiting Lowest quintile Highest quintile Senegal 2001 12 34 Guinea 2000 12 49 C6te dIvoire 1993 10 38 Ghana 1992 16 21 Kenya 1992/93 17 21 Madagascar 1993 9 44 SouthAfrica 1993 14 40 Tanzania 1993 13 23 Sources: Dayton(1995); Demery (1995); Bemier, Chao, andDemery (1994); Castro-Leal(1995a, 1995b); DemeryandVerghis (1994); World Bank (1995) 38Becauseof data constraints, these are rough estimatesfor eachincome quartiles. 113 5.43 The results o f incidence analysis by rural and urbanregions given inTable 5-15, indicates that per capita expenditures for primary and secondary education are biased towards urban areas. At the primary level, urbanexpenditures for primary are almost 1.2 times as high as inrural areas. Table 5-15: Unit cost by regionand educationlevel(CFA), 2000 Education Indicator Primary Secondary1 Subsidy per capita inurban areas 5045 693 Subsidv ver capita inrural areas 4202 368 Source: Ministryo f Education and Household Survey data 'Includes lower and upper secondary and technicallvocational 5.44 At the secondary level, the difference is infavor o furban areas by a factor o f 1.9 when compared with rural areas. These differences can be explained primarily by lower enrollments in rural areas. Further analysis by different regions in Senegal has shown that, although the regional expenditures are distributedfairly equally at the primary level, this changes markedly at the secondary and higher education level, where expenditures are highly concentrated in urban areas, especially Dakar. If absolute expenditures are examined, there i s not much disparity inprimary education where expenditures in urban areas are 52 percent o f total expenditures (the per capita expenditures in Table 29 are lower because o f lower rural enrollments). The situation i s drastically different in secondary education where 63 percent o ftotal spendingi s done inurban areas. Table 5-16: Unit cost by gender, regionand education level(CFA), 2000 Education Indicator Primary Secondary1 Subsidy per capita inurban areas Boys 11588 1899 Girls 11011 1205 Subsidy per capita inrural areas Boys 11606 1266 Girls 9620 593 Source: Ministryof Education and Household Survey data 'Includes lower and upper secondary and technicallvocational 5.45 Gender disparities persist across urbanand rural areas (Table 30). Inurban areas, the per capita expenditures are biasedtowards boys inboth primary (1.1:1) and secondary education (1.6: 1). The degree o f bias is much higher inrural areas where expenditure on boys is 1.2 times that o f girls in primary and nearly 2.1 times in secondary education. Thus gender bias persists across urbadrural divide with stronger bias inrural areas. 114 5.46 In sum, there is strong evidence that the incidence of government spending on education i s biased towards the richest segment o f the population, against girls and in favor o f urban areas. The strength o f these biases i s low for primary education and gets progressively higher as one moves up the education ladder. The inefficiencies in the education system are exacerbated by these inequities, reinforcing the need for public expenditure management reform to accelerate progress in Senegal's education system. SECTION111: IMPROVINGCONDITIONSFORVULNERABLE GROUPS The Government o f Senegal in its Poverty Reduction Strategy Paper (PRSP) underscores the importance o f improving the living conditions o f vulnerable groups. It intends to do so alongside its other objectives o f wealth creation, and the promotion o f human capital and basic social services. The next chapter will examine a particular instrument o f transfers-pensions-to protect one set o f vulnerable people, namely, the elderly. This chapter examines how fiscal policies could be used to protect the poor. 6. COMMODITY TAX REFORMFOR POVERTY ALLEVIATION A. INTRODUCTION 6.1 One of the most important and powerful instruments at the disposal o f the government i s its fiscal policy. Inpoor countries with large informal sectors and poorly functioning tax administrations, indirect taxation, i.e. the taxation o f goods and services, i s not only the main source o f revenue for the government but also the most effective fiscal vehicle to impact individuals' welfare. Typically, indirect taxation in developing countries accounts for two thirds o f the government's fiscal revenue. Revenue from indirect taxation i s also in most developing countries equally split between trade and domestic taxes. The most invoked rationale for the imposition o f tariffs on imported goods i s the government's desire to protect local industries. 6.2 This chapter describes how tax policy-designed to protect industries and raise revenue for the government-can be reformed to protect poor people as well. Targeted transfers and subsidies can also help reduce poverty. But these policies can be costly, are usually unsustainable in the long-run, and frequently leak to the non-poor. By focusing on revenue-neutral tax reform, this chapter describes a set o f market-based policies that could bothprotect the poor and be sustainable, politically as well as financially. 6.3 InSenegal, several goods have been for a longtime, and still today, protected for diverse reasons among which the protection o f the producers is often invoked. This chapter is concerned with a subset o f these goods, notably rice, meat, sugar, milk, wheat flour, refined oil, tomato paste and cigarettes. Protection, however, means higher domestic prices, and has an impact on the living standards o f the population. Although, most o f these goods have been under protection for long time, no serious assessment o f the impact o f this protection on the welfare or poverty o f the Senegalese has been made to date. 116 6.4 The new available household survey Enquzte Sknkgalaise Auprk des Mtnages (ESAM 11) overlapping between years 2001 and 2002 is used in this study. However, only the first round o f ESAMI1i s yet available. Information on the main characteristics, consumption, and income o f Senegalese households for the months o f May to August only has been collected. Using ESAM 11, this study addresses the question o f the impact on Senegal's poor, o f lifting or reducing protection for these different commodities. Changing the level o fprotection for these commodities i s equivalent to changing the rates at which they are taxed. The question this study intends to address then is: How can taxes be reformed so that the same level o f revenue i s collected, but poor people are better off? 6.5 Indeed, protection as well as taxation or any other type o f barrier, works as a consumer's price shifter. The ultimate objective o f this study i s the impact o f imposed barriers on the specific commodities cited above on the welfare o f the poor. The barrier that will be under study here i s the taxation regime o f those goods. The choice o f the commodities that are studiedhere stems fkom the facts that not only a sizeable share o f poor households' budget goes to those commodities, but also poor households consume a relatively important proportion o fthe aggregate consumption for those commodities. 6.6 Let us for the sake o f illustration quickly describe the taxation regimes o f sugar, milk, and rice. As stated inchapter 2, there is a monopoly for the production o f sugar in Senegal, and mark-ups have been high, ranging from 50 to 100 percent. The sugar industry has benefited and still benefits from various tariff barriers even after the liberalization. At this moment, the tariff on sugars is at a 20% rate, and the VAT rate on sugar is 10%. The sugar also takes advantage o f a specific WAEMU supposedly temporary protection tariff the "Taxe Conjoncturelle a l'hportation" (TCI) which aims at protecting from the volatility o f world prices, this tariff is rated at 10%. The nominal protection rate i s therefore o f 32% whereas adding the V A T rate makes the consumers face a price that i s 45.2% higher than the world price. The sugar industry had at times also benefited from the fixation o f a reference price at the entry point. Reference prices can sometimes double the world price. In 2000, with a world price at 142 CFAF/kg the reference price for granulated sugar was fixed at 289 CFAFkg to which a customs duty of 7% was applied implyinga nominal protection rate o f 118%. 6.7 Unsweetenedmilk face a 20% customs duty and a 20% VAT rate too. A "taxe interieure" o f 12% i s also applied. The wedge between its world price and the price at which consumers can buy it at a store is around 60%. A major liberalization o f the rice market was implemented inthe 1990s followed by elimination o f a special excise tax in 2000. The customs duty for broken rice is now set at lo%, and rice i s a necessity good to which a zero-rate V A T i s applied. 6.8 All these mechanisms drive awedge betweenthe world price and the consumer's price and hurt the poor. The purpose o f this chapter is to study the ways in which the taxation regime can be reformed, to alleviate the burden the poor bear while at the same time keeping constant government's fiscal revenue. 117 6.9 An appendix presents the methodology used for the evaluation o f tax reform and the information required to carry out such an exercise. Because o f serious data limitations this studyuses a very simplemethodology. Tax reformsbenefit or hurt people depending on how much they buy or sell o f the commodities inquestion. Then to analyze the impact o f the envisaged reform, both the perspective o f the producers or net sellers, and consumers or net buyers should to be taken into consideration. Unfortunately, because o f a lack o f reliable data on households' sources o f income, producers are not considered in this study. Only the perspective o f the consumers will be taken. The impact o f the reducing taxes or lifting protection on producers, and/or workers inthe affected sectors i s omitted. If producers' poverty i s greater than consumers' poverty, then the impact o f reducing protection will certainly be an increase in poverty. However, for the commodities that are dealt with here, one can be confident about the fact that consumers' poverty i s much greater. B. AGGREGATE EXPENDITURE POVERTYPROFILES AND Some basic statistics 6.10 We begin with a profile o f poverty and household expenditures in Senegal. The data we use i s the first round o f the recent and on-going Senegalese household survey, ESAM 11. Table 6-1 presents basic information on the population and expenditure distributions in Senegal based on the survey. 6.11 There are slightly more than one millionhouseholds in Senegal, more than half o f whom live inrural areas. An average o f 9.1 persons live in a household. Since the size o f the household is also, on average, bigger in the rural areas, only 40% o f the Senegalese population dwells inurban cities3'. 6.12 The average Senegalesehouseholdspent 1,813,820 CFA Francs in2001. Average household's expenditure in Dakar i s almost twice the average rural household's expenditure4'. 39The sample is balanced between urban and rural households. Because Senegal is mostly rural, urban households are over-represented in the survey if weights are left out. Therefore, throughout the study all variables are properly weighted. 40After the removal of few outliers, there are 6,570 households in the ESAM I1sample. Because only the fmt round o f the survey is used, the annual expenditure figures in table 1 correspond to three times the results o fthe computations carried out using E S A M 11. 118 Table 6-1: Basic expenditure statistics for Senegal, 2001 Units Other Senegal Dakar Urban Rural Population Persons 9,666,633 2,178,529 1,748,924 5,739,180 Shares 22.54% 18.09% 59.37% Households Households 1,061,476 273,033 197,139 591,304 Shares 25.72% 18.57% 55.71% Average household size 9.12 8.17 8.96 9.77 Sample size Households 6,570 1,982 1,359 3,229 Persons 59,916 16,199 12,174 31,543 Annual Expenditure Millions CFAF 1,925,331 726,060 383,595 815,676 Annual Expenditure per household CFAF 1,813,819 2,659,237 1,945,808 1,379,452 Daily Expenditure per capita CFAF 553.26 925.78 609.26 394.79 Daily Expenditure per household CFAF 5,038.40 7,386.77 5,405.01 3,831.81 Total ExpenditureDistribution Deciles 1 1.9% 2.0% 2.0% 2.2% 2 3.3% 3.4% 3.6% 3.7% 3 4.3% 4.4% 4.7% 4.8% 4 5.3% 5.6% 5.7% 5.9% 5 6.3% 6.8% 7.0% 7.0% 6 7.6% 8.2% 8.3% 8.1% 7 9.4% 9.7% 9.9% 9.5% 8 11.9% 12.0% 12.2% 11.6% 9 15.9% 16.1% 16.1% 15.1% 10 34.1% 31.8% 30.5% 32.0% ~~ Source: World Bank staffs calculationsbased onpreliminary results fromESAMI1 6.13 Senegal is a highly unequal country where the highest quintile's spending equals half o f the total expenditure. The Lorenz curve, a graphic representation o f the cumulative distribution o f expenditures in table 6-1 plotted against the cumulative proportion o f households i s presented in figure 1. The country i s divided in three "regions" which are Dakar, other urban cities, and rural areas. The Lorenz curves for the urban and rural areas overlap almost perfectly, meaning that consumption distributions within region are almost identical. Inequality measures for the different regions must therefore be close. Indeed, the Gini coefficients are 0.42, 0.4, and 0.4 for Dakar, the other urban cities, and the rural areas respectively. However, the Lorenz curve for the entire country lies slightlybelow the regional curves with a national Gini o f 0.44. There i s thus a substantial amount o f inequality between regions, with urban areas much richer than rural ones. 119 Figure 6-1: Expenditure Lorenz curves for the regions 100% 2 90% .- + S 80% a, 70% + Q, &Senegal 0 60% K 0 Dakar B 50% g 40% Urbain .-2 1 Rural =5 30% c m 20% 0 10% 0% 0 0.2 0.4 0.6 0.8 1 Currulative proportionof households Poverty Profiles 6.14 In his 2002 PRSP, the Government of Senegal uses a poverty line of 143,080 FCFA per adult equivalent. For the sake o f simplicity this line is set at 178,800FCFA per capita per mum for this study's purposes, and considered as the extreme poverty line. Another poverty line will be set at 255,500 CFA francs per capita per annum4'. Throughout the study a household will be deemed very poor and poor if its per capita consumption i s below the first and second poverty line, respectively. 6.15 Although it would be interesting to have complete poverty profile for Senegal, i.e. by region, by socio-economic status, by ethnic group, by education level o f the head o f household, etc, regional poverty profiles will suffice for the purpose o f this study. The population is divided into three distinct groups: the very poor whose per capita expenditure i s below the extreme poverty line; the poor whose daily expenditure i s above the extreme poverty line but still below $1; and finally the non-poor who spend more than $1 a day. We will sometimes refer to them respectively as the low, medium, and high expenditure groups. Table 6-2 gives the distribution of income groups by region from the survey. According to table 6-2, 53.9% o f the population live inextreme poverty, and almost 70% o f extremely poor dwell inrural areas. 41These correspond respectively to 490 and 700 CFA Francs a day. Taking an average exchange rate o f 700 CFA Francs for one U S Dollar, the extreme poverty line is at 70 cents a day, andthe secondpoverty line i s the now standard 1$ a day line. 120 Table 6-2: Expenditure groups size by region Dakar Other Urban Rural Senegal Number %age Number %age Number %age Number %age Very Poor 481 7.3 674 10.2 2394 36.3 3539 53.9 Poor 396 6.0 278 4.2 453 7.2 1137 17.3 N o nPoor 1105 16.8 407 6.2 382 5.8 1894 28.8 All 1982 30.1 1359 20.6 3229 49.3 6570 100 Source: World Bank staffs calculationsbasedonpreliminary results from ESAM I1 6.16 Table 6-2 allows the computation o f the poverty headcount index, i.e. the proportion o f extremely poor andor poor households in a given region. More general Foster, Greer, and Thorbecke poverty indices are given in Table 6-3 below. Table 6-3 presents the poverty profile with the headcount ratio; the poverty gap, i.e. by how much on average every poor household should be given to eradicate poverty; and finally the severity or depth o f poverty which gives extra weight to households that are farther below the poverty line. For all indices the rural sector scores the highest, e.g. at $1 a day poverty line, 88.1% rural households are considered as poor and the poverty gap is almost halfthe poverty line. Table 6-3: Regionalpoverty profiles Extreme Poverty Line Poverty Line Dakar Urban Other Rural Senegal Dakar Urban Other Rural Senegal Headcount Index 24.2% 49.3% 73.9% 53.9% 44.2% 70.0% 88.1% 71.2% Poverty Gap 6.90% 17.9% 32.0% 21.5% 15.4% 30.8% 47.2% 34.2% PovertySeverity 2.81% 8.75% 17.3% 11.2% 7.16% 17.0% 29.2% 20.0% Source: World Bank staff`s calculations basedon preliminary results from ESAMI1 An Heuristic Description of theMethodology 6.17 The methodology used to evaluate which commodities should be taxed or subsidized for revenue-neutral poverty alleviation can be heuristically described as follows (for a slightly less heuristic description, see Box 1). Suppose there are only two goods, 1and 2, and the poor spend a larger share o f their income on good 1than on good 2. Lowering the tax on good 1 would help the poor most. Since the reform has to be revenue-neutral, we would have to raise the tax on good 2. But what if good 1 i s an "efficient" revenue-raiser? That is, what if it will take a sizeable increase in the tax rate on good 2 to replace the revenue lost by lowering taxes on good l ? Inthat case, it might not be inthe poor's interest to lower good 1taxes if it entails increasing good 2 taxes by such a large amount that it leaves the poor worse off. 6.18 Thus, in seeking poverty-reducing directions o f tax reform, we need to balance two factors: (i) distributional characteristic o f the commodity i.e. the importance o f the 121 the commodity in poor people's budget and the importance o f poor relative to non-poor consumption o f the commodity; and (ii) the revenue-generating "efficiency" o f a tax on a commodity. The former i s captured by the concentration curve o fthe commodity whereas the latter i s associated with the marginal efficiency cost o f funds (MECF) for the commodity. 6.19 A concentration curve for, say meat, is the cumulative share o f aggregate meat expenditure plotted against the cumulative proportion o f households rankedinincreasing order o f welfare which i s measured here with total expenditure. A concentration curve42 is constructed in the same way than a Lorenz curve by replacing on the vertical axis (total) income by commodity expenditure. The only information one needs to draw the concentrationcurve o fmeat i s households' expenditure on meat andhouseholds' measure of welfare, here total expenditure. 6.20 For tax reform purposes efficiency considerations o f the commodities must be taken into account besides distributional characteristics. Changing the taxation regime o f a given commodity induces households to adjust their consumptionbasket. Indeed, when say chicken becomes more expensive because o f heavier taxation, households may buy less chicken and allocate this extra money to meat or fish for instance. They might also consume more chicken and less o f all other commodities. The behavioral adjustment has therefore an impact on the total tax revenue raised through the change o f the consumption basket. Commodity 1i s a more efficient revenue raiser than commodity 2 if to raise one additional unit o f revenue the cost to the society i s lower by taxing commodity 1 rather thancommodity 2. The cost to society is capturedbythe commodity's MECF. 6.21 The MECF i s the ratio between the tax base i.e. aggregate quantity consumed o f the commodity and the change inoverall tax revenue resulting from a small change o f the tax rate o f that same commodity (the marginal revenue). The higher the MECF o f a commodity the more costly, in efficiency terms, it i s to tax that commodity. The information one needs to estimate the MECF i s the tax base and the marginal revenue which requires knowledge o fdemand elasticities. 6.22 Much o f the information needed to gauge this balance can be gleaned from the difference o f two concentration curves. Ifthe MECF o f the two commodities is the same, the desirability o f lowering taxes on good 1 (say, edible oil) and raising taxes on good 2 (meat) over the whole range o f poor people amounts to requiring the concentration curve o f edible oil to be above the concentration o fmeat throughout that range. This means that meat consumption i s systematically more concentrated among the non-poor than consumption o f edible oil. If the MECFs o f the two commodities are not the same, then we need to look at the modified concentration curves, which is just the concentration curve multiplied by the MECF o f the commodity. A poverty improving tax reform requires the modified concentration curve o f commodity 1 to be above the modified 42Concentration curves are also interesting in their own right because one can infer the total expenditure elasticity o f a commodity from the curvature o f the concentration curve. This elasticity equals one if the concentration curve and the Lorenz curve are identical, is greater (resp. less) than one ifthe concentration i s below (resp. above) the Lorenz curve. An inferior good has a Ginitotal expenditure elasticity less than one. 122 concentration curve o f good 2 for the range o f all poor households. Ifthis condition fails, no "Dalton improving" tax reform-Le., revenue-neutral tax reforms where poorer households benefit more than richer ones-exists for these two c~mmodities~~. NecessaryI nformation 6.23 To be able to gauge a reform the government needs to have at its disposal information on the distribution o f household consumption or expenditure on existing commodities, the aggregate expenditure on each commodity, and for each tax instrument, its marginalrevenue or marginal efficiency cost o fpublic funds. For Senegal, distribution of expenditure i s available from ESAM11, therefore aggregate expenditures (or tax bases) also are available. Unfortunately, marginal revenue i s not readily available and one needs to compute them using a simulation tax model or a computable general equilibrium model. The information needed is the matrix o f demand elasticities (own and cross) and the tax rates. Since this information cannot be derived with the information currently available, most o f the following analysis will assume that MECFs are equal or assume the MECFs for different commodities from existing studies. This is a cautionary note for the conclusions drawn inthis study. Aggregate Patterns of Expenditure and Impact of TaxReforms 6.24 We begin by looking at how households allocate their budget among eleven aggregate categories o f expenditure detailed below. The first group Food includes expenditure on food, drinks, and tobacco and concerns the bulk (66%) o f all market transactions; the second group Own Food accounts for own-produced consumption o f agricultural goods; the third group, Clothing, concerns expenditure on clothes and shoes, the fourth group Housing, concerns rents, water and electricity bills; the fifth group Equipment concerns mostly housing equipment and repairs; the sixth group and seventh groups concerns Health and Education expenditure respectively; the eighth group concems expenditure on transportation, and communication; the ninthgroup expenditure on Leisure and Cultural Activities; the tenth group i s expenditure on Other Services; and finally the eleventh group i s spendingonMiscellaneous items. 6.25 Although the primary interest o f this study i s on reforming the fiscal regime that applies to a chosen set o f commodities that fall within the food group, it i s important to assess the impact o f taxation changes on all other categories o f expenditure. A good understanding o f general household consumption and possible substitution effects with price changes is therefore important. The relative size o f the share o f food expenditure i s also an important determinant o f the efficiency o f this policy. Ifthe commodities whose protection or taxation i s reduced constitute a very small part o f the transactions, then that 43 Necessary and sufficient conditions for the existence o f an improving reform that involves several commodities are fortunately available. Without getting into cumbersome details, the fundamental condition is that the weighted sum o f modified concentration curves be positive throughout the range o f poor. The weights are the tax proceeds that accrue because o f a change inthe taxation rate o f a particular commodity, it is the marginalrevenue times the marginal change intax rate. Onthis see Mayshar andYitzhaki (1995). 123 policy will not be very effective inalleviating poverty. For illustrative purposes, we show below the results o ftax reform simulations usingthis aggregate group o f commodities. Figure 6-2: Household and RegionalExpenditure Patterns 100% 90% Miscellaneous 80% Other Services 70% Leisure Transport 60% Wucation 50% Health 40% 6quipment Housing 30% Clothing 20% Ow n Food 10% Food 0Yo VeryPoor Poor Non-Poor Rural Other Dakar Urban 6.26 Figure 6-2 shows that the lion's share o f expenditure goes to food. Every income group spends way more than half o f its budget on food, with the ratio as high as 68% for very poor households. On top o f buying goods in the market, households also consume from their own production. This phenomenon almost exclusively happens in rural areas, where own-produced consumption is around 6% o f the total household outlay. Because very poor households have a greater share o f own-produced consumption, one would want the production side to be introduced in the study. Indeed, reducing protection on some commodities could then benefit only the non poor-and harm the very poor who produce more than they consume-thereby aggravating poverty. However, a quick look at the data shows that the mean share o f own produced consumption infood consumption i s 22% and only 13% o f total consumption, meaning that including production should not alter the results that much. 6.27 Not surprisingly, per capita food expenditure i s decreasing with respect to per capita expenditure, meaning that holding household size fixed, the richer the household, the less it spends on food relative to the other spending categories. The same pattern i s also observed depending on the region. Households living in Dakar spend on average slightly more than 60% o f their budget on food. This is probably a consequence o f more than 55% o fthe households inDakar belonging to the non-poor group, as shown intable 2, and within that, they probably belong to the highest part o f the expenditure distribution. The average household inDakar is above the poverty line. Rural households, 124 who are also the poorest, spend on average (including own produced consumption) more than70% oftheir budgetonfood. 6.28 Within the category o f non-food expenditure, clothes and shoes constitute the most important group. Spending on health i s also quite important for very poor households, who allocate about 15% to health. In absolute terms, however, richer household pay more for health. Indeed, whereas the average very poor household spends around 1,800 CFA Francs on health, the average non-poor household spends 11,500 CFA Francs. Education expenditure constitute a modest share o f the budget and i s almost negligible for the very poor. Transport and communication constitute a sizeable share o f the budget for urbanand nonpoor households. 6.29 As stated above, it i s important to have information on the consumption patterns by expenditure groups, and by regions as well. Table 6-4 shows for Dakar, the other urban cities, and the rural areas, the consumption pattem for the low, medium, and high expenditure groups that have been defined above. Note that budget shares in the table concern all households, not only those reporting positive consumption in a particular group o f goods. ESAM I1 defines eleven groups o f goods, which are afterwards disaggregated by sub-groups andparticular items. Table 6-4: Aggregate expenditure pattern of income groups within regions Dakar Other Urban Rural Very Poor Non Total Very Poor Non Total Very Poor Non Total Poor Poor Poor Poor Poor Poor Food 69.3 62.0 58.6 61.8 71.9 68.1 63.6 68.5 73.5 69.5 72.6 72.7 of whichown 0.0 0.0 0.0 0.0 0.7 0.4 0.4 0.6 7.4 4.9 7.5 6.9 Clothingand Shoes 7.3 9.8 9.2 8.9 10.1 10.3 9.3 9.9 8.5 9.4 7.4 8.5 Housing 8.3 10.2 10.3 9.8 4.6 6.0 7.2 5.7 3.3 4.1 3.9 3.5 Equipment 1.9 3.3 3.2 2.9 2.1 2.7 3.3 2.6 2.8 3.4 3.2 3.0 Health 6.4 6.5 6.1 6.3 5.5 5.2 5.6 5.4 5.6 5.7 4.6 5.5 Education 1.6 1.8 2.4 2.1 0.9 0.8 1.0 0.9 0.6 0.5 0.2 0.5 Transport 2.2 3.1 5.1 4.0 2.8 3.7 5.1 3.7 2.7 3.0 3.4 2.9 Leisure 0.1 0.1 0.2 0.1 0.1 0.0 0.2 0.1 0.2 0.4 0.2 0.2 Others Services 0.6 0.7 1.8 1.3 0.5 0.8 1.6 0.9 0.5 0.6 0.7 0.6 Miscellaneous 2.4 2.6 3.0 2.8 1.7 2.4 3.3 2.3 2.3 3.4 3.8 2.6 Total ~. .~ 4.6 5.9 27.2 37.7 5.7 3.9 10.3 19.9 23. 7.7 11.3 42.4 6.30 Table 6-4 above concerns the aggregate consumption patterns, and presents evidence on how income groups allocate their consumption budget among the different groups o f goods in any given region. The last line tells us how much each income group in a region has spent relative to total expenditure in the country, e.g. 37.7% of total expenditure i s recorded in Dakar, and very poor rural households account for 23.4% o f total private expenditure. 6.31 Food, clothing, housing, health, and transport altogether account for about ninety percent o f all households' budget inall regions. The subsequent analysis will be restricted to these five categories. For illustrative purposes, let us first treat each o f these aggregate 125 groups as a single commodity. The question one would like to answer i s then: is there any change o f taxation policy that can reduce poverty? To answer this question, let us first draw the concentration curves o f the different aggregate groups along with the Lorenz curve, which i s the concentration curve for per capita expenditure. As could be anticipated, the curve for food i s above all other curves. Some are above, and others below, the Lorenz curve. Since the concentration curve for food is above all other curves and does not intersect with any o f them, this means that financing a tax-reduction (or subsidy) on food by a tax-increase on any other category can bepoverty reducing. This is true for any two curves that do not intersect. When on the other hand two curves intersect, the conclusion i s not straightforward. Inthat case, it will depend on the point of intersection. If the intersection i s before the poverty line, then those two commodities cannot be ranked, if the intersection point i s after the poverty line then the tax on the commodity whose curve that is above should be reduced. Figure 6-3: Concentration Curves 1 0.9 0.8 0.7 0.6 -Lorenz -Food 0.5 Clothing 0.4 Transport 1 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 Cumulativeproportion ofhouseholds 6.32 Transportation and communication seem to be the activities on which taxes should be increased. Ifwe are only interested in the impact o f changing two tax rates on aggregate poverty, then we should look at the effect on those below the poverty line only. Table 6-5 records the results when the government aims at reducing the severity or depth o f poverty, i.e. a=2. The very first line o f the table displays the Gini expenditure elasticities o f the commodity groups. With an elasticity of 0.87, only food seems to be a necessity, even though clothing and health are close to one. Transport and education stand here for luxury goods with very high Gini elasticities. The columns in Table 6-5 stand for commodities whose taxes are increased, and the row the commodities whose taxes are decreased (or are subsidized) to preserve revenue. Since the coefficient o f proportionality has not been defined, note that the magnitude o f the result cannot be interpreted; only the sign o f this expression matters. When it is positive then the policy is 126 poverty reducing and when it i s negative it would be poverty increasing. For example, line 2 column 3 i s a policy where one CFA franc o f additional taxes i s raised through an increase in the tax rate on housing, and the proceeds are usedto subsidize clothing. This policy i s poverty reducing: it reduces the measure o f depth o fpoverty by 2.20 percent. Table 6-5: Illustrationof aggregate gains in poverty reduction Food Clothing Housing Equipment Health Education Transport Gini Exv El. 0.87 1.02 1.27 1.44- 1.10 2.97 2.21 Food Clothmg Housmg Equipment Health Education Transport Note All coefficientsare multipliedby 100 Tax decrease (subsidy) oncommodity mrow financed by tax mcrease oncommodity mcolumn Source World Bank staff's owncalculationsbasedon p r e h m a r y results from ESAM11. 6.33 So far, all marginal efficiency costs o f public funds are considered equal. The story would be different if the cost at the margin o f raising one CFA franc o f public revenue from a food tax i s 1.03, 1.15 from tax on clothing, and 1.90 from transport tax for instance. With this example, all concentration curves intersect and no improving reform can be found as was the case ifall MECFs were equal. 6.34 Because the cost to society o f taxing transport i s higher than taxing food, one would be tempted to tax food to subsidize transport. This would indeed be efficient, would it be equitable? As figure 4 shows, the cost to the four poorest deciles from an additional franc in revenue by taxing food i s almost 18 cents, while it is 13 cents for transport. Therefore taxing transport to subsidize food although costly in overall efficiency is more suitable ifredistributionmatters. 127 Figure6-4: IntersectingModified ConcentrationCurves 2 1.8 1.6 1.4 1.2 ~ -Lorenz __ 1 Food Clothing 0.8 Transport 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 Cumulative proportion o f households 6.35 Note the change in scale o f the modified concentration curves. Taking into consideration the marginal efficiency cost o f taxing a commodity, the modified concentration curve starts at the origin and ends at the value o f the MECF. For instance the endpoint for the modified concentration curve o fthe transport sector i s (1, 1.9). 6.36 The equivalent for the modified concentration curve when considering the Lorenz curve i s the Generalized Lorenz curve which is the Lorenz curve multiplied by the mean income. 6.37 Another interesting instrument to gauge the impact o f the reform i s the curve o f cumulative gains. It is just the difference between the modified concentration curves. Figure 5 i s the marginal cumulative gains for a reform consisting o f taxing clothing and subsidizing food, using the marginal efficiency cost o f funds defined above. As figure 5 shows, this reform does not benefit society as a whole since the total gain is infact a loss of twelve cents. However, since the curve increases at the bottom o f the distribution and starts declining after the fourth decile, almost all poor people (headcount is 44.8%) gain inthat reform. Theburdenofthis reformis infact bornebythe nonpoor which canmake itdesirable to the govemment. 128 Figure6-5: Cumulativegainfor households Food vs Clothing 0.02 0 -0.02 -0.04 -0.06 -0.08 -0.1 -0.12 Cumulative proportion of households Expenditure Patternfor Protected Commodities and Impact of TaxReforms 6.38 Let us now tum to the impact on poverty for the commodities that enjoy tariff protection, which are broken rice, sugar, wheat flour, milk, refined oil, tomato paste, cigarettes, and poultry which i s proxied here by meat. Table 6-6: Expenditurepatternfor selectedcommodities ExpenditureGI ups Regions Very Poor Poor NonPoor Dakar Other Urban Rural share %c share %c share %c share %c share %c share %c BrokenRice 13.7 87.1 9.75 91.1 7.15 80.8 7.42 80.8 9.74 91.5 14.1 87.5 Wheat Flour 1.64 26.5 1.15 42.9 0.98 43.5 0.20 51.8 0.26 57.3 0.42 16.8 Sugar 5.65 95.8 4.76 99.0 3.61 97.3 3.15 97.7 4.39 98.4 5.98 96.2 Milk 2.56 69.0 3.41 84.9 3.80 92.1 3.77 97.4 3.45 94.0 2.39 63.1 RefinedOil 8.56 96.3 7.66 99.0 5.99 97.6 6.45 98.3 7.96 99.0 8.42 96.1 Meat 6.56 62.1 8.18 80.1 9.75 86.2 8.92 87.3 8.92 85.1 6.73 60.8 Tomato 1.90 76.8 2.05 89.2 1.87 82.9 1.66 87.9 1.93 92.2 1.28 73.7 Cigarettes 2.56 24.3 2.64 21.6 2.71 2.73 20.6 2.05 18.2 2.68 27.0 Total 43.1 39.6 35.8 34.3 38.6 42.0 Note: %c is the percentage of household that report positive consumption o f the commodity Source: World Bank staffs calculations based on preliminary results from ESAMI1 129 6.39 These commodities constitute a very high share o f expenditure for all households. Very poor households allocate 43.1% o f their budget to these goods. Expenditure on sugar and refined oil are recorded for almost all households. Broken rice is the single most important commodity for the very poor, whereas it i s meat for the non poor. Inthe regions, this group o f commodities represents 34.3% o f total expenditure in Dakar, and 42% in rural areas. Given these high numbers, it i s expected that protection-which means higher prices for consumers-comes at a high cost in terms o f poverty. Most o f the analysis will, however, be focused on broken rice, sugar, refined oil, and meat. As already stated, one o f the determinants o f the suitability o f a tax reduction for a commodity i s the share o f it that is consumed by the targeted population, namely the poor. Figure 6 shows the share o f the low, medium, and high expenditure groups in the different commodities. Figure6-6: ExpenditureSharesof ProtectedCommodities 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 6.40 Except for rice, sugar, and refined oil, the non poor group consumes at least half of all commodities. This i s even more striking for meat for which over 60% goes to the non poor, whereas the very poor consume only 10%. Only 62% o f the very poor households record a purchase o f meat, this proportion increase rapidly to 80% for the poor and 88% for the nonpoor. The commodities inwhich the verypoor have a relatively highsharearericewith morethan25%, sugar, refinedoil, cigarettes, andwheat flour. 130 Table 6-7: GiniExpenditureElasticitiesfor selected commodities Broken Wheat Refined Meat Tomato Rice Flour Sugar Milk Oil Paste Cigarettes GiniExp 0.56 1.25 0.66 1.31 0.65 1.34 0.90 4.94 6.41 The expenditure elasticities suggest that that rice, sugar, refined oil, and tomato paste are necessities. This has been confirmed econometrically. The budget share o f each good i s regressed on the logarithm o f per capita household expenditure, the logarithm o f the household size, and demographic characteristics of the household. Note that, because of the limitation o f our data inwhich no information on price and quantities is available, we cannot include price or even unit values. Only prices collected at Dakar exist, and these cannot be included in the regression because each price will be merely a constant. The regression also showed that a household's demographic composition does not matter inthe consumption pattern of these selected goods. This provides some assurance that considering per capita expenditure, or adult equivalent per capita expenditure to account for the demographic structure of the household, as our welfare indicator would give similar results. 6.42 Looking more closely at meat, sugar, rice, and oil, the concentration curves o f these goods are displayed in figure 7. Rice i s way above all other curve, not intersecting with any o f them, meat is way below, whereas oil and sugar almost overlap in the low range o f the distribution. Were the marginal efficiency costs o f these commodities equal, it would always be poverty reducing to finance a subsidy on rice by a tax on any other good. The benefit would be greater ifmeat is taxed more heavily. The analysis cannot be conclusive on which between sugar andrefinedoil is dominant, but both dominate meat. Figure6-7: ConcentrationCurvesfor Meat, Sugar, Oil, andRice 1 0.9 0.8 0.7 0.6 Broken Rice 0.5 Sugar Refined Oil 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 Cumulative proportionofhouseholds 131 6.43 The marginal efficiency costs o f these commodities are for the time being assumed to be equal. Let us examine revenue neutral reforms involving a change in the taxation regime o f two commodities at a time. Figure 8 represents reforms such that a decrease o f one CFA franc from the taxation o frice i s financed through an increase o fthe taxation o f the other commodities. The "high-peaked" curve for instance considers meat as the financing commodity. These curves represent the cumulative gains for the households from such reforms, and are simply the difference between the concentration curve o f rice and that o f the other commodities. Note also that since concentration curves start at zero and end at 1, the cumulative gain curves start and end at zero, meaning that on aggregate the reforms are neutral. Gains from these reforms are distributional. As we have already seen from figure 5, marginal gain curves do not display this feature when commodities have different MECFs. 6.44 The reforms involving commodities other than rice can be readily visualized from figure 8. Indeed, the difference between the meat curve and, say, the oil one is the reform that finances a decrease in revenue from oil taxation with an increase in meat taxation. All other two-by-two reforms can therefore be obtained from figure 8. All curves show that rice i s a commodity whose taxation should be reduced. The gains from reducing the taxation on rice are highest and shared with a higher proportion o f the society when they are financed with meat taxation. Indeed, as long as the curves are on an increasing path, the gains to the households are positive. They reach a plateau when those household gain i s nil, and they decrease when the household's loses inthe reform. Figure 6-8: Cumulative gain curves for Rice vs Meat, Sugar, and Oil 0.35 0.3 0.25 0.2 -RicevsMeat 0.15 -RicevsSugar Rice vs Oil 0.1 0.05 0 -0.05 Curmlativeproportionof households 132 6.45 The curves for sugar and oil meet and start decliningjust after the seventh decile, meaningthat the bottom 70% o fhouseholds would gain from the reform. The curve when meat is the financing commodity climbs very fast meaning that households' gain increases at a high rate, and starts curbing downward only after the 8th decile. The transfer from the richest quintile to the rest o f the nation i s quite substantial with these reforms. It is important to remember here that the impact on the producers i s not taken into account. Senegal has many small rural households who are rice producers. The proposed reform may impoverish these household by reducing income in their pocket. This mightreduce the benefit o fthis policy. 6.46 Another feature that can reverse the conclusions i s the introduction o f differential marginal efficiency costs. As stated above, we unfortunately do not have the analysis to allow a proper estimate o f the MECFs. Therefore, they will be assumed in this section, again for illustrative, purposes, to be 1.02, 1.08, 1.15, and 1.42 for rice, oil, sugar, and meat respectively. In that case, the reform envisaged above will have efficiency losses. The aggregate loss for each reform will equal the difference between the marginal efficiency costs o f funds, e.g. taxing meat to subsidize rice will have a loss o f 40 francs for each hundred francs o f subsidy. The losses are very big. However as inthe previous section the distributional impact can be poverty reducinginthe sense that the loss will be borne almost entirely by the highest quintile. 6.47 So far we have only looked at reform involving pairs o f commodities. The methodology can however be applied to any number o f commodities. For instance the revenue neutral reform, where rice i s subsidized for 1.5 CFA francs financed by a tax o f 1 CFA franc on refined edible oil, 5 cents o f francs from meat, and the remaining 45 cents of francs by sugar i s shown in figure 9. This reform i s costly to society but beneficial to the poor. This comes from the highMECF for meat. Figure6-9: Cumulativegaincurvefor areformwith manycommodities 0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 Cumulative proportionofhouseholds 133 C. CONCLUSIONS 6.48 For protection and revenue-raising reasons, the Government o f Senegal pursues a policy o f taxing commodities. Inits PRSP, the Government has also committed itself to protect the vulnerable groups in society. This chapter has examined how commodity tax policy can be reformed to protect the most important vulnerable group, the poor. It explored directions o f tax reform that would preserve government revenue, while reducing the burden on the poor, especially the very poor. We find that there are promising opportunities for tax changes that could reduce poverty in Senegal. In particular, reducing taxes on (or subsidizing) a commodity that i s disproportionately consumed by the poor (rice), and financing this with tax increases on commodities such as meat, which are favored by the non-poor, could make a substantial contribution to poverty reduction in Senegal, while protecting government revenues. Even if it were more efficient to tax rice than meat, the differential in efficiency would have to be extremely large to compensate for the dramatic equity gains from such a policy. 134 7. PENSIONREFORMINSENEGAL 7.1 This chapter summarizes the mainfindings o f a comprehensive study o f Senegal's pension system titled "Building a Secure, Sustainable and Modern Retirement-Income System in Senegal." The report was produced upon request o f the Minister o f Finance o f Senegal and is a product o f collaboration between the World Bank and the Government o f Senegal. 7.2 Untilrecently, pension policy has not been accorded highpriority in Senegal or for that matter in most o f Africa. This has begun to change in large part due to the natural maturation o f pension schemes set up in the 1960s where this term refers to the risingratio o fpensioners to contributors. Since pensionpromises have not been matched bythe accumulationo freserves, this process inevitably leads to a greater fiscal burden or the need to default on promises or both. The financial deterioration has been exacerbated by the failure to effectively invest the reserves that did emerge inthe early years o f the scheme and most importantly, bythe largely exogenous shock o f a declining formal labor force (andtherefore, contributor base). 7.3 The focus today on pension policy in Senegal is due to at least three problems facing policymakers. The most urgent has been the need to contain the fiscal pressure caused by burgeoning expenditures on civil service pensions. Given limitations faced with regard to runningdeficits, policymakers must either cut other important programs or raise taxes to pay pensions to this small segment o f the population. Reducing this liability, at least inthe medium term, has become an important policy objective since the late 1990s. 7.4 A second, more direct cause for concern is the increasing recognition o f the failure to provide adequate pensions to formal sector workers not in the civil service. IPRES pensions have stagnated and the low or negative implicit rate o f return on contributions to the scheme has reduced its credibility. While there i s evidence of increasing demand for instruments that could ensure old age income security, the public policy response inthe form o f P R E S is rapidly losing credibility. This situation presents a major obstacle to other public policy objectives, such as extending coverage o f social insurance institutions to new groups o f workers and expanding the size o f the formal ` sector ingeneral. 7.5 Finally, Senegal's economy, like most others inthe region, suffers from a dearth o f long term savings and institutions that can intermediate these effectively. Under the right conditions, pension reserves could contribute to long term growth by providing the demand for long term assets. While this effect cannot be expected without parallel reforms in other areas, some regional in nature, the alternative o f continued reliance on short-term savings and external assistance does not seem viable. 7.6 Thus, the motivation for pension reform emanates from a variety o fpublic policy objectives including improved social protection and labor markets, fiscal sustainability and capital market development. 136 7.7 The report provides the first comprehensive assessment o f the formal pension system in Senegal. We find that despite low coverage - about five percent o f the labor force - the system has amassed large unfunded liabilities inthe form o fpensionpromises without assets to back them. The pension debt to civil servants alone is equivalent to one quarter o f GDP even after a partial reform in 2002. Another important finding is that pension promises to private sector workers have been systematically eroded through ad hoc changes, thus compromising its role as a vehicle for social protection. Finally, the lack of pensionreserves precludes the potential role o f the system ingenerating long term savings that can have a positive impact on economic growth. 7.8 Inorder to remedythese problems, we recommenda series ofparametric reforms within the current institutional framework that would lead to two outcomes: First, the measures would make the overall pension system more sustainable. This would mitigate the crowding out o f other expenditures due to pension liabilities and improve the overall fiscal position o f government. Second, the recommended measures would remove the high degree o f uncertainty that pervades the pension system. In the past, these ad hoc changes to key parameters have introduced `policy risk' to what should be a secure mechanism for long term savings. Tying the development o f pension levels to objective indices, rather than political decisions, also removes the inequities that arise when two workers with identical contribution histories receivedbenefits o fvery differentvalue. 7.9 This first phase o f reform i s a prerequisite to more fundamental, medium term reforms. Stakeholders in Senegal are already discussing a paradigm shift that would assign a new role to private, funded pensions generated through individual account accumulations. The report reviews emerging proposals, lays out the main policy choices for any systemic reform and highlightsthe particular challenges o f this kind o f reform in the Senegalese context. 7.10 Many o f the Stage 1 reforms could be implementedtoday. The report, produced in conjunction with staff at both pension schemes, includes concrete proposals for changes to benefit formula and indexation procedures as well as institutional changes that would improve administrative efficiency and governance. Systemic reforms incontrast, would take several years to formulate andprobably would involve a transition period that would include temporary institutional arrangements and probably a regional approach to matters o f financial sector supervision andpension fund investmentrules. 7.11 The report i s only the first step towards defining a strategy for pension reform in Senegal. It envisions a two-stage reform process. First, the FNR and IPRES must be placed on a sound financial footing and inequities be removed. For the m,this - intra and intergenerational - should can be achieved by extending the reforms o f 2002, specifically by: 0 Movingto price indexationo fpensions 0 Increasingthe number o f years o f earnings usedto calculate the initial benefit 0 Shiftingpro-natalbenefits out o fthe pensionscheme 137 7.12 A package o f parametric reforms was recommended for the IPRES scheme that would eliminate the uncertainty o f the past regarding pension calculation and indexation. These measures, which could be adopted immediately include: 0 Shift to automatic indexation o f benefits to earnings growth until the cumulative increase has reached fifty percent in real terms (compensation for past erosion o f real benefits) at which time pensions in progress would beprice indexed 0 Index all other parameters to wage growth permanently 0 Index the ceiling on earnings subject to contributions to salary growth 0 Increase the normal retirement age to 60 for men and women over say, a 10- year period 0 Introduce lump sum payment for workers that reach retirement age with fewer than 5 years o f contributions 7.13 A number o f institutional changes were also recommended for PRES. Detailed implementationplans would require further study and discussion within P R E S and with the Government. The measuresinclude the following: 0 Overhaul the governance structure andinvestmentpolicy for pension reserves 0 Introduce annual independent audits 0 Improve information systems and integrate database for use ininternal analysis 0 Introduceprofessionalcriteria for some Boardmembers or establish professional advisory body; train Boardmembers 0 Improve disclosure andreporting practices 0 Eliminate duplication o f functions with Caisse Securite Sociale for efficiency gains 7.14 Finally, we would recommend that serious consideration be given to integrating the two existing schemes in the long run. This could be achieved for example, by requiring all new entrants into the civil service to join the IPRES scheme combined with a more modest supplementary pension. 7.15 We also recommend further analytical work to better inform the medium term policy discussion. The most pressing topics include: 0 Living standards o f elderly people andthe role o f social assistance 0 The role o f informal old age income support systems and their future 0 Labor market impact o fpension (and social insurance) taxes 0 Transition issues relatedto the integrationo f civil servants pension provisions 138 7.16 The immediate changes recommended will stabilize the formal pension system in Senegal and prevent current and future taxpayers from having to bail out a pension scheme that covers a very limited part o f the population. However, the underlying problems associated with an unfunded, defined benefit scheme in both an economic and political sense would still bepresent. Encouragingvoluntary private and funded pensions i s a legitimate alternative and one that already has support among stakeholders in Senegal, as has the objective to gradually extend coverage o f the pension system. Additional thought will also need to be given to the role o f redistribution in a new system. 7.17 A systemic reform would require careful preparation over several years, especially with regard to institutional design. Questions to be answered upfront include the choice between a mandatory and a voluntary system, the regulatory environment, coverage, treatment o f civil servants, fund management rules, and the financing o f the transition. 7.18 Emerging proposals in Senegal include a mandatory supplementary pension system and the voluntary solidarity fund. The latter includes a govemment subsidy for contributions and provides for an investment o f 20-30 percent o f funds in small and medium-sized enterprises, with concomitant challenges as regards 'the management and investment decisions. A third proposal is the expansion o f P R E S by adding on a funded component o f individual accounts. 7.19 Shifting to a multipillar system that includes a mandatory, fimded, and privately managed component involves a number o f policy decisions. A strategic choice must be made on the size o f the funded pillar in absolute terms and in relation to the unfunded pillar. Simply adding on a funded pillar to the existing unfunded one implies raising contributions rates, which under current circumstance seems undesirable. Carving the funded pillar out o fthe unfundedpillarimplies lowering benefits o f the unfunded system inthe future and, since current pensions are financed with current contributions, finding the resources to cover the gap to pay existing pensioners. The overall size o f future pensions o fthose workers movingto the new multipillar system would depend on the rate o freturn o f the funded pillar and how previous contributions to the unfundedpillar would be valued. Since the benefits o f switching are generally rising with age, it would be useful to make switching to a funded second pillar voluntary for all those who have already contributedto the existingpension scheme to bolster public support for reforms. 7.20 Many countries have financed the transition with a mixture of tax increases and higher debt. The currently comfortable financial position o f P R E S implies that this could be a feasible strategy, with the amount involved being manageable. Moreover, it would seem reasonable to also offer FNR participants the switch to a funded second pillar and close the FNR for future civil servants, who could be integrated into the same two-pillar systemas workers inthe formal private sector. 7.21 Close attention would need to be given to the administration of funds, investment policies, withdrawals, and regional considerations. Regulatory choices on the administrative structure can determine the administrative costs o f the system, a crucial 139 concem for a relatively small-scale system. These choices also determine the ease to contribute, and the attractiveness for pensionproviders to offer their services. Inthe long- run, a system with specialized fund managers would seem preferable, with an interim solution o f a professional and arms-length centralized asset management. A specialized regulator may need to be created for the pension activities. An early decision would have to be taken on how to withdraw funds on the basis o f insurance sector capacity. Regional restrictions o f foreign investments outside the CFA zone and the limited size o f regional financial markets may require a concerted reform effort to allow a broad diversification o fpotential pension savings. 7.22 Successful implementation o f this complex reform will require concerted effort with support from the highest level o f government over a multi-year period. It will also be necessary to take into account regional factors, especially as relate to financial sector supervision and investment o f funds. But the potential payoff i s large: A secure, modem pension scheme will directly benefit those that participate by ensuring a reasonable stream o f income intheir old age. 7.23 A pension scheme that credibly provides income security for old age will be worthy o f extension so that over time, the scheme will provide social protection to an increasing proportion o f the population. This will also leadto a growing accumulation o f long term savings in the Senegalese economy. A significant proportion o f these funds can be invested in regional capital markets increasing the demand for securities and contributing to much needed liquidity. A successful pension reform in Senegal could provide an example for other West African countries facing a similar situation. SECTIONIV: PRINCIPLES FORIMPLEMENTATIONOF THE STRATEGY The discussions inthe previous sections indicate that reaching PRSP goals could require a number o f politically difficult reforms. The historical culture o f rent seeking has seeped from production sectors into the social sector programs, and will prove difficult to root out o f the economy. Difficult reforms -- such as resource management inagriculture and fishing, changes in tax and subsidy programs, changes in public expenditure management systems - will take time and require community involvement in order to build sufficient ownership. However, one type o f reform should pose fewer difficulties and aid the process o f building ownership on the rest: Increasing transparency and information sharing. The greater accountability that comes with transparency can improve the efficiency and effectiveness o f PRSP programs, and can ensure a role for civil society inmonitoring progress and shaping the next PRSP. The chapters in this section review several initiatives the government has undertaken to improve transparency and information sharing: Expenditure tracking surveys, PRSP monitoring systems, and budget systems reforms, where the last two chapters were drafted by the government. The government's contribution o f the chapter on budget reforms i s especially significant. World Bank research and experience has systematically shown that government ownership i s essential, and the main factor determining the success o f public expenditure management reforms. Therefore, it i s critical to consider the diagnosis and recommendations from the government's point o f view. 8. EXPENDITURETRACKINGSURVEYINSENEGAL: THE HEALTHSECTOR A. INTRODUCTION 8.1 The health sector i s among the priority sectors for the Government o f Senegal (GoS) as stated in its PRSP 2002. For the last five years, many countries have increased allocations to priority sectors with little or no impact on outcomes. Senegal belongs to this group. Indeed, the public expenditure reviews conducted by the government in 1998 and 1999 reveal that the budget o f the ministry o f health (external funding included) has increased at a steady pace, from 29 billions in 1995 to 47 billions in 1999, a 12% annual increase rate.44This trend i s reinforced today in the era o f PRSPs the health budget from the government in 2003 has been voted at 35 billions from 30.5 billions in 2002. Outcomes such as infant or maternal mortality, access to health services, immunization rates, or the rate o f deliveries assisted by a qualified medical staff, have improved but at a 44See Plan StratCgique ((Pauvrett Santt D,October 2001 141 much slower pace during the same period. There is therefore a weak correlation between budgetaryallocations at the centrallevel, and outcomes at the grassroots. 8.2 Two reasons can be thought o f to rationalize this disconnection betweenallocated public resources and outcomes. First, there might exist some delays and leakage in the system. Delays will hamper service delivery and have a bad impact on outcomes, because ifresources are received late inthe fiscal year, the beneficiaries cannot absorb them all. The adverse effect o f leakage i s more straightfonvard to grasp, indeed the fact that resources are diverted during their transit, means that allocations as recorded in the central budget do not really reflect the "true" level o f resources inthe sector. Outcomes are expected to be correlated with that true level o f resources the sector gets, and which is to date unobserved in most developing countries. Second, even in the absence o f delays and leakage, worse outcomes than would be expected can be a consequence o f mismanagement or just inefficient use o f resources at the service delivery point. The reality certainly lies somewhere in between these two extreme scenarios, each reason beingpart o fthe story. 8.3 The purpose o f this study i s to try and assess quantitatively the separate roles o f delays and leakages, as distinct from inefficient service provision in Senegal. Although data has been collected to address both reasons each inits own right, the focus here will, however, be on delays and leakages. To that end, a tracking expenditure exercise has been conducted. In Senegal, this exercise has been designed to track public funds, from their origin to their destination and analyze how each hierarchical administrative step influences the size and the celerity o f the traveling amount o f resources. The questions we intendto answer are: 1)what percentage o f the initial amount that left the central level reaches the final beneficiary, 2) how long does it take to that resources to reach its final destination. Answering these questions will intum help to inform policy for an improved budgetary system. Enhancing the efficiency o f the system o f resources flows i s paramount inthe era o fPRSP duringwhich more resources will be injected inthe priority sectors. Indeed, if no relation between allocations and outcomes exists, the rationale for increasing priority sectors' budgets vanishes, since more resources would be put in a leaking bucket. Fixing the leak in the system is therefore a prior action, that can be well taken once bottlenecks and holes have been clearly identified. Trackingmethodology and sample 8.4 The methodology of the study involved surveys o f government officials at each decision point in the transfer o f health resources, to measure the flows o f funds passing through the relevant stages and the amount o f time taken at each stage. In sum, the surveys provide information on the flow o f funds from origination to destination. Inthe investigation, the surveys were set at two levels. 8.5 At the central level, the survey was comprehensive in the sense that questionnaires were administered to each administrative level involved at a givenpoint in the system o ftransfer of public funds from the central level to health structures. 142 8.6 At the periphery, a sampling strategy has been devised. Ten districts and 100 health centers have been sampled and administered comprehensive questionnaires. A stratified sampling methodology has been used. First, five regions have been randomly selected, then in each region two departments have been chosen. Within each department 10health posts and one district have been selected for interviewing.The sample must be representative at the urbdrural level nationally. The list below presents regions and departments in which the survey has been carried out. Selected regions are underlined and selected departments are underlined and in bold. The list o f all facilities is given at the end o f the document. 0 Dakar: Rufisque,Dakar, Pikine and Gutdiawaye 0 Saint-Louis: Dagana, Saint-Louis, and Podor 0 Tambacounda: Bakel, Ktdougou, and Tambacounda 0 Thiks : Mbour, Thi&s,and Tivaouane 0 Kaolack: Kaffrine, Kaolack, and Nioro du Rip. 8.7 In 1996 the GoS launched decentralization reforms that transferred nine responsibilitiesnamely: (i) use, (ii) land environment, and natural resources management, (iii) population, health, and social action, (iv) youth, leisure, and sports, (v) culture, (vi) education, (vii) planning, (viii) land development, and (ix) urbanization and habitat, to the local governments. To compensate local governments for the transfer o f these responsibilities, the central government also transferred resources to the local governments, inthis process most o f the budget for the functioning o f the health centers budget is now managed by local authorities. The local governments are therefore a central part o f this study and needed to be interviewed. A total o f 37 local governments are included in the sample. It i s worth stressing that local governments have been included on the basis o f the sampled districts and facilities. All local governments managed the funds o f at least one district and/or facility inthe sample. 8.8 There are three types o f local administrative units, (i)the region, (2) the commune, and (3) the rural communitys. We will sometimes use the generic name o f communes to name local administrative entities. Because the administrative decentralization did not take into account the already existing decentralized health system, a set o f health posts that are managed by a single district may belong to different communes. This can be the origin o f frictions and conflict. Indeed, the district is the direct hierarchical unit above the health posts, giving training, supervising, etc, but the financing part rests on the hands o f the commune. All local governments are fully independent, there is no vertical relationship between them. Even if rural communities are located within a region, the mayor o f that rural community is not hierarchically inferior to the president o f the regional council o f the region in which the rural community is. The mayor takes the decisions that concern his locality without any need to refer to a superior, except the central government, at the Direction o f Local Governments, located inthe Ministry o f Local Governments delegated to the Ministryo f Interior. 143 8.9 Inthe decentralization reform, the following prerogatives have been given to the local authorities concerningthe health sector: The Region: 1) management and maintenance o f regional and departmental hospitals; 2) management, maintenance, and equipment o f rural health centers located in rural communities; and 3) implementationof preventionand hygiene policies. The commune: 1) management, maintenance, and equipment of urban health centers; and 2) building, management, maintenance, and equipment o furbanhealthposts The Rural community: 1) building, management, maintenance, and equipment of rural healthposts, rural maternities and healthhuts (cases de santk). The communes inour sample are given inthe following table: Table 8-1: Communesand districtsinthe sample of the survey REGION DEPARTMENT COMMUNE RURALCOMMUNUY DISTRICT P1lun& Pilune Pilune DAKAR Rufisque Rufisque BarGy Sangalcam Rufisque Sebikotane Kaolack Kaolack Thiar6; Ndiaffatte Ndoffane Keur So&; Ndiebel Kaolack KAOLACK Nioro Nioro-'6 10ruralcommunities listedin footnote Nioro Notto Thies Thies Fandem Thies THIES Diender Guedj Mbour Mbour Malicounda Mbour Tambacounda Tambacoundaa7 6 ruralcommunities listed in footnote Tambacounda TAMBACOUNDA Kothiany; Koulor; Bake1 Goudiry Bala; Bele; Kdra; BaniIsrael; Goudiry Dougu6 SaintLouis SaintLouis Mpal Gandon SaintLouis Fanaye SAINT-LOUIS Podor NdiayenePendao Podor Ndioum Mboumba Podor Gollere GaloyaToucouleur 8.10 We will only study the after-decentralization period which covers the years 1997 to date. The government finances the local government through two main channels, the 45There i s no rural community in the department of Pikine which is totally urban. 46The tenruralcommunities in the district of Nioro are: MedinaSabakh; Paos Koto; Prokhane; Kaymor; Ngayene; KeurMadiabel; KeurMaba Diahou; NdramCEscale; Wack Ngouna; TaYbaNiasskne 47The six rural communities inthe district of Tambacoundaare: Sinthiou MalCne; Koussanar; Koumpentoum; NBtCboulou;Missirah; Maka 144 Decentralization Fund (Fonds de Dotation a la Decentralisation (FDD)), and the Equipment Fund (Fond d'Equipement des Collectivitks Locales (FECL)). The FDD covers current non-salary expenditures, whereas the FECL covers for the investment program. B. THETRACKINGEXERCISE: THE FLOW OF FUNDS 8.11 The circuit followed by flow o f funds from the central level to the frontline service provider depends on the nature o f the resources, which has also a bearing on the administrative actors involved inthe system. To better understandthat system let us first quickly recall the main features the decentralization reforms have introduced into the health sector. 8.12 The health ministry has organized the supply o f health services in a pyramidal way with three distinct levels o fprovision: primary health care i s delivery by the healthcenters, posts and huts.These facilities are under the formal authority o f the district for the Ministry o f Health's administration; 0 secondary health care i s provided at the level o f departmental and regional hospitals under the authority o f the medical region; tertiary healthcare At the top o f the pyramidi s provided at national hospitals and hospital with a research component directly under the authority o f the Ministry o f Health. 8.13 The head o f both the district and the region is called he medical chief o f the district (region). The chain o f command within the organizational structure o f the M o H goes from the Cabinet o f the Minister, to the medical chief o fthe region, then the medical chief o f the district, and finally the head o f the facilities. The administrative local government counterparts are as stated above the region, the commune, and the rural community, with their delegated responsibilities. 8.14 To date, the central government provides support to the local government for the nine transferred responsibilities only for non-wage recurrent expenditure. The budget support comes through two main channels, the decentralization fund (FDD or Fonds de Dotation de la Dkcentralisation) for recurrent non capital expenditures, and the local government's equipment fund (FECL or Fonds d'Equipement des Collectivitks Locales) for capital andinvestment expenditures. 8.15 The wage bill is still centralized and received by the health agents directly. Although it would be interesting to track salaries, this study only focuses on non-wage expenditures. A similar study will be conducted in Education where salaries will be tracked mainly for the purpose o f detecting ghost workers, i.e. paid civil servants who no longer exist or work elsewhere, and reconciling the Ministryo f Finance and the Ministry o f Education human resources databases. This exercise might also be prove very interestinginthe health sector, inwhich anecdotal evidence suggests than there are many 145 workers whose position inremote area i s vacant while those agents live inDakar. This i s a serious inefficiency that will need rectification in a subsequent exercise. This type o f wastage also explains part o f the weak correlation between budget and outcomes. Indeed, these wage expenditure that plunder the Ministry o f Health budgethave a negative, or at best zero, payoff for the health sector. 8.16 Through the decentralization fund, the health providers are supposed to receive resources to cover if not all, at least part, o f their running costs. Since the beginning o f cost recovery, private expenditure partially crowded out public expenditure since the central government reduces its budget support to health facilities that were deemed to have enough resources in their revolving funds for drugs. However, since the decentralization, the healthbudget i s managed by the communes. The region receives and manages the budget o f the departmental and regional hospitals, and also the budget o f the rural district; the commune receives and manages the budget o f the district; the rural community i s supposed to receive the budget o fthe health posts. c. TRAJECTORYTHE FLOWSOFFUNDS OF 8.17 The path o f any resources is contingent on its type. Capital and wage expenditure do not follow the same circuit and are even not executed by the same governmental agency although they both depart from the central government budget and are intendedto reach the same beneficiary. Table 8-2 describes for the different resources the path they follow. Table 8-2: Flow of ResourcesTrajectory Expenditure Type Source of Finance ExecutingAgency Final Recipient Wage State Budget: current expenditure MoH-MoF Recurrent Expenditure FDD LG HealthPosts, @on-wage) LGown resources IDPH HealthCenters, Subsidies Hospitals P T Capital Expenditure FECL LL CIB AGETIP CSP Note: CIB= consolidated investment budget; CSP = commune support program; I D P H (PDIS) = integrated development program for health; 8.18 This study will be mostly interested inthe decentralizationfundwhich constitutes the main resource for the day-to-day running o f health posts that supply most o f health services inthe country. The other types o f expenditure will also be quickly analyzed. 8.19 One o f the major drawback o f the design o f this fund is its linking to VAT receipts which depend on a number of factors completed out o f line with the health needs of the country. The percentage o f VAT that goes to the FDD has, however, not yet been fixed by the government. A better rule would be to decide on the amount to allocate to 146 this fund depending on expressed needs o f the beneficiaries and the government's priorities in the health matter. Notwithstanding the way it i s decided, a good feature o f the FDD i s that its global annual amount i s at least known to all actors since it is recorded in the State budget. Table 8-3 shows the evolution o f the FDD since the 1996 decentralizationreforms: Table 8-3: FDD trends Structure and Allocations of the FDD InCurrent CFAFrancs ~~~~ 1997 1998 1999 2000 2001 2002 Regions 3,026,809,255 3,026,809,255 3,548,785,255 3,698,048,164 3,931,976,258 4,011,976,258 Communes 1,622,240,355 1,622,240,355 2,020,264,385 2,121,830,914 2,508,329,692 2,924,581,682 Rural communitys 142,692,666 142,692,666 222,692,666 242,494,962 338,963,470 652,979,060 Support to deconcentrated services 97,794,724 97,794,724 97,794,724 527,162,960 255,133,795 200,000,000 RegionalAgency of Development 255,133,795 300,000,000 TOTALFDD 4,889,537,000 4,889,537,000 5,889,537,030 6,589,537,000 7,289,537,010 8,089,537,000 variation 20% 12% 11% 11% 8.20 It should be emphasized here that this concerned all nine delegated domains o f responsibilities. A share o f the FDD i s allocated to the deconcentrated services o f the State to finance their costs in the technical help they provide to local authorities in the management o f their new responsibilities. Around 10% i s also allocated to regional services and agencies such as the regional agency o f development. The lion's share o f the FDDbelongs to the eleven regions probably commensurate to their relative charges. The share o f health is, however, earmarked which considerably facilitates the tracking exercise. For instance, the FDD-Health i s 2.48 and 2.55 billions o f CFA francs in 1999 and 2000 respectively. FDD-Health is therefore around 40% o f the total FDD underscoring the importance the government attaches to this component. In absolute terms, however, the amount o f the FDD i s quite small, even though it i s supposed to cover only current non capital and nonwage expenditures. Themobilization circuit of the FDD 8.21 Now we know the level to which the FDD i s fixed every year, let's turnto the real issue o f how it i s mobilized and how it reaches the intended beneficiaries for its health component. There are seven main steps for the mobilization o f the FDDresources: 147 1) the amount o f the FDDis fixed by the MoF and voted by the National Assembly in the Law of Finance. The FDD is accounted for inthe budget of the Ministry of Interior (MoI) that i s the authority for the local governments; 2) the National Council for the Development o f Local Governments (CNDCL Conseil National de Dkveloppement des Collectivitks Locales) meets and decides on the allocation o f the FDD among the different local governments. This council is presided by the President, members o f the government, two members o f the parliament, two govemors, two delegates from the Mayors' Association, and two president o f regional council sit inthis high level council which meets once a year if ever. A decree o f repartition i s taken as an outcome o f this meeting, the decree specifies the aggregate amounts allocated by function in the FDD, e.g. support to deconcentrated services; 3) the inter-ministerial committee composed o f the MoF and the M o I meets to decide on the final repartition o f the FDD, by local government and according to the different credit lines as health, education, support to deconcentrated services, etc, the outcome i s an inter-ministerial argte; 4) each local government on the basis o f its allocated budget in the FDD, writes its global budgetwhich i s transmittedto the MoI; 5) the M o I submits the commitments o fthe local governments to the MoF; 6) the MoF after intemal operations o f control and puts into place lines o f credit for the local governments andnotifies them through "mandats de prkposks"; 7) the local governments have their budget and can spend it on the behalf o f the health facilities as earmarked. 8.22 Having disaggregated the circuit o f the FDD it i s now possible to precisely evaluate this system and pin down those steps where delays are most important. A further step will be need to reach the front line provider, indeed, the seventh step stops at the local government level. We are therefore still at the central level. Our ultimate goal is to know how much resources the local government sends to the beneficiary provider. This questionwill be answered thanks too the questionnaires, when data has been collected. The evaluation of the FDD circuit 8.23 The tracking is interested in two main issues, 1) delays and 2) leakages. Let us first have a look at potential delays inthe mobilization o fthe FDD. Mobilization time 8.24 Because each step in the mobilization o f the FDD is followed by an administrative act, it i s fairly straightforward to quantify the time it takes for the FDD to be available to the local governments. Unfortunately, this time might hugely vary from one year to another. Indeed, although these seven steps must follow each other (or quite), there i s no safeguards that can avoid sliding schedules, and therefore induce variations in 148 the length o f the mobilization time. Before getting into more details, let us first quickly give a flavor of the average year with table 8-4: Table 8-4: the average year for the FDD circuit Step Number Periods 1. Adoption of the Finance Law by the Parliament December n-1to January n 2. Meetingof the CNDCL: Decree of repartition of the FDD March of the current year 3. Inter-ministerial meeting MoF and MoI: Inter-ministerial Arrete Aprilof the currentyear 4. Local governments budgets sent to the M o I May 5. M o Isends local governments' budget to MoF May - June 6. Line of credit in place and nohfication to the localgovernments July - October 7. Budgets for recurrent non-wage expenditure at the frontline health services providers are in September - October place 8.25 It takes on average ten months for the FDDresources to be at the disposal o f the providers. This leaves only two months to the facility to absorb those resources. There i s clearly an imbalance in the system. The meeting o f the CNDCL that only decides on aggregate allocations and hands out the work to the inter-ministerial meeting seems out of purpose. Indeed, it takes three months for this meeting to take place mostly because it i s presided by the President whose time i s very limited. This step has no benefit, and imposes long delays on the FDDresources, hampering service delivery. 8.26 Let us quickly present from 1997 to 2001 the different dates at which the CNDCL met, ifit did, the decree promulgation dates, and all the different steps resources ought to go through before being at the disposal o fthe health services providers. 149 Table 8-5: Steps for the trajectory of FDD resources Identification Step 1997 7 Year 2000 2001 units Adoption of the Finance Law date 12/31/96 12/8/97 12/31/98 12113199 12/27/00 Law No 96-32 97-01 98-51 99-88 2001-13 ('000) F 4,889,537 4,889,537 5,889,537 6,589,537 7,289,537 3,500,000 3,500,000 3,500,000 3,500,000 3,820,000 CNDCL Decision Coxnit6Nationalde Dbveloppement des CollectivitesLocales - CNDCLMeeting date 01/26/00 01/30/01 - Decree ofFDD allocation date 12/27/96 06/10/98 03/04/99 03/15/00 Decree 98-500 99-181 Inter-ministerial Meeting Distribution of FDDbetween LGs date LGs send their budgets to the M o I Ministry of Interior -Dakar Region Region date 13-Ma 8-Aug 5-Jm Commune of Pikuie date Commune of Rufisque date 18-Se Commune of B a r p y date 28-Se~ 29-Mar 22-Jar 15-Feb 29-Ja place for LGs LocalGovernments - Dakar Region 9-Oct 16-Auf 24-Oct 31-Aug 31-0~1 29-SeE 31-0ci 29-Se~ Conmiunaute rurale de Yenne 1 31-0~1 Actual Receipt of << mandats de prepose by N Local Governments- Thies date Region Region M a l August Augus Commune of "hips Ma) August Augus Commune of Mbour Augusi August Augus Ruralcommunity of Notto Augusi September JUl) Ruralcommunity of Tassette Augusi September JUlj 150 8.27 Let us discuss more in detail each o f the first six steps o f the FDD resources mobilization process. These steps almost exclusively involve the central level agencies only. Stepl: TheFinance Law 8.28 It is regularly voted by the National Assembly inDecember o f the year preceding the actual fiscal year. Table 8-5 shows that the Law i s voted at the latest on 31 December. The amount of the FDDandthe FECLare recorded inthe Finance Law. Step 2: Meetina of the CNDCL 8.29 The CNDCL meets and decides on the allocation o f the FDD among the nine delegated responsibilities. The allocation is then recorded in the "Decree o f Repartition" (DoR). All line ministriesalong with the Ministryo f Decentralization and the Ministryo f the Budget seat inthe CDNCL. Apart form 1997, the first year o f the decentralization, the decree o f repartition i s available at the earliest on March. In 1998, the decree has been promulgated inJune and in2001 the decree has even not been promulgated. The CNDCL is in fact a very cumbersome step to pass, indeed it is presided by the President whose time is very limited. However, after three unsuccessful attempts for the President's availability, a restricted committee i s put in place to decide on the repartition, this has been the case in 1997 and 2000. Because it only decides on rough aggregate allocations, and imposes huge delays on the process, the usefulness o f the CNDCL meeting can be questioned. In 2001 for instance, the Decree o f Repartition has even not been promulgatedbut the process did follow its course. Step 3: The inter-ministerial meeting 8.30 This i s the key meeting inthe whole process o fthe FDD allocation. Once the DoR has been adopted by the CNDCL and aggregate allocation for the delegated responsibilities decided, the Ministry o f Budget and the Ministry o f Decentralization meet with other line ministries who have a stake inthe FDD and decide on the allocation to each commune. A list o f the local governments with the precise amount o f FDDthat i s allocated to each o f them along with the breakdown for health, education, and so on is produced during this meeting. The inter-ministerial meeting normally takes place once the DoR has been promulgated. It takes on average one and a halfmonth after the decree for this meeting to take place. The date the inter-ministerial meeting in 1998 i s not available but since the DoR has been available in mid-june one can estimate this date around mid-august. Again, except for the very first year o f the decentralization in which the inter-ministerial meetinghas been held in february, this meeting is held in the best case scenario inapril. In2001 the meeting took place injune without the DoR. 151 Step 4: Local Governments send their budaet to the Ministry ofhterior 8.31 This is merely a technical step that has almost no bearing on the process. The local government's budget must be in conformity with the prescriptions o f the Ministry o f Finance in terms o f allocations to the different sectors that reflect the government's overall objective. This i s reminiscent o f a centralized system and therefore somehow in contrast with the whole idea o f decentralization that transfers policy orientation at the local level. The only rationales one can find for this procedure is the infancy o f the decentralization process, and the lack o f competence at the local level. The budget o f the local government i s approved by the authority that represents the central government at the local level, namely the Governor for the region, the Prkfet for the commune, and.the Sous-Prkfet for the rural community. Step 5: TheMinistw of Interior seizes the Ministw of Finance for credit lines settlement 8.32 This is also a technical step. After the reception o f the local governments budget, the Ministry o f Interior requires the Ministryo f Finance to credit the account o f the local governments with the allocation they have been given at the inter-ministerial meeting. This procedure is handled by the Procurement and General Administration Directorate (DAGE) o f the MoI. Because o f the high number o f local governments, and the lack o f humanresources at the DAGElevel, this step can be very long. Step 6: TheMinistr?,of Financeputs credit lines in place and notifies local governments 8.33 The Ministry o f Finance after receiving the request from the Ministry o f Interior engages an internalprocedure to make available to the local governments their funds. The Ministry issues notifications to the local governments once their credit is in place, the notification instrument is a "mandat de prkposk". For a commune, the reception o f its "mandat" means that funds are available at the regional treasury. H o w fast the "mandats" are issued depends on the effort the agents o f the Ministry devote to this task, and on the availability o f funds at the central treasury. From Table 8-5, the local governments can receive their "mandat" and the earliest in may and sometimes as late as October, which leaves them only two months to send the resources to the health services providers that are supposed to absorb all these resources before the fiscal year ends. The date o f availability o f the funds at the local level varies between the regions. Unfortunately, due to very bad record keeping practices, data have been gathered only for the regions o f Dakar and Thibs. Communes in Dakar seem to receive their funds later than communes inThibs. Recommendationsfor reducing the mobilization time of the FDD 8.34 The process o f the mobilization o f resources o f the FDD can take up to ten months form the adoption o f the Finance Law to the notification o f the "mandats de prkposCs". Because the FDDconstitutes an important source o f funding for the providers, this is a real constraint for their day-to-day functioning. The steps that have been described above all take place at the central level and solutions for a much shorter 152 mobilization time can be found. For any o f these steps, except the adoption o f the Finance Law which is timely, it i s possible to find a way to reduce delays. 8.35 The first and most obvious action that jumps to mind i s the suppression o f the CNDCL meeting. Indeed, this meeting imposes very long delays on the process and does not have any added value. The aggregate repartition o f the FDD, as decided by the CNDCL, can be done either at the Parliament level and included inthe Finance Law, or during the inter-ministerial committee that would then decide on both aggregates and specific allocations. This would represent a gain o f three months and relieve the President's schedule o f the CNDCL meeting which he does not really need to attend. 8.36 A second way to reduce even further the delays on the FDD reconstitution due to rigidities at the central level i s to decide on a deadline for the inter-ministerial committee meeting. Ifthe CNDCL meeting i s suppressed, this deadline can be fixed around mid- or end-january which i s a reasonable date. Leakages 8.37 After the end o f the mobilization process o f the FDD, the resources o f the health providers are managed by the local government officials. There i s therefore still one step remaining before the resources arrive at the intendedbeneficiaries. It i s also at that level that leakages are possible. Let us analyze the amounts received by the communes in the regions o f Dakar and Thiks, for which data are available, and the amounts they transferred to all health providers for which they hold the FDDbudget. 153 Table 8-6: Health FDD receivedby the local Governments and health providers I Identification I UnitsI 1997 7 Amounts received for FDD/Health Region 343,324 Dakar Communeof Pikine K-FCFA 53,756 53,756 II 66,945 II 292,841 II 70,906 70,325 37,668 I 40,975 I 37,991 I 49,344 1 8 ' Allocations to healthservices providers ~ Medical Regionof Amount K-FCFA 1 Dakar Availability Date 28-Mar District of Pikine Amount 26,878 35,162 Availability ~ 1 1 !:? 1 District of Rufisque Amount K-FCFA 30,943 34,675 34,678 , I Availabilitv Date ~. ~ I I I 7 Amounts received for FDD/Health 1I 1 I Region I K-FCFA I 115,751 Thies Communeof Thi& K-FCFA 7,500 8,000 8,000 1 Region I I ~ II CommuneofMbour K-FCFA 5,000 5,000 6,000 6,000 10,000 I 8 Allocations to health ~ MedicalRegionof I I 11,102 11,102 11,102 11,102 ~ I Thies July July July August District of Thies Amount 10,811 10,811 10,811 10,811 10,811 Availability July July July August 4 Tulv , District of Mbour Amount 15,297 15,297 15,297 Date K-FCFA 15,297 15,297 Availabilitv Tulv July July July August 8.38 For the administrative region o f Dakar, only the amounts in 1998 and 2000 are available. Among the communes, the regionsreceive larger amounts o fresources than the communes. This is commensurate with their responsibilities because they have the departmental and regional hospitals. The communes in the region o f Dakar also receives more resources than communes o f the same level inthe region o f Thibs. 8.39 Let us first note that local governments can allocate some o f their own resources to health. Therefore the amount o f health expenditure at the local level may be higher than the FDD allocation. The contrary i s also possible, Le. expenditure smaller than FDD allocations. However, because o f the hngibility o fthe FDD allocation it is not possibleto tell how much o f this allocation the local government gives to providers. There i s here also a serious problem of asymmetric information. Indeed, the health providers do not know how much has been allocated to the commune under whose responsibility they are to spend on health. 8.40 It is interesting to note the variation inthe dates o f availability o f resources at the provider level. Inthe region o f Dakar, although the region receives its FDD resources at 154 the earliest in august (see Table8-5), it transfers resources to the medical region between february and march. The region therefore takes on its own budget to transfer to the health providers and probably recollects its money once the "mandat de priposC" i s received. The same pattem i s also observed for the commune o f Rufisque. On the contrary, the commune o f Pikine always awaits hnds to be available (or nearly) to disburse to its health providers whose resources are available at the earliest in august. The region o f Dakar and the commune o f Rufisque try to protect their health sector from the volatility o f the date o f availability o f funds - a risk imposed by the central government - whereas Pikine completely transfers this risk onto its providers. For the region o f Thibs, a strange regular pattem appears in which all communes make resources available to their health providers at the same time and sometimes after sometimes before they themselves receive the funds. Allocations vs Actual Receipts 1997 1998 1999 2000 2001 Mean Region of Dakar I I 343,324 I I 292,841 I Medical Region 2.508 10,824 10.824 13,824 9,495 Ratio A/AR II0.0315 II II 0.0472 II 3.9% I1 Commune of Pikine 53,756 53,756 66,945 70,325 70,906 63,138 District of Pikine 26,878 26.878 33.473 35.162 34.453 31.369 Ratio A/AR 0.50 0.50 I 0.50 I 0.50 1 0.49 I 50% 1 Commune o f Rufisque 27,791 37,668 40,975 37,991 49,344 38,754 District of Rufisque ND 30,943 34,675 34,678 33,432 Ratio A/AR 0.76 0.91 0.70 I I I I I 86% Thih Region II Region of Thies - I 115,751 132,416 153,562 159,321 151,800 142,570 Medical Region 11,102 11,102 11,102 11,102 11,102 11,102 Ratio A/AR 0.104 0.119 0.138 0.I43 0.136 12.8% Commune of Thies 7,500 7,500 7,500 8,000 8,000 7,700 District of Thibs 10,811 10,811 10,811 10,811 10,811 10,811 Ratio A/AR 1.44 1.44 1.44 1.35 1.35 I41% Commune of Mbour 5,000 5,000 6,000 6,000 10,000 6,400 District of Mbour 15,297 15,297 15,297 15,297 15,297 15,297 Ratio A/AR 3.06 3.06 3.06 2.55 2.55 285% 8.41 Table 8-7 gives the amounts received by the local governments for the FDD health versus the amounts they sent to the health provider, there i s a row which gives the ratio of these two terms, for the different communes inDakar and Thibs. For both Dakar and Thibs, the medical region receives very little from the resources allocated to them and managed by the president o f the regional council. InDakar, for the two years both numbers are available, the region allocates on average only 3.9% o f the resources earmarked to the medical region, whereas the ratio climbs to only 12.8% for Thibs. The picture i s sharply different at the commune level. Indeed, in Dakar the communes of 155 Rufisque and Pikine give less than they receive to their health providers, whereas in Thies the communes o f Thies and Mbour give more to the providers than the resources allocated to them bythe FDD. 8.42 Some local governments give resources in advance to the frontline service providers whereas others wait for funds to be available, some give more resources than normally allocated whereas others divide allocated resources by two or even 25 for some regions. How can one rationalize these huge variations in the same system? There are at least two possible determinants one can think of. First, the local governments that are now responsible for the health policy have their own fiscal revenue apart from what i s transferred to them by the central government. Therefore, those that have health as a priority might allocate more resources and make those resources available to the health providers before the central government send the FDD resources. Second, inthe process of decentralization the local governments have been transferred nine responsibilities and therefore have other priorities than health. Health i s allocated the highest share in the FDD leaving other areas o f responsibility with small resources. If the priorities o f the local governments is, say, urbanization and habitat they can divert resources from health and allocate them to their number one priority. This effect is exacerbated by the fact that only the local governments have the information on the breakdown o f the FDD. And even if the providers had the information, they have no handle to induce the local authority to disburse the money to them. However, having this information will anyway put them in a better position to claim their resources and increases their bargaining power. 8.43 The fact that region are bigger, and have more priorities probably explains the very low share o f health resources that get to the health providers. The regions use most of the resources earmarked to health for other purposes. 8.44 InDakar, leakages are very important in the communes o f Pikine and Rufisque that allocate less resources to the health providers than the FDD requires. Incontrast, the communes o f ThiBs and Mbour in the region o f ThiBs take from their own resources to add to the FDD allocations which are however much smaller than the allocations for Pikine andRufisque.Note also that throughout the period o f 1997 -2001 the communes o f ThiBs and Mbour allocate a fixed amount to the health sector. Therefore, increases in the FDD allocation do not affect the policy o f the local governments, they merely crowd out existing resources. D. CONCLUSION 8.45 This study has mainly looked at the trajectory o f the resources o f the Decentralization Fund from the central level to the service providers. Apart from user fees, the resources from the FDD are the main revenue at the disposal o f the health facilities and therefore receiving these resources in full and in a timely manner can help greatly improve the quality o fpublic health services. 156 8.46 A number o f constraint have however been identified in the process o f the mobilization o f the FDD resources. These constraints can induce both delays and/or leakage o f the resources. The first and very important constraint to be underscored i s the very late reconstitution o f the FDD resources. Indeed, it takes between eight and ten months for the FDD resources, irrespective o f the size o f the amount, to reach the frontline service providers. The process i s burdenedby a number o f usually useless steps as for instance the CNDCL meeting which takes three to six months and has no value added. 8.47 Another important ingredient that is crucial mostly for leakage is the asymmetry o f information between the local governments and the frontline service providers on the amount o f the FDD health allocated to the local governments. To reduce this asymmetry the list produced as a result o f the inter-ministerial meeting should be sent to all health providers. 8.48 The decentralization process has introduced major changes in the health sector. Among them, the most important i s the transfer o f their resources to the local governments. There i s a misalignment between the objectives o f the health centers and the detainees o f their financial resources, namely local authorities such as the president o f the regional council, the mayor, or the president o f the rural community. The central budget allocated to the health sector i s transferred to the local governments that also have other priorities and have been transferred other responsibilities as well. If asymmetries o f information are important between the managers o f the health centers and posts and the local governments, financial means might never reach the facilities. It i s also crucial to assess the "true" level o f accountability o f the local governments to their constituencies to be able to gauge the impact o f the effectiveness o f the decentralization process which is still inits infancy. 8.49 Health centers and posts have no handle on the local government to get the resources allocated to them by the central government. Moreover, local governments have no incentive to channel up the money to the health services because they believe that health facilities have already enough money through the revolving fimd o f drugs, that they themselves manage with complete autonomy and discretion since the Bamako Initiative. 157 TheSample of interviewed Facilities 158 9. MONITORING PRSPIMPLEMENTATION 9.1 The Government has recognized the critical importance o f developing a systematic monitoring strategy to ensure the successful implementation o f the PRSP. It has continued the participatory procedures used in developing the PRSP to construct a strategy to monitor the implementation process and evaluate the impact o f PRSP programs. This strategy will involve all segments o f society and regions o f the country. The documents describing the strategy have been discussed at a validation seminar with broad participation, and havebeen adopted by the Cabinet. 9.2 Monitoringthe progress o f implementationwill require a major effort to improve the gathering, storing, and processing o f data. The national statistical systems will need to promote better communication among all relevant actors, at all stages o f the collect and diffusion o f information. 9.3 Monitoring PRSP implementation will also require certain budget system reforms. First, poverty-reduction spendingwill be specifically identified and tracked. Second, the practice o f program budgeting will be extended beyond the Ministries o f Health and Education so that monitorable objectives are more easily identified. Third, the Government will adopt a Medium Term Expenditure Framework to ensure the medium term financing for PRSP programs, and to provide the financial framework for sectoral programbudgets. A. INNOVATIONSINTHE MONITORING POVERTYINSENEGAL OF THE 9.4 Poverty reduction i s one o f the Government's main objectives. To this end, an overall strategy has been developed using a broad participatory approach and i s to be implemented in2003. The participatory approach used to draw up the PRSP involved all stakeholders at the local and national levels, inthe public and private sectors, within civil society and among the development partners. 9.5 The strategy's mainobjectives reflect three priority themes: (i) doubling o fper the capita income by 2015 in an environment o f strong, balanced and more equitably- distributed growth; (ii)generalized access to essential social services along with the accelerated introduction o f basic infrastructures in order to enhance human capital by 2010; and (iii) eradication o f all forms o f exclusion within the nation and the the establishment o f gender equity at all levels o fprimary and secondary education by 2015. 9.6 Over the short term, and assuming that no major exogenous shocks occur, the aim will be: to ensure average growth o f seven to eight percent inorder to reduce the ranks of the poor by at least 15 percent over the 2003-2005 period; to bring about a significant increase in the rate o f access to health care and education, at the same time improving service quality and gender parity in education; and to meet, as quickly as possible, the WHO'Srecommended standards regardingdrinkingwater consumption, i.e., 35 liters per inhabitant per day. 160 B. SUPPORT TO THE STATISTICAL SYSTEM 9.7 Over the last five years, Senegal has benefited fi-om substantial technical assistance from the IMF, the World Bank and others in the area o f statistics and data management. Five missions contributed to improving the Government's fiscal system through technical support in the areas o f tax policy, macro-fiscal management, and the budgetary aspects o f the petroleum sector. Within the framework o f the IMF-BCEAO cooperation, Senegal also benefited fiom technical assistance in banking supervision, financial sector assessmentprograms andbalance o fpayments statistics. Further missions provided assistanceand recommendations on institutional issues andreal sector statistics. 9.8 Economic databases in Senegal are considered fairly comprehensive, although they require improvement in the area in the timeliness and quality o f national accounts, fiscal and balance o f payments data. Remaining weaknesses are related to the coordination o f data collection and processing, delays in data dissemination, lack o f social indicators and data, and non-availability o f adequate human and financial resources. To address some o f the above issues, an IMF statistics mission recommended inApril 2001 the strengthening ofphysical andhumancapital allocated to data collection and maintenance. With a view to supporting Senegal to comply with international standards interms o f data availability and timeliness, ajoint Bank-Fund mission visited Senegal in September 2001 to prepare a report on the observation o f standards and codes (ROSC) and to give advice with regard to Senegal's recent participation in the General DataDissemination System. 9.9 Senegal is a member o f the West African Economic and Monetary Union (WAEMU), and member countries o f the above institution have agreed, among other things, to comply with convergence criteria interms o f fiscal andinflation ratios. Beyond the above criteria, the WAEMU countries have identified key economic and social indicators, which are considered to be building blocks in the progressive walk towards stronger regional economic integration. This project aiming at strengthening the Directorate o f Forecasting and Statistics would help Senegal monitor its PRSP, its own development plan and most importantly ensure compliance with the regional criteria o f economic convergence and integration. 9.10 Inaddition, the following support is on-going: The World Bank i s supporting the Directorate o f Forecasting and Statistics in conducting the household budget and consumption survey for the period 2001-02. It is also providing support quality enhancement o f sector and macro statistics, production and monitoring o f socio-demographic and household data, strengthening data management and dissemination, and modelling. The United Nations Fund for Population Activities (UNFPA) i s supporting the Directorate o f Forecasting and Statistics in the preparation o f the population census. 161 The Fund for local development o f the Canadian International Development Agency i s providing financing for the functioning o f the Permanent Secretariat o fthe Technical Committee set up to preparethe poverty reduction strategy paper for Senegal. On the basis o f European Union financing, the Directorate o f Forecasting and Statistics intendsto launch a computer-based framework whose objective is to produce upgraded national accounts statistics and conduct a survey on informal employment in urban areas, with technical expertise from AFRISTAT over the period 2001-02. c. INFORMATIONSYSTEMSAND SURVEYS TO SUPPORT THE PRSP 9.11 Inorder to cut the poverty rate inhalf by 2015, information systems are required for monitoring progress. The Government o f Senegal has therefore opted to enhance its system for generating data and monitoring household living conditions, inorder to lay a better foundation for its policies and programs, light the way for its interventions and oversee their effects. 9.12 Inthis regard, the infomation systemdeveloped shouldmake it possible, through effective monitoring o f the impact o f the poverty reduction strategy in Senegal, to strengthen or refocus policies aimed at improving the population's living conditions. These information systems are built upon tools that provide insight into both the direct and indirect manifestations o fpoverty in Senegal. 9.13 The direct manifestations o f poverty can be observed by means o f databases and surveys. The databases and surveys developed thus far are: (1) the Live Database (LDB); (2) the Social Indicators Database (Banque de Donne'es des Indicateurs Sociaux, or BADIS); (3) the second Senegalese household survey (Enqukte se'ne'galaise auprks des me'nages 11,or ESAM 11); and (4) the 1-2-3 survey. Inaddition, studies and research are underway to develop (5) models such as 1-2-3 and the IntegratedMacroeconomic Model for Poverty Analysis (IMMPA), and (6) a monitoring indicator based on the most easily observed or quantified manifestations o fpoverty. TheLiveDatabase (LDB) 9.14 The LDB is as much as system as it is an approach. The key objective o f this approach i s to get the right information to the right people inthe most appropriate format for regular monitoring and evaluating o f progress towards the country's goals in the PRSP. Developed initially by the World Bank for its own internal use, it has since evolved to address the needs o fnational statistical systems and regional organizations and i s being implemented in a dozen countries and institutions worldwide. It addresses the data management needs o f the statistical system by focusing on the problems o f storage, access, documentation, quality control, analysis, and dissemination o f economic, sector and social indicators. 162 9.15 The LDB is a comprehensive system based on a SQL server database with a host of tools for uploading, disseminating and analyzing data. It requires the availability o f a local area network and has capability for web dissemination. The system can also operate using On-Line Analytical Processing (OLAP) for more complex analysis and business intelligence. The design structure o f the databases and its tools have been benchmarked to global best practice and in2001 earned it the DMReview "World Class Solutions Award for Business Intelligence). 9.16 The major business challenges the LDB attempts to address are: 1) key decision makers do not have access to timely, reliable and up-to-date information; 2) data are scattered across multiple databases that are often unlinked, located in various ministries, invarious formats, and arenot readily accessible; 3) generating standardreports, bulletins and annexes is time consuming, costly and often inaccurate; 4) the process o f information gathering, monitoring, analysis, and dissemination is largely manual; 5) staff are not adequately trained and work in isolation and finally 6) the quality and documentation on the data i s questionable. 9.17 The LDB approach is centered around the idea o f consolidating all the indicators needed for all possible publications and models in one database. It i s not designed to replace all the databases inDPS, but rather to extract the key strategic information from existing databases, surveys, and censuses, both within DPS as well as from other ministries for improved accessibility for publication, analysis, and web dissemination. As such, the LDB is as much about the process o f linkingusers with producers, internal change management, training, and bringing technological solutions to address existing needs. 9.18 This approach requires initial sustained support to the DPS inimproving their data administration, quality control and the establishment o f new policies and procedures for gathering and capturing o f information. A World Bank Grant from the Trust Fund for Statistical Capacity Building has as one o f its major components the recruitment o f a consulting firm to support the DPS in the implementation o f this LDB solution over an extended period. While the LDB tools have been installed at the DPS, they have not yet been internalized and linkedwith existing databases and models due to the narrow focus on storage and dissemination. 9.19 Among the key benefits of this project are 1) increased access to high-quality, up- to-date information customized to various constituencies; 2) reduction in the cost and time o f producing and dissemination reports, 3) increased demand for information leading to a higher profile for the DPS, and 4) linking users and producers o f statistics together for improved monitoring o f the PRSP goals. A key indicator o f success will be the sustainability o f this project with minimal external support. 9.20 The LDB is also expected to be implemented in a number o f neighboring countries with regional support from Afristat. This regional approach, along with the sharing o f experiences from other parts o f the continent and the World are expected to produce direct benefits to the DPS. 163 TheSocial Indicators Database (BADIS) 9.21 The Social Indicators Database (BADIS) i s a compilation o f over 2,000 time series pertaining to economic and social conditions in Senegal over more than 25 years. The social indicators used to develop a social welfare performance chart (Tableau de bord de la situation sociale, TBSS) for Senegal are extracted from BADIS. The TBSS makes it possible, on the one hand, to measure Senegal's socio-economic performance and, on the other, to monitor standard-of-living and quality-of-life trends on an ongoing basis. Inaddition, the TBSS i s an essential element inthe development o f policies aimed at meeting the population's social welfare needs. 9.22 BADIS also provides information on the following factors: population, prices and deflators, national accounts, employment and unemployment, foreign aid, health, food security, educatiodliteracy, and housing. Each o f these factors corresponds to social data covering an extended time period. Second Household Survey (ESAMII 9.23 The second Household Survey, which ended inJune 2002, i s essential to diagnose and monitor the population's living conditions. The survey will make it possible to: (i) determine the profile o f poverty and analyze its determinants; (ii)observe the distribution of annual household expenditures according to certain household characteristics, as well as the disparities between the various socio-economic groups within the population; (iii) analyze household consumption expenditures and, finally; (iv) develop indicators for employment, education, health and nutrition, property and access to housing. Inorder to further refine the data on poverty in Senegal, ESAM I1includes a special module called the Core Welfare Indicator Questionnaire (CWIQ) (Questionnaire UnzjG sur les Indicateurs de De`veloppement, QUID). It makes it possible to monitor, at regular intervals, key poverty indicators and the impact o f policies, programs and projects upon living standards. The CWIQ covers household characteristics relating to health, education, employment, child nutrition, property and housing. It also evaluates the population's access, utilization and satisfaction with regard to basic social and economic services, and particularly educational and health care infrastructures. The 1-2-3 survey 9.24 The 1-2-3 survey, currently underway, i s designed as an alternative way o f overcoming the sampling difficulties encountered in the informal sector. The immediate goals o f this operation are (i) to obtain, by means o f specific surveys, reliable and comparable data on urban employment (inboth the modern and informal sectors), on the main characteristics o f the informal sector and on household consumption and poverty; (ii)to analyze the roleofthe informal sector insatisfyinghouseholdneeds; (iii) to analyze, at the regional level, consumption structures and the scope o fpoverty within the national economy; and (iv) to analyze, at the regional level, the weight and role o f the informal sector in the national economy. The medium-term goal is to establish a mechanism to monitor employment and the informal sector, andto improve the statistical 164 services' data collection and analysis capabilities on questions relative to employment, the informal sector and household living conditions. 9.25 This survey i s beingimplemented inthree phases. Phase 1 i s along the lines o f an `employment survey'. Its purpose is two-fold, namely: (i) to provide the main indicators usedto describe the situation o f individuals and households on the labor market; and (ii) to serve as a filter for phase 2 (which deals with the selection o f informal production units (PUS).Phase 2 is ofthe `enterprise survey' type, and deals with PUS.The function of phase 2 i s to: (i) provide the main characteristics o f PUS(e.g., their performance, positioning, and outlook) and (ii) to identifierthe constraints and needs o f IPUs. Phase 3 is a type o f household"consumption budget" or "living conditions" survey, the objectives o f which are: (i) to determine the level and structure o f householdconsumption according to place o f purchase; and (ii) to quantify and analyze poverty from both the classical monetary standpoint as well as in terms o f its more recent dimensions (e.g., subjective perceptions o fpoverty, vulnerability, exclusion, etc.). The 1-2-3 model 9.26 The 1-2-3 model, which i s already inplace and was used for the PRSP, i s based upon two basic economic principles: optimization and equilibrium. The economy i s divided into two production sectors. The 20 equations that make up the structure o f the model describe the behaviors o f actors, the constraints they face and the conditions o f market equilibrium. 9.27 This model can quickly quantify the consequences o f a change in the terms o f trade and the impact o f Government expenditures on the national economy, particularly the most vulnerable households. The second iteration o f the 1-2-3 model is already operational and was used in connection with the PRSP to simulate the impact o f a fluctuation in the terms o f trade on GDP growth, consumption and household income. Simulations have also been done to determine these impacts on various population deciles and on the Gini coefficient. TheIntegrated Macroeconomic Modelfor Poverty Analysis (IMMPA) 9.28 The IMMPA was developed by a World Bank team. IMMPA is a macroeconomic evaluation and analysis technique for quantifying the impact o f policies and external shocks upon income distribution, employment and poverty in low- and middle-income highly-indebtedcountries. Thus, IMMPA is a dynamic model for analyzing the short-, medium- andlong-term effects o f certain shocks. 9.29 In its present version, the IMMPA model, which is currently being adopted in Senegal, makes it possible to evaluate the impact o f shocks on: incomes and consumption for the categories o f households under consideration; income distribution and consumption, via the GIN1 coefficient and the THEIL index; rural and urban poverty thresholds and employment. 165 Synthetic indicator of standard of living, excluding income or expenditure 9.30 The standard o f living is generally measured on the basis o f data on household income, expenditures or consumption. This data usually comes, inturn, from expenditure surveys or budget/ surveys (such as ESAM). Such surveys, although indispensable for , monitoring the well-being o f the population, are cumbersome, time-consuming to implement and, therefore, performed infrequently. They are repeated every four or five years. During this interval, poverty must be assessed at least once a year in order to gauge the efficacy o fthe policies and programs beingimplemented. 9.31 Without presumingto replace the budget/consumption surveys, the Directorate for Forecasting and Statistics (Direction de la Prevision et de la Statistique, DPS) i s seeking to construct a synthetic standard-of-living indicator that could be obtained from a simplified survey or from surveys that it routinely carries out. This indicator would be based on variables that are easy to collect and that recur inmany surveys. They have to do with housing characteristics and the ownership o f certain goods (e.g., radio, television, refrigerator, car, bicycle, etc.). 9.32 Such an indicator would make it possible to obtain poverty quintiles as well as measures o f inequality. It could be cross-referenced with the survey's relevant variables (e.g., health, education, etc.) in order to ascertain household behaviors depending on standard o f living. Ifthe indicator is constructed from the same variables as were usedin a precedingsurvey, it could serve to assess the direction o f the poverty trend. 9.,33 The indirect manifestations o f poverty are detected through the behavior o f the economy, as measured by (1) the industrial production index; (2) observation o f the business cycle; and (3) the national accounts. Industrial Production Index (IPr) 9.34 The industrial production index, which had been based on figures from 1976, has been updated. Due to changes inthe economy caused by the creation and elimination o f enterprises and industrial restructuring operations, the sample was no longer representative and the weighting no longer reflective o f current industrial structures. In order to be prepared in advance for the planned introduction o f a unified industrial production index for the WAEMU countries, the DPS i s updating the monitoring sample by re-focusing the index on the year 1999, following a survey done between September and October 2001. The new sample covers 71 enterprises that account for 90.8 percent o f the country's industrialvalue added. 9.35 Two types o f indices are presented : (i)variable-base index (1999=100) that compares the level o f production for the current quarter (Le., the first quarter o f 2002) to that o f the same quarter o f the baseline year (i.e., the first quarter o f 1999); and (ii) the fixed-base index that compares the level o f production for the current quarter to the quarterlyaverage for the baseline year. 166 Improved monitoring of economic trends 9.36 The desired changes in monitoring economic trends via the economic outlook report are justified by the concern with identifying the impact o f the trends on the population's living conditions. The changes pertain to form and content, giventhe nature o f users' demands for economic trend data. As far as content i s concerned, the scope o f reporting has been broadened. For the primary sector, subchapters have been added on livestock products (e.g., meat, eggs, milk, etc.) and forest products subject to quotas. For the secondary sector, the plan i s to monitor, on a regular basis, a representative subsampling derived from the new IPI. Reporting on the tertiary sector takes into account telecommunications; the activities o f the Regional Stock Exchange; land, rail and maritime transportation o f passengers; and the activities o f travel agencies. The frequency o f the report's publication has also been improved. A chapter on employment i s planned. Updating of national accountsfor 1997-2001 9.37 Senegal's National Accounts using SNA1987 as the baseline year were no longer reflective o f Senegal's economic realities. The effort is designed to comply with the new National Accounting rules (SNA 93) that have been progressively put into practice in most countries o f the region. An updating o f Senegal's national accounts is underway. The new nomenclature conforms to the Nomenclature for Activities o f AFRISTAT Member Countries (Nomenclature des Activith des Etats Membres d 'AFRISTAT, NAEMA) and to the Nomenclature for Products o f AFRISTAT Member Countries (Nomenclature des Produits des Etats Membres d'AFRISTAT, NOPEMA) which are, in turn, in accord with the third revision o f the International Standard Industrial Classification o f all Economic Activities (ISIC). 9.38 The improvements brought about by this update have harmonized Senegal's national accounting system with SNA 1993, which i s the system adopted by the WAEMU and CEMAC countries. This allows for international comparisons in the context o f multilateral oversight. The main corrections had to do with the evaluation o f agricultural production and o f the public administration. The development o f revised national accounts were also an opportunity to set up a database that makes it possible to mobilize new sources, facilitate comparisons, and improve economic data management and the procedures usedto develop national accounts. 9.39 The DPS subsequently embarked upon the adoption o f AFRISTAT's ERETES data-processing module, which was developed usingprinciples that have already inspired this new series. The baseline year selected for introduction o f the ERETES module is 1999; it will be used to produce a new series to be revised every two years. Thus, this series o f national accounts initiates a process of revision occurring in parallel with the improvement o f Senegal's statistics and calculation tools. The revised series will ultimately be used for retrospective extrapolation back to the beginning o f the preceding decade. 10. IMPROVING THE EFFICIENCYOFPUBLIC EXPENDITURES Executive Summary 10.1 The devaluation o f the CFA franc in 1994 enabled Senegal to return to the path o f economic growth and contributed to considerable easing in the financial constraints affecting public expenditure. Despite this improvement inpublic finances, public investment execution continues to be characterized by a sizable deviation from programmed levels, which has negative repercussions on Senegal's economic growth and socioeconomic development. Thus, over the five-year period 1996-2000, the average execution rate was only a bit over 68 percent? with a peak at 77 percent in 1999 and a low o f 60 percent in2000.49The greatest constraints appear to relate to investments financed against external resources, for which it i s rare when halfthe annual appropriations are put to use. 10.2 This situation i s worrisome to the public authorities as they are confronted by a rising tide of impatience from the people who want their pressing needs for economic and social infrastructure to be satisfied, particularly following the change in political leadership which occurred in2000 and because o f the demands arising from the forthcoming implementation o f the new poverty reduction strategy and the introduction o f medium-term expenditure frameworks (CDMTs), or budgetingby objectives. 10.3 It is inthis context that it is necessary to situate the work ofthe group established by the Ministry o f Economy and Finance, whose conclusions are reflected in this report. The purpose o f the report, after providing a brief diagnosis o f the various constraints that bear on the project life cycle, is to propose to the authorities a plan o f actions that can ease the obstacles observed inproject and expenditure management, and the implementation o f which will significantly improve public investmentperformance. Shortcomingsof thepublic expenditure management mechanism in Senegal 10.4 The difficulties noted in the implementation o f the public finance management system are nothing new. The most obvious dysfunctions are found: - first, in the capacities, tools, and work methods o f the Ministryo f Economy and Finance (MEF) and the various technical ministries (TMs) for designing, executing, and monitoring public investmentprojects on the institutional, technical, and financial planes. - second, in the complexity and unwieldiness o f the administrative and budgetary procedures for expenditure execution. The delays observed in this regard are principally the result of 48Or project executioninthe amount of CFAF 1,057.6 billionas comparedto appropriations opened inthe amount ofCFAF 1,564.4 billionover theperiod. 491999was ayear of exceptionalachievementsbecauseof the mobilization ofthepolitical authorities on the eve ofthe 2000 elections, while performancesuffered in2000 from the political and institutional changes taking place inMarch2000. 168 the form known as the "technical and financial execution form," the completion o f which was a prerequisite for any resource mobilization; the long times required to close fiscal years and the difficulties o f carrying investment appropriations forward when they were committed but no payment orders hadbeenissued, or when they had not beencommitted; constraints affecting the awarding o f government contracts, in particular the absenceof deadlines conveyedto the various and multiplepartiesconcerned in the operations of calling for bids, opening submissions, and awarding the successful bidder inpublic biddingprocedures; payment delays, themselves also attributable to the lack of deadlines by which concerned staffmust process materials, as well as to cash flow problems; the selection o f enterprises with limited technical capacities, financial resources, and access to bank credit for undertaking the work entrusted to them sometimes in priority sectors (construction and refurbishingof roads, classrooms, health centers, wells, etc.); and the multiplicity and complexity of disbursement procedures for foreign donors, each o f which imposes its own methods and procedures which differ from one organization to the next and from the methods and procedures used by the government. - third, as regards deficiencies in the organization and operation of public investment projects. The shortcomings relate to: 0 the lack of a harmonized legal status for projects, exposing the MTs and project leaders to incurring risks harmful to the proper conduct o f project activities; 0 a lack o f rigor in recruitment, which i s not always carried out in keeping with efficiency and transparency criteria; non-existent ex-post control (and, as a consequence, the absence o f a penalty system) for project operations and performance, in particular projects financed against domestic resources; 0 imperfect control by project leaders o f the planning o f activities and procedures for awarding contracts. - fourthand last, as regards limitsto the information system on projects and expenditure. These pertain for the most part to: 0 the fragmenting into multiple, autonomous subsystems which do not share information with one another and are managed by different units (DTAI, DDI, Treasury); 0 limited automation since the DDI subsystem is semimanual while the others (DTAI and Treasury) are semiautomatic, raising the problem of the exhaustiveness and rapid andregular availability o fthe data recorded; limited integration of the new public expenditure software (DTAI) even though it represents significant advances in the management of the expenditure process; persistent delays in the operational functioning of the AIDA system (Automated System for Investments, Debt, and Aid), which i s supposed to be 169 usable for managing the investment process in its entirety, whether financed against externalresources or fundeddomestically; 0 and the obstacles incurred with introducing the project accounting and financial framework, which is essential for harmonizing and standardizing the accounting systems o f projects and for providing a means of more effectively monitoring and controlling the financial operations carried out by projects (ongoing control and control o f closing statements). 10.5 This diagnosis illuminates the range o f constraints to be eased for there to be any resulting improvement in the resource absorption capacities o f development projects. Consequently, the recommendations set forth inthe following paragraphs represent the range of the priority action plans that must be carried out in order to achieve the objectives o f the PRSP, inparticular those relating to increasingproduction and satisfying social demand. Actions to be carried out to speed the absorptionof resources bypublic investment programs 10.6 The action plan revolves around four lines o f action addressing the major concerns identified inthe diagnosis above. 10.7 It focuses first on strengthening the capacities of the MTs and the MEF inmanaging public investment projects and programs. The proposed solutions are functional, institutional, and methodological innature, and concern: 0 following its updating, continued implementation o f the action plan resulting from various recommendations, including those formulated by the latest Superior Council on the Plan and updatedinthe context o f the diagnostic studies o f the capacities o f the technical ministries (DP/LTNDP, 1990; DCEF/GTZ, 1999); 0 the introduction o f a new ad-hoc structure responsible for examining the bottlenecks to the proper execution o f investments and identifying solutions that will improve the efficiency o fpublic investment. This structure will be patterned after the Ad Hoc MonitoringCommittee for the 1999 Budget Reconciliation Law referred to above. It will be chaired by the Prime Minister with secretariat services provided by the MEF, which will thereby have an appropriate channel for guaranteeing the physical and financial monitoring o f the public investment programs carried out by some ministries; 0 the gradual extension, in particular in the sectors identified as priorities by the PRSP, o f the methodology for preparing CDMTs or budgeting by objectives, already inuse inthe education and health sectors, with a view to facilitating the preparation, programming, and ex-ante and ex-post evaluation o f programs and projects; 10.8 The second area o f recommendations relates to easing and improving the administrative and budgetary procedures for expenditure execution, with a focus on: 0 confirmation o f the elimination o f technical and financial execution forms for the budget lines listed individually inthe budget law (for the most part consisting o f projects under BCI 2002), in implementation o f the new WAEMU budget nomenclature; 170 0 accelerating the procedure for the notification of allocations under the BCI. It i s proposed that notification materials be prepared in advance, that is, once the budget has been passed by the National Assembly, and sent directly to the technical ministries concerned, with copies to project leaders; 0 the introduction o f a special procedure for Decrees on carrying appropriations forward; conceivably, one could envisage the DDIpreparation and the signature o f a first Decree to carry forward by January 31 at the latest, covering all commitments for which no payment orders have been issued as o f that date. Appropriations for which there are no commitments, after arbitrage by the Minister, would be covered by a second and final Decree on carrying appropriations forward, which will have to be finalized by March 31 o f each year; 0 the establishment o f processing deadlines so as to eliminate the slowdowns observed with respect to expenditure execution procedures and the government contracting process; 0 the practice of stopping payments and making advance payments inthe absence o f the physical file. The installation o f new expenditure software affords an opportunity to reduce processing lags, inparticular as regards stopping payments and makingVAT advance payments; 0 the early preparation o f bidding materials once the decree on allocating appropriations has been approved, so as to allow for a provisional award taking into account the lengthening o f the submission deadlines, which increase from 10 days to 15 days (urgent procedure) or from 20 days to 30 days (normal procedure); 0 rigorous implementation o f the criteria for selecting co-contracting enterprises for administration, so as to eliminate those that present few guarantees for the performance o f contracts. To this end, the drafters o f the new general guidelines on technical and financial rules now being prepared should devote particular attention to the question o franking andpre-qualification o f enterprises; 0 Preparation o f the manual of procedures for investment expenditure execution, and especially for enteringinto government contracts inaccordance with the new Contracting Code. These documents would be valuable both for officials involved in the management o f public spending and for the personnel o f co- contracting enterprises inadministration. 10.9 The third area for proposals involves improvement in the organization and operation of public investment projects. Inthisarea, the paths toward findingsolutionsinvolve: 0 the preparation, following studies, o f a harmonized legal framework for public investment projects, clarifying the status o f personnel and their compensation modalities, rules on organization and operation, reporting requirements, internal and external controls, etc.; 0 systematization o f (in particular strict compliance with) recourse to the procedure for calls to nominate candidates for project director positions; 0 strengtheningo fthe qualifications o fproject leaders, who must systematically be familiarized with modernproject management tools; 0 and the implementation o f an annual training plan for project employees responsible for the procurement o f goods and services on the procedures for national expenditure and the procedures o f the main donors, in particular as 171 regards contracting and disbursements (World Bank, AfDB, European Union, etc.). 10.10 Finally, as regards the information system, progress and fine-tuning should involve: 0 completion o f the draft accounting and financial framework for projects, which also constitutes a work area in which it is essential to invest, given the spilloverd effects it could have on the efficiency o f project management; 0 Effective implementation of the I-AIDA and D-AIDA sofONare modules, which couldbegin operations once the work o f validation and loading o f historical data has been completed. Later, the Ministry should study and then meet the conditions for implementingthe other modules completing AIDA, in particular E-AIDA(DCEF), R-AIDA and P-AIDA (DP); 0 Modernization o f the information system for public expenditure as a whole (wages, current budget, and investment budget) and government revenue, from the "forecasting" phase to the "settlement" phase, so as to be in a position to efficiently manage the execution o f poverty reduction programs. This will assume an organization built around vast databases and the use o f a network linking the various agencies and the various central and regional sites, making use o f the opportunities affordedby new information technologies. 172 A. INTRODUCTION 10.11 For years, public investment execution has deviated considerably from programmed levels. A substantial share o f the extemal and domestic resources available for financing public investment cannot be consumed, as reflected in extremely low public investment and overall investment ratios. This de facto situation has negative repercussions on Senegal's economic growth and socioeconomic development. This situation i s worrisome to the public authorities as they are confronted by a rising tide of impatience from the people who want their pressingneeds for economic andsocial infrastructureto be satisfied. 10.12 However, Senegal i s one o f the African countries that had very early access to a national planning system for purposes o f designing and managing public investment globally and in detail. This framework provides a means o f ensuring the preparation, programming, monitoring, and appraisal o f public investment programs. In contrast, the financial execution o f projects relies upon the public finance management system, itself founded on a legal framework made up o f Law No. 2001-09 o f October 15, 2001, taken in the context o f the harmonization o f fiscal legislation and procedures in the Member States o f WAEMU,jOthe aim of which is the overall regulation of the procedures and modalities for approving, executing, and controlling public revenue and expenditure. 10.13 These arrangements precisely define the procedures, conditions, and executing agents for the current budget and investment budget inparticular, as well as the circuit followed by each expenditure, including investment expenditure, regardless o f whether it is financed domestically or with extemal resources. 10.14 Despite this administrative and technical arrangement, which by all appearances i s complete, progress in respect o f the physical and financial management and execution o f projects has been particularly slow if one takes into account the serious dysfunctions that continue to affect the project life cycle, regardless o f the sector concerned or the origin of the resources used. Thus, over the five-year period 1996-2000, the average execution rate was only a bit over 68 ~ercent,~' with a peak at 77 percent in 1999 and a low of 60 percent in 2000.52In some sensitive sectors, such as Health and Education, project execution results are frankly despairing given the poverty reduction objectives. O f the 250 health posts planned over the term o f the PDEF (1998-2002), only 50 will be completed, or 20 percent of the targeted level! Work has yet to begin at the construction site for Ministry headquarters, eight months from the closing date o f the World Bank credit, and the three regional hospitals financed against the budget reconciliation laws (LFRs) o f 1997 and 1999 have yet to be formally accepted, whereas they were expected to be operational in 1999 and 2000, respectively. For the construction o f classrooms, the rate o f execution in the past for years Replacing Law No. 75-64 of June 28, 1975 handing down the Organic Law on Finance and its implementing decrees, in particular Decree 66-458 of June 30, 1966 establishing government accountingregulations. Or project executioninthe amount of CFAF 1,057.6 billion as comparedto appropriations opened inthe amount ofCFAF 1,564.4 billionover the period. 52 1999 was a year of exceptionalachievementsbecauseof the mobilization of the political authorities on the eve of the 2000 elections, while performance suffered in 2000 from the political and institutional changes takingplace inMarch2000. 173 (1998-2001) i s 67 percent (4,316 classrooms completed, as against 6,433 planned), but domestically funded investments were executed at a rate o f 90 percent while projects with partners peaked at 40 percent! Generally speaking, the greatest constraints appear to relate to the mobilization o f extemal resources, where annual appropriations are rarely half spent, while national budget ceilings are often consumed at a rate exceeding 90 percent. 10.15 This situation would benefit from rapid analysis and correction owing to the demands arising from the forthcoming implementation o f the new poverty reduction strategy and the introduction o f medium-term expenditure frameworks (CDMTs) or budgeting by objectives. Indeed, the government o f Senegal, as a beneficiary o f the enhanced Heavily IndebtedPoor Countries Initiative (HIPC Initiative), has begun the process o f drafting and implementing its Poverty Reduction Strategy Paper (PRSP). Based on the medium- and long-term orientations of the Xth Economic and Social Development Plan, the poverty reduction strategy adopted in December 2001 serves as the basis for future economic reform programs. The identification, formulation, evaluation, selection, execution, and monitoring o f public investment projects aimed at achieving the objectives set forth in the Orientation Plan and the PRSP constitute necessary prerequisites to the efficiency o f investments and enhancement o f resource absorption capacities. 10.16 It i s inthis context that it i s necessary to situate the work o f the group established by the Minister o f Economy and Finance, whose conclusions are reflected in this report. The purpose o f the report i s to propose to the authorities a plan o f actions able to lift the obstacles observed in project management and to significantly improve performance in the area o f public investment. Beforehand, it would be advisable to provide a diagnosis o f the various constraints that bear on the project life cycle, in light o f the analyses initiated by the various unitsinvolvedinthe management ofpublic investments. A.l STATUSOFTHECAPACITIES, TOOLS, AND METHODSOFTHE MEFAND MTS 10.17 A number o f factors explain the weaknesses o f the parties involved inthe management of public investments. These deficiencies may be identified in the various stages o f budget execution, inparticular inconnectionwith its monitoringo fpublic expenditure. 10.18 The authorities' decision to adopt the PRSP and the medium-term expenditure frameworks (CDMTs) or budgetingby objectives (see Box 1) in some key social sectors will surely influence the new method for programminghudgeting expenditure, as well as program and project monitoring. Monitoring 10.19 Program monitoring i s carried out within the overall framework o f monitoring instrumentsintroducedby the SNP reform. The network supporting investment monitoringis tending to disappear owing to the weakened human resources o f the units concerned in the technical ministries and the DCEF. This monitoring function i s carriedout at two levels: - Day-to-day monitoring o f the projects and programs under way; and - Targeted monitoring when preparing the report on investment execution inyear n-1. 174 Day-to-day monitoring 10.20 Day-to-day monitoring is organized on an ongoing basis by the sectoral units o f the DCEF during various meetingswith the projects and contacts with their respective managers. The quarterly progress reports produced by some projects are also used for monitoring purposes, as are the statements generated by the financial information system managed by the DDI, which report on the payment orders issued and disbursements against the financing allocatedto the various projects. 10.21 Simultaneously, the project supervision missionsthat donors regularly send to Senegal provide an opportunity to organize working meetingswith the various stakeholders inprojects and to obtain information about them. Indeed, these missions are finalized by the signing of an aide-mkmoire which reports on the progress made and the problems encountered. Corrective measures are generally proposed, and the various sectoral units are called upon to ensure their implementation incooperationwith the offices concerned. 175 Targeted monitoring 10.22 Targeted monitoring takes place over a period o f about three months, prior to launching the PTIP n+l,with a view to the drafting of the annual report on public investment for year n-1. For the past 5 years, the DCEF has engaged the services o f at least three local accounting firms to gather information on certain projects which could not be obtained completely during the preceding fiscal year. Since most o f the projects are located in the regions, the use o f accounting firms allows for greater flexibility in travel to project sites to verify the physical and financial performance o f those projects. Generally, the projects whose monitoring i s turned over to the accounting firms represent 80 percent to 85 percent o f the financing under the PTIP envelope and a third of the projects listed. 10.23 Recently, inresponse to the adoption of the budget reconciliation law for 1999 and the approaching elections, the public authorities decided to replace the CRI, and a replacement has beennamed. Execution 10.24 Controlling the execution o f public investment projects and programs by the MTs has always been a problem in Senegal owing to the limitations imposed by administrative, technical, and financial procedures, as well as by the level o f expertise and reduced motivation o f the human resources involved in management. To rectify this situation, one solution tried by the government and donors was the creation o f executing agencies alongside some of the technical ministries since the late 1980s. 10.25 These advantageously took the place o f the government's technical officers as regards ensuring the execution o f public investment programs, in particular by replacing the central government's lengthy and complicated financial execution procedures by more flexible and more effective internal procedure^.^^ At the same time, they made it possible to resolve the problems o f some MTs' capacities in the area of project preparation and programming, as well as questions ofpersonnelmotivation. 10.26 It must be noted, however, that these agencies are continuing to evolve inthe absence of a cohesive overall legal framework that would have made it possible to avoid some slippages observed. Also noteworthy are the more and more troubling delays at the sites entrusted to the Public Works and Employment Agency (AGETIP), which appear to be the outgrowth o f the agency's excessively broad portfolio, a situation favorized by the agency's virtual monopoly over this works execution market. 10.27 Apart from the capacities o f SNP participants, the obstacles to proper public investment execution mostly pertain to the procedures for mobilizing resources and executing expenditure. 53 As inthe case of the Public Works andEmploymentAgency (AGETIP), the Municipal DevelopmentAgency, the Autonomous RoadWorks Agency, etc. 176 A.2 ADMINISTRATIVEANDBUDGETARYCOMPLEXITYAND UNWIELDINESSFOR EXPENDITURE EXECUTION 10.28 As regards the administrative and budgetary procedures for the execution o f expenditure financed by domestic resources, the delays observed are principally the result O f 0 the form known as the "technical and financial execution form," the completion o f which was a prerequisite for any resource mobilization. With the harmonizationo f the budget format and the adoption o f a common nomenclature for all WAEMU countries, this constraint o f completing technical and financial execution forms has outlived its usefulness for projects inasmuch as the presentation o f investmentexpenditure by nature takes all information needs into account; 0 the long times requiredto close out the fiscal year, inparticular the problems o f carrying investment appropriations forward when they were committed by no payment orders had been issued, or when they had not been committed.. Owing to the constraints imposedby the financial framework and the problems involved indetermining the precise financial position ofprojects, signature of the decree on carrying appropriations forward generally occurs after a considerable delay and poses risks to the continuation o f work between the end o f the year and the first quarter o f the following fiscal year; 0 the constraints affecting the awarding of government contracts, in particular the absence of deadlines conveyed to the various and multiple parties concerned in the operations of calling for bids, opening submissions, and awarding the successful bidder in public bidding procedures. While the government procurement procedures have been reformed and eased inthe context o f the new draft decree setting forth the Contracting Code,54 there are grounds for apprehension that such problems will persist if the structures and parties involvedare not requiredto performtheir functions byreasonable deadlines; 0 payment delays, themselves also attributable to the lack o f deadlines by which concerned staff mustprocess materials, as well as to cash flow problems; for this reason, the recipient o f the support i s quite often forced to halt work. This ' situation leads to additional delays ininvestment program execution; e the selection o f enterprises, particularly SMEdSMIs, that have insufficient mastery o f the administrative procedures and the management o f government contracts, and in particular lack the necessary financial resources or access to bank credit needed to prefinance the works entrusted to them. These SMEs/SMIs are involvedmore and more inthe priority sectors (construction and refurbishing o f roads, classrooms, health centers, wells, etc.). 10.29 In the case o f expenditure against external resources, one o f the obstacles faced in the early 1990s, which related to the government payment o f counterpart funds, has virtually disappeared. This said, delays inmobilizing funding continue to have a significant impact on ~~ j4 Inthisnewcode, submissiondeadlineshavebeenlengthened,butthe government contracting thresholdhasbeenraised(CFAF 15million for supplies and CFAF 25 million for construction contracts). Itis also plannedto establish fees for ministries to examinebids ministry by ministry, insteadof usingthe single nationalcommissioninplacepreviously. 177 the execution o f projects financed by foreign donors (execution rates are far lower than for domestically financed projects). The difficulty arises, inter alia, from the multiplicity and complexity o f disbursement procedures, as each donor imposes its own methods and procedures which differ from one organization to the next and from the procedures used by the government. 10.30 Taking into account the constant turnover at these levels and despite the training efforts made by some donors, delays in mobilizing external financing continue to handicap the proper conduct of projects, which often suffer as well from the effects o f their poor organization. A.3 DEFICIENCIESTHE ORGANIZATION ANDOPERATIONOFPUBLICINVESTMENT IN PROJECTS 10.31 Most o f the problems observed in the operation of projects arise from the conditions under which they were established, their organization, and their internal procedures. 10.32 On the institutional level, despite their long-standing role in the administrative mechanism and their importance inthe execution of the MTs' missions, projects do not have any harmonized status. They are subcomponents o f the administration, without their own legal personality, and their legal status poses serious problems inrespect o f aspects relating to personnel (modalities for recruitment, compensation, motivation, advancement, and contract termination). From the time this kindo f structure i s set up untilit ceases to function, there are many situations and operations in which the MTs and project managers are hesitant in the conduct of activities owing to the legal uncertainties surrounding projects. Each project, evolving in an unregulated framework, ultimately adopts a special status that i s prejudicial to supervisors controlling and guiding activities. 10.33 Moreover, such a context fails to promote an objective choice o f project personnel, and the selection process does not always adhere to criteria o f efficiency and transparency. Evenin cases where a competitive procedure is used for hiring senior management, it often occurs that individuals who did not even participate in the selection examinations are ultimately appointed. As a consequence, project personnel are not always qualified to successfully carry out the activities and tasks for which they were recruited (goods and services may be procured at terms that are disadvantageous to the State as regards duration and cost, the project timetable canbe affectedby additional delays, etc.). 10.34 More serious still is the fact that ex-post controls (such as annual audits) o f project operations and performance, inparticular when they are financed against domestic resources, i s virtually non-existent owing to the lack o f obligations in this regard. There i s no individual accountability, apart from that customarily applicable to all government employees, affecting project leaders, who do not regard themselves as subject to any reporting obligation even though their activities nevertheless have a specific nature. 10.35 Finally, one o f the major obstacles to public investment projects and programs is the imperfect control by project leaders over the planning o f activities and procedures for awarding contracts. This situation is compounded by the fact that projects rarely have (or comply with) codified technical and financial procedures for their day-to-day operations. This i s particularly true for accounting and financial rules (see theparagraph on the information 178 system below) which often intermix public accounting and private accounting owing to the sources o f their resources. 10.36 All these impediments attributable to administrative and financial procedures and flaws in project organization and structures pose an even greater burden in an environment markedby the structuralweakness o f the information system, which i s not organized ina way that allows for adequate project supervisionandmonitoring. A.4 LIMITSTOTHEINFORMATIONSYSTEM 10.37 Information i s o f crucial importance in the process o f financial steering and monitoring o f public investment projects. The information and financial management system must provide the public authorities with a means of ensuring the ongoing and ex-post control of investment expenditure, the effective and efficient use o f the resources made available to projects, and the analysis o f expenditure with a view to developing and carrying out future programs and projects. A proper information system integrates all these aspects by facilitating information sharing within a system of databases residing on a computer network that allows for access in real time by all users concerned, but also by the authorities responsible for task supervision or steering. 10.38 Public investments in Senegal are financed through the national budget and by sizable contributions from donors, and consequently, the existing financial information system on project execution is based principally upon the information provided by the government's expenditure chain (managed by the DTAI), the accounting system o f the Disbursement Authorization Officer (DDI), the accounting operations recorded at the Public Treasury, and the data reportedby donors. 10.39 To be sure, the DTAI sequence tracks operations relating to the opening o f appropriations and the commitment and issuance o f payment orders for expenditure committedagainst national appropriations. Inpractice, however, there areanumber ofconcerns raisedbythischain: 0 the difficulty o f adaptingthe selective carrying forward o f files at the end o f the year, thus makingthe status o f some balances unpredictable; 0 the unavailability o f information on the budget situation on a real time basis; 0 processinglags that are often quite lengthy, largely owing to the nature o f the administrative procedure; 0 no identification o f the economic nature o f the expenditure committed by projects; 0 no indication o f the effective settlement o f expenditure by the Treasury. 10.40 In the case o f development projects financed in whole or in part by donors, the DDI's information system is based, until such time as the I-AIDA software becomes operational, on the accounting system o f the Disbursement Authorization Officer, who uses MS Excel to produce all the statements on drawings against development projects. The statements thus produced relate to all public projects currently being executed, that is, those projects for which the DDIissues payment orders as well as the smaller number o f projects that are monitored directly by donors, such as the subsidies granted in the context o f the 179 technical cooperation provided mostly by the GTZ, the priority Solidarity Fund(formerly the FAC), USAID, the UNDP, CIDA, and JICA. 10.41 At present, these statements canbe made available not more than 15 days after the end o f the period, generally quarterly, or on demand. They can also be prepared "by project,'' "by donor," "by sector," or as "summary statements." 10.42 However, as may be observed, the information system remains fragmented into multiple, autonomous systems that are each managed by different units (DTAI, DDI, Treasury). Some are still semimanual (DDI) and may pose the problem o f exhaustiveness and the rapid and regular availability of the data recorded, while others are semiautomatic (DTAI and Treasury). Ina more fundamental sense, however, these are subsystems that do not share information among themselves. 10.43 The public investment expenditure sequence has been incorporated into the introduction o f the WAEMU's new harmonized budget nomenclature. The new public expenditure management software envisages the use of a database and a network including all parties to budgetary operations and proposing considerably more user-friendly technical functionalities. 10.44 While it does present some improvements (in particular the real-time availability o f information), this new system nonetheless continues to incorporate some o f the inadequacies o f the system it i s called upon to replace. 10.45 For investments, for example, only the execution o f operations charged against domestic resources will be tracked, and the payment operations carried out through the Treasury will continue to be excluded from the database. Computerization will thus continue to be confined to operations carried out at the MEF, whereas it could have been introduced to the spendingministries. 10.46 Another solution initiated as part o f the PAI-BAD project (1991-2000) i s the AIDA (Automation of Investment, Debt, and Aid) information system. This system was designed to encompass the entire process of appraisal, financial execution, and physical and financial monitoring o f investments, and management o f aid flows and debt. Many technical, financial, andhuman obstacles have prevented completion o f the process o f introducing AIDA. 10.47 Along these lines, the DDIi s endeavoring to relaunch the introduction o f the project accountingand financial framework, an essential prerequisite for ensuringthe cohesive and comprehensive monitoring o f investment expenditure. This framework would have the effect o f harmonizing and standardizing project accounting systems and providing the means for more effective control o f the financial operations carried out by projects (ongoing control and control o f closing statements). This i s an interesting prospect deserving o f close attention as regards improving the effectiveness o f public investment. For this reason, the DDI's efforts shouldbe accommodated with a view to the rapid implementation o f this valuable tool. 10.48 The foregoing diagnosis encompasses all phases of the project life cycle, from the information system through the technical, administrative, and budgetary procedures, to investment structures and management services. It illuminates the range o f constraints to be lifted for there to be any cohesive improvement in the resource absorption capacities o f development projects. The rapid implementation o f the recommendations set forth in the following chapter constitutes a priority inasmuch as the successful achievement o f the 180 objectives o f the PRSP, in particular those relating to increasing production and the satisfaction of social demands, requires a significant improvement in the efficiency o f investments. B. OUTLINE OF POSSIBLESOLUTIONS 10.49 In keeping with the findings o f the foregoing diagnosis, the recommendations proposed have been grouped into the four following headings and themes: 1) strengthening the capacities, tools, and methods o f the MEFand various MTs for developing, executing, andmonitoring public investmentprograms and projects; 2) simplifyingthe administrative and budgetaryprocedures for expenditure execution; 3) strengthening the project management mechanism; 4) and finally, the introduction of a high-performance information system on public investment projects and expenditure. B.l STRENGTHENING THE CAPACITIES AND IMPROVINGTHE TOOLSAND METHODS OF THE MEFAND MTS 10.50 To strengthen the capacities o f the MTs and the MEF for managing public investment projects and programs, the proposed solutions are functional, institutional, and methodologicalinnature, and concern: a) following updating, continued implementation o f the actionplanresulting from various recommendations, including those formulated by the latest Superior Council on the Plan and updated in the context o f the diagnostic studies o f the capacities o f the technical ministries (DPILTNDP, 1990; DCEF/GTZ, 1999); b) the introduction o f a new ad-hoc structure responsible for examining the bottlenecks to the proper execution o f investments and identifying solutions that will improve the efficiency o f public investment. This structure will be patterned after the Ad Hoc Monitoring Committee for the 1999 Budget Reconciliation Law referred to above. It will be chaired by the Prime Minister with secretariat services provided by the Ministry o f Economy and Finance. The MEF will thereby have an appropriate channel for guaranteeing the physical and financial monitoring o f the public investment programs carried out by some ministries. It would concentrate its efforts on major programs and should, if possible, provide the periodic coordination required to address the identified obstacles to project execution by the prescribed deadlines. This secretariat should have broadened jurisdiction as well as adequate financial resources from the national budget. The mode o f operation as well as the composition o f this ad hoc structure should be delineatedby decree; c) the gradual extension o f the use o f medium-term expenditure frameworks (CDMTs) inkeeping with the budgets by objectives previously established in the education and health sectors. With regard to the methodology for preparing CDMTs (see Box I),this extension should make it possible to improve the definition o f sectoral activities and the preparation o f sectoral budgets. 181 B.2 EASINGADMINISTRATIVEANDBUDGETARYPROCEDURESFOREXPENDITURE EXECUTION The recommendationsrelate to: a) the elimination of technical and financial execution forms for the budget lines listed individually inthe budget law (for the most part consisting o f projects under BCI2002), inimplementation ofthe new WAEMUbudgetnomenclature; b) accelerating the procedure for notification of allocations under the BCI. It is proposed that notification materials be prepared in advance, that is, once the budget has been passedby the National Assembly, and sent directly to the project leaders, with copies to the technical ministries concerned; c) the introduction of a special procedure for Decrees on carrying appropriations forward; conceivably, one could envisage the DDIpreparation and the signature o f a first Decree to carry forward by January 31 at the latest, covering all commitments for which no payment orders have been issued as o f that date. Only those appropriations for which there are no commitments would then be submitted for arbitrage by the Minister on the basis o f the financial constraints set forth inthe Government Financial Operations Table (TOFE) for our international obligations (such as the poverty reduction strategy) and analysis o f the documentation to be provided by the MTs. These lines would be covered by a second and final Decree on carrying appropriations forward, which will have to be finalized by March 31 o f each year, with mandatory performanceby the DDIand ultimate penalization o f any MTs that failed to produce the background materials necessary for taking action on their files; d) the preparation of bidding materials once the decree on allocating appropriations has been approved, so as to allow for a provisional award taking into account the lengthening o f the submission deadlines, which increase from 10 days to 15 days (urgent procedure) or from 20 days to 30 days (normal procedure); e) the rigorous implementation o f the criteria for selecting cocontracting enterprises for administration, so as to eliminate those that present few guarantees for the performance o f contracts. To this end, the drafters of the new general guidelines on technical and financial rules now being prepared should devote particular attention to the question o franking and prequalification o f enterprises. If necessary, it may prove worthwhile to examine the possibility o f subcontracting the ranking to private entities (such as VERITAS) which would be held liable in the event o f established errors; f, the practice of stopping payments and making advance payments in the absence o f the physical file. The installation o f the new expenditure sofhvare affords an opportunity to reduce processing lags, inparticular as regards stopping payments and making VAT advance payments. Indeed, the on-line availability of all the information relating to the expenditure file (recipient, amount, file references, nature, etc.) should enable the units concerned (Office o f the General Paymaster at the Treasury and Tax Department) to carry out the payment stoppage and advance payment operations usinga form edited on-line each day, on the basis of simple consultation of the proposed payment orders issued by the Controller o f Financial Operations (COF), without havingto request the physicalpayment files; 182 g) the establishment o f processing deadlines so as to eliminate the slowdowns observed with respect to expenditure execution procedures and the government contracting process. With the introduction of the COF and the new expenditure software, the processing lags for the files should, in principle, be reduced considerably. Indeed, the objective for this reform o f the expenditure procedure i s for a file to remain at any given control point for no more than 48 hours; h) Preparation of a manual of procedures for investment expenditure execution, and especially for entering into government contracts in accordance with the new Contracting Code. These documents would be valuable both for officials involved in the management of public expenditure and for the personnel of cocontracting enterprises inadministration. B.3 IMPROVEMENTIN THE ORGANIZATION AND MANAGEMENT OF PUBLICINVESTMENT PROJECTS As regards day-to-day project management, the paths toward finding solutions involve: a) the preparation, following studies, o f a harmonized legal framework for public investment projects, clarifying the status o f personnel and officials and their compensation modalities, rules on organization and operation, reporting requirements, internal and external controls, etc.; b) systematization o f (in particular strict compliance with) recourse to the procedure for calls to nominate candidates for project director positions; c) strengthening o f the qualifications o f project leaders, who must systematically be familiarized with modern project management tools such as MS Project for improvedplanning and monitoringo f day-to-day activities; d) Implementation o f an annual training plan for project employees responsible for the procurement o f goods and services on the procedures for national expenditure and the procedures o f the main donors, in particular as regards contracting and disbursements (World Bank, AfDB, EuropeanUnion, etc.). B.4 ORIENTATIONSREGARDINGTHE INFORMATIONSYSTEM 10.51 The government has a budgetary framework for public investment that i s sufficiently relevant to support an effective information system, even if it i s necessary to review the application of some provisions, inparticular those applicable to the production o f reports by project leaders. Apart from this problem, efforts shouldbe focused on implementing a number o f specific short- and medium-term measures with a view to overcoming existing shortcomings. 10.52 Inthe short term, efforts should be focused basically on the resources to be mobilized in order to (i) the systems that have been developed and have already consumed make substantial resources operational, and (ii) complete the other remainingmodules. 10.53 The I-AIDA and D-AIDA software could effectively begin operations once the work o f validation and loading o f historical data has been completed. This obviously assumes that the necessary resources are mobilized, in particular for the purchase o f additional 183 workstations, for consultancy fees (public debt), and for the recruitment or assignment o f additional staff and the training o f the employees concerned. 10.54 Later, the Ministry should study and then meet the conditions for implementing the other modules completing AIDA, inparticular E-AIDA (DCEF), R-AIDA and P-AIDA (DP). Getting the AIDA Project moving again would appear to be o f the greatest possible importance for ensuring a sound basis for the information system on public investments in general. 10.55 The accounting and financial framework for projects i s another work area where investment i s essential given the induced effects it could have on the efficiency o f project management. The DDI will intensify the steps already taken vis-a-vis donors with a view to securing financing for continuation o f the project. Ifneedbe, the government or other donors will have to be called upon to contribute to covering residual financing requirements. Be that as it may, it will first be necessaryto update this accounting framework, inparticular to bring it into line with the newbudgetary and accounting systemrecently introducedby WAEMU. 10.56 In the medium and long terms, the information system for public expenditure as a whole (wages, current budget, and investment budget) and government revenue, from the "forecasting" phase to the "settlement" phase, will have to be modemized so as to be in a position to efficiently manage the execution o f poverty reduction programs. This will assume an organization built around vast databases and the use o f a network linking the various agencies and the various central and regional sites, making use o f the opportunities afforded by new information technologies. With this inview, a prior analysis of the existing situation should be conducted with a view to identifyingthe strengths and weaknesses of the existing information system. At the same time, experts/consultants should study the changes to be made inour existing system (at the institutional, technical, human, material, and other levels) to make it comparable with those used inthe present day by the major enterprises and banks in Senegal. This ambitious program will require the intervention of donors, particularly the World Bank, which has appropriate resources for supporting the modernization o f public expenditure management systems. 184 REFERENCES Adam, C.S, D.L. Bevan and G.Chambas. 2001. `Exchange Rate Regimes and Revenue Performance in Sub-SaharanAfrica' Journal of Development Economics, vol64 173- 213. Adenikinju, A., L.Soderling, C.Soludo, and A.Varoudakis.2002.'Manufacturing Competitiveness in Africa: Evidence from Cameroon, Cote d'Ivoire, Nigeria and Senegal', mimeo. Badiane,O.1993.'Extemal Trade Pessimism and the Role o f Regional Markets in the GroundnutTrade' Journal ofAfrican Economies, vol. 22, no1. Balassa, B.1989.'Comparative Advantage, Trade Policy and Economic Development' London, Harvester Wheatsheaf. Becker, G.1975.'Human Capital: A Theoretical and Empirical Analysis, with Special Referenceto Education' Universityo f Chicago Press. Benhabib,J. andM.Spiegel.1994.'The Role ofHumanCapital inEconomic Development: Evidence from Cross NationalAggregate Data' Journal ofMonetary Economics, vol. 34, 143-173. Cohen, D. andM.Soto.2001.'Growth andHumanCapital: GoodData, GoodResults' CEPR DiscussionPaperno 3025. Devarajan, S. and D.Rodrik.l992.'Do the Benefits o fFixedExchange Rates Outweigh Their Costs?The CFA Zone inAfrica' inIan Goldinand L.Alan Winters, eds., Open Economies: StructuralAdjusment and Agriculture. Cambridge; New York and Melbourne: Cambridge University Press, PP 66-85. Diagne, Abdoulaye andDaffk, Gaye. 2002. `Le Sbnkgale en Quete d'une Croissance Durable' CREA et kditions KARTHALA: Paris, France. Filmer,D.,Fever and its treatment among thepoor and lesspoor in sub-Saharan Africa. 2002, World Bank: Washington DC. Filmer, D.andL.Pritchett, Estimating WealthEffects without Expenditure Data -- or Tears: An Application to Educational Enrollments in States of India. 1998, World Bank: Washington DC. Filmer,D.andL.Pritchett, Theeffect of household wealth on educational attainment: evidencefrom 35 countries. Population andDevelopment Review, 1999. 25(1): p. 85- 120. Freeman,R. 1993."Labor Market Institutions andPolicies: Helpor Hindrance to Economic Development?" Proceedings of the WorldBank Annual Conference On Development Economics, 1992.World Bank: 117-44. Gelband, H.andS. Stansfield. 2001. "The Evidence Basefor Interventions to Reduce Under Five Mortality in Low and Middle-Income Countries". CMH Working;Paper Series: PaperNo. WG5:9. page 15. 185 Golub, S. and A. Mbaye. 2000. "Obstacles and Opportunities for Senegal's InternationalCompetitiveness: Case Studies ofthe Peanut Oil, Fishing and Textile Industries." Swarthmore College andUniversite Cheikh Anta Diop de Dakar. Processed. Groupe Serdha. 1997.Etude Retrospectiveduns le Secteur Population et Sante` : Composante Quantitative.- Dakar Gulhati, R. andR.Nallari.1990.'Successful Stabilization andRecovery inMauritius' ED1 Development Policy Case Series, World Bank. Golub,S. andA.A. Mbaye. 2002 "Obstacles andOpportunities for Senegal's International Competitiveness:Case Studies ofthe Peanut Oil, Fishing andTextile Industries" Africa Region Working Paper Series number 37. Gwatkin, D., S. Rutstein,K.Johnson, R.Pande, andA. Wagstaff, Socioeconomic Differences inHealth, Nutrition andPopulation. 2000, The World Bank: Washington DC. Hill,K.,R.Pande, M.Mahy,andG.Jones, Trendsin ChildMortality in theDeveloping World: 1960to 1996. 1999, New York: Unicef. Hinkle, L.and A. Herrou-Aragon. 2002. `How Far DidAfrica's First Generation Trade Reforme Go? An Intermediate Methodology for Comparative Analysis of Trade Policy' World Bank. International Monetary Fund.2001. "Senegal: Selected Issues." IMFCountry ReportNo. 011188. .2002. "Senegal: Second Review Underthe Third Annual Arrangement Underthe PovertyReduction andGrowth Facility ...,' Processed. Jalan, J. andM.Ravallion, Does Piped Water ReduceDiarrheafor Children in Rural India? 2001, World Bank: Washington DC. Kakwani, N.C., A. Wagstaff, andE. VanDoorslaer, Socioeconomic inequalities in health: Measurement, computation and statistical inference. Journal of Econometrics, 1997.77(1): p. 87-104. Jones, P.2001.'Are EducatedWorkers Really More Productive?' Journal of Development Economics, vo1.64 57-791. Krueger, A. andM.Lindah1.2000.'Educationfor Growth: Why and for Whom' NBER WorkingPaperno 7591. Latreille,T. andA.Varoudakis.1996. Croissanceet Compttitivitt de 1'Industrie Manufacturikre au Skntgal' OECD, Document Technique no 118. Lavy, V., J. Strauss, D.Thomas, andP. de Vreyer, Quality of care, survival and health outcomes in Ghana. Journal of HealthEconomics, 1996. 15(333-357). Lindauer, D. 1999. "Labor MarketReforms andthe Poor." A BackgroundPaper for the WorldDevelopmentReport, 2000/2001. World Bank. Processed. 186 Mbaye, A. A. and S. Golub. 2002 "Explaining Manufacturing Growth and Competitiveness inSenegal: A Sectoral Analysis ofRelative Costs andPrices" The World Bank Processed. Mbaye, A.A.. 2003 "An IndustryLevelAnalysis of ManufacturingProductivity in Senegal" Africa Region Working Paper Series, number 41. Ndiaye, Mansour. 2002 `Explaining African Economic GrowthPerformance: The Case of Senegal'. World Bank. Mincer, J. 1974.'Schooling, Experience and Earnings. Columbia UniversityPress. Mulaga, G. and J. Weiss. 1996. `Trade ReformandManufacturing Performancein Malawi 1970-91' World Development, vol24 no7. Mosley, W. and L.Chen, An analyticalframework for the study of child survival in developing countries. Population andDevelopment Review, 1984. 10: p. 25-45. Noguet, M.2001. `Stabilization Orthodoxe ouHetedoxesdans les Economies en Transition : les Exemples de la Hongrie et de laPologne' Revue d'Economie Politique, vol 111,no5. Oaxaca, R., Male-female wage dEfferentials in urban labor markets. International Economic Review, 1973.14: p. 693-709. O'Connell, S. andB.Ndulu.2000.'Afiica's Growth Experience. A Focus on the Sources of Growth'AERC GrowthProject. Ozatay, F.2000.'The 1994Currency Crisis inTurkey' Policy Reform, vo1.3 327-352. Pison, G. et als (eds). 1995.Population Dynamics of Senegal. NationalAcademy of Science. Washington DC. Pritchett, L.1996.'Where Has all the Education Gone?' World Bank Policy Research Working Paperno 1581. Rama, M. 1998. "Wage Misalignment inCFA Countries: Are Labor Market Policies to Blame?" PolicyResearchWorking Paper, 1873. World Bank. Processed. Ravallion, M.and S. Chen. 1997. "What CanNew Survey Data Tell U s about Recent Changes inDistribution andPoverty?" WorldBank Economic Review 1l(2): 357-82. Sacerdoti, E., S.Brunschwig andJ.Tang.l998.'The Impact ofHumanCapital on Growth: Evidence from West Africa' IMFWorking Paper WPl981162. Teal, F. 1996. "The Size and Source of Economic Rents inaDeveloping Country Manufacturing Labor Market," Economic Journal 106(437): 963-76. Terrell, K.and J. Svejnar. 1989.The Industrial Labor Market & Economic Performance InSenegal. Boulder: Westview Press. UNICEF, Progresssince the World Summit for Children: A Statistical Review. 2001, UNICEF: New York. Velenchik, A. 1995. "Apprenticeship Contracts, SmallEnterprises, and Credit Markets in Ghana," WorldBank Economic Review 9(3): 451-745. 187 . 1997. "Govemment Intervention, EfficiencyWages, andthe Employer SizeWage Premium inZimbabwe," Journal of Development Economics 53(2): 305-38. Wagstaff, A., P. Paci, and E.van Doorslaer, On the measurement of inequalities in health. Social Scienceand Medicine, 1991.33: p. 545-557. Wagstaff A andFengZhao, (forthcoming). In Pursuit of the ChildHealth Millennium Development Goals: The Case of Senegal, The World Bank. Wagstaff, A. andN.Nguyen, Poverty and survivalprospects of Vietnamese children under Doi Moi, in Vietnam. 2001, World Bank: Washington DC. Wagstaff, A., Socioeconomic inequalities in child mortality: comparisons across nine developing countries. Bulletin of the World HealthOrganization, 2000. 78(1): p. 19- 29. World Bank, 2002. `Senegal. Integrated Framework Diagnostic Trade Integration Study' Draft Report. .2002. African DevelopmentIndicators. Washington, D.C.: World Bank. . 1995. "Senegal: An Assessment ofLiving Conditions, Volumes Iand11.Report No. 12517-SE.World Bank. World Economic Forum. 2001. TheAfrica CompetitivenessReport, 2000/2001. New York: Oxford. WB159149 P:\SENEGALWACRO\CEM\Senegal CEM - DraftSFinaLdoc November 12,2003 10:09AM