Report No. 25185-LCR Saint Kitts and Nevis OECS Fiscal Issues Policies to Achieve Fiscal Sustainability and Improve Efficiency and Equity of Public Expenditures December 29, 2003 Caribbean Country Management Unit Poverty Reduction and Economic Management Unit Latin America and the Caribbean Region Document of the World Bank IOCR InstitutionalandOrganizationalCapacity Review MDGs Millennium DevelopmentGoals MIS Managementinformationsystems MoF Ministryof Finance MoH Ministryof Health MOW Ministryof Communications, Works, PublicUtilities andPosts MTESP Medium-TermEconomic Strategy Paper NHC NationalHousingCorporation NIA Nevis IslandAdministration OECD Organizationfor Economic CooperationandDevelopment OECS Organizationof EasternCaribbeanStates PAC PublicAccounts Committee PRGF Poverty ReductionGrowthFacility PSIP PublicSector InvestmentProgram PSIPoc Public Sector InvestmentProgramOperating Committee PSIPPCC PublicSector InvestmentProgramProjectCooperatingCommittee RGSM RegionalGovernment Securities Market SEDU SmallEnterpriseDevelopmentUnit SELF StudentEducationandLearningFund SIGFIS StandardizedIntegratedGovernmentFinancialInformationSystem SIGTAS StandardizedIntegratedGovernmentTax AdministrationSystem SOEs State-ownedenterprises SSMC St. Kitts Sugar ManufacturingCompany sss SocialSecurity Scheme UDC UrbanDevelopmentCorporation UK UnitedKingdom UNESCO UnitedNationsEducational, Scientific, andCulturalOrganization VAT Value-addedtax WAEMU WesternAfrica Economic andMonetaryUnion WBI WorldBankInstitute Preface This report was completed inJune 2003 and reviews St. Kitts and Nevis public expenditures untilthe year 2002. The proposedpolicy recommendationsreflect the situation of the country at the time of the completion of the report. Acknowledgments This report i s based on the contributions of a team consisting of Lorraine Blank (social protection), Andreas Blom (education), Patrick Grady (public sector investment program and public sector employment and compensation), Ted Paterson (budget management), Helen Saxenian (health), and Claudia Sepdlveda (fiscal sustainability and fiscal and business cycles). As Task Manager, Claudia Sepdlvedawas responsible for the overall preparation of the report. Shahrokh Fardoust and David Rosenblatt were peer reviewers. The team would like to thank Antonella Bassani and Mauricio Carrizosa for their helpful comments and their invaluable collaboration to improve this report. The report also benefited from the accomplished research assistance of Laura dos Reis and Kevin Tomlinson and the excellent assistanceof Anne Pillay. The assistance, cooperation, and input of many Government departments in St. Kitts and Nevis-too numerousto mention individually-is gratefully acknowledged. In particular, extensive assistance was received from the Budget and Planning Unit of the Ministry of Finance. The report also benefited from the assistance andcooperation of the EasternCaribbean Central Bank, the Caribbean Development Bank, and the International Monetary Fund. Without their help, preparation of this report would not have been possible. Table of Contents EXECUTIVE SUMMARY .............................................................................................................................. i 1. FISCALDEFICIT. PUBLICDEBT.AND FISCALSUSTAINABILITY ............................... 1 ............................................................. I1. DOES . I ST.KITTSANDNEVIS . FISCAL PERF~RMANcE.1990-2002 2 ST KIPSAND NEVIS FACEFISCALA SUSTAINABILITY PROBLEM? ..................................... 4 2.I11. FISCALCONSOLIDATION INST KITTS ANDNEVIS'S . PUBLIC SECTOR .......................................... 11 FISCALPOLICY AND BUSINESSCYCLES ......................................................................... 17 TRENDS . FISCALACCOUNTS. 18 I1. I. AND CYCLES INST KIPSAND NEVIS 1980-2001.......................... THE STRUCTURALBUDGET SURPLUS INST.KlTTS AND NEVIS .................................................... 23 3. BUDGETMANAGEMENT ....................................................................................................... 27 I. FRAMEWORK ................................................................................ 29 I1. CONSTITUTIONALAND LEGAL BUDGET PREPARATION. EXECUTION. AND MONITORING .............................................................. 30 37 4.I11. BUDGET ACCOUNTABILITY ........................................................................................................... THE PUBLICSECTORINVESTMENT PROGRAM ............................................................ 41 I ST.KITTSANDNEVIS . FEDERAL GOVERNMENT PSP:PREPARATION. EXECUTION. AND MONITORING .................................................................................................................................. 43 51 5.I1. SIZE AND COMPOSITIONOFTHE NEVIS ISLAND ADMINISTRATION PSIP...................................... PUBLICSECTOREMPLOYMENT AND COMPENSATION ............................................. 53 I. RECENTTRENDSINEMPLOYMENT COMPENSATIONINTHE ST KIPS AND NEVIS AND . FEDERAL GOVERNMENT ................................................................................................................ 54 I1. RECENT TRENDSINPUBLICSECTOREMPLOYMENT COMPENSATIONINTHENEVIS ISLAND ADMINISTRATION............................................................................................................. AND 63 6. THE HEALTHSECTOR ........................................................................................................... 67 . ............................................................................... 68 I1. THEHEALTHSYSTEM: COVERAGEANDDELIVERY I. HEALTHOUTCOMESINST KITTS AND NEVIS ...................................................................... 72 75 7.I1. USESOFHEALTH SPENDING INST KIrlYrSAND NEVIS. . 1995-2002 ............................................... THE EDUCATIONSECTOR .................................................................................................... 79 I. THEEDUCATIONAL SYSTEM INST KITTSAND NEVIS . .................................................................. 81 I1. EDUCATION . ......................................................................... 82 I11. USESOFEDUCATION OUTCOMES INST KITTS AND NEVIS EXPENDITURES .KITTSANDNEVIS.1994-2002 ................................. INST 84 8 . SOCIALPROTECTIONPROGRAMS .................................................................................... 93 I. ST.KITTS ANDNEVIS SOCIAL PROTECTIONSTRATEGYANDPROGRAMS .................................... 95 I1 COMPOSITIONANDEVOLUTION . NEVIS.1996-2001 ....................................................................................................................... OF SOCIALPROTECTIONSPENDINGINST KITTS AND 101 I11 SOCIALPROTECTIONANDRISKMANAGEMENT INST.KITTSANDNEVIS ................................. 102 BIBLIOGRAPHY ..................................................................................................................................... 107 APPENDICES .......................................................................................................................................... 111 Appendix A The Hodrick-PrescottDetrendingProcedure.......................................................... 111 Appendix B StructuralFiscal Balance Methodology ................................................................. 112 Appendix C Social Protection Programs in St.Kitts and Nevis.................................................. 115 ANNEXES Table A1 Central Government FiscalAcounts in St.Kitts and Nevis 1980-2002 .............................................................................................................................................. 117 .......................................................................................................... Table A2 Central Government Fiscal Acounts inSt.Kittsand Nevis 1980- 2002 (InmillionsEC$) 120 Table A3 Central Government Fiscal Acounts in St.Kitts and Nevis 1980-2002 (InmillionsEC$95) ..................................................................................................... 123 Table A4 MainEconomic IndicatorsSt.Kitts andNevis 1980-2002 ......................................... (As percentageof GDP)............................................................................................... 126 Table A5 Operations of Public Enterprises in St.Kittsand Nevis 1995-1998............................ 126 128 Table A6 Social Protection Recurrent Expenditures by Ministries and Programs (InThousandsEC$)..................................................................................................... 129 ListofTables Table E.S. 1 Summary of Policy Recommendations.............................................................................. xiv Table 1.1 Customs Revenue Forgone from Tax Concessions (as percentageof GDP) .......................... 13 Table 2.1 Long-Term GDP. Central Government Revenue. and Expenditure GrowthRates. 1981-2002 ...................................................................................................................................... 20 Table 2.2 Cyclical Properties of GDP. Central Government Revenue. andExpenditure. 1980- 2002 ............................................................................................................................................... 22 Table 3.1 Main Items Responsible for the Variance BetweenActual and BudgetedExpenditures Table 3.2 Ministries Consistently Over or Under Budget in St.Kitts and Nevis. 1995-2000 inthe Ministryof Finance andPlanning. St.KittsandNevis. 1995-2000 (inmillionsEC$) .......35 Table 4.1 Medium-Term PSIF' for St.Kittsand Nevis. 2001-06 (inthousandsEC$) ........................... (average percentage)...................................................................................................................... 36 Table 4.2 Sectoral Composition of Actual CapitalExpenditures for St.Kitts andNevis Federal 44 Table 4.3 Sectoral Composition of Actual Capital Expenditures for St.KittsandNevis Federal Government (as percentage % GDP)............................................................................................. 47 Table 4.4 Source of Financing of St.Kittsand Nevis Federal Government Capital Expenditures Government (as % total capital expenditures) ............................................................................... 47 Table 4.5 RealizationRate of St.Kitts andNevis Federal Government Capital Expenditures. (as % of total capital spending)...................................................................................................... 49 Table 4.6 RealizationRate of St.Kitts and Nevis Federal Government Capital Expendituresby 1994-200 1...................................................................................................................................... 49 FundingSource. 1994-2001 .......................................................................................................... 50 Table 4.7 Actual Public Sector InvestmentExpenditures and Realization Rate for Nevis Island Table 5.1 Employment inthe St.Kitts and Nevis Federal Government. 1995-2002 ............................. Administration (in thousandsEC$)................................................................................................ 51 Table 5.2 Distribution of Established Positions by Ministry inthe St.Kitts and Nevis Federal 55 Table 5.3 St.Kitts and Nevis Federal Government Compensation. 1995-2002 .................................... Government. 2000-01 .................................................................................................................... 56 Table 5.4 PersonalEmoluments and Wages by Ministry inthe St.KittsandNevisFederal 59 Table 5.5 Public and Private Sector Wages inSt.Kitts and Nevis. 2001 (EC$) .................................... Government (EC$) ......................................................................................................................... 60 61 Table 5.6 Employment inNevis IslandAdministration. 1995-2002 ..................................................... 63 64 Table 5.8 PersonalEmoluments and Wages inthe NIA. 1997-2002 ..................................................... Table 5.7 Established and Non-Established Positions. by Ministry. inthe NIA. 2001-02 .................... 65 Table 6.1 Health Indicators inSt.Kittsand Nevis Compared with Barbados. Jamaica. and Table 6.2 Health Indicators in St.Kittsand Nevis. 1995 and 2001........................................................ other OECS Countries. 2000.......................................................................................................... 69 70 Table 6.3 Leading Cause of Mortality. 19962001 (number of deaths per year) .................................. Table 6.4 Characteristics of the Health System. St. KittsandNevis. 1991-2001 ................................. 70 Table 6.5 Uses of Health Facilities in St.Kitts and Nevis. 1999 (inpercentage) .................................. 73 Table 6.6 Health Spending in St.Kitts and Nevis Compared to Barbados. Jamaica. and 74 Table 6.7 RecurrentHealth Spendingin St.Kittsand Nevis. 1995-2002 (inEC $ 95) ........................ other OECS countries. 2000 (as percentage) ................................................................................ 76 Table 7.1 EducationInstitutions in St.Kitts and Nevis by Ownership. 2000-01 .................................. 77 81 Table 7.2 Distribution of Education Subsidies by Quintile. 1999 ......................................................... 88 Table 8.1 Social ProtectionPrograms and Expenditures in St.Kitts and Nevis. 2000 .......................... 98 Table 8.3 St.Kitts and Nevis Risk Indicators by Age Cohort. Existing Programs. Table 8.2 Incidence of Poverty. Indigence. and Other Related Indicators. 1999 (percentage) ............102 and SuggestedInterventions ........................................................................................................ 103 ListofFigures Figure 1.1 St.Kittsand Nevis Central Government FiscalPerformance 1990-2002 (as % of Figure 1.2 St.Kitts and Nevis Central Government Primary Expenditure, Interest Payment and GDP) ............................................................................................................................................... 2 Figure 1.3 St.Kitts and Nevis Central Government Debt andPrimary Surplus 1990-2002 (as % Revenue 1990-2002 (as % of GDP) ................................................................................................ 3 Figure 1.4 St.KittsandNevis Central Government Short-Term Fiscal Gap 1991-2002 (as % GDP) ............................................................................................................................................... 3 Figure 1.5 St.Kitts andNevis Central Government ProjectedShort-Term Fiscal Gap 2003-2005 GDP) ............................................................................................................................................... 6 Figure 1.6 St.Kitts and Nevis Central Government Medium-Term Fiscal Gap 1991-2002 (as % (as % GDP) ..................................................................................................................................... 7 Figure 1.7 St.Kitts and Nevis Central Government Revenue Composition 1990-2002 GDP) ............................................................................................................................................... 8 Figure 1.8 St.Kittsand Nevis Central Government Expenditure Composition 1990-2002 (as % Total Tax Revenue).............................................................................................................. 11 Figure 1.9 St.Kitts and Nevis Central Government Expenditure Composition 1990-2002 (as % of GDP) ............................................................................................................................... 14 ....................................................................................................................... Figure 2.2 Fiscal Cyclical Adjustment in St.Kitts andNevis 1983-2002 (as percentageof GDP) .......23 Figure 2.1 St.Kitts and Nevis GDP, GDPTrend and Cyclical Component 1980-2002 ....................... (in millionsEC$ 95) 15 19 Figure 2.3 Cyclical Adjusted Primary, Overall, and Current Balance inSt.Kitts andNevis, 1983-2002 (as % GDP) ................................................................................................................ 24 Figure 3.2 The Budget Cycle in the St.Kittsand Nevis Federal Government ...................................... Figure 3.1 St.Kittsand Nevis Government Structure ........................................................................... 29 32 Figure 3.3 Variance between Actual and BudgetedRecurrentExpenditures, Ministry of Finance and all other Ministries St.Kittsand Nevis 1995-2000 (inpercentage) ....................................... 34 Figure 3.4 Variance betweenActual and BudgetedRecurrentExpenditures, Ministry of Works, and all other Ministries St.Kitts and Nevis 1995-2000 (in percentage) ....................................... 35 Figure 5.1 Government Employment as Percentageof Population, 2001 (Selected Caribbean Countries) .................................................................................................... 56 Figure 5.2 Average Wage as a Percentageof GDP per capita. 2001 Figure 6.1 Shareof Recurrent HealthExpenditures inGDP. St.Kittsand Nevis 1995-2000 (Selected Caribbean Countries) .................................................................................................... 61 Figure 6.2 Shareof HospitalExpenditures inTotal Recurrent HealthExpenditures. St.Kitts (percentage) ................................................................................................................................... 76 and Nevis 1995-2002 (percentage) ............................................................................................... 77 Figure 7.1 Gross Enrollment and CompletionRates inPrimary Education inCaribbean Figure 7.2 Shareof EducationExpenditures in GDP. St.Kitts and Nevis 1994-2002 (as %) ............... 82 Countries.2001 (in percentage) ..................................................................................................... Figure 7.3 EducationExpenditure in St. Kitts and Nevis 1994-2002 (millions of EC$ 95) ..................84 84 Figure 7.4 Shareof Recurrent EducationExpenditures inRecurrent Expenditures. Figure 7.5 St.Kitts and Nevis Recurrent EducationExpenditures by subsector 1996-2002 St.Kitts andNevis 1994-2002 (as %) ........................................................................................... 85 Figure 7.6 Pupil-Teacher Ratio and Costs of Education in St.Kitts and Nevis. 2000 (inUS$) ...........85 (as % Total) ................................................................................................................................... Figure 7.7 Shareof Trained Teachers inPrimary and Secondary School. St.Kitts and Nevis. 89 Figure 8.1 Social Protection Expenditures in St.Kitts and Nevis (percentage of GDP) ..................... 1999 (inpercentage) ...................................................................................................................... 91 Figure 8.2 Social ProtectionExpenditures in St.Kitts and Nevis (millions of EC$ 95) ..................... 101 102 ListofBoxes Box 1.1 The Convergence Criteria in the ECCU and WAEMU ........................................................... 10 Box 3.1 FinancingExpenditures `Below the Line'. The Use of Advance and Deposits Accounts......26 Box 2.1 Chile's Fiscal Rule ................................................................................................................... 33 48 Box 5.1 Civil Service Downsizing in Guyana ....................................................................................... Box 4.1 PortZante Cruise Ship dock .................................................................................................... Box 6.1 Strengthening the National HIV/AIDS Programover 2003 to 2007: The St.Kitts and 58 Box 7.1 Financingof Tertiary Education inLatin America and the Caribbean .................................... Nevis HIV/AIDS Preventionand Control Project ........................................................................ 72 Box 8.1 A Closer Look at the Social Assistance Program in St.Kitts ................................................ 87 100 OECS FISCAL ISSUES: ST. KITTSAND NEVIS POLICIESTO ACHIEVE FISCALSUSTAINABILITYAND IMPROVEEFFICIENCY ANDEQUITY OFPUBLIC EXPENDITURES EXECUTIVE SUWIMARY Overview 3. Notwithstanding a common central bank and a common currency, St. Kitts and Nevis, like the 1. St. KittsandNevis isa two-islandfederation other ECCU member countries, conducts fiscal in the Eastern Caribbean with a combined policy independently. While limited progress population of about 45,000 inhabitants (of whom has been made on a common fiscal policy in the 75 percent live in St. Kitts) and a per capita region, the deterioration of the ECCU fiscal gross national income (GNI) of US$ 6,980 in position in the recent years, notably in Antigua, 2003 (World Bank Atlas Methodology, current Dominica, and St. Kitts and Nevis, has US$). This small, open economy has reinitiated discussion among its members of the traditionally depended on the production and need to coordinate fiscal policies.* export of sugar. However, as a result of declining trade preferences for sugar, in the last 4. During 199G95, the St. Kitts and Nevis decade the structure of the economy has shifted economy showed a strong growth performance, toward services, notably tourism. The economy with a growth rate of about 4.3 percent per year. is highly vulnerable to exogenous shocks, such However, since the mid-1990s, economic as hurricanes, which have inflicted substantial growth in St. Kitts and Nevis slowed to 3.6 damage to the capital stock and reduced growth percent per year as a result of exogenous shocks inrecent years. (declining trade preferences for sugar, hurricanes, the September 11 events, and the 2. As a member of the Eastern Caribbean recent global economic slowdown), and Currency Union (ECCU), St. Kitts and Nevis declining productivity. Moreover, in2002 the St. shares a common central bank, the Eastern Kitts and Nevis economy contracted by 1.0 Caribbean Central Bank (ECCB), and a common percent, the first economic contraction currency, with the other five independent experienced by the country in almost 20 years. members of the Organization of Eastern This slowdown and contraction have been Caribbean States (OECS) and three dependent accompanied by increasing fiscal deficits and territories (Anguilla, Montserrat and the British rising public debt and debt service obligations. Virgin Islands).' Their currency is the Eastern The medium term growth potential of St. Kitts Caribbean (EC) dollar, which has been pegged and Nevis i s in the range of 3-4 percent, to the U.S. dollar since 1976 at EC$2.70 per reflecting the completion of most of the major US$I.The commitment to a fixed exchange rate infrastructure projects, a pick up in the global and a prudent monetary policy under the ECCU economy inducing restoration of tourism to it umbrella has been a key element in maintaining an inflation rateclose to internationallevels. * The communiqut of the Special Meeting of the ' The OECS Authority, Castries, St. Lucia, September 28, OECS consists of six independent members: 2001 alludes to the need to examine in the medium Antigua and Barbuda, Dominica, Grenada, St. Kitts term public sector reform, aimed at increasing and Nevis, St. Lucia, and St. Vincent and the efficiency, and fiscal reform including matters related Grenadines, in addition to three dependent territories to revenue, expenditure, financing, and debt of the United Kingdom, Montserrat (full member), management, as well as common approaches to Anguilla, and the British Virgin Islands (associate taxation and incentives to be applied to cruise members). shipping. pre-September 2001 levels, and the near living on less than US$1 per day. With respect doubling of the room capacity in the country as a to the health MDGs, the under5 mortality and result of the opening of the 900- room Marriott HIV/AIDS goals are the most relevant to St. Resort. Main exogenous risks to the medium Kitts and Nevis. Malnutrition (defined as the term scenario are posed by a prolonged global prevalence of underweight children), maternal recession, potential uncertainty engendered by mortality, and tropical diseases (other than Nevis's plans to secede, a natural disaster such dengue, which is endemic) are not major issues. as a tropical storm or hurricane, and risks of a The under-5 mortality rate in St. Kitts and Nevis drying up of the regional capital market fuelled dropped from an estimated 36 per 1,000 in 1990 by high fiscal deficit and debt ratios. to 24 per 1,000 in 2001. This is a decline of about 3.7 percent per year. The MDG target-a 5. Despite high levels of per capita incomes reduction of two-thirds from 1990 to 2015- and good social service provision, poverty would require that the rate fall to about 12 per remains a persistent problem in St. Kitts and 1,000 by 2015. According to trend projections, Nevis. The national 1999 survey data indicate St. Kitts and Nevis will not reach but will come that approximately one in three individuals are close to this MDG target, since the present trend poor (that is, living below a locally defined would result in a rate of 14 per 1,000 by 2015. poverty line). In St. Kitts 30.5 percent of the The rate of decline is however likely to slow population lived below the poverty line, as did from 2001 to 2015, because as child mortality 32 percent in Nevis. Moreover, approximately rates get lower, incremental declines get harder 11 percent of St. Kitts's population and 17 to achieve. It is difficult to assess the likelihood percent of Nevis's population is indigent; that is, of reaching the HIVIAIDS targets, because they do not have sufficient resources to meet baseline data on HIV prevalence and incidence their basic dietary needs.3 Income are poor. Adult HIV prevalence is between 1.5 insecurity/volatility at country and household and 2 percent. The Government has launched a levels as a result of climatic and external comprehensive HIV/AIDS prevention and economic shocks is particularly pronounced in control program with the support of the World St. Kitts and Nevis, like in the rest of the Bank with the objective of reversing its spread. Caribbean, given its small size, limited economic diversification and high degree of 7. To improve competitiveness, restore rapid openness, and its extreme vulnerability to economic growth, and ensure its medium-term hurricanes. The impact of this extremely high sustainability in the context of the currency volatility is especially negative for the poor and union, the main challenges to the St. Kitts and near poor who may not be able to rely on Nevis government are to (a) tighten fiscal savings or on the government's social protection policy, notably through expenditure cuts; (b) programs to smooth consumption during times increase the efficiency of public investments and of hardship. public service delivery; (c) strengthen regulation and efficiency of public utilities and sea/air 6. With respect to the Millennium transport; and (d) promote education and skills Development Goals (MDGs), the primary development to prepare the population, notably education goal has been attained-primary the poor, to take advantage of new opportunities education is universal for boys and girls. It is inthe globalenvironment. likely that the goal of eradicating extreme poverty and hunger will be achieved; St. Kitts and Nevis i s estimated to have few habitants We cannot estimate with certainty the share of the population living on less than US$1per day because In 1999, the adult equivalent poverty and indigence information is unavailable regarding the purchasing lines were equal to EC$3,361 and EC$2,135 in St. power of one U.S. dollar inSt. KittsandNevis. Kitts, andEC$3,941 andEC$2,453 inNevis. .. 11 8. This report concludes that the fiscal deficit 9. The report is organized as follows: Chapter in St. Kitts andNevis is unsustainableand poses 1 discusses fiscal sustainability in St. Kitts and a risk to the stability of the currency board Nevis and presents options for fiscal arrangement. Most of the adjustment needed to consolidation; Chapter 2 discusses the role of the achieve a sustainable fiscal policy will have to business cycle in the design of fiscal policy by come from expenditure cuts, in particular a reviewing the cyclical components of the fiscal reduction in the number of established and non- accounts; Chapter 3 discusses the budget established positions? containment of capital management system in place and its effect on expenditures to projects geared to growth and budgetary outcomes. Chapter 4 examines the poverty reduction and funded largely by grants Public Sector Investment Program (PSIP); and concessional loans, and the closure of the ,Chapter 5 discusses public sector employment unprofitable St. Kitts Sugar Manufacturing and compensation; Chapters 6 and 7 examine Company (SSMC). A reduction in tax public expenditures and outcomes in the exemptions and discretionary concessions is also education and health sectors; and Chapter 8, the essential, as is a restructuring of the revenue final chapter, discusses social protection base through the introduction of the value-added programs. tax (VAT). While social indicators are good for a country at this level of gross domestic product Fiscal Sustainability (GDP), there is ample room for more efficient modalities of education and health service 10.At the Central Government (CG) level, since delivery that can generate significant savings 1995 the fiscal position of St. Kitts and Nevis and improve effectiveness. In addition, the has weakenedprogressively (from a CG primary government should strive to ensure that its social deficit of 3.3 percent in 1995 to 10.8 percent of programs, in particular social protection, are GDP in 2002). The main causes of this provided in a well-targeted and effective manner deterioration have been increases in the wage inorder to providemore effective safety nets for bill and capital expenditure (to a degree in the possible impact of sugar sector restructuring, response to hurricane repair), and increases in fiscal consolidation and the rationalization of expenditures linked to legislative elections that education and health service delivery, and took place in 1995 and 2000. The government continued vulnerability to external shocks. has relied on both external and domestic Specific recommendations on how to achieve borrowing, mostly at commercial terms, to fiscal sustainability,. improve social service finance its fiscal deficit. As a result, the CG delivery, and reform social protection are debt-to-GDP ratio increased from 51.5 percent provided inthe attachedmatrix. See Table E.S. I. of GDP in 1995 to 104.4 percent of GDP in 2002, and the interest expenditures increased from 2.1 percent of GDP in 1995 to 7.6 percent of GDP in 2002. The losses of public sector There are two types of government positions inthe enterprises, notably the SSMC and the St. St. Kitts and Nevis Federal Government: established Christopher Air and Sea Ports Authority, put and non-established. Employees in established further pressure on the government deficit and positions are appointed by the Governor General public debt. upon formal request by the Public Service Commission, with the approval of the Establishment 11.The short-term fiscal sustainability exercise Division in the Prime Minister's Office. Established (one year) shows that since 1995 the fiscal positionsare permanentpositions listedinthe budget, policy conducted by St. Kitts and Nevis (both and their compensation is classified as Personal the CG and the public sector) has been Emoluments. Employees in non-establishedpositions unsustainable. The short-term adjustment are hireddirectly by line ministries, their numbers are not presentedin the budget, and only their wages are required to achieve fiscal sustainability of the includedas a line item. CG in 2003 is about 14 to 16 percent of GDP, depending on the assumptions of real interest ... 111 and GDP growth rates. Since the CG guarantees increasing coverage, including establishing an the debt of public enterprises, the government adequate regulatory framework to ensure that would need to adjust by additional 2 to 3 public monopolies are not replaced by percentage points of GDP to ensure the fiscal unregulated and inefficient private monopolies. sustainability of the public sector. FiscalPolicy and the Business Cycle 12. This size of adjustment is significant (equivalent to a primary surplus of between 3 15. Economic volatility tends to discourage and 4 percent of GDP), and it would be difficult long-term investments in physical and human to achieve it in the short-term, even with a major capital, and thus leads to a lower growth fiscal reform. The ECCB's Monetary Council potential. The poor segments of the population meeting that took place in February 2003 are generally the most affected becausethey lack reconfirmed the agreed regional fiscal the means to protect themselves from adverse framework that includes ceilings on debt and employment and income effects. Governments fiscal performance of the CG (see Chapter 1, have a broad set of instruments available to paragraph 1.28). The aim of the ECCB in setting reduce aggregate volatility and its effects, one of these fiscal convergence criteria is to ensure which is fiscal policy. In developed economies, long-run fiscal sustainability and the stability of fiscal policy is usually designed to stimulate the currency union by limiting negative output when the economy moves into recession spillovers. Based on its fiscal ratios in 2002, the and to contract output when an expansion takes deadline for fiscal convergence (a four-year place. A fiscal policy designed in this way leads horizon) requires a fiscal adjustment higher than to a strongly procyclical budget surplus the one discussed above if the objective is to (countercyclical fiscal policy). meet the proposed 60 percent of GDP ceiling, and will require an even faster public sector 16. The St. Kitts and Nevis GDP cyclical reform and gearing up of the social protection component shows no clear link to its major programs. sources of economic volatility, hurricanes, but several expansions and contractions are worth 13. Adoption of the IMF tax policy noticing. Between 1980 and 2002, St. Kitts and recommendations would reduce the short-term Nevis was hit by six hurricanes. In three out of fiscal gap only slightly. Hence, the fiscal six hurricane years, St. Kitts and Nevis's GDP adjustment would have to rely mainly on decreased below its trend, and in the other three expenditure cuts, notably a reduction in the years increased above its trend. The deepest number of established and non-established contraction took place in 1983, a year after the positions, containment of capital expenditures, debt crisis and the greatest expansion in 1989, and the closure of the unprofitable SSMC. The the year of Hurricane Hugo. increasing financial difficulties that the SSMC is facing are the result of a drop in productivity and 17. All components of the fiscal accounts are the phasing out of trade preferences. The report more volatile than GDP. The highest volatility is recommends that the government close the for grants, capital expenditure; and capital SSMC as soon as possible to avoid further revenue. Current expenditure, all its deterioration of an already unsustainable fiscal components, as well as capital expenditure are situation. acyclical (that is, in good times or bad times the government does not increase or reduce its 14.The sales of government assets can be an expenditures). This result is similar to the one option to soften the adjustment required to observed in Organization for Economic achieve fiscal sustainability. However, any Cooperation and Development (OECD) privatization should not be driven by revenue countries, but not in developing countries. considerations alone, but rather by the objective Current and tax revenue are procyclical (not of improving quality and efficiency and iv surprisinggiven that most taxes are in some way move the country to a countercyclical fiscal proportional to economic activity). policy we propose a fiscal rule based on a structural surplus of about 2 to 4 percent of 18.Finally, all measures of fiscal balance such GDP. This rule is more austere than the Chilean as primary, current, and overall balance are fiscal rule of a structural surplus of 1 percent. acyclical, that is, they do not co-move with the However, considerations of credibility and high GDP cycle. In good times the government does GDP volatility in small islands would require not run a higher or lower fiscal surplus than its that small countries like St. Kitts and Nevis trend. This is an interesting result and contrary adopt tighter fiscal rules. This fiscal rule implies to the empirical evidence for developing a fiscal adjustment of about 14 to 16 percent of countries. However, most of this evidence does GDP, in line with the adjustment recommended not include small islands like St. Kitts and to achieve fiscal sustainability. Nevis, in which the procyclicality of tax revenues is high enough and expenditures are 22. However, successful implementation of a acyclical. fiscal rule based on a structural surplus will require, as a first step, timely and reliable fiscal 19.The cyclical adjustment to the government information. As discussed in the report, the budget balances in St. Kitts and Nevis varies current presentation of the fiscal information from -1.8 percent of GDP in 1983 and 2002 does not follow international practices and (recession years) to 1.9 percent of GDP in 1989 presents a distorted picture of the country's (expansion year). For example, in 2002 the fiscal situation. overall fiscal deficit after grants was 18.3 percent of GDP, but factoring out the cyclical Financial Management impact of lower tax collection, the cyclically adjusted overall fiscal deficit was only 16.5 23. St. Kitts and Nevis has introduced percent of GDP. The St. Kitts and Nevis cyclical substantial changes to the budget structure and adjustment compared to other countries in the processes in the last few years, and its timely region is in the low range. The cyclical reporting is commendable. These changes, adjustment in Chile is between -4.0 and 2.5 however, have not yet fully translated into percent of GDP. improved resource allocation and effective expenditure controls, becauseof the lack of firm 20. Once the cyclical effects are factored out, resource envelopes at the aggregate and ministry the country has run a structural primary, current, level underpinned by a medium-term economic and overall deficit for most of the 1990s. In strategy and ministries' corporate plans nested in 1994, the structural primary deficit was 0.5 a consistent macroeconomic framework. Both percent of GDP and kept increasing until it the Strategic Country Outlook (a yearly reached 8.9 percent of GDP in 2002. The same document that presents the economic prospects pattern is observed for the structural overall of the country for the next fiscal year) and the fiscal balance. Since 1993, the country has run a Medium-Term Economic Strategy Paper structural overall deficit that has increased until (MTESP, a four-year economic plan) last it reached 16.5 percent of GDP in 2002. This prepared for the Caribbean Group for confirms the conclusion that the business cycle Cooperation in Economic Development has had little effect on the sharp deterioration of (CGCED) in June 2002, suffer from the same St. Kittsand Nevis's fiscal position. shortcomings: (a) top-down estimates of aggregate resources available for public 21. Fiscal rules have two main objectives: (a) to expenditure not always consistent with enforce fiscal responsibility, and (b) to ensure a macroeconomic stability, (b) lack of bottom-up countercyclical fiscal policy. In this regard, St. estimates of the cost of carrying out ongoing and Kitts and Nevis is on the right path with an new programs, and (c) no attempt to reconcile acyclical fiscal policy. However, in order to the aggregate resources with the cost estimates V of ongoing and new programs. Most of the 27. The Audit Department does a commendable ministries' corporate plans are out of date or job. Audit Reports are current for the federal cannot provide a link among objectives, government. Unlike in other OECS countries, expenditure required, and outcomes. reports in St. Kitts and Nevis no longer detail hundreds of oversights and minor transgressions, 24. The budget provides an accountant's view of focusing instead on material findings. Also, the government transactions rather than a Audit Department continues a modest program presentation of the government's fiscal policy. of value-for-money audits to supplement The budget adheres to the traditional compliance audits. However, while in principle administrative, program, and economic structure independent because of authority specified inthe format. However, this presentation is insufficient Constitution, the Audit Department has no to assess the fiscal position of the government guaranteed operating budget or staffing and to follow up on key programs supposedly complement, and must obtain approval from the linked to policy priorities. The Ministry of Minister of Finance via the annual Estimates Finance does, however, have the capacity to process. Additional factors that have affected its present the budget by standard functional efficacy are the automation of critical financial classification and it is recommended that it do so management functions without adequate input infiscal year 2004. from the Audit Department in the design and implementation stages, the nonfunctioning of the 25. Moreover, sizeable expenditures (mainly Public Accounts Committee, and the lack of capital expenditures) have been financed from training for staff. "below the line" using advances and deposit accounts without being brought to balance and 28. The financial statements of St. Kitts and reflected in the budget in the years in which they Nevis present a partial view of government occur. Thus, the recurrent and overall balances, activities. Significant fiscal operations are as well as other line items presented in the excluded. The operations of statutory bodies, budget, do not represent an accurate picture of including state-owned enterprises (SOEs), are the government's fiscal situation. It is not covered, and several contingent liabilities are recommended that the government bring to not presented in the financial statements. Most balance all "below the line" accounts starting in of the SOEs are significantly in arrears in the 2003. preparation of their financial statements and the submission of these for audit and for 26. The debt management function is the Parliament's review. Although the Director of weakest component of the financial management Audit hasthe rightto require submission of SOE information system. Until 2002, the debt audit reports and to supplement them by further management function was split between a small investigations if hehhe considers it necessary, unit in Finance, which uses the Commonwealth he/she has not done so. Finally, following the Secretariat Debt Recording and Management last election, the members of the Public System (CS-DRMS) to maintain data on foreign Accounts Committee (PAC) were appointed in debt, and the Accountant General's Department, November 2000, but the Committee did not which handled domestic debt (mainly treasury meet subsequently. The Committee did meet a bills). These units have now been merged and few times during 1992-94, but took little action renamed the Debt and InvestmentUnit under the of substance. Weak PACs are a common feature Accountant General. However, this new unit in the OECS countries, and the report presents lacks qualified personnel to keep adequate recommendations to improve their functioning records of external and domestic debt to analyze to become a vibrant component of public the data and advise the government on debt financial oversight. management issues. 29. The formulation of the budget receives inputs from the social partners, but this vi undoubtedly could be made more systematic. which the capital spending intentions included in The recent agreement by OECS countries to the estimates do not represent a firm investment establish National Economic Councils with plan. No cost-benefit analysis is done by the line public and private sector members may lead in ministries to assess whether the projects this direction. However, the capacity of the proposed are cost-effective (with the exception social partners-most of which are small of those financed by external donors). Several organizations-to go beyond their immediate investment projects included inthe PSP and the concerns and address complex national policy capital budget should be classified as recurrent issuesremains limited. expenditures. Moreover, a significant amount of the capital expenditures incurred by the Public Sector Investment Program government are recorded "below the line" and not captured in the PSIP. 30. In St. Kitts and Nevis, on average, during 1995-2002, capital expenditures as a fraction of 33. The implementation of the PSIP is carried GDP increased about three times compared to out by the line ministries and monitored by the 1990-94, including "below the line" transactions Planning Unit. However, the coordination not presented in the annual budget estimates. among these agencies is weak. Post-evaluation This increase in expenditures was mostly of projects is usually conducted only for projects financed through loans (commercial and financed by external funds. Because no concessional) and is one of the main reasons examples of such evaluations were provided for behind the difficult fiscal position facing the local and foreign-funded projects, it is government today. Although the damage caused impossible to assert how useful they have been by Hurricanes Luis in September 1995, Lenny in for future project planning. November 1998, and Georges in November 1999 explains part of the increase in capital Public Sector Compensation expenditures, a large portion was not hurricane related and its aim was not necessarily to 34. Given the small size of its economy and stimulate growth or reduce poverty. population, St. Kitts and Nevis has limited resources to carry out the whole range of 31.The public sector investment program of St. functions performed by any government. As Kitts and Nevis does not include investment several studies have shown, small countries tend projects by statutory agencies, such as SOEs. to have bigger public administrations in This is a serious gap because these investments response to the higher cost of supplying public can represent a sizeable portion of the public goods and the need to provide a stabilization sector investment program and can expose the role to ameliorate the effect of external shocks. government to serious fiscal liabilities. The However, in the case of St. Kitts and Nevis the government should ensure an accurate public administration is too large and expensive representation of the public sector investment for sound fiscal management. programby including inthe PSIP the investment projects of all statutory agencies. Investment 35. The main finding of this report is that the projects by private corporations but guaranteed highwage bill inthe St. Kitts and Nevis Federal by the government should be disclosed in the Government (15 percent of GDP) is employment government's financial statement as government driven rather than wage driven. During 1995- liabilities. 2002, the number of established positions increased in total by 38 percent, and the number 32. Although the PSIP is prepared as a four-year of non-established positions in full-time rolling medium-term investment program, in equivalent basis increased in total by 124 practice the focus is the one-year capital budget. percent. At the same time, government real The relatively low realization rate (about 25 wages decreased slightly. St. Kitts and Nevis's percent) provides an indication of the extent to ratio of government employment to population vii (14.2 percent) is at least three times higher than 38. There is considerable scope for the that observed in other economies in the government to divest some of its functions to the Caribbean. Although the federation structure of private sector with the objective of improving the country and the fact that in St. Kitts and quality and efficiency and increasing coverage. Nevis many public utilities are part of the So far, only telecommunications has been Central Government explains to some extent the privatized and garbage collection contracted out. high ratio, the main cause has been an increase The electrical utility, water and sewerage, port in the numbers of established and non- handling, the cruise ship terminal, and the established positions since 1995. government printing office could all be considered potential candidates for privatization. 36. In 2000, the Government ,imposed a freeze Some services where it is easy to specify output on new hiring and identified the wage bill as an can be more efficiently performed by the private area where cutbacks were needed as part of a sector. These include road maintenance, airport strategy of fiscal consolidation. However, services, laundry and catering services in although the freeze has been implemented with hospitals, and cleaning and maintenance services some success for established employees, non- in public buildings. The Post Office could be established employment head count has corporatized to improve efficiency. Finally, increased since 2000 in total by 31 percent, or some public sector activities in the areas of 3.8 percent in full-time equivalent. In the short social services, and education could be term, it is critical that the government transferred to nonprofit organizations, as is strengthens controls on non-established being done to a limitedextent in St. Lucia. employment and reports it in the estimates similar to the way it reports established 39.The government has so far delayed the employment. Over the medium term, the decision to downsize its civil service and reform downsizing in public service employment its public sector because it hopes that the new required to close the fiscal gap is more than the Marriott Hotel, scheduled to operate at full 10 percent of the current number of employees capacity in September-October 2003, will be foreseen by the government. Downsizing the able to absorb some of the potential laid-off civil service to the levels of 1995 by cutting 800 public servants, as well as some of the workers established positions and 500 non-established that may lose their jobs if the sugar company positions (head count) would reduce the wage (SSMC) is closed down. bill by EC$42 million, or 5 percent of GDP. The downsize will have to be across all line 40. A public sector reform program would need ministries with National Security, Education, to address the issue of the existing fragmentation Health and Environment taking a significant cut and duplication of work among ministries, because they represent about 70 percent of all departments, and agencies, leading to a diffusion government employment. of responsibilities, and the possibility of the government divesting some of its functions to 37. The St. Kitts and Nevis government lacks a the private sector. The reform would need to be public sector human resource strategy. Job informed by a human resource management descriptions and performance appraisals, which strategy, and a performance appraisal system are required to effectively manage personnel, do should be established with pay or bonuses to not exist even for established positions. No reward good performance. It would also need to systematic analysis of training needs has been be accompanied by the introduction and/or carried out. Training in essential skills is upgrade of information systems in order to deficient and there is no training plan for accurately cost its implications (pension and individual employees to enable them to advance severance payments) and monitor public along their career paths. Training is often employment (established and non-established offered in an unsystematic way, depending on positions) on a regular basis. the availability of donor financing. viii 41. A Public Sector Reform Unit was recently scale, (b) duplication across islands of planning established with the mandate to lead the process and monitoring activities, (c) health services of public sector reform, which is directed at mostly publicly provided, (d) health financing modernizing the public service. This will include m,ostly through general revenue, and (e) a high taking an inventory of the human resource ratio of nurses to physician. The health system capacity of each ministry, identifying the also faces the challenge of adapting to a relevant training to address any deficiencies changing disease profile, including the need to discovered, and taking a critical look at the vigorously address HIV/AIDS. There is procedures and systems in use in the public coordination, but Nevis has its own ministry service to make sure they have not become responsible for health, and its own chief impediments to efficiency. I t is encouraging that secretary responsible for the administration of the government has established such a unit; local health services. however, it is not clear whether the unit will be adequately funded and has the support of the 44. As the disease patterns change over time and Cabinet to carry out its mandate. with the goal of achieving greater economies of scale, it may be possible to rethink the Health configuration of health facilities in the federation, primarily by reorganizing and 42. The Federation of St. Kitts and Nevis has a consolidating some of the primary health center health system with a strong primary health care services. The government should also be able to focus. It is dominated by publicly financed and achieve greater economies of scale by better publicly provided health care. There are no coordinating health services with its OECS private clinics or hospitals, but there are neighbors and other countries in the region, physicians in private practice. Public spending including specialized laboratory tests and on health during 1995-2002 was about 3.3 procurement of medical supplies, an excellent percent of GDP, on average. Public health example of exploiting economies of scale at the spending in real terms grew for the same period regional level that can be extended to other at about 4.4 percent per year. Given the tight services is the Eastern Caribbean Drug fiscal constraints, this growth is unlikely to be Procurement Service. maintained over the short to medium term. Moreover, wages absorb about 80 percent of all 45. Public health financing is mainly through health spending in St. Kitts and 73 percent in the consolidated fund. Revenue from user fees is Nevis. These figures are on the high side, and minor. User fees at public facilities amount to raise questions as to whether there is sufficient only about US$6 per capita per year, or 3 allocation for pharmaceuticals, medical supplies, percent of health recurrent spending in 2000. and maintenance of equipment and buildings. Facilities do not retain fees, so have little Capital expenditures in health are highly donor incentive to put effort into collection. Fees revert dependent, with a very large share from the to the consolidated account. Furthermore, there European Community. are broad exemptions and fees are modest. While fees may seem low on a per service basis, 43. Public facilities appear to provide user fees for a catastrophic health problem could reasonably good coverage, with a good level of put an unacceptably high burden on a poor consumer satisfaction. Nevertheless, the health household's financial situation, because the system faces many challenges in the future. exemptions are not well targeted. The user fee Most immediately, the public system is facing policies would benefit from a review in terms of strong budgetary pressures as part of the overall equity and efficiency. The government should fiscal crisis. St. Kitts and Nevis's small size and also consider how fees might be significantly federation structure also pose many challenges increased to cover costs for physicians who treat to the administration of public health programs, their private patients in public facilities, and for such as (a) unexploited potential economies of overseas visitors. ix 46. The government is considering adopting a among individuals with secondary, national health insurance system based on postsecondary, and university education, payroll taxes. Given that there is a general respectively. Hence, there are strong economic revenue-financed system in place, it is not clear motives for the individual and the government to why it would make sense to add a payroll-tax invest ineducation. system of social insurance. The main motivation may be simply to find an additional source of 49. Education expenditures in St. Kitts and revenue for the health system. However, payroll Nevis increased by about 40 percent in real tax systems have disadvantages. They are a tax terms between 1994 and 2000. However, this on labor. In addition, given that many increase in public resources has not always been households are outside the formal sector, the accompanied by an improvement in education system would not cover the full population. outcomes, equity, and efficiency. The fiscal Working with the existing financing system, but crisis that the country is facing will require a cut finding ways to improve efficiency to ensure in expenditures for all sectors, but this should that the country is getting better value for its not preclude the government from completing money could be a better choice in the long run the educational agenda by making better use of for such a small country. The main measures to the available resources and increasing the improve efficiency over the short to medium nongovernmental provision in education, for term are (a) reorganizing and consolidating the example, at the tertiary level. primary health center services, (b) substituting nurses with nurse assistants where possible, and 50. Except for pupils in early childhood, (c) increasing the reliance on well-targeted user Kittitians and Nevisians receive education fees for financing. predominantly from public institutions. The education system in St. Kitts and Nevis stands 47. Public recurrent spending on hospitals in St. out in a Caribbean context in two ways: (a) the Kitts increased dramatically during 1995-2002. dual ministerial structure for one country due to Its share in total recurrent expenditures its federal structure, and (b) the automatic increased from about 52 percent in 1995 to close progression in basic education, where no child to 59 percent in 2002. This finding may be repeatsa class unless the parents directly express partly due to definitional changes. The budget the wish. The dual ministerial structure increases categories in use on a year-to-year basis had administrative costs. The automatic progression frequent changes. However, it is likely that system leads to impressive educational hospitals are puttingpressure on health spending indicators in terms of enrollment, graduation, in St. Kitts, and this is certain to continue unless and repetition (with the millenniumdevelopment (a) changes are introduced to improve hospital goals related to primary education having been efficiency, and (b) chronic disease prevention attained). However, it hides major quality and and early diagnosis and treatment programs are inefficiency problems. In this system, the use of strengthened to keep people healthy and out of an effective monitoring and evaluation system is hospitals. critical for the system to perform adequately. Education 51. During 1994-2001, on average, St. Kitts and Nevis spent about 5.2 percent of GDP in 48. The close link between education and labor recurrent and capital expenditures on education, market earnings implies that the lack of with a rising trend toward the end of the period. education is one of the primary sources of This compares with the average for Caribbean poverty in St. Kitts and Nevis. Data from the countries of 6.6 percent of GDP and the Latin 1999 Survey of Living Conditions indicate that American average of 4.1 percent of GDP. 27 percent of those with no education or a primary education are poor, while only 18 52. Since the mid-1990s, administration costs percent, 12 percent, and 5 percent are poor and tertiary education have taken an increasing X share of recurrent education expenditures. Given 55. The most important educational barriers for the large fixed costs involved in education low-income families are limited access to, and administration-for instance developing of inconsistent quality of, the education system. exams, information systems, strategies, and The School Feeding Program is traditionally curricula-it is expected that smaller states perceived as an expenditure line that benefits allocate a higher share to administration.6 low-income students. However, in St. Kitts and However, there are considerable savings to be Nevis at least 65 percent of the well off, or half gained from deeper within and cross-country of the student population, receive free meals. A cooperation, which, moreover, would stimulate similar pattern is observed with free textbooks. quality improvements through international The lack of targeting for schoolbooks is sharing of best practices. especially worrisome given that 1 out of 5 poor children has no textbook, while this happens for 53. The share of tertiary education in recurrent only 1 in 12 among children from the richest education expenditures increased from 14 quintile. The government could hence either percent in 1996 to 18 percent in 2002. In a Latin limit the School Feeding Program and textbook American context, this share is below average, provision to low-income families only, or but so is coverage of tertiary education in St. introduce fees for higher-income families. Kitts and Nevis. The combination of limited access to tertiary education and a low cost- 56. Salaries are, by far, the largest spending recovery ratio for tertiary education makes this item, taking up about 90 percent of all recurrent budget line inequitable. Further, given that a education expenditures in 2000. At the nonnegligible share of highly educated secondary level, there is especially reason for individuals migrates abroad, the public loses a concern because the wage share in total corresponding share of the returns to this public secondary education spending increased from an investment. It i s recommended that fees be already high92 percent in 1996 to 99 percent in increased and regional cooperation between 2000. This leaves exceptionally little money to colleges be strengthened to ensure efficiency other crucially important learning inputs, such as improvement through specialization. writing paper, textbooks for poor students, and school maintenance. This skewed allocation of 54. The share of secondary education expenditure toward salaries unnecessarily expenditures in total recurrent education reduces the effectiveness of public resourcesand expenditures in the St. Kitts and Nevis Federal results in inefficiently low learning. The main Government has decreased since the mid-1990s. reason for the large share of salaries in recurrent In the mid-l990s, 4 out of each 10 education education expenditures is the low pupil-teacher EC$ went to secondary education. At the start of ratio. St. Kitts and Nevis spends the most per the new millennium, 3 out of 10 reached this student per year in primary and secondary subsector. The adjustment reflects a serious education among the Caribbean countries, squeezing of nonsalary expenditures. The share US$860 and US$1,080, respectively. In primary of recurrent expenditure to secondary education and secondary education there are, on average, should not be allowed to fall further, and a 18 and 13 pupils per teacher, respectively. The reallocation of recurrent expenditures to inefficiency involved in this poor teacher nonsalary expenditures at this level is deployment is tremendous. It is recommended recommended. that these indicators be raised to the average Caribbean level, 25 and 18 pupils per teacher in primary and secondary education, respectively. 57. Quality of instruction, and especially teacher The average cost of administration in other education systems in the Caribbean reaches 10 qualification, is key for increasing effectiveness percent, twice the Latin American average of 5 of learning. Teacher qualifications are percent. inadequate in St. Kitts and Nevis. Only 56 xi percent and 29 percent of the teacher corps in coordination among programs within or between primary and secondary education are adequately islands. With the exception of the Social trained to perform their task. The lack of Security Program, programs are financed and qualifications i s a severe and long-term barrier administered separately by the two islands. for increased efficiency in the education sector. Although the government has expressed its The inefficiency i s partially a result of the rule intent to better coordinate its social protection of admission at the teachers college that programs, it has yet to establish a mechanism to stipulates two years of teaching experience for oversee the development of an integrated social eligibility. It is recommended that teachers be protection strategy and a rationalized set of eligible for training before the commencement programs. The result is that social protection of teaching. priorities are not clearly identified, and fiscal prioritization of programs is lacking. This is Social Protection reflected in the budgetary process during which budgets are essentially determined by previous 58. Social protection programs in St. Kitts and budget allocations and not in response to Nevis emphasize risk-coping strategies; that is, changing needs and priorities. programs to help households cope with the effects of risk (especially poverty) rather than 60. The large number of programs, each with prevention and mitigation strategies. Moreover, separate administrative systems and procedures, the existing social protection programs are not strain an already overburdened public sector. well poised to respond to either significant The lack of coordinated efforts across ministries economic downturns or natural disasters. Since and departments has resulted in overlaps, the mid-1990s, social protection expenditures in duplication, inefficiencies, and administrative St. Kitts and Nevis have been on average about waste. To the degree to which they serve the 3.5 percent of GDP. Although below the average elderly, Social and Public Assistance and for the Latin American region (4.7 percent of noncontributory Social Security target the same GDP), any effort to strengthen social protection group. Moreover, Social Security, the Social and cannot include costly new programs, given the Public Assistance programs, the Student difficult fiscal position that the St. Kitts and Education and Learning Fund (SELF), and the Nevis government is facing. Rather, what are Ministry of Health all have different systems for required are improved targeting and distribution identifying beneficiaries. This increases systems that can provide a safety net for possible administrative costs and time costs for retrenchment in the public sector and the sugar beneficiaries. There is a need to strengthen the industry and that can be geared up quickly inthe planning and executing capacity for these case of external economic or climatic shocks. programs and to put in place mechanisms for their fiscal prioritization. 59. The government of St. Kitts and Nevis has adopted a social protection strategy designed to 61.Several programs, including the school ensure access to health care and primary lunch program for primary students and a school education and to expand access to secondary and bus service for secondary students, are not tertiary ed~cation.~The government also targeted and are provided cost-free irrespective attempts to ensure that other programs and of income. In St. Kitts, health care subsidies are services (including income support, food, targeted based on categorical criteria (children, clothing, housing, and utilities) are available for elderly, disabled) rather than income. Reducing the poor, This stated, the existing social program leakage and introducing targeting of protection strategy is not clearly articulated or school feeding and a few waivers for health well coordinated. There is an array of programs services to the poor would considerably ease in both St. Kitts and Nevis, with little pressure on the budget. Divestment of programs to private providers (profit or nonprofit) could 'Both as a federation and as separate island states. also result in cost savings. However, much- xii needed capacity building and institution 63. Budgeting and management information strengthening, including development of systems also need to be strengthened. Several targeting, management information systems programs, including cash transfer and labor (MIS), and survey capacity, would add market programs, are not identified in the development and administrative costs. government's budgets as separate programs. Expanding program coverage with provision of This limits effective planning and weakens fiscal benefits conditional on activities that promote accountability. Elevating social protection human capital also come with their own activities to program status would mean that investment costs. The net fiscal impact of the expenditures, including expenditures by object above reforms would need to be carefully code, could be easily tracked over time and assessed. evaluated with respect to program inputs and outcomes. In addition, management information 62. Effective planning will require updated systems needto be computerized. poverty and labor market information at regular intervals. Timely and reliable information on 64. Inadequate MIS systems and capacity poor and vulnerable groups is essential for the constraints with respect to data collection, design and implementation of social safety net monitoring, and evaluation are common across programs. The Caribbean Development Bank the OECS. Regional initiatives to MIS (CDB)-financed Poverty Assessment that was development and monitoring and evaluation conducted in 1999 provided useful information; capacity building would address the human however, data on poverty and labor markets resource constraints in individual countries and need to be collected on a regular basis. would be more cost effective than developing systems on a country-by-country basis. ... Xlll E .Y s> .->x EE E b E 3 Y Y Y c 8 c 5 Y c8 CA 2Y Yl o 0 99 4 I 2: 0 2 Y > X U 0 3 k! L Y il c 0 n r 5Y P 8 L 4zm 3 Y .-> x E e, Y 5E Y F 8 v) .-0ELd Y 8Y rr G Y m E b Y .3 BG 1 -x Y Q 0 m ."E U 0 e, Ee, iE Y Y M 3c 3c M 0 5 v3 1. FISCALDEFICIT,PUBLICDEBT,ANDFISCALSUSTAINABILITY 1.1 St. Kitts and Nevis is a two-island federation in the Eastern Caribbean with a combined population of about 45,000 inhabitants (of whom 75 percent live in St. Kitts) and a GN1)of US$ 6,980 in 2003 (World Bank Atlas Methodology, current US$). The economy has traditionally depended on the production and export of sugar. However, in the last decade traditional economic activities such as agriculture have given way to a service-based economy. 1.2 St. Kitts and Nevis, as a member of the Eastern Caribbean Currency Union (ECCU), shares a common central bank, the Eastern Caribbean Central Bank (ECCB), and a common currency with the other five independent members of the Organization of Eastern Caribbean States (OECS) and the two U.K. territories of Anguilla and Montserrat.' Their currency is the Eastern Caribbean dollar, which has been pegged to the U.S. dollar since 1976 at EC$2.70 per US$1. The commitment to a fixed exchange rate and a prudent monetary policy under the ECCU umbrella has been a key element in maintainingan inflationrate close to international levels. 1.3 Notwithstanding a common central bank and a common currency, each member country conducts fiscal policy independently, and scant efforts have been made for a common fiscal policy in the region. Nevertheless, the deterioration of the ECCU fiscal position inrecent years, due mainly to the highfiscal deficits inAntigua, Dominica, and St. Kitts and Nevis, has reinitiated the discussion among its members of the need to coordinate fiscal policies to ensure the stability of the currency union.' 1.4 St. Kitts and Nevis's fiscal positionhas deteriorated sharply since the mid-1990s. The main reasons for this deterioration are increases in the wage bill and capital expenditure in response to hurricane repair, a weakening of the finances of the state- owned St. Kitts Sugar Manufacturing Company (SSMC), increased expenditures linked to legislative elections that took place in July 1995 and March 2000, and increased debt service. 1.5 Inlight of the precedingdiscussion, the mainobjective of this chapter is twofold: first, to evaluate the fiscal sustainability of the St. Kitts and Nevis Central Government (CG) and public sector in the short and medium term, and second, to propose various options for the adjustment requiredto achieve fiscal sustainability. 1.6 The mainfindings and recommendations canbe summarized as follows: * The OECS consists of six independent members: Antigua and Barbuda, Dominica, Grenada, St. Kittsand Nevis, St. Lucia, and St. Vincent and the Grenadines, and three dependent territories of the United Kingdom, Montserrat (full member), Anguilla, and the BritishVirgin Islands (associate members). The communiqut of the Special Meeting of the OECS Authority, Castries, St. Lucia, September 28,2001 alludes to the need to examine inthe medium term public sector reform, aimed at increasing efficiency, and fiscal reform including matters related to revenue, expenditure, financing, and debt management, as well as common approaches to taxation and incentives to be applied to cruise shipping. 0 The fiscal policy conducted by the government since the mid-1990s has been unsustainable inthe short and medium term, and will continue to be unsustainable in the next years unless the government adjusts its expenditures to the revenues mobilized. The size of the adjustment required to achieve fiscal sustainability of the CG in 2003 is about 14 to 16 percent of GDP, depending on the assumptions of real interest and growthrate. 0 The proposed fiscal adjustment relies mainly on expenditure cuts (in particular in the wage bill and containment of capital expenditures). The adoption of the tax policy recommendations suggested by the International Monetary Fund (IMF) would yield an increase inrevenue collection of only about 2 percent of GDP, in contrast to a required adjustment of about 14 to 16 percent to achieve fiscal sustainability . 0 If the public enterprises are included in the exercise, then the government will need to adjust additional 2 to 3 percentage points of GDP to ensure the fiscal sustainability of the public sector. A prompt decision by the government on the possible closing of the SSMC would be a step forward inthe fiscal sustainability of the public sector. I. ST.KITTSANDNEVIS FISCAL PERFORMANCE, Figure 1.1St Kittsand NevisCentralGovernmentFiscal 1990-2002 Performance 1990.2002 (as % of COP) 1.7 At the Central 5.00 Government level, since 1995 the 0.00 fiscal position of St. Kitts and -5.00 Nevis has progressively -10.00 weakened." As Figure 1.1shows, the St. Kitts and Nevis CG -15.00 followed a cautious fiscal policy -20.00' from 1990 to 1994. During this Swce:ECCEandhF period, the CG ran a primary surplus (after grants) in 1990-93, 0PrimaryFiscalBalance(aftergrants) r!T Current FiscalBalance and a small primary deficit in -Overall Fiscal Balance (after grants) 1994. However, from 1995 onward the fiscal situation worsened. In 1995, the CG ran a primary deficit of 3.3 percent of GDP; it was contained in 1996 and 1997, but rose again in 1998, and by 2000 was 9.4 percent of GDP. In2001, the CG reduced its primary deficit to 6.7 percent of GDP, but in 2002, the primary deficit reached 10.8 percent of GDP-the highest primary deficit since 1990. ~~~ ~ lo The source of the data used inthis chapter is the ECCB for 1990-1995, and the IMFfor 1996-2002. Two different sources are used because from 1996 onward the St. Kitts and Nevis government has recorded a significant amount of capital expenditures"below the line," transactions that the MF has moved "above the line" intheir fiscal figures, but which the ECCB has done only inselectedyears. 2 1.8 The main cause of the Figure 1.2 St Klttsand NevisCentralGovernmentPrlmaryExpenditure, deterioration of the Central InterestPaymentandRevenue 1990-2002 60,00 (as X of GDP) Government finances since 1995 has been an increase in 50.00 - HurncaneGustav September1995 government spending. At the August lggo 40.00 core of the surge observed in Central Government 30.00 expenditureare: 20.00 10.00 0 Increases in the wage bill and capital 0.00 expenditure in response 1990 1991 1992 1993 1994 1 -10.00 to hurricane repair. Since 1995, St. Kitts -20'oo Sourca ECCB and IMF and Nevis has been hit 0TotalPrimaryExpenditure E3InterestPayments . I FiscalBalance(abrgrants)+Total Pnmaty Revenue by three hurricanes: Luis in September 1995, Georges in November 1998, and Lenny in November 1999. 0 Increases in expenditures linked to legislative elections that took place in July 1995 andMarch2000. As Figure 1.2 shows, total non-interestexpenditures increasedfrom 22.4 percent of GDP in 1990to 32.8 percent of GDP in 1995,34.3 percent of GDP in 1998,39 percent of GDP in 1999, and to 44.7 percent of GDP in2002, while total revenue only increased from 24.7 percent of GDP in 1990to 29.1 percent of GDP in 1995, and about 31 percent for 1998-2002. 1.9 Because of the weakening of the CG fiscal Figure 1.3 St. Kittsand Nevis CentralGovernmentDebtandPrimarySurplus 1990-2002 position, the government (as % Of GDP) has turned to both external 120.00 and domestic borrowing, 110.00 100.00 mostly at commercial 90.00 80.00 terms, to finance its fiscal 70.00 deficit. Hence, the ratio of 60.00 50 00 CG debt to GDP increased 40.00 from 52.6 percent of GDP in 30 00 20.00 1990 to 104.4 percent of 10.00 1 + S I GDP in 2002. As Figure 1.3 0 00 -10.001 2 shows, from 1990 to 1994, -20.00 the ratio of CG debt to GDP -30 00 Sourca ECCBandIMF showed a negative trend, 0 DomesticCentralGovernment 0 ExternalCenbalGovernment reaching its lowest value in 0 PrimaryFiscal Balance (aRer grants) 1994 (46.4 percent). A year later in 1995, the debt-to-GDP ratio increased to 54.1 percent, and kept rising to peak at 104.4 percent in 2002. As a result, interest expendituresincreased from 2.1 percent of GDP in 1995 to 7.6 percent of GDP in 2002. 3 As will be illustrated inthe next section, inorder to maintainfiscal solvency, the CG will need to generate a sizeable primary surplus to pay the interest on the debt. 1.10 On average, during 1995-2001, domesticdebt representedaround 60 percent of total CG debt. Most domestic debt is of short-term maturity. Most public external debt is denominated in U.S. dollars and has a maturity of five or more years. Public domestic debt, on the other hand, is composed mainly of treasury bills that are by nature short term with high interest rates. In2002, domestic interest payments were 4.0 percent of GDP compared to 3.5 percent of GDP for foreign interest payments, even though both external and domestic debt representedabout 50 percent of GDP. InNovember 2002, the government of St. Kitts and Nevis raised EC$75 million in domestic debt through the issue inthe RegionalGovernment Securities Market(RGSM) of a 10-year debenturewith an interest rate of 7.5 percent per year, payable semiannually. The stated purpose of this bond was to refinance existing short-term debt and create debt of longer maturity. 1.11 The public sector fiscal performance is even inferior to that of the CG.' The ' financial operations of the public enterprises, mainly the losses in the SSMC and the St. Christopher Air and Sea Ports Authority, increased the CG overall deficit from 7.7 percent of GDP in 1998 to 12.6 percent of GDPfor the public sector (1998 is the last year with available reliable information). Public sector debt-to-GDP ratio in 1998 was around 101 percent of GDP compared to about 61 percent of GDP for the CG. This debt accumulation of the public sector reflects the losses experiencedby the public enterprises rather than investment. 11. DOES KITTSANDNEVISFACEFISCAL ST. A SUSTAINABILITY PROBLEM? 1.12 The figures on the primary deficit and debt presented above have raised serious concerns about the fiscal solvency of the government. We answer these concerns using an intertemporal approach to fiscal solvency. This approach, broadly speaking, defines a fiscal policy as sustainable if the present value of all future primary surpluses is enough to repay the current outstanding debt.12 1.13 In this chapter, following Blanchard, Chouraqui, and Hagemann (1990),the government repayment capacity is indicated by the gap (t*- t), where t* is the sustainable tax rate (defined as the tax rate which would achieve a stable debt-tc+ GDP ratio) and t is the ratio of tax revenue to GDP.13If the gap (t*-t) is positive, it I'The public sector is defined as theCGpluspublic enterprises. l2This approach includes seignorage as a source of revenue; however, the apportioning of seignorage to national economies has not been discussed in the ECCU, and it is not taken account in the fiscal sustainability exercise at the country level. However, any fiscal sustainability exercise at the ECCU level should incorporateseignorageas a revenue. r 1 l3t* is defined as t*=(r- 1 where E , is the mathematical expectation operator, gi is the ratio of non-interest government expenditure to GDP, r is the real interest rate, 8 is the GDPreal growth rate, bo is the initial debt-to-GDP ratio, and ( r - 0 ) is the discount rate. 4 signals the need for either tax increases and/or spending decreases to achieve a stable debt-to-GDP ratio. Conversely, if the gap (t*-t) is negative, then the intertemporal budget constraint is not binding and the government canreducetaxes and/or increase ~pending.'~ 1.14 We calculate gaps for two horizons: the short-term gap associated with a one-year horizon, and the medium-term gap associated with a four-year h~rizon.'~ The main advantage of the short-term gap is that it canbeconstructed easily, without the use of forecasts, and can serve as a useful benchmark. Its main weakness is that it is myopic and does not take into account future changes in government spending. For instance, the 2002 short-term gap indicates the adjustment required to stabilize the debt- to-GDP ratio at its 2001 level given the primary spending and tax revenue in 2002. The medium-term gap corrects this myopia by using forward-looking indicators for government spending and revenue. For instance, the 2002 medium-term gap indicates the adjustment required to maintain the debt-to-GDP ratio at its 2001 level given the primary spending and tax revenue pathfor 2002-05. 1.15 Both the short- and medium-term gaps show that since 1995 the fiscal policy conducted by the St. Kitts and Nevis CG and the public sector has been unsustainable. Moreover, the short-term size of the adjustment required to achieve fiscal sustainability of the CG in 2003 is about 14 to 16 percent of GDP, depending on the assumptions of real interest and GDP growth rate.16 Since the CG guarantees the debt of public enterprises, the government will need to adjust an additional 2 to 3 percentage point of GDP to ensure the fiscal sustainability of the public sector. l4 The index is symmetric in the sense that the treatment of taxes and spending is equivalent. The adjustment can come from tax increases, spending cuts, or a combination o f both. IsInterms of the formula presented infootnote 5, the short-term gap corresponds to i= 1, and the medium- term gap to i= 4. For the medium-term gap, a horizon o f four years was chosen because it is the time span of the Medium-Term Economic Strategy Paper prepared by the country. l6 Indeed for the fiscal sustainability exercise the key concept is the discount rate, defined as the real interest rate minus real GDP growth rate. Inthis fiscal sustainability exercise two discount rates are used, 1 percent and 3 percent. Both are consistent with an annual nominal interest rate of 7.5 percent, an annual inflationrate o f 2 percent, and a path for real GDPgrowth of about 2.5 to 4.5 percent. 5 Central Government 1.16 We can distinguish Figure1.4 S t Kilts andNevisCentralGovernmentShort-TermFiscalGap 1991-2002(asZ GDP) two periods in the short-term (one- year horizon) fiscal 16M 1 1400. policy conducted by the CG. E The first period, 1991-94, is 8 l 2 O 0 - characterized by a ae n 31000- sustainable fiscal policy and consequently decreasing 8- am1 ; debt-to-GDP ratio." Figure 4oo 1.4 shows that the size of the f fiscal adjustment required to c s 2M maintain a stable debt-to-GDP O M ratio was on average negative at 200-e a discount rate (real interest rate I-, -9 OShot.Tem Gap (Discountrate3 5)UShohTennGap (DiscountRate 1 8)- T O ~ I CentralGovernment Deb1 minus real GDP growth rate) of 1percent, or positive but small at a discount rate of 3 percent. That is, the CG could have reduced taxes and/or increased expenditures slightly at a discount rate of 1percent. 1.17 The second period, 1995-2002, is characterized by an unsustainable short- term fiscal policy and therefore an increasing debt-tWGDP ratio. Every year since 1995 the fiscal adjustment required to achieve a sustainable debt-to-GDP ratio was positive and sizeable at discount rates of 1 and 3 percent. In particular, for 2002 the adjustment required to stabilize the CG debt-to-GDP ratio (to its 2001 level) was about 12 percent of GDP at a discount rate of 1percent, and 13 percent of GDP at a discount rate of 3 percent. That is to say, government expenditures neededto be cut andor taxes to be increasedby 12 to 13 percent of GDP inone year inorder to achieve a constant debt- to-GDP ratio.18 1.18 For 2003-05, the short-term fiscal gap projections show that the fiscal policy will remain unsustainable if the CG keeps its primary expenditure and tax revenue at the level observed in 2002.19 As Figure 1.5 shows, in 2003 the adjustment required from the CG to stabilize its debt-to-GDP ratio at its 2002 level is between 14 and 16 percent of GDP, depending on discount rate assumptions. That is, government expenditures need to be cut and/or taxes to be increased by 14 to 16 percent of GDP. If by 2005 no adjustment has taken place, the debt-to-GDP ratio will increase to about 154 percent of GDP, and the adjustment requiredwill be 14 to 17 percent of GDP, depending on discount rate assumptions. "Theshort-termgapisconstructedbyusingthefollowingequation:(t*-t) ( = r -8 )bo-si. Blanchard, Chouraqui, and Hagemann(1990) present estimates of the short-term and medium-termgap for OECD countries that fluctuate between -3.5 percent of GDP in 1989 for the U.K., and 6.4 percent of GDP in 1985 for Italy. l 9The projections for the scenario without IMF tax recommendations assume a primary level of expenditures and revenues in 2003-05 similar to the 2002 level. The scenario with IMF tax recommendationsassumes a tax revenue-to-GDPratio higher by 2 percentagepoints in2004 and 2005. 6 1.19 The adoption of the Figure 1.5 St. Kitts and Nevis CentralGovernmentProjected Short-Term IMF tax policy FiscalGap 2003-2005 (as% COP) recommendations will reduce 18M Wimout IMFT u Recommendations the short-term fiscal gap only Wlth IMFTax Recommendations slightly. Inorder to achieve a d ISM 0 14M sustainable fiscal policy the ae212M fiscal adjustment will have to !!low Q rely mainly on expenditure g 8M cuts. As Figure 1.5 shows, the E 2 E 600 adjustment required in 2005, if 4M the IMF tax policy a e 2M recommendations are adopted, om is between 12 and 16 percent 2003 m m 2033 Source Authots calculanons of GDP compared to 14 to 16 percent of GDP without the adoption of the IMF tax recommendations. This size of adjustment is high (equivalent to a primary surplus of between 3 and 4 percent of GDP) and it would be difficult to achieve inthe short term even with a major fiscal reform.20 1.20 The proposed adjustment for the CG, 14 to 16 percent of GDP, will only stabilize the debt-to-GDP ratio to about 100 percent of GDP in future years. This type of debt sustainability analysis does not address the optimal level of indebtedness, its composition, or politicaleconomy considerations. Nevertheless,today St. Kitts and Nevis has one of the highest debt-to-GDP ratios in the world, and interest expenditures on its debt amount to about 7 to 8 percent of GDP. 1.21 The target of debt-to-GDP ratio for the CG of 60 percent has been proposed by the ECCB's Monetary Council and would require that the CG generate consistent primary surpluses of about 20 percent of GDP for the next 4- 5 years. However, these figures for primary surplus may be politically unfeasible, even with growth rates of 6 to 7 percent of GDP. 1.22 The use of government liquid assets (cash plus tradable shares) can only partially contribute to reduce the debt stock. This is becauseas of December 2001, the ratio of liquid assets to debt was only 7 percent.2'0ther assets (land, port, etc.) could be used to reduce the debt stock. But any privatization should not be driven by revenue considerations alone, but rather by sectoral efficiency criteria, including establishing an adequate regulatory framework to ensure that public monopolies are not replaced by unregulated and inefficient private monopolies. The short-term gap in 2003 with and without implementation of the IMF tax recommendations is the samebecause we assume that the increaseinrevenuecollectionwill start in2004. *' The liquid assets considered were cash (EC$ 14.8 million as of December 2002) and shares in the St. Kitts-Nevis-AnguillaNational Bank, the St. Kitts Tourism Development Corporation, the Caribbean Shoe Manufacturers,the Leeward Island Air Transport, CARICOMCorn SoyaBean Company, the Development Bank of St. Kitts and Nevis, the St. Kitts and Nevis Telecommunications, the St. Kitts and Nevis Cable Communications, and the St. Kitts Urban Development Corporation, all with a total nominal value of EC$37 millionas of December 2001. 7 1.23 Debt restructuring is an alternative option to reduce the CG debt stock, but the composition of the debt complicates any possible restructuring effort. In 2002 about 50 percent of the debt was domestic and of short-term maturity, mainly treasury bills, and the main holders were the financial sector of St. Kitts and Nevis and the regional domestic banks.External debt has a maturity of five or more years and is heldby multilateral institutions. 1.24 The medium-term Figure 16 St Kitts and Nevis CsntralGovernmentMedium-TermFiscal Gap 1991-2002 fiscal (four-year horizon) (as% GDP) 18m gap is consistently above the short-term gap, c 16m showing the severity of $ 14M the fiscal situation ahead 812, of the one-year horizon.22 Figure 1.6 shows that -8 ; contrary to the short-term '0° gap, the size of the fiscal e5P c 6~ adjustment during 1991-94 Om was positive and sizeable at around 4 percent of GDP at E 2(JD a discount rate of 1percent, om 1931 1992 1993 IS94 1935 1596 1997 1998 1% 2000 2m1 202 and almost 5 percent of Source ~ ~ tcalculations ~ h ~ f GDP at a discount rate of 3 UMedium-TerrnGap (Dircount rat^ 3 %) E3Medium-Term Gap (Dacounl Rate 1 %] -Total Central Government Deht percent. For 1999, the last year that a fair comparison can be made between the short- and medium-term gap, the medium-term gap is about 4 percent above the short-term gap.23 That is to say, in 1999, given the primary government expenditure path from 1999 to 2002, CG expenditures needed to be cut and or taxes to be increased by 11percent of GDP at a discount rate of 2 percent, instead of the 9 percent of GDP suggested by the short-term gap. 1.25 How sensitive are the estimated gaps to the underlyingassumptions?Clearly, the medium-termgap is only as good as the forecasts on which it i s based. However, both gaps depend on the value of the discount rate, (r-0). Increasesinthe real interest rate or slowdowns in growth increase the size of the adjustment and have a bigger impact if the initial level of debt is high. 22 The medium-term gap is given by (t*-t)=average over the current and next three years of gi+(r -0 )bo- t. The lastyear that actual values can be used to estimatethe medium-term gap is 1999. Medium-termgaps from 2000 to 2002 use forecasted values for the primary expenditure in2003-05. 23 In constructing the historical series for the short-term gap, actual values of non-interest spending and taxes have been used until 2001;beginning with 2002 forecasts are incorporated. The historical medium- term gap uses actual values of non-interest spending and taxes until 1998; beginning with 1999 forecasts are incorporated. 8 Public Sector 1.26 If the public enterprises are included in the exercise, in particular the losses of the S M C C and the St. Christopher Air and Sea Ports Authority, then the government will need to adjust in the short term an additional 2 to 3 percentage points of GDP to ensure the solvency of the public sector. For 1998, the last year with available reliable information on the primary balance of the public enterprises, the adjustment required to achieve fiscal sustainability in the public sector is about 2 percentage pointshigher than the one requiredby the CG at a discount rate of 1percent, and around3 percentagepoints at a discount rate of 3 percent. Fiscal Rules in the ECCU 1.27 The ECCB's Monetary Council that took place in February 2003 reconfirmed the agreed fiscal framework that includes ceilings on debt and fiscal performance of the CG.24The aim of the ECCB in setting these fiscal rules i s to ensure long-run fiscal sustainability and the credibility of the currency union by limiting negative spillovers. However, as we discussed in paragraph 1.3, the ECCB has no mandate over fiscal policy inany member country, and the fiscal rules are not binding at this moment for the member states. Box 1.1 presents the convergence criteria in the ECCU and the Western Africa Economic and Monetary Union (WAEMU). The- WAEMU, like the ECCU, comprises developing countries with a currency pegged to only one single currency, the Euro. 1.28 In 2002, St. Kitts and Nevis did not meet any of the four debt and fiscal performance ceilings proposed by the ECCB's Monetary Council. Every year since 1991, the CG has run a current account surplus below the 4 to 6 percent of GDP target. The target for overall deficit has been reached inonly 3 of the last 12 years. The debt-to- GDP target was achieved every year Until 1998, and the debt service-to-current revenue target has not beenmet since 2000. 1.29 At the time of the reconfirmation of the fiscal framework agreement, the ECCB's Monetary Council regarded 2007 as a critical date for fiscal consolidation. This deadline for convergence seems unrealistic based on the fiscal performance of St. Kitts and Nevis in2002: (a) current deficit of about 5 percent of GDP, (b) overall deficit of 20 percent of GDP, (c) debt-to-GDP ratio of 104.4, and (d) debt service to current revenue of 24 percent of GDP. Indeed, this four-year horizon would require a fiscal adjustment higher (primary surplus of about 20 percent of GDP) thanthe one discussedin the previous section (primary surplus of 3 to 4 percent of GDP) if the objective is to meet the proposed 60 percent of GDP debt ceiling by 2007. Moreover, the WAEMU experience to meet the primary criteria in a three-year span was unsuccessful and it was extended three additional years (to end-December 2005). 24The methodology used to set the deficit and fiscal ceilings is discussed in two ECCB documents, "Developing Sustainable Debt Indicators for the ECCB," July 2001; and "The Targeted Public Sector Savings inthe member countriesof the ECCB" (undated). 9 Box 1.1The Convergence Criteria inthe ECCUand WAEMU EasternCaribbean CurrencyUnion(ECCU) The ECCB has put forward a proposal that comprises the following four targets to be achieved by 2007: Central Government current account surplus of 4 to 6 percent o f GDP; Overall Central Government budget deficit o f no more than 3 percent of GDP; Total outstanding Central Government debt of no more than 60percent o f GDP; and Debt service payments by the Central Government of no more than 15 percent o f current revenue. Western African EconomicandMonetaryUnion(WAEMU)* The regional Convergence, Stability, Growth, and Solidarity Pact adopted in December 1999 by the WAEIvfU has four primary convergence criteria and four secondary criteria. The norms established by these criteria had to be met by 2002. The primary criteria are: Ratio o f the basic fiscal balance to nominal GDPmust be 0 percent or more; Ratio o f outstanding domestic and foreign debt to nominal GDPmust not exceed 70 percent; Average annual inflation cannot be more than 3 percent a year; and Nonaccumulation of domestic and external payment arrears in the current fiscal period. The secondary criteria are: Ratio of the wage bill to tax revenue cannot exceed 35 percent; Ratio of domestically financed public investment to tax revenue mustbe at least 20 percent; Ratio of current external deficit excluding grants to nominal GDPcannot exceed 5 percent; and Tax-to-GDP ratio must be 17 percent or more. However, because o f the limited progress achieved by the member co,untries in meeting the convergence criteria at end-December 2002, the WAEMU Commission proposed to the heads of state to extend the timetable to end-December 2005. * The eight members of the WAEMUare Benin, BurkinaFaso, CBte d' IvBire, Guinea-Bissau, Mali, Niger, Senegal, and Togo. 1.30 The successful achievement of debt and fiscal ceilings needs the simultaneous introduction of a system of incentives and penalties against noncompliance, and a mechanism for monotoring and surveillance of the fiscal performance. The ECCB has not yet specified any system or mechanisms to ensure compliance with the proposed fiscal ceilings. Moreover, the WAEMU example suggests that survaillance of the fiscal ceilings and close monitoring against annual targets during the convergence phase is key to achieving fiscal convergence. 1.31 The ECCB-proposed fiscal ceilings only cover the CG, while public enterprises and government guaranteed debt have been excluded. However, as the fiscal sustainability exercise conducted in the last section suggests, the public sector position is critical and should be included in any proposal of deficit and debt ceilings. Moreover, the inclusionof the public sector into the fiscal framework would ensure that the government would produce accurate and timely information on the financial position of the public enterprises and its debts. 10 1.32 The deficit and debt ceilings do not take into account the cyclical nature of the fiscal indicators in small countries subject to frequent external shocks. In this regard, we propose to move from debt and fiscal performance ceilings to a countercyclical fiscal policy based on a simple fiscal rule for the structural balance. This rule has been applied successfully in Chile, a country much more vulnerable to external shocks than St. Kitts andNevis, and is the topic of Chapter 2 of this report. 111. FISCAL CONSOLIDATIONINST. KITTSAND NEVIS'S PUBLIC SECTOR 1.33 In the previous section, we showed that the public sector primary surplus needs to be increased in the next years to ensure fiscal sustainability. W e propose that the fiscal adjustment rely primarily on expenditure cuts, in particular in the wage bill, and containment in capital expenditures. Tax increases are seen only as a small fraction of the adjustment. As discussedbelow, this type of adjustment is based on the observed pattern of government expenditures and on the assumption that a tax reform will yield an increase in tax collection of only about 2 to 4 percent of GDP, and on several studies of how the composition of the fiscal adjustment influences its success.25 TaxRevenues 1.34 The main characteristics of the St. Kitts and Nevis tax structure in the last decade can be summarized as follows (see Figure 1.7): Current revenue (tax and non-tax revenue) has remained constant at about 30 percent of GDP in the second part of the decade compared to 22 percent of GDP during 1990-94. About 80 percent of Figure 1.7 St. Kitts and Nevis CentralGovernment Revenue Composltion 1990. tax revenue i s 2002 (as %Totallax Revenue) derived mainly lm% from two types of 90% taxes: taxes on a% international trade 70% and taxes on 60% income and profits. yJ% 40% The tariff reduction 30% under the 1993 zo% agreement on the introduction of a ox common external 1990Source1991 ECCB and IMF 1992 1993 1994 1995 1996 1997 1998 1999 2XC XOl 2002 tariff (CET) Taxes on InternationalTrade fiTaxesonIncome8, Profits reduced the share of Taxes on Dornedc Goods &Services Taxes on Property import duty taxes on tax revenue from 26 percent during 1990-94 to 21 percent during 1995-2002. 25See, for example, the studies by Alesina and Perotti (1996) and Perotti (1996) for the OECD countries. 11 0 Taxes on income and profits represented 26 percent of tax revenue during 1990- 2002. There is no tax on personal income, although income from employment is subject to taxation through the social services levy, which is a payroll tax. The absence of a personal income tax, although rare in other developing countries, is a feature present in a few other OECS countries. Consequently, the main source of revenue inthis category is corporate income tax, about 17 percent of tax revenue, and the payroll tax. 0 Taxes on domestic goods and services increasedas a percentage of tax revenue from 16 percent during 1990-94 to 20 percent during 1995-2001. This category includes a number of excises and licenseson specific goods and services; the largest ones are the tax on hotels and services, the gasoline levy, and the stamp duty. The gasoline levy, contrary to other taxes on domestic goods and services, is the residual of the retail price (fixed by the government) minus the cost, insurance, and freight (CIF) price of the imported product and the fixed margins for importers and retailers. This net tax can be negative and fluctuates with the CIF price if the retail price is not adjusted. The experience inSt. Kitts and Nevis is that after an increase inthe CIF prices, retail prices are kept fixed and net tax absorbed the increase. The government should introduce an automatic pricing and taxation mechanismfor adjusting petroleum prices inline with international price movements. 0 Property taxes on the rental income from houses and on land contribute a modest 2 percent of the total tax revenueintake. Non-tax revenue representedroughly 8 percent of GDP between 1990 and 2002. Capital revenue and grants contributed only 1 percent of GDP inthe same period. 1.35 During 1996-2000, the revenue forgone by the St. Kitts and Nevis Central Government in customs duties, due to tax concessions, was on average 18 percent of of tax concession^.^^ As Table 1.1 shows, hotels and tourism received the highest GDP?6A key feature of the presenttax systeminSt. KittsandNevis is the pervasive use proportion of tax credits followed by special agreements and fiscal incentives. In the extreme case, for example, where all exemptions to hotel and tourism were repealed, a revenue gain of 6 to 8 percent of GDP would be expected, assuming zero elasticity of imports with respect to tariffs. 26 ''Anelaborate All tax concessionspay the so-calledconsumptiontax onCIFimports. system of import duty exemptions provides for very generous concessions for taxes on imports for the development of industry, tourism, agriculture, social, educational, cultural, and other purposes. The main exemptions are provided under the Customs Act, the Fiscal Incentive Act, and the Hotel Aid Ordinance.Also, some investorsbenefit from income tax holidays. 12 Table 1.1Customs Revenue Forgone from Tax Concessions(as percentage of GDP) Fiscal and Government and 1996 4.59 8.22 0.57 0.19 4.29 0.74 0.59 19.21 1997 4.27 7.80 0.61 0.26 4.5 1 0.34 0.40 18.19 1998 4.26 6.84 0.62 0.20 4.20 1.23 0.65 18.01 1999 3.94 6.94 0.94 0.21 5.21 0.37 0.41 18.01 2000 3.61 7.98 0.56 0.18 4.46 0.61 0.40 17.79 Source: St. Kitts and Nevis Customs. 1.36 The main purpose of tax concessionsinSt. Kitts and Nevis, as in other OECS countries, is to attract foreign investment. However, the empirical evidence suggests that taxes are not among the main determinants of foreign investment. Foreign investors place more value on a stable economic environment with transparent rules than on tax incentives.28 The most suitable approach to address the concerns of tax competition among OECS countries is to support more forcefully a program of fiscal policy convergencethat includes tax harmonization among member states. 1.37 A complete discussion of a medium-term tax policy strategy is beyond the scope of this report. The IMF Fiscal Affairs Department has produced two reports on the subject and estimates that the revenue implications of its tax recommendations will result in an increase in revenue collection between 2 and 4 percent of GDP (taking into account the introduction of phase 4 of the Common External Tariff [CET])?9 The main recommendations of these reports can be summarized as follows: Consumption taxes. Strengthening the existing consumption tax structure by converting the present consumption tax on goods into a full manufacturer-level sales tax, operated with a credit mechanism (as in a VAT). Creating new excise taxes for alcohol, tobacco, and fuels, andmakingsome adjustments to the other taxes leviedon services. Direct taxation. Reduction inthe rate of corporate income tax, increase inthe rate of the social services levy, the introduction of a simple final withholding tax on rents, royalties, and interest, and extension of the corporate income tax by making it a tax on enterprise profits (including those of larger unincorporated enterprises). Tax concessions. The efficiency of concessions granted should be analyzed and appraised on a regular basis; tax concessions on import duties and the consumption tax on imports should be significantly curtailed, and broadly restricted to imports by the government and imports of capital equipment for well-defined projects. 28See Zee, Stotsky and Ley (2002) and the references therein. 29See Kingand Fulford (2001); and Norregaard and Isaac (1996). 13 e Tax administration. Strengthening the areas of enforcement and procedures, audit, and coordination betweenthe revenuedepartments. The major implication on the revenue side is that there is scope for an increase in revenue collection of only about 2 to 4 percent of GDP; the rest of the adjustment inthe primary surpluswill needto come from expenditure reduction. Central Government Expenditures Figure 1.8 St. Kitts and Nevis Central Government ExpenditureComposition 1930-2002 1.38 CG expenditures, as a (as % of GDP) percentage of GDP, increased 60.00 from about 25 percentof GDP 50.00 during 1990-94 to 40 percent of GDP during 1995-2002. In 40,00 real terms, the CG 30.00 expenditures doubled in the 20.00 same period. As Figures 1.8 and 1.9 show, the largest part of the increasewas due to personal emoluments and wages (6 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 percentage points of GDP) and 0 PersonalEmolumentsandWages mOther Goodsand Seriices 13 capital expenditures (5 Tmsfers andSubsidies 0CapitalExpenditure 0 InterestPayments percentage points of GDP). These increaseshave ledto an unnecessarilylarge government. 1.39 Personalemoluments and wage expenditures remained stable from 1991to 1994, at aroundEC$48 million (in constant 1995 EC dollars, or 9 percent of GDP). However, during 1995-2002, the wage bill doubledto EC$103 million (in constant 1995 EC dollars, or 15 percentof GDP). At the core of the large increasesobserved in the wage bill in real terms and as a percentage of GDP i s an increase in the number of established and non-established positions in the Central G~vernment.~'Government's real wages remained flat during this period. Thus, any effort to reduce the wage bill will have to focus on a rationalization of the number of public servants in the CG. The existing number of established and non-established positions is unnecessarily large because much work is duplicated across ministries, departments, and agencies. In addition, there is considerable scope for savings stemming from divesture and subcontracting to the private sector. (See Chapter 5 of this report.) 30 There are two types of governmentpositions in the St. Kitts and Nevis Federal Government: established and non-established. Employees in established positions are appointed by the Governor General upon formal request by the Public Service Commission, with the approval of the Establishment Division in the Prime Minister's Office. Established positions are permanent positions listed in the budget, and their compensation is classified as Personal Emoluments. Employees in non-established positions are hired directly by line ministries, their numbersare notpresentedinthe budget, and only their wages are included as a line item. 14 1.40 Capital expenditure Figure 1.9 St. Kittsand Nevis CentralGovernmentExpenditureComposition 1990-2002 presents a pattern similar to (in millionsECS95) the wage bill. From 1990 to 450.00 1 1994, capital expenditure 400.00 - fluctuated around 3 percent of 350.00 - GDP (approximately EC$17 300.00 - 250.00 - million in constant 1995 EC 200.00 - dollars). As in the case of the 150.00 - wage bill, a ruptureoccurred in 100 00 1995. During 1995-2002, 50.00 capital expenditure as a 12!:e2cMri.92 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 fraction of GDP increased 2.5 0PenonalEmolumenkandWages 0OtherGoodsandServices times, and in real terms 3.5 0 TransfersandSubsidies nInterestPayments times, compared to 1990-94. m Capital Expendibrre The main explanation for these increaseshas been the damage causedby Hurricanes Luis inSeptember 1995, Lenny inNovember 1998, andGeorges inNovember 1999.However, a close look at the Public Sector Investment Program (PSIP) shows that hurricane-related expenditures were not the only cause of the increases in capital expenditure. Investment projects inconsistent with the growth strategy defined in the Medium-TermEconomic Strategy Paper (MTESP) and projects beyond the ones required to repair the damage caused by the hurricanes were carried out without any cost-benefit analysis and financed domestically at high interest rates. Therefore, tighter control of the PSIP is crucial to improve the efficiency of the public investment and will help the return of the government to a fiscally sustainable path. (See Chapter 4 inthis report.) 1.41 Expenditureson goods andservices during1990-2002 were stable at around 9 percent of GDP, but transfers and subsidies doubled from 1.6 percent of GDP during 1990-94 to 3.6 percentduring1995-2001. The latter is due inpart to subsidies givenby the CG to the SSMC. 1.42 Finally, interestpaymentsas a percentageof GDPalmost doubledfrom 1990 to 2002. This reflects two factors: first, the higher indebtednessof the St. Kitts andNevis government, and second, the increased reliance of the government on borrowing domestically at higher costs. Moreover, if the government continues its fiscal policy on an unsustainable path, interest payments will start to crowd out the other components of government expenditures. Public Enterprises and the Social Security System 1.43 St. Kitts and Nevis has a number of public enterprises,but a detailed list of all of them and the law that created them has been impossible to obtain from the authorities?'Up-to-date financial statements for these enterprises are not available, and the National Housing Corporation (NHC) has not made public its financial statements 31We know of the existence of at least six public enterprises: SSMC, the Central Marketing Corporation (CEMACO), the Frigate Bay Development Corporation, St. ChristopherAir and Sea Ports Authority, the DevelopmentBank, and the NationalHousingCorporation(NHC). 15 since its creation in 1996.32Nevertheless, transfers from the CG to the public enterprises have been substantial andunsustainablegiven the fiscal position of the CG. 1.44 Despite this scant information, in 1998 public enterprises added 5 percentage points to the overall CG fiscal deficit. Thus, the overall public sector fiscal deficit including public enterprises was 12.6 percent of GDP compared to 7.7 percent of GDP for the Central Government. The main contributors to the deficit of the public enterprises in 1998 were the St. Christopher Air and Sea Ports Authority and SSMC. The St. Christopher Air and Sea Ports Authority is estimated to have run a deficit of 2.3 percent of GDP, and the SMMC a deficit of 2.8 percent of GDP. The increasing financial difficulties that the SSMC is facing are the result of a drop in productivity and adverse external shocks. Inlight of the bleak financial prospectsfor the industry, the government is discussing the options presentedina World Bank-funded study that examines the costs and benefits of maintaining the industry, closing the industry, or intermediate options. It i s essential that the government make a decision as soon as possible with respect to the possible closing of the SSMC before it further compromises the already unsustainable fiscal positionof the government. 1.45 The finances of the Social Security Scheme (SSS) remained sound with a surplus of roughly 4.6 to 5 percent of GDP during 1995-98. The SSS is a pay-as-you- go system. The benefits are financed by current contributions, 10 percent of wage income, with 5 percent paid by the employer and 5 percent by the employee. The overall SSS surpluses are indirectly offsetting the deficit of the public enterprises. However, this is a myopic view because inthe mediumterm the finances of the SSS will become more strained as the number of retirees begins to rise. In the medium term, a review of the social security system will be needed for St. Kitts and Nevis, as well as for the other OECS countries. 32 The NHC assists low-income families by constructing dwellings on land obtained from the government at no cost. 16 2. FISCAL POLICY AND BUSINESS CYCLES 2.1 The Latin America and the Caribbean (LAC) Region, like other developing regions, faces a high degree of volatility inreal output. Inthe last 40 years, the standard deviation of the growth rate of real GDP inthe LAC regionhas been twice as high as in industrial economies. Moreover, smaller economies such as the ones considered in this report tend to experience, on average, higher volatility than larger economies.33 2.2 The importance of economic volatility in long-term growth and its effect on the poor segments of the population are at the center of the need to identify policies and institutions that can ameliorate their impact. Volatility tends to discourage long-term investments in physical and human capital and thus leads to a lower growth potential. In addition, the poor segments of the population are generally the most affected because they lack the means to protect themselves from adverseemployment and income effects. 2.3 The government has a broad set of instruments available that it can use to reduce aggregate volatility and its effects. One of these instruments is fiscal policy. Inindustrial economies, fiscal policy is usually designedto stimulate output when the economy moves into recession and to contract output when an expansion takes place. A fiscal policy designed inthis way leads to a strongly procyclical budget surplus (countercyclical fiscal policy). 2.4 Inindustrialeconomies, a countercyclical fiscal policy is usually accomplished in two ways: first, by having components in the budget that respond automatically to the business cycle, such as tax revenues (which respond positively) or unemployment benefits (which respond negatively), and second, by using discretionary components in the budget to provide a stimulus duringbadtimes. 2.5 This chapter first describes the cyclical features of government finances in St. Kitts and Nevis by applying the Hodrick-Prescott methodology, an algorithm that decomposes a time series into cycle and trend. Second, it suggests the implementation of a countercyclical fiscal policy based on a fiscal rule that uses the structural (cyclically adjusted) balance as the indicator of fiscal policy. The application of this simple fiscal rule can allow the government to accumulateresources ingood times and runthem down inbadtimes. 2.6 The mainfindings andrecommendations can be summarized as follows: 0 The cause of the sharp deterioration in the fiscal position of St. Kitts and Nevis since the mid-1990s i s a long-term growth rate in expenditures, specifically personal emoluments and wages and capital expenditures above revenues, and not cyclical factors such as hurricanes. 33See De Ferranti and others (2000); Crucini (1997); and Head(1995). 17 0 St. Kitts and Nevis's fiscal policy is acyclical. That is, the primary, overall, and current fiscal balance does not co-move with the GDP cycle. In good times the government does not run a higher or lower fiscal surplus than its trend. This finding is contrary to the empirical evidence for developing countries (procyclical fiscal policy). 0 The structural fiscal balance measures confirm that the deterioration in the fiscal position in St. Kitts and Nevis was not due to the business cycle. In order to enforce fiscal discipline and move the country to a counteryclical fiscal policy, we propose a fiscal rulebasedon a structural surplus of about 2 to 4 percent of GDP. This rule implies a fiscal adjustment similar to one suggested inChapter 1(14 to 16percent of GDP). I. THENUSANDCYCLESIh'ST.KITTSANDNEVISFISCAL ACCOUNTS, 1980-2001 2.7 I s the businesscycle the main cause of the fiscal sustainability problemfacing the St. Kitts and Nevis government? This section attempts to answer the question by examining the cyclical fluctuations inGDP and government fiscal accounts from 1980 to 2002 using yearly data.34The methodology used to disentangle trend and cycle is the Hodrick-Prescott (HP) filter.35 The empirical evidence confirms the results obtained in Chapter 1: the main cause of the deterioration of the fiscal position inSt. Kitts and Nevis i s a long-termgrowth rate in expendituresabove current revenue. The business cycle had little influence on the sharp deterioration of the fiscal position. TlieBiisiness Cycle in St. Kitts nrzd Nevis 2.8 Since the 1990s' St. Kittsand Nevis's long-term GDP growth rate has slowed to 4.2 percent per year from 6.3 percent in the previous decade. As Table 2.1 shows, St. Kitts and Nevis's GDP trend grew by 5.1 percent per year from 1981 to 2002. This long-termgrowth rate is the result of two distinct periods. From 1981to 1990, GDP trend grew by 6.3 percent per year, while from 1991 to 2002 GDP trend grew at only 4.2 percent. This reduction in the long-term growth rate is a fundamental change in the economic environment that the government should take into consideration in future preparationof the budget andthe Country Strategic Outlook. 34 The choice of the sample period was largely driven by the availability of data. The ECCB provided us with disaggregated governmentfiscal accounts and real GDPonly since 1980. 35 In layman's terms, the HP filter can be thought of as removing a smooth trend (like one would draw freehand) from the data. A completediscussionof this methodcan be found inAppendix A. 18 2.9 The St. Kitts Figure2.1 St. Kith and Nevis GDP,GDP Trend and CyclicalComponent1980-2002 and Nevis GDP cyclical component 750 7 GDP and GDPTrend (In millionsECS 95) shows no clear link to hurricanes, but several expansions and contractions are worth noting. As Figure 2.1 shows, between 1980 and 2002, St. Kitts and Nevis was hit by six hurricanes. In three out of six hurricane years, St. Kitts and Nevis GDP Deviationsfrom Trend the St. Kitts and Nevis (as percentage) GDP decreased below 10% its trend, and in the other three years 6% increased above its trend. The deepest contraction took place in1983,ayear after the debt crisis and the -4% i If U greatest expansion in Hurricane Klaus,November1984 HurricaneLuis, September 1995 1989, the year of the HurricaneHugo,November1989 HumcaneGeorger,November19 1 Hurricane Lenny,November1999 Hurricane Hugo. In Source:Author`s calculationsbasedon ECCBand IMFdata 1983, St. Kitts and 0cycgdpreal Nevis's GDP decreased by almost 6 percent in comparisonto its trend, and did not recover until 1986. From 1986 to 1990, the St. Kitts and Nevis economy grew by about 2 percent above the GDP trend, with the highest expansion in 1989, 3.6 percent. From 1990 to 2002, St. Kitts and Nevis's GDP followed closely its trend with the exception of 1991, 1992, 1997, and 2002. In 1997, GDP grew about 2.3 percent above its trend, and in 1991, 1992, and 2002 about 2.1 to 2.6 percent below its trend. The contraction in 2002 mainly reflects the effect of September 11 on tourism andthe global economic slowdown. 2.10 GDPvolatility inSt. Kittsand Nevis was lower inthe 1990scompared to the 1 9 8 0 ~Volatility, measured by the standard deviation of the cyclical component of . ~ ~ GDP, was about 2.1 percent during 1980-2002,2.7 percent during 1980-89, and only 1.7 percentduring 1990-2002. 36This reduction in volatility in the 1990shas also been observed inLatinAmerica and the Caribbean by Loayza, Fajnzylber, and Calder6n (2002) for output per capita, and by the "Caribbean Economic Overview 2002: Macroeconomic Volatility, Household Vulnerability, and Institutional and Policy Responses" for gross nationaldisposableincome. 19 2.11 The main cause of the deterioration in St. Kitts andNevis's fiscal position is a long-term growth rate in expenditures higher than revenues, and not cyclical fluctuations. As Table 2.1 shows, during 1981-2002, the current and capital expenditures trend grew on average at 5.9 percent and 19.1 percent per year, respectively, while current revenue trend increased at only 4.9 percent per year. Moreover, the data show two distinct periods. From 1981 to 1990, the long-term growth rate in current revenue (3.1 percent) was slightly above current expenditure (2.9 percent), but during 1991-2002 the pattern was reversed. From 1991 to 2002, the long-term growth rate in current revenue (6.4 percent) was significantly below current expenditure (8.3 percent), leadingto the difficult fiscal positionthat the government faces today. Table 2.1 Long-TermGDP, CentralGovernmentRevenue, andExpenditureGrowthRates, 1981-2002 Average GrowthRate 1981-2002 1981-1990 1991-2002 Current Revenue 4.9 3.1 6.4 Tax revenue 4.9 3.1 6.4 Tax on income andprofits 8.6 9.4 7.9 Tax on property 13.6 20.8 7.5 Tax on domestic goods and services 12.2 16.0 9.0 Tax on international transactions 2.5 -0.3 4.8 Non-tax revenue 4.9 3.1 4.3 Capital Revenue 11.2 24.1 I.5 Grants -2.7 -12.0 5.1 Current Expenditure 5.9 2.9 8.3 Expenditure on personalemolumentsand wages 6.6 4.1 8.9 Expenditure on other goods and services 2.1 -0.6 4.4 Expenditure on transfers and subsidies 11.3 -1.5 13.4 InterestPayments 13.7 15.7 11.9 Capital Expenditure 19.1 26.5 14.8 GDP 5.1 6.3 4.2 Note: Statistics are basedon Hodrick-Precotttrend data. Growth rates are first differences inthe trend of the variable. 2.12 On the expenditure side, personal emoluments and wages and capital expenditures drove the highlong-term growth rate of total expenditure. As Table 2.1 shows, during 1981-2002 the trend of personal emoluments and wages grew by 6.6 percent per year and capital expenditures by 19.1 percent per year, with personal emoluments and wages representing about 37 percent of total expenditures and capital expenditures about 12 percent. Thus, these results validate the conclusions of Chapter 20 l-personal emoluments and wages and capital expenditures are the main sources of the fiscal problem facing the country. Moreover, Chapter 5 will restrict the large increase in the long-term growth inpersonal emoluments and wages to an increase inthe number of civil servants rather than an increase in wages. The third determinant of the high long- term growth rate of total expenditure was interest payments. The trend of interest payments grew at a rate of 13.7 percent during 1981-2002 and represented about 10 percent of total expenditures.The same conclusions apply to the subperiod 1991-2002. 2.13 On the revenue side, during 1981-2002 the main determinants of the long- term growth rate of current revenue were taxes on income and profits, domestic goods and services, and non-taxrevenue. As Table 2.1 shows, during 1981-2002 the trend of taxes on income and profits grew by 8.6 percent per year, taxes on domestic goods and services by 12.2 percent, and non-tax revenue by 4.9 percent per year with taxes on income and profits, domestic goods and services and non-tax revenue representing about 17 percent, 12 percent, and 28 percent of current revenue, respectively. Moreover, although taxes on international transactionsrepresentedabout 41 percent of total current revenue, the long-term growth rate was only 2.5 percent and therefore not a key determinant of the long-term growth rate of current revenue. This is not a surprising result, because the tariff reduction under the 1993 agreement on the introduction of a CET reduced import duty revenues. Any tax reform should take these findings into account. 2.14 Grants present a negative trend during 1980-2002, and capital revenue shows a positivetrend for the same period. Grants have decreased at a long-termrate of about 3 percent per year, while capital revenue has increased at a rate of 11.2 percent per year in the same period. The negative trend in grants is one of the reasons why in Chapter 4 we suggest that the government should stop relying on grant funds for the financing of capital expenditures. 2.15 All components of the fiscal accounts are more volatile than GDP. The highest volatility belongs to grants, capital expenditure, and capital revenue. As Table 2.2 shows, grants volatility is 35 times higher than GDP volatility, and capital expenditure and capital revenue 33 times higher. Current revenue volatility is higher than current expenditure volatility-6.4 percent compared to 5.8 percent. 2.16 Current expenditure, all its components, and capital expenditure are acyclical. In other words, in good times the government does not increase its expenditures, and in bad times does not cut them. The correlation between the cyclical component of expenditure and GDPfluctuates from -0.13 to 0.3 1, but is not significant at 10 percent. This is an interesting result similar to the one observed in OECD countries, but not in developing countries. However, Talvi and VCgh (2000) show the same result for the case of the DominicanRepublic. 21 Table 2.2 Cyclical Propertiesof GDP, CentralGovernment Revenue,and Expenditure, 1981)-2002 Standard Correlation Deviation (%\ with GDP Current Revenue 6.4 0.54*' Tax revenue 7.8 0.74** Tax on income and profits 8.2 0.50% Tax on property 27.2 -0.04 Tax on domestic goods and services 12.4 0.30 Tax on international transactions 8.9 0.73** Non-tax revenue 8.4 -0.16 Capital revenue 69.4 -0.20 Grants 74.3 -0.09 Current Expenditure 5.8 0.19 Expenditure on personal emoluments and wages 8.8 0.3 1 Expenditure on other goods and services 7.6 -0.13 Expenditure on transfers and subsidies 16.9 0.00 InterestPayments 10.7 0.07 Capital Expenditure 69.5 0.09 Primary Deficit (after grants) 13.0 0.05 Overall Deficit (after grants) 19.3 -0.04 Current Deficit 5.5 -0.10 GDP 2.1 1.oo ** * Significant 5 Significant at 1 percent level. at percent level. Note: Statistics are basedon Hodrick-Precott filtered data. 2.17 Current and tax revenue were procyclical during 1980-2002, due to the procyclicality of taxes on income and profits and international transactions. All other tax components are acyclical. The correlation between current revenue and tax revenue with GDP is 0.54 and 0.74, respectively, and significant at the 1 percent level. Taxes on international transactions and income are highly procyclical, with a correlation of 0.73 and 0.50, respectively, and significant at the 1 percent level. The procyclical nature of current and tax revenue is not surprising, because most taxes are in some way proportional to economic activity. 2.18 Finally, all measures of fiscal balance, such as primary, overall, and current balance, are acyclical. That is, the primary, overall, and current fiscal balance does not co-move with the GDP cycle. In good times the government does not run a higher or lower fiscal balance than its trend. The correlation between primary surplus and overall surplus with GDP cyclical component is -0.01 and -0.04, respectively, but not significant at 10 percent. This is an interesting result and in contradiction with the empirical evidence for developing countries. However, most of this empirical evidence does not includesmall islands like St. Kitts andNevis. 22 2.19 Inthis section, we will constructstructural budget surplusmeasuresto factor out cyclical effects of conventional budget surplus measures. Once this is done, the adjusted measures are taken to be indicators of the stance of fiscal policy. The methodology used here is the same as the one usedby the European Community. It uses the HP filter-based trend inGDP to measurepotential output. For the case of the St. Kitts and Nevis economy, cyclical adjustments to only three tax revenuecategories were made using their respective tax revenue elasticity. No cyclical adjustments were made for property tax and government e~penditure.~' A complete discussion of the methodology is presentedinAppendix B. Structiird Budget Estimtite,rfbr St. Kitts crnd Neiis, 1983-2002 2.20 The cyclical adjustment to the government budget balances in St. Kitts and Nevis varies from -1.8 percent of GDP in 1983 and 2002 (recession years) to 1.9 percent of GDP in 1989 (expansion year). For example, as Figure 2.3 shows, in 2002 the overall fiscal deficit after grants was 18.3 percent of GDP, but factoring out the cyclical impact of lower tax collection, the cyclically adjusted overall fiscal deficit was only 16.5 percent of GDP. The latter concept is the so-called structural fiscal deficit. The St. Kitts and Nevis cyclical adjustment compared to other countries inthe region i s inthe low range. The cyclical adjustment inChile is between -4.0 and 2.5 percent of GDP. 2.21 International Figure 2.2 Fiscal Cyclical Adjustment in St. Kitts and Nevis 1983-2002 transaction taxes ( as percentage of GDP) account for more than half of the cyclical adjustment. (See Figure 2.2) As we discussed in paragraph 2.17, this is due to the high correlation \q hpS! 4 S9 0' \Q + +6 between the cyclical component of taxes in international transactions and GDP, -2.50 ' Source Author's CalCUlaPOnSbased on ECCB and IMFdata 0.73, and the 0 CyclicalAdjustment in InternationalTransactionTax importance of taxes in 0CyclicalAdjustment inGoodsTax international 0CyclicalAdjustment in Incomeand ProfitsTax transactionsintotal revenue (40 percent). 37 The elasticities were 2.0 for income and profit, 2.3 for domestic goods and services, and 3.6 for international trade. No adjustment was made for property taxes and government. 23 Figure2.3 Cyclical AdjustedPrimary,Overall, and Current BalanceinSt. Kittsand Nevis, 1983-2002 (as o/c GDP) Piimaiy Balance, CYCIIF~I Adjusted PIlznawEildrl~eand Cycllc~liAUju*ent (as pmwntag* of GDP) O0 wTotalCyclicalAdjustment ~ o t eAUVors calcuIaI~onsDesed on ECCB and IMF data - CyclicallyAdjustedPnmary Balance - c Pnmary Balance Overail Balance. CyclicalAdjustedPrimaryBalanceandCyclicalAdJustment (as percentageof GDP) looo 1 0TotalCyclicalAdjustment-Overall Balance -- CyclicallyAdjusted Overall Balance Cumen1 Balance , Cy611cd AdjustadCurrent Balance and Cyclical Adjustment (as percentageorGDP) 3.00 7 -Ooo -7 00 1 Note Authors calcuiaOonsbasedon ECCB end IMFdata nTotalCyclicalAdjustment-Current Balance -* Cyclically Adjusted Current Balance 2.22 Once the cyclical effects are factored out, the country has run a structural primary, current, and overall deficit for most of the 1990s. As Figure 2.3 shows, the government has runa cyclically adjustedprimary deficit every year since 1994. In 1994, the structural primary deficit was 0.5 percent of GDP, andkept increasinguntil it reached 8.9 percent of GDP in2002. The same pattern is observedfor the structural overall fiscal surplus. This confirms the main message of Chapter 1-the business cycle had little effect onthe sharp deteriorationof the St. Kitts andNevis fiscal position. 24 2.23 Can we design a fiscal rule for St. Kitts and Nevis based on our structural budgetmeasures?Any fiscal rulehas two objectives: (a) to enforce fiscal responsibility, and (b) to make certain that the government runs a countercyclical fiscal policy. In this regard, St. Kitts andNevis is on the right path with an acyclical fiscal policy. However, in order to move the country to a countercyclical fiscal policy we propose a fiscal rule based on a structural surplus. 2.24 We proposea structuralsurplus of about 2 to 4 percentof GDP for St. Kitts and Nevis. This rule is more austere than the Chilean fiscal rule of an overall surplus of 1percent. However, considerations of credibility and highGDP volatility in small islands may require that small countries such as St. Kitts and Nevis adopt tighter fiscal rules. This fiscal rule implies a fiscal adjustment of about 14 to 16percent of GDP, similar to the adjustment proposed in Chapter 1. However, as we discussed in Chapter 1 this fiscal rule will only stabilize the debt to GDP ratio to about 100percent of GDP in the next years. Box 2.1 presents indetail the fiscal rule inplace inChile since 2000. 2.25 However, successful implementationof a fiscal rule based on a structural surplus will require, as a first step, timely and reliable fiscal information. As we discuss at length inChapter 3, the current presentation of the fiscal information does not follow international practices and presents a distorted picture of the country's fiscal situation. Unless St. Kitts and Nevis is willing to solve this shortcoming, fiscal mimics can sabotage a fiscal rulebasedon a structural surplus. 25 Box 2.1 Chile's Fiscal Rule In 2000, the Chilean government introduced a fiscal rule based on a structural surplus of 1 percent of GDP. The objective of this fiscal rule is twofold: (a) to strengthenChile's commitment to fiscal responsibility, and (b) to develop fiscal policy indicators that will serve as a tool to formulate medium-term fiscal policy rules, avoiding the procyclical bias inpublic finances. The fiscal rule is a self-imposedmeasureenforcedby the government to guide fiscal policy during 200 1- 0.5. The computation of the structural balance follows the IMFmethodology and reflects the amount revenues and fiscal spending would reachif the GDP were at its potential level and the price of copper were at the medium-term price. Consequently, it excludes the cyclical effects of economic activity and the price of copper. The institutional coverage used inthe structural balance is the Central Government. The figure below summarizesthe results of the application of this methodduring 1989-2000. I The estimate of the structural balance consistsof three steps: Adjustment of the actual balance according to the concept of net worth variation of the Central Government (adjustedbalance). I Chile: Stucturaland Overall Balance 1989-2000 (as % GDP) Estimation of the cyclical impact of each budget component. It includes the impact of the economic cycle on tax revenues and the impact of cyclical fluctuations of the price of copper on fiscal revenuesfrom copper. The cyclical component of spending has not been taken into considerationbecause there is no significant relation between public spending and GDP in the Chilean economy. The cyclical component of tax revenues is obtained by adjusting observed tax revenues using an estimated output elasticity of 1.05. The cyclical adjustment to the copper price is based on the gap between actual export price and the estimatedreferenceprice. Estimation of the structural balance by subtracting cyclical components (tax revenues and incomefrom copper) from the adjustedbalance. Source: Marcel and others (2001). 26 3. BUDGETMANAGEMENT. 3.1 Recent literature has emphasized the role that budget procedures have in influencing fiscal outcomes.38Alesina and others (1996) have found evidence that, in LatinAmerica, budgetary institutions have had an important effect on primary deficits. 3.2 Since 1995 St. Kitts and Nevis's, increasing levels of public debt and fiscal deficits have been of concern. Public spending, taxation, and borrowing are the result of the government's budget management. This chapter discusses the rules governing the decision making that leads to the formulation of the budget, its passage through the legislature, and its implementation, execution, and monitoring. 3.3 St. Kitts and Nevis has introduced substantial changes to the budget structure and processes in the last few years, and its timely reporting is commendable. These changes, however, have not yet fully translated into improved resource allocation and control of the overall envelope allocation. That is because the main problem is the absence of firm resource envelopes consistentwith macroeconomic stability. 3.4 The mainfindings and recommendations canbe summarized as follows: OnLegislation 0 The new Finance (Administration) Act (FAA), now being reviewed by the government before submission to the legislature is a step forward in fiscal discipline. The new FAA is to contain more stringent provisions governing the use of advances and deposit accounts. In addition, this new FAA presents an excellent opportunity for the government to extend it into a fiscal responsibility law by introducing explicit fiscal targets to be met. 0 OnBudget Preparation, Execution, and Monitoring 0 Strengthen the Medium-Term Economic Strategy Paper (MTESP) and the Country Strategic Outlook as a tool to anchor the budget by making certain that the following elements are thoroughly addressed: (a) estimates of aggregate resources available for public expenditure consistent with macroeconomic stability, (b) bottom-up estimates of the cost of carrying out ongoing and new policies based on reliable assumptions, and (c) reconcile the aggregate resources with the cost estimates of the ongoing and new policies. 0 The capital budget prepared by the Planning Unit needs to be an integral element of the Country Strategic Outlook and the MTESP, and not an addendum with a desired list of investment projects. A complete discussion of the Public Sector Investment Program (PSIP) and policy recommendations are the focus of Chapter 4. 38 Untilrecently, this literaturehadfocused on the OECDcountries. Von Hagen(1992) andVon Hagenand Harden (1995) find that budget institutions have a significant impact on debt ratios and on deficits in the European Union countries. Poterba (1994), among others, has studied the effects of fiscal restraints on fiscal outcomes for the United States, reaching qualitatively similar conclusions. 27 The government should stop using the "below the line" account to finance permanent expenditures that are never brought up to balance in the current fiscal year. Moreover, the government should move, as soon as possible, to present the budgetaccording to international practices, includingfunctional classification. * The current program presentation inthe budget is insufficientto link expenditures to policy priorities such as growth and poverty reduction. Therefore, the government will be well advised to disaggregate expenditures at the program level in a way that will be useful for monitoring outcomes, such as in health, education, and social protection. a The debt management function is the weakest component of the financial management information systems. The government has formed a new Debt and Investment Unit under the Accountant General's office to monitor and manage domestic and external debt. However, the unit needs to be staffed with qualified personnel that can keep adequate records of external and domestic debt, analyze the data, and advise the government on debt managementissues. o The Standardized Integrated Government Financial InformationSystem (SIGFIS) inTreasury hasresultedintimely reporting. However, there are still shortcomings that curtail the effective operation of SIGFIS. First, skilled and experienced personnel remain in short supply. Second, a commitment control system is needed. Third, not all line ministries have terminals to access the SIGFIS system. To address the second problem, the Ministry of Finance has commissioned the development of a commitment control module for SIGFIS, and plans are for this systemto be inplace by 2003. 0 OnBudget Accountability a The financial statements of St. Kitts and Nevis should cover all government activities. The operations of statutory bodies, including state-owned enterprises, and several contingent liabilities should be included in the financial statements, such as government loanguarantees to finance private investment projects that the government will have to honor if the recipient of the loan is no longer able to do so, or bailout operations for failed enterprises(for example, SSMC). Weak Public Accounts Committees (PACs) are a common feature in the OECS countries. The recent Country Financial Accountability Assessment (CFAA) report for the OECS recommended a number of measures, including the following, to beginaddressingthis common problem: a Given that opposition parties are often insufficiently represented to form a functioning PAC, each country should consider enacting the appropriate constitutional or other changes to allow the appointment of no parliamentary members. Members of Parliament and other interested parties should participate in the Commonwealth Parliamentary Association, which, in conjunction with the World Bank Institute (WI),has organized a Study Group on Public Accounts Committees andplans follow-up seminars. 28 OECS countries should draw on the advice of neighboring countries such as Jamaica, where the PAC is a vibrant component of government oversight. A media campaign should be developed to publicize PAC activities to the public and to create demand from the public for strong public financial management. I. CONSTITUTIONALANDLEGAL FRAMEWORK 3.5 The Federation of St. Kitts and Nevis attained independence in 1983 as a Westminster-style parliamentary democracy. The country has a unique asymmetric federal constitution and a unicameral legislature, with 11elected representatives and up to 8 appointed senators sitting inthe NationalAssembly. Elections are held at least every five years, the last election took place in2000 with the incumbent government winning a new mandate, and the next elections are scheduled for 2005. Nevis has its own subnational island assembly and administration, with the federal government also serving as the island administration for St. Kitts. The Nevis Island Administration (NIA) has significant autonomy over island affairs. The Constitution stipulates a revenue sharing formula between the federal government and the NIA. Figure 3.1 presents the government structure. Figure 3.1 St. Kitts and Nevis Government Structure Social Security Scheme 3.6 The Constitution (Saint Christopher and Nevis Constitution Order 1983) includesgeneral provisionson financial matters. These are amplified in the Finance Act of 1990, which is about to be replaced by a new Finance (Administration) Act (FAA). This will provide the basic law on government financial management and will empower the Minister of Finance to issue rules and regulations for more detailed guidance. The First Schedule to the FAA contains new Financial Regulations and the Second Schedule contains new Procurement and Stores Regulations. Section 51 of the FAA allows the Minister to amendthese regulations. 29 3.7 The new FAA is a step forward in fiscal discipline and transparency in the use of advances and deposit accounts. In addition to updating provisions to reflect current practices, including the use of electronic media and the introductionof automated tools for vote book accounting, the new FAA is to contain more stringent provisions governing the use of advances and deposit accounts.39For example, in cases where an advance warrant has been used to authorize expenditures in anticipation of repayment or the raising of a loan, the new act will contain an obligation to bring such expenditures to account should it be determined that the loan or repayment is not forthcoming. Also, there will be a provision to restrict deposit accounts from going into overdraft. The basic framework of the consolidated fund, parliamentary authorization of annual estimates, accountability of permanent secretaries and heads of departments to Parliament, and the overall responsibility of the Ministry of Finance (MoF) for public financial management, all remainthe same. 3.8 While in the past, Central Government borrowing was governed by specific acts, since 1992 it has been governed under only one act: the Development Loans Act. This act currently allows the Central Government to borrow up to a ceiling of 60 percent of GDP via loans and debentures. Moreover, this act gives ample power to contract debt and guarantee loans contracted by other agencies to the MoF without the need of parliamentary approval for each loan. 3.9 Other legislation that affects financial management in St. Kitts and Nevis includes: e Revenue legislation and fiscal incentives legislation, chiefly the Fiscal Incentives Act and the Tourism Incentives Act, which define the conditions in which exemptions from taxes, customs duties, service charges, and fees are granted. e Acts, which establish statutory bodies (including state-ownedenterprises). e The Audit Act of 1990. This requires the Director of Audit (DOA) to report to Parliament at least once a year on the work of the office and key findings related to the audit of the government's accounts and records. The law empowers the Audit Department to undertake value-for-money audits. The DOA also audits the accounts of the NIA. 11. BUDGETPREPARATION,EXECUTION, MONITORING AND 3.10 St. Kitts and Nevis has a dual budget system. The Budget unit in the MoF is responsible for recurrent expenditures and for monitoring overall expenditures and revenues, while the Planning Unittakes the lead with respect to capital projects. St. Kitts and Nevis has a calendar fiscal year. The budget is always passedbefore January 1.The budget cycle is a continuous 24-month process, as presentedinFigure3.2. 39A final review of the draft legislation by all permanentsecretaries is scheduledfor May, following which the Legal Affairs Department will make any last revisions before submission to the legislature in the summer. 30 3.11 The budget is prepared based on the Country Strategic Outlook, the current MTESP,the capital budget, and eachministry's strategy outlook. However, allthese documents fall short of presenting a reliable framework for the budget. Both the Strategic Country Outlook (a yearly document that presents the economic prospects of the country for the next fiscal year) and the MTESP (a four-year economic plan that provides the information for the next budgets) last preparedfor the Caribbean Group for Cooperation in Economic Development (CGCED) in June 2002, suffer from the same shortcomings: (a) top-down estimates of aggregate resources available for public expenditure not always consistent with macroeconomic stability, (b) lack of bottom-up estimates of the cost of carrying out ongoing and new policies or based on unreliable assumptions (for example, tax revenues), and (c) no attempt to reconcile the aggregate resources with the cost estimates of ongoing and new policies. On the ministries' strategic outlook, most of them are out of date or cannot provide a link among objectives, expenditure required, and outcomes. 3.12 Given the dual budget system, the capital budget prepared by the Planning Unit is appended to the country strategic outlook and the MTESP, but it is not an integrated element of these documents. Indeed, both the Public Sector Investment Program (PSIP), which presents the investment plan for four years, and the capital budget, that presents the current investment plan, are a desiredlist of investment projects rather than a well-thought-out investment plan with growth and poverty reduction as main objectives. Therefore, they cannot be used with confidence to project the fiscal outcomes in the current or next fiscal years. The PSIP is the focus of Chapter 4 of this report. 3.13 The budget provides an accountant's view of government transactions rather than a presentation of the government's fiscal policy. The budget adheres to the traditional administrative, program, and economic structure format. However, this presentation is insufficient to assess the fiscal position of the government and to follow up on key programs supposedly linked to policy priorities. For instance, as we will discuss in detail in Chapter 8, it is almost impossible to obtain from the budget the amount allocated to crucial safety net programs and to follow them over time. 31 I Box 3.1 FinancingExpenditures"BcIow the Line": The Use of Advance and DepositsAccounts I Inrecentyears, the St. Kitts andNevis governmenthas beenmakingsizeable expendituresoutside its budgetprocess usingbelow-the-line accounts. Traditionally, the biggest item has been the treatmentof revenue transfers from the Federal Government of St. Kitts and Nevis to the Nevis Island Administration (NIA). For the most part these are a Constitutional requirement, but from 1983 to 1993, the Federal Government treated them as advances rather than transfers and, hence, did not show them in the estimates or in its statement of revenues and expendituresprepared for audit each year. By 1994, the total value of advances outstandingto the NIA hadreachedover EC$21million, at which point the balancesfor other types of advances were EC$990,000 to other governments, EC$620,000 to public officers, EC$IO million pending the raising of loans, and EC$17 million for all other advances (for example, SSMC). These balances remained reasonably constant until 1999, when "all other advances" increased by over EC$13.5 million, followed by another increaseof almost EC$10 millionthe following year. The outstanding balance of "all other advances" reached over EC$40 million-about 5 percent of GDP-by the end of 2000. Few of these "advances" have been made with the expectation of repayment: some went to the SSMC, some to the Air and Seaports Authority, andover EC$lO million went to various capital projects. Duringthe same period, the FederalGovernment also useddeposit accounts to finance post-hurricane and other expenditures. Deposit accounts are properly usedfor a variety of purposes, such as holding security deposits so that importers can obtain their goods before the final assessment of duties. Successive governments have also established such accounts to lodge receipts in response to hurricane relief appeals. In these cases, unbudgeted emergency expenditures-whether capital or recurrent-are financed directly from the relevant account. Once the crisis has passed, Parliamentaryapproval should be obtained via a supplementaryappropriation, at which point both the expenditures and the donated revenues would be brought to account. However, since the string of hurricanes in 1998 and 1999, these deposit accounts have not been analyzed, approved via appropriation bills, and brought to account. Therefore, millions of dollars in expenditures made in this fashion have not been reflected in the government's reported expendituressince 1998. Transfers and expendituresfinanced from below-the-line accounts are coveredby the Financial Act and are reflected inthe statementof assets andliabilities presentedfor audit as partof thegovernment's annualfinancial statements. As such, these expendituresare not improper and there are some safeguards againstimpropriety. However, when such substantialitems are not brought to accountfor the years inwhich they are incurred, the statementof revenues and expendituresno longer accuratelyreflects the government's operations, impairing parliamentary oversight and accountability, and greatly complicating any analysis of fiscal policy and the impact of government activities on the economy. I 3.14 Ministry of Finance does, however, have the capacity to present the budget by standardfunctional classification, andshould start doing so infiscal year 2004. St. Kitts and Nevis uses to some extent the functional classification to make its economic and revenue projections. Thus, it should be able to move quickly to present the budget according to international best practices, such as the ones contained in the IMF's Government Financial Statistics (GFS). This will also facilitate investors in the Regional Security Market to more accurately assess the fiscal positionof the OECS countries. 3.15 Untilrecently, the St. KittsandNevis budget scored well incomprehensiveness. But in recent years, more sizeable expenditures (mainly capital expenditures) have been financed from "below the line" using advances and deposit accounts, without being brought to balance and reflected in the budget in the years in which they occur. Thus, the recurrent and overall balances as well as other line items presented inthe budget do not represent an accurate picture of the government's fiscal situation. The government needs to bring to balance all "below the line" accounts starting in2003 (see Box 3.1). 33 3.16 Actual recurrent expenditures Figure3.3 Variance betweenAchial and BudgetedRecurrentExpenditures, exceeded the estimates for 1994-2000, MoFandallother ministries h.~ittandNevis1905-2000 s with the exception of 1998, by about 6 (inpercentage) percent (see Figure 3.3). This 15.0% 7 outcome could be the result of expenditure control problems andlor inaccurate forecasting. The latter could be due to significant shocks, such as the hurricanes of 1995, 1998, and 1999, or to weaknesses inplanning. The evidence 1998 49st' presented below suggests weakness in -5.0% planning as the maincause. 3.17 The Ministry of Finance and Planning (MoF) accountedfor most of the variance. The MoF systematically spent above its budget during 1995-2000. In contrast, all other line ministries were within budget from 1996 through 1999, andoverspent only in 1995and2000 (both electionsyears). 3.18 Within the MoF, the key item responsible for the variance was debt servicing. As Table 3.1 shows, debt servicing was consistently underestimatedand accountedfor 52 percentof the total overexpenditure by the MoF from 1995 to 2000, and about 75 percent of the total overexpenditurein 1999. Bonus salaries (for which no budget provision is made) were granted in three of the six years and, in 1996, the wages paid to those on the Short-Term Work Experience Program (also unbudgeted) amounted to over 33 percent of the total MoF overexpenditureinthis period. The other items responsiblefor the overexpenditurewere pension benefits in 1995 and 2000, transfers to the SSMC in 1995, and significant claims against the government (related to disputes with contractors andexpropriations)paidin 1997and2000. To examine allocations across sectoral ministries, this analysis is based on Estimates Books for the Federal administration, supplemented with information from annual reports from the Director of Audit. The treatment of certain items (for example, debt repayment) does not accordwith IMFstandardsfor government accounting, but that does not affect the main issues arising from the analysis. 34 Table 3.1 Main ItemsResponsiblefor the Variance BetweenActual and BudgetedExpendituresin the Ministry of Financeand Planning,St. Kittsand Nevis, 1995-2000 (in millionsEC$) 1995 1996 1 1999 2000 Total DiscrepancyMoF 9.9 15.0 3.7 4.4 29.9 7.8 Debt services -0.5 4.7 2.0 5.1 22.4 3.1 Extra month bonus salaries & wages 5.3 5.8 7.2 Short-term work experience 5.5 Pensionbenefits 2.0 1.3 Claims against government 1.4 3.7 SSMC subvention 2.1 All other 0.5 0.9 0.3 -0.7 0.3 0.2 Source: St. Christopher and Nevis, "Report of the Director of Audit," various years. 3.19 The Ministry of Communications, Works, and Public Utilities (MOW) overspent every year with the exception of 1998 and 1999 (hurricane years). Moreover, as Figure 3.4 shows, the variance between budgeted and actual expenditures was higher in 1995, 4.5 percent, and in 2000, 4.9 percent, both election years. All line ministries collectively spent below their recurrent budget except inelection years. 3.20 Aside from the M o F and Figure3.4 Variance betweenActual andBudgetedRecurrentExpenditures, MOW, since 1995 the Governor st KIUS andN+VIS ~99s-zaaa MOWandall other ministries General and the Office of the Prime (inpercentage) Minister have been consistently over 6.0°h budget and the social sectors 4,0% consistently under budget. Table 3.2 divides the ministries into three groups: 2.0x (a) those consistently over budget, (b) 0.0% those consistently under budget, and (c) those with no consistent pattern.41 -2.0% Persistent patterns of under or -4.0°h overexpenditure at the line ministry -6,0x level suggests weaknesses in work 1 planning and budgeting in some line Source Author'r C~ OCommunications,Work and PublicUtilities -All other Ministries ministries, which, ifrectified, would improve budget accuracy and expenditure control. , 4'Consistently over or under budgetmeans over or under budget for all or for all but one of the years for which data are available. 35 Table 3.2 MinistriesConsistentlyOver or UnderBudgetinSt. Kittsand Nevis, 1995-2000 (average percentage) ConsistentlyOverBudget Governor General (21.6%) Education (-4.0%) LegalAffairs (4.6%) Office of the PrimeMinister (I1.6%) Health (-5.7%) Culture, Youth, & Sports (3.2%) Finance andPlanning(24.1%) Agriculture (-8.0%) National Security (-0.3%) Audit (-8.5%) Tourism (-5.4%) ForeignAffairs (-12.2%) Comm. & Social Affairs (-6.0%) Parliament (-12.8%) Source: St. Christopherand Nevis, "Report of the Directorof Audit," variousyears. 3.21 The Standardized Integrated Government Financial Information System (SIGFIS) in Treasury is the key system for maintaining timely, accurate, and comprehensive financial data. The system has resulted in timely reporting. Trial balances are produced on the first working day following the end of each month, which allows the Budget Unit to compile its Revenue-Expenditure Analysis Reports for the Cabinet within two weeks of month's end. The Annual Accounts are submitted for audit within the stipulated period (six months following the end of the financial year). 3.22 There are still shortcomings that curtail the effective operation of SIGFIS. First, skilled and experienced personnel remain inshort supply. Second, a commitment control system is needed. Program managers have a tendency to order goods and services even when commitments exceed the free balance available for the account.42Third, not all line ministries have terminals to access the SIGFIS system. To address the second problem, the MoF has commissioned the development of a commitment control module for SIGFIS to block the issuanceof purchaseorders when there is inadequate available balance, and a computerized vote book, which will allow agencies to post allocations, warrants, purchase orders, and payroll and other vouchers to their vote book accounts. This should reduce manual errors and simplify reconciliationagainst the SIGFIS printouts.43Plansare for these systems to be inplace in2003. 3.23 Concerning revenues, the Standardized Integrated Government Tax Administration System (SIGTAS, Inland Revenue) and Automated System for Customs Documentation and Administration (ASYCUDA) systems computerize most revenue accounting tasks for the two major revenue departments. Unfortunately, the two systems are not integrated, which would facilitate the maintenance of a single account per taxpayer covering all taxes. Revenues from the Electricity andWater Departments are becoming more significant becauseof recent rate increases, and a new computerizedwater billing system was announced in2002. 42This resultsinabuildup of payablesbecausethe government is committed to pay for goods andservices delivered on the basis of an order from a responsible public officer, even if that officer has exceeded his or her authority authorized by the allocations issuedto date. To curtail the ability of spendingunits to place orders with suppliers and then delay submission of the documentation to Treasury for processing, the MOF should notify suppliers that only the official purchase order will be recognized. 43The initialversion will not incorporatecomputer-assistedreconciliation features, and will not be directly linkedto SIGFIS. Ministry of Finance officials are aware that these enhancements are needed, but have not identified financing for its implementation. 36 3.24 The debt management function is the weakest component of the financial management information systems. Until 2002, the debt management function was split between a small unit inFinance, which uses the CS-DRMS to maintain data on foreign debt, and the Accountant General's Department, which handled domestic debt (mainly treasury bills). These units have now been merged and renamed the Debt and Investment Unit under the Accountant General. However, this new unit lacks qualified personnel to keep adequate records of external and domestic debt, analyze the data, and advise the government on debt management issues. 3.25 With the announcement of a fiscal consolidation program in 2000, additional controlshavebeenimplemented.The MoFhas reserved 10percent of the appropriatedbudgets for all discretionary expenditures (that is, excluding personal emoluments, wages, debt servicing, and statutory obligations). A hiring freeze was instituted and overtime payments now require management approval.44Inaddition, a Central Purchasing Unit was established in March 2000. This unit initially purchased all government stationery, but its role has been expanded to the procurement of furniture and equipment. Eventually, it will purchase all goods and services required by government, incorporating the procurement function of the General Supplies Office, the Pharmaceutical Procurement Unit, the Public Works Department, and the Electricity Department. A computerized procurement module is now being tested. 3.26 All these additional controls have not helped to restrain the sizeable fiscal deficit. That is becausethe main problem is the absence of firm resource envelopes at the aggregate and ministry level underpinned by an MTESP, strategic outlook, and ministries' corporate plans nestedina consistent macroeconomic framework. 111. BUDGETACCOUNTABILITY Siipretne Audit Institution 3.27 A number of factors havehinderedthe potentialbenefitsaccruingfrom the work of the Audit De~artment.4~ Audit Department, while in principle independent because of The authority specified in the Constitution, has no guaranteed operating budget or staffing complement, and must obtain approval from the Minister of Finance via the annual Estimates process. Inaddition, the automation of critical financial management functions without adequate input from the Audit Department inthe design andimplementation stages, the nonfunctioningof the Publics Accounts Committee, lack of training for staff, particularly on computer-assisted auditing tools, auditing of computer systems, and value-for-money auditing have affected the efficacy of the Audit Department. 44 In 2001, the Audit Department undertook a special audit of overtime at the Public Health Department and determined:"...procedures inplace ...have proven ineffective incontrolling payment for overtime .. There are no effective means of verifying and certifying the number of hours worked ...There . was no evidence of attempts to certify works for which overtime was paid ... The number of irregularities ... cause the computation of such paymentsto be inherently flawed." (Audit Reportfor 2000, pp.44-48) 45 The World Bank's recent OECS Country Financial Accountability Assessment (CFAA) contains a detailed assessment of the problems facing SupremeAudit Institutions inthe OECS countries. 37 3.28 In spite of these problems, the Audit Department does a commendablejob. Audit Reports are current for the federal government.& Unlike Audit Reports in other OECS countries, reports in St. Kitts and Nevis no longer detail hundreds of oversights and minor transgressions, and focus instead on material findings. Also, it continues a modest program of value-for-money audits to supplementcompliance audits. 3.29 The financial statementsof St. KittsandNevis presenta partialview of government activities. Significant fiscal operations are excluded. The operations of statutory bodies, including state-owned enterprises, are not covered. Inaddition, several contingent liabilities are not presentedin the financial statements, such as government loan guarantees to finance private investment projects that the government will have to honor if the recipient of the loan is no longer able to do so, or bailout operations for failed enterprises (SSMC). 3.30 Most of the state-owned enterprises (SOEs) are significantly in arrears in the preparation of their financial statements and their submission for audit and, eventually, before Parliament. Audits of SOEs are contracted to private accounting firms in accordance with their enabling acts. The Director of Audit has the right to require submission of their audit reports and to supplement these by further investigations if hehhe considers it necessary. He/she has not done so. 3.31 Followingthe last election,the membersof the Parliamentand the Public Accounts Committee (PAC) were appointed in November 2000, but the Committee has not met ~ubsequently."~ The Committee didmeet afew times from 1992to 1994, but took little action of substance. Unlike other OECS countries, the majority of PAC members (three of five) are supposed to come from the opposition. However, inthe current Assembly the government holds all eight seats from St. Kitts, with the three seats from Nevis split betweenthe two mainNevisian par tie^.^' Given that these two parties are rivals in elections to the Nevis Assembly, and both advocate secession, it is perhaps unreasonable to expect effective work on a PAC that examines only the accounts of the federal government. 3.32 Weak PACsare a commonfeatureinthe OECS countries.The recent CFAA report for the OECS recommended a number of measures, including the following, to begin addressingthis commonproblem: e Given that opposition parties are often insufficiently represented to form a functioning PAC, each country should consider enacting the appropriate constitutional or other changesto allow the appointment of nonparliamentary members. 46However, the Annual Accounts for Nevis, for which the federal Director of Audit also has audit responsibility, have not been submittedsince 1996due to a problem with the accounting software installed. 47The World Bank's recentCountry Financial Accountability Assessment (CFAA) contains adetailed assessmentof the problems facingPublic Accounts Committees inthe OECS countries. 48There are also four appointed senators-two from the governing party and two from opposition parties-who sit inthe National Assembly. 38 0 Members of Parliament and other interested parties should participate in the Commonwealth Parliamentary Association which, in conjunction with the World Bank Institute (WBI), has organized a Study Group on Public Accounts Committees and plans follow-up seminars. 0 OECS countries should draw on the advice of neighboring countries such as Jamaica, where the PAC is a vibrant component of govemment oversight. 0 A media campaign should be developed to publicize PAC activities and create demand from the public for strong public financial management. 3.33 The formulation of the budget receives inputs from the social partners, but this undoubtedly could be made more systematic. The recent agreement by OECS countries to establish National Economic Councils with public and private sector members may lead in this direction. However, the capacity of the social partners-most of which are small organizations- to go beyond their immediate concerns and address complex national policy issues remains limited.There have also been extensive consultations with private sector firms and associations concerning reforms to taxes and fiscal incentives. The Prime Minister and the Minister of Commerce also meet monthly with private sector representative^.^^ 49The Governor of the ECCB also chairs a tripartite (government, business, and labor) committee, but its focus has beenon getting all parties to better understandthe links betweenwages and productivity, and to agree on measures to enhanceproductivity. 39 4. THE PUBLICSECTORINVESTMENTPROGRAM 4.1 The conventional wisdom of the economics profession is that countries need capital to grow, and that a direct relationship exists between capital spending and growth. Consequently, when economists evaluate the allocation of public resources between current and capital spendingingovernment budgets, they tendto becritical of countries that allocate alarge share of government expenditure to current spendingand to applaud countries that spend on capital. 4.2 This bias is embedded inthe "golden rule," that is, only current expenditureneeds to be balanced by local revenue, but a country can (within limits) run a fiscal deficit equal to the capital spending of the government. This rule is implicit in many OECS budget presentations, where the focus is on the current fiscal balance even though for fiscal sustainability the key concept is the primary fiscal balance (overall balanceminus interest payments). 4.3 Because most current spending by a government reflects entitlements or previous commitments, such as salaries, payments on public debt, and so forth, governments have in the short run limited scope to influence it. In contrast, the capital budget and its composition are highly discretionary. Informulating the capital budget, government makes the decisions that at the end will determine the size, composition, and in some cases the geographical location of capital budget. 4.4 Why does this matter? Because in St. Kitts and Nevis, on average, during 1995-2002, capital expenditures as a fraction of GDP increased about three times compared to 1990-94, including "below the line" transactions not presented in the annual budget estimates. This increase in expenditures was mostly financed through loans (commercial and concessional) and i s one of the main reasons behind the difficult fiscal position facing the government today. Although the damage caused by Hurricanes Luis in September 1995, Lenny in November 1998, and Georges in November 1999 explained part of the increase in capital expenditures, a large portion was not hurricane related and its aim was not necessarily to stimulate growth or reduce poverty. 4.5 The purpose of this chapter is first to describe and draw attention to the limitations of the processes in place to prepare, implement, and monitor the Public Sector Investment Program (PSIP) for the Federation of St. Kitts and Nevis and the Nevis IslandAdministration (NIA), and second, to assess the size and composition of the PSIP. 4.6 The mainfindings andrecommendations canbe summarized as follows: On Capital Expendituresand PSIP Preparation, Monitoring, and Execution o The capitalbudget and the PSIPpreparedby the PlanningUnitinthe Ministry of Finance (MoF) needs to be an integral element of the Country Strategic Outlook and the Medium- Term Economic Strategy Paper (MTESP), and not an addendum presenting a desired list of investment projects. The Budget and Planning Unit should work closely to anchor the capital budgetand the PSIPto the country strategy. Moreover, the BudgetUnitshould set the overall envelope for capital expenditures (consistent with macroeconomic stability) 41 early in the year so that the Planning Unit can attempt to prioritize the projects early in the process insteadof at the end of the capital budget preparation. o The government of St. Kitts and Nevis needs to establish clear selection criteria for investment projects. First, only investment projects, independently of their source of financing, with all the required information (in particular projected recurrent costs) should be included in the capital budget and the PSIP. Second, only investment projects prioritized based on some type of cost-benefit analysis and in accordance to the overall resources available for capital expenditures should be included in capital expenditures and the PSIP. Third, all projects with a cost above EC$2 million should be subject to a formal cost-benefit analysis before being included in the capital budget or the PSIP. Fourth, given the high debt burden, only projects financed with concessional financing should be included. o At a technical level, documentation needs to be prepared on the process of preparing the PSIP, and staff needs to be trained in the procedures and tools of project analysis for preparingthe PSIP. o The Planning Unit should require the line ministries to report capital spending monthly and should stop disbursements to those projects that do not provide the informationat the end of the month. e On the Coverage of Capital Expenditures and the PSIP o Some investment projects currently included in the PSIP and as capital expenditures in the budget should be classified as recurrent expenditures. International practice should be followed inthis respect. o A significant amount of the capital expenditures incurred by the St. Kitts and Nevis government are recorded "below the line" and are not captured inthe budget or the PSIP. The government should stop this practice. o All investment projects by statutory agencies, such as state-ownedenterprises, or private corporations but financed through loans guaranteed by the government, are currently excluded from the PSIP. These investment projects can represent a sizeable portion of the public sector investment program and can expose the government to serious unexpected fiscal problems. The government should ensure an accurate representation of public sector investment by including the investment projects of all statutory agencies in the PSIP. Investment projects by private corporations but guaranteed by the government should be disclosed inthe government's financial statement as government liabilities. 42 I. ST.KITTSANDNEVIS FEDERAL GOVERNMENT PSIP: PREPARATION,EXECUTION, AND MONITORING" 4.7 St. Kitts and Nevis has a dual budget system. The St. Kitts and Nevis Federal Government PSIP is coordinated by the Planning Unit in the M o F with the support of two committees, the PSIP Operations Committee (PSIP OC) and the PSIP Project Coordinating Committee (PSIP PCC)." The PSIP OC meets monthly and has the dual function of monitoring and coordinating the current capital spending program and identifyingprojects for inclusion in the next PSIP. The PSIP PCC is expected to meet quarterly and its function is to provide additional oversight and to make senior ministerial officials accountable for the capital spending program. However, since it was establishedinthe fall of 2000, it has not met regularly. 4.8 The public sector investment program of St. Kitts and Nevis is in practice the investment program for the Federal Government. All investment projects by statutory agencies, such as state-owned enterprises, are excluded from the PSIP. This is a serious gap becausethese investment projects can represent a sizeable portion of the public sector investment program and can expose the government to serious fiscal liabilities. The government should ensure an accurate representation of the public sector investment program by including in the PSIP the investment projects of all statutory agencies. Investment projects by private corporations but guaranteedby the government should be disclosed inthe government's financial statement as government liabilities. A good example of the fiscal difficulties this can cause is provided by the hurricane-related costs stemming from the Port Zante cruise ship dock, the reconstruction of which was financed by the government although the port i s ownedby a private corporation (see paragraph4.19). 4.9 The St. Kitts and Nevis Federal Government prepares a four-year rolling medium- term PSIP as part of the annual budget process. However, in practice the focus is the one- year capital budget. As Table 4.1 shows, the capital spending intentions are the highest for the budget year (estimates). Infuture years, planned capital spending drops off. The reason for this pattern is twofold: lack of a well-thought-out medium-term country strategy and sectoral strategies that allow line ministries in coordination with the MoF to prepare a medium-term PSIP, andthe fact that donors' objectives and financing become explicit only closer to the budget year. 50 All the figures used in this chapter do not include "below the line" transactions because to do an analysis at the sectoral level we had to rely on the figures in the PSIP, which do not include these transactions. Therefore, the figures inthis chapterdiffer from the numbersused inChapter 1of this report and by the IMF. 51 The PSIP OC is formed by the financial officers from the line ministries and chaired by the Director of the Planning Unit. The PSIP PCC is formed by all Permanent Secretaries and chaired by the Prime Minister with the Deputy Chair being the Minister of Communications, Works and Public Utilities. The responsibilities of the PSIP PCC in St. Kitts and Nevis are more focused on strategic decisionmaking than the Project Monitoring Committee in St. Lucia, which is chaired by the Prime Minister and occupies itself with all aspects of public sector investment, including monitoring and implementation. It is similar in responsibilities to the Priority Planning Consultative Committee inGrenada, but chaired at a higher level by the Prime Minister insteadof the Chief Planning Officer. 43 4.10 At the beginning of the year, the Planning Unit sends a call letter to all line Ministries requesting proposals for investment projects. But the information required and received in these proposals is inadequate to assess them. For each project (ongoing or new), which is submitted for inclusion in the PSIP, a project form is supposed to be prepared containing a description of the project, its objectives, financing source, and rationale. Information on projected recurrent expenditures is also supposed to be included on the project form, but the ministries do not regularly provide the required recurrent expenditure information. In addition, since no cost-benefit analysis is done by the line ministries to assess whether the projects proposed to the MoF are cost-effective, (with the exception of those financed by external donors), such informationis rarely provided. Table 4.1 Medium-Term PSIPfor St. Kithand Nevis, 2001-06 (inthousands EC$) EconomicServices 4,288 17,465 604 288 1,622 941 AgricultureForestryFisheries 4,079 17,27 1 604 128 522 0 Tourism 209 194 0 160 1,100 941 EconomicInfrastructure 11,455 8,325 17,477 46,031 14,687 4,667 Transport.& Communications 8,680 5,550 4,963 16,000 11,373 3,078 Energy 50 596 5,729 12,706 2,805 1,420 Water & Sewerage 2,725 2,179 6,785 17,325 509 169 SocialServices 21,492 36,919 19,437 19,695 19,808 19,601 Education 9,585 5,148 10,043 9,074 10,953 11,384 Health & Community Affairs 11,285 31,591 9,394 10,62 1 8,855 8,217 Housing/Lands 622 180 0 0 0 0 General Administration 4,779 6,086 15,110 2,321 2,112 4,082 Public Administration 4,779 6,086 15,110 2,32 1 2,112 4,082 Total 42,014 68,795 52,626 68,335 38,229 29,291 New Projects 24,734 19,351 7,677 6,677 OngoingProjects 27,892 49,984 30,552 22,614 Financing Revenue 5,180 1,979 38,229 29,291 Loan 38,195 60.986 2,976 2,412 Grant 9,252 5,371 26,897 24,016 Note: 1.The figuresin this table differ from those inChapter 1 becausethe figures inthe PSIPdo not include"below the line" capitalexpenditures. 2. Sectoraland total discrepanciesinthe table were presentinthe datagivento the Bank's team. Source: St. Christopher andNevis, DevelopmentandPlanning (Planning Unit), February2003. 4.11 Project proposals received are reviewed and prioritized first in April by the PSIP OC and then in September by the PSIP PCC, which is supposed to decide which investment projects will be included in the budget. But in practice limited prioritization takes place. The PSIP OC ranks each project using as criteria a combination of sector development objectives and national budgetary constraints. However, since cost-benefit analysis i s only commonly done for projects financed by external funds, there i s no consistent criteria that 44 can be applied to decide among investment projects. The investment projects rankedby the PSIP OC are then sent to the PSIPPCC to make the final decisions on which projects will be included in the PSIP. First priority is given to ongoing projects regardless of their merits, because presumably they already have funding. The new projects are supposed to be selected based on the ranking attached to them by the PSIP OC. However, inpractice the PSIP PCC has not been capable of making the required tough choices among investment projects, with the result that most of the new investment projects are included in the budget. Those that are not deemed a current priority, but that are still important, are heldfor the next year's budget. While the PSIP is labeled as a rolling four-year plan, it is really a one-year plan for the budget year, which also includes projects that are not expectedto be implemented. 4.12 The Ministry of Finance does not coordinate contacts with donors, which is an important part of the priority-setting exercise. Line ministries can contact donors directly. In some cases, this can result in the funding of new projects that were not expected to be implemented and therefore were not included inthe capital budgetfor the year and the PSIP. 4.13 The government's stated objective of the PSIP is to stimulate growth and reduce poverty. Nevertheless, some investment projects only partially fulfill these objectives. Good examples of this include: in 1997 about 6 percent of total capital expenditures (EC$1.8 million) was allocated to upgrade netball and tennis facilities, in 1997 about 13 percent of capital expenditures (EC$5 million) was allocated to football stands inWarner Park, and in2000 about 12percent of total capital expenditures (EC$6.2 million) was allocated to Carifesta VII. Carifesta i s a Caribbean arts and crafts festival that takes place regularly inthe Caribbean, organized each time by a different country. In2000 it was organized by St. Kitts and Nevis, and the government, inresponse to the alleged lack of capacity to make housing arrangements for all participants, builthousesto accommodatethe participants, which were subsequentlysold. 4.14 Several investment projects included in the PSIP and the capital budget should be classified as recurrent expenditures. Moreover, a significant amount of the capital expenditures incurred by the government are recorded "below the line" and not captured in the PSIP. Investment projects such as maintenance of roads, the purchase of books and magazines, paintingof schools, census and household surveys, and technical assistancefor a law project and the constitutional commission are examples of projects that should be moved to the recurrent budget. The reason why they have been kept in the capital budget may have been to avoid an increase in the deficit inthe current fiscal balance, because it is the key fiscal variable targetedby the government, and to provide an opportunity to present them to the donors with the hope of obtaining grants for their financing. Moreover, since 1995 a large portion of capital expenditures has been financed "below the line" and therefore not recorded inthe PSIP or inthe capital budget. The Planning Unit in the MoF has made an effort to track down the capital expenditures financed from "below the line," but it has not been very successful. The government should end the practice of financing expenditures from "below the line." 4.15 The following steps can be taken to improve the preparationof the PSIP: e The capital budget and the PSIPprepared by the PlanningUnit inthe Ministry of Finance need to be made an integral element of the country strategic outlook and the MTESP and not merely an addendum containing a wish list of investment projects. The Budget and 45 Planning Unit should work closely to anchor the capital budget and the PSIP to the country strategy. Moreover, the budget unit should set the overall envelope for capital expenditures (consistent with macroeconomic stability) early in the year so that the Planning Unit can attempt to prioritize the projects from the start instead of at the end of the process. e The government of St. Kitts and Nevis needs to establish clear selection criteria for investment projects. First, only investment projects, independently of their source of financing, with all the required information (in particular projected recurrent costs) should be included in the capital expenditure and the PSIP. Second, only investment projects prioritized based on some type of cost-benefit analysis and in accordance with the overall resources available for capital expenditures should be included in capital budgetand the PSIP.Third, all projects with acost aboveEC$2million should be subject to a formal cost-benefit analysis before being included inthe capital budget or the PSIP. Fourth, given the high debt burden, only projects financed with concessional financing should be included. e At a technical level, documentation needs to be prepared on the process of preparing the PSIP, and staff needs to be trained in the procedures and tools of project analysis for preparing the PSIP. e The Planning Unit should require the line ministries to report capital spending monthly and should stop disbursementsto those projects that do not provide the information at the end of the month. 4.16 During 1996-2001, capital expenditures were on average around 5.0 percent of GDP?*As Table 4.2 shows, capital expenditures fluctuate from a low of about 4.3 percent of GDP in 1998 to apeak of about 6 percent of GDPin 1999 and 2000, when reconstruction inthe aftermath of Hurricane Lenny increased spending. On average, during 1996-2001, the largest amount of capital expenditure was spent on infrastructure, at about 2.5 percent of GDP, followed by social services, 1.7 percent of GDP, and general administration, 0.5 percent of GDP. 4.17 During 1996-2001, half of capital expenditures were concentrated in the area of economic infrastructure, including energy and water services. The latter could be diverted to the private sector in the future. Capital expenditures on infrastructure include roads, water, sewerage, ports, and other infrastructure investments, which are largely under the responsibility of the Ministry of Communications, Works and Public Utilities. As Table 4.3 shows, on average about 16 percent of total capital expenditures during 1996-2001 was allocated to energy and water investment. However, if these services could be privatized in the medium term, with an adequate regulatory framework, some resources could be freed up for investment in social services. 52Not including "below the line" transactions. 46 4.18 Social services including education, health, sports, community affairs, land, and housing accounted for about 33 per cent of capital spending. General administration accounted for around 12 percent of capital spending, and economic services (agriculture, tourism, and manufacturing) about 6 percent. Table 4.2 Sectoral Compositionof Actual CapitalExpendituresfor St. Kitts and NevisFederalGovernment (as percentage % GDP) EconomicServices 0.07 0.19 0.15 0.16 0.74 0.46 0.34 AgricultureForestryFisheries 0.04 0.18 0.02 0.01 0.03 0.44 0.12 TourisdCultureEnvironment 0.03 0.01 0.05 0.04 0.71 0.02 0.18 Manufacturing 0.00 0.00 0.08 0.12 0.00 0.00 0.03 EconomicInfrastructure 2.28 2.50 2.95 4.05 2.25 1.24 2.54 Transport. & Communications 1.84 1.85 2.45 0.71 2.07 0.94 1.64 Energy 0.04 0.00 0.08 3.16 0.00 0.01 0.55 Water & Sewerage 0.40 0.65 0.42 0.18 0.18 0.29 0.35 SocialServices 1.73 1.36 0.74 1.62 2.54 2.32 1.72 Education 1.28 0.40 0.33 0.75 1.20 1.03 0.83 Health, Sports & CommunityAffairs 0.39 0.96 0.18 0.63 1.04 1.22 0.74 Housing/Lands 0.06 0.00 0.23 0.24 0.30 0.07 0.15 GeneralAdministration 0.56 0.95 0.50 0.24 0.42 0.52 0.53 Public Administration 0.56 0.95 0.50 0.24 0.42 0.52 0.53 Total 4.6 5.0 4.3 6.1 6.0 4.5 5.1 Source: St. ChristopherandNevis, Developmentand Planning(PlanningUnit). Table 4.3 Sectoral Compositionof Actual Capital Expendituresfor St. Kitts and NevisFederalGovernment (as 9%total capital expenditures) EconomicServices 1.59 3.81 3.56 2.64 15.73 10.21 6.26 AgricultureForestryFisheries 0.95 3.62 0.43 0.11 0.55 9.71 2.56 TourisdCulturelEnvironment 0.64 0.19 1.20 0.62 15.19 0.50 3.06 Manufacturing 0.00 0.00 I.93 1.92 0.00 0.00 0.64 EconomicInfrastructure 49.18 50.04 67.89 66.69 36.37 27.26 49.57 Transportation& Communications 39.56 36.98 56.39 11.71 33.53 20.66 33.14 Energy 0.95 0.00 1.73 52.05 0.00 0.12 9.14 Water & Sewerage 8.66 13.05 9.77 2.94 2.85 6.49 7.29 SocialServices 37.21 27.25 17.13 26.75 41.04 51.15 33.42 Education 27.66 8.00 7.59 12.39 19.38 22.8 1 16.30 Health, Sports & CommunityAffairs 8.35 19.23 4.15 10.40 16.87 26.86 14.31 Housing/Lands 1.21 0.02 5.40 3.96 4.79 1.48 2.8 1 General Administration 12.31 20.86 11.41 5.25 9.35 11.37 11.68 Public Administration 12.01 18.89 11.41 3.91 6.85 11.37 10.74 Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Source: St. Christopher and Nevis, Development andPlanning(Planning Unit). 47 Box 4.1 Port Zante CruiseShip Dock The new Port Zante cruise ship dock in Basseterre, St. Kitts is an example of a private project developed with loans guaranteed by the government, outside the PSIP process, but ended up with the government paying at least the interest on the debt contracted by a private corporation. The new dock, which everyone including the government and the cruise ship companies agreed was badly needed, was originally launched as a private development, but with government-guaranteedfinancing and some government participation. The new Port Zante, scheduled to open in June 1997, was designed with a deep harbor to accommodate cruise ships, luxury yachts, and pleasure sailing vessels with a maximum draught of up to 30 feet. It has a 400-foot sheet pile marginal wharf, roll-on-roll-off facilitate transit sheds, high mast lighting, and a moving buoy to facilitate berthing vessels between 400 and 750 feet long. It also has a 75-boat marina. The administration building, located on the 26-acre surrounding site, houses information and tour operator counters, immigration, customs, and emergency information. The new site will also offer luxury accommodations and a casino, as well as commercial, retail, and professional offices. The total cost of the project was US$16.5 million, financed commercially through the UrbanDevelopment Corporation (UDC) with a loan guaranteed by the government. The new port (dock, marina, and the administration building) opened in June 1997. Nevertheless, in September 1998, Hurricane Georges struck and the dock suffered structural damage. The problem was that the dock was built with a "Y" design that turned out to make it very vulnerable to hurricane damage, as some local people had predicted. The facilities were not fully insuredand the settlement was not sufficient to cover the reconstruction of the dock to the required higher standards necessary to resist future hurricanes. The new reconstructed port was scheduledto open inNovember 1999, when Hurricane Lenny destroyedthe port with heavy structural damage. After this, the rebuilding of the dock and constructing a larger breakwater of boulders was undertaken to prevent wave damage, with a cost of about US$25 million financed commercially with a government guarantee. Consequently, the governmenthad to assume a higher share of the financing, which had to be raised on commercial terms, and to take a bigger shareof the ownership of the project. The needfor repeatexpendituresfor a cruise ship docking facility that was destroyedtwice by hurricanes and that had to be financed commercially by the governmentraisesserious questionsabout the original design of the facility and about the appropriateness of the government's initialinvolvement. 4.19 The extent to which the capital expenditures have been driven by hurricane recovery and reconstruction is cause for concern because a sizeable fraction of those expenditures has taken place outside the PSIP. Significant infrastructure damage was caused by the passage of Hurricane George in 1998, and in particular by Lenny in 1999. Public infrastructure, such as roads and ports, was particularly hit, as was private infrastructure such as hotels, houses, and agricultural lands. Consequently, the government was required to invest substantial resourcesto rehabilitate damagedpublic and private infrastructure. However, because these expenditures were financed "below the line" and outside the PSIP process, it is extremely difficult to assess their cost-effectiveness. For example, the need for repeat expenditure in the new Port Zante cruise ship dock inBasseterre, which was destroyed twice by hurricanes and had to be financed by borrowing on commercial terms by the government, raises questions about the design and implementation of the project, and the tradeoffs made at the intersectoral and intrasectoral level. Box 4.1 discussesthe case of Port Zante indetail. 4.20 Since 1997,capital expenditures have been mostly financed through loans instead of revenue. As Table 4.4 shows, since 1997 more than 50 percent of capital expenditures have been financed through loans (both commercial and concessional) compared to only about 25 percent in 1996. Grants financing, on the other hand, shows a mixed pattern. In 1996, grants financed 48 about a quarter of all capitalexpenditures, in 1999only 6 percent, andin2001 25 percentof total capital expenditures. The irregular pattern of grant financing highlights the volatility of grant availability to finance the PSIP. Table4.4 SourceofFinancingof St. KithandNevisFederalGovernmentCapitalExpenditures (as % of totalcapitalspending) 1996 1997 1998 1999 2000 2001 003: Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Revenue 53.0 21.3 20.2 19.7 12.3 9.0 11.3 9.8 Loan 24.3 70.7 70.4 76.0 73.1 66.5 49.7 72.6 Grant 22.7 8.0 9.4 6.4 14.6 24.6 39.0 17.6 *All figures from 1996-2001 are based on actual expenditures. Figures in 2002 and 2003 are based on revised and estimated figures, respectively, dependingon available data. Source: St. Christopher and Nevis, Development and Planning (Planning Unit). 4.21 The estimatedfigures for capital expenditures presented in the budget have to be used with caution due to the low realization rate of planned capital expenditure (25 percent). On average, during 1994-2001 only about 25 percent of planned capital expenditure was actually implemented. As Table 4.5 shows, the realization rate varied from a low of 13.4 percent in 1994to a high of 34 percent in2001.53The relatively low realization rate provides an indication of the extent to which the capital spending intentions included in the estimates only represent a wish list andnot a firm investmentplan. Table 4.5 RealizationRateofSt. Kittsand Nevis FederalGovernment CapitalExpenditures, 1994-2001 Approved Estimates 140.4 145.4 112.1 105.2 99.3 104.7 142.3 103.43 (EC$Mill.) Actual (EC$Mill.) 18.8 36.1 29.4 29.3 32.4 24.2 45.6 35.466 Rate(%) 13.4 24.9 26.3 27.9 32.7 23.1 32.1 34.3 26.8 Note: The figures inthis table differ from those inChapter 1 becausethe figures inthe budget do not include "below the line" capital expenditures. Source: St. Christopher and Nevis, Estimates for the Year. 4.22 By funding source, grants have the lowest realizationrate, with an average of 7.2 percent during 1994-2000 (Table 4.6). This is because the government has a propensity to include more projects inthe PSIP that would be funded by grants in the hopes that the funding would be forthcoming over the course of the year. Over time, because of the relatively high level of per capita income, grants have become a decreasingly important source of financing, with 53 The higher ratio observed since 1995 does not necessarily imply that the government has been more successful in implementing planned capital spending. The higher ratio can merely indicate a higher level of spending on unplanned projects due to hurricane relief. 49 much of the gap being taken up by commercial and concessional loans. This decreased availability of grants hasnot beenfully taken into consideration inthe preparation of the PSIP. Table 4.6 RealizationRate of St. Kithand NevisFederal Government Capital Expendituresby Funding Source, 1994-2001 1994 1995 1996 1997 1998 1999 2000 Loans 8.5 8.7 13.2 24 19.1 50.3 4.5 18.3 Grants 1.9 20.8 9.9 2.0 4.5 8.6 2.4 7.2 Source: St. Christopherand Nevis, Estimatesfor the Year. 4.23 Coordination among donors does not appear to be a problem in St. Kitts and Nevis, but a dependency of the government on donors to select projects does. Donors seem to keep in close touch through meetings sponsored by the Caribbean Development Bank or other key donors. The regular, usually biennial meetings of the Caribbean Group for Economic Development also provide another useful forum for coordination. However, the St. Kitts and Nevis government can sometimes be pressuredby particular donors to pursue specific projects. The problem here is that the government does not carry out adequate economic analysis of particular projects and is thus not able to select the most appropriate ones with the highest rate of return to the country. Instead, it relies on the donors for analysis and selection, as well as financing of the projects. 4.24 Implementation of the PSIP is carried out by the line ministries and monitored by the PlanningUnit. However, the coordination between line ministries and the PlanningUnit i s weak. Officers from the Planning Unit periodically are sent to the field to collect information on the physical completion of projects because the requiredmonitoring form that must be filled out monthly by the line ministries does not always provide reliable information. There is also a monthly statement from the Treasury on capital expenditures to date. This information on spending is compared with the informationon physicalcompletion to make sure the projects are on track.54 4.25 Post-evaluation of projects is usually conducted only for projects financed by external funds. Because no examples of such evaluations were provided for local and foreign- funded projects, it is impossible to assert how useful they have been for future project planning. 54The only times when the informationon spendingis notavailable is when donor agencies makepaymentsdirectly to consultants.This is not likely to be a major problembecauseof the relatively small size of the resultingpayments. Foreign-fundedprojects usually have their own requirementsfor monitoringand reportingon physical completion andexpenditure.This helpsensure that the funds are properly spent. 50 11. SIZE AND COMPOSITIONOFTHENEVIS ISLANDADMINISTRATION PSIP 4.26 The Nevis Island Administration does not have a formal PSIP. No analysis or conceptual work is done in the Department of Planning and Development. Most projects are domestically financed andthere are hardly any externally financed projects. 4.27 The preparation, implementation, and monitoring process is very fragmented. In addition to the Department of Planning and Development, which was originally supposed to be responsible for investment planning, and the Project Management Office in the Ministry of Communications and Works, which is responsible for implementation, there is now a new Ministry of Physical Planning, Infrastructure, Development and Environment that has been created and given responsibility for physical planning of investment. Exactly how its responsibilities differ from those of the other two ministries is not clear. Such a separation of physical and economic planning is not likely to facilitate PSIPcoordination. 4.28 The prioritization of investmentprojects is supposed to be made by the Estimates Committee. However, these decisions are based on a simple list provided by the Department of Planningand Development. No analysis is provided. Sometimes projects are initiated based on the ad hoc requests of ministers, and sometimes contractors are asked to do other small projects that come to ministers' attention. Table 4.7 Actual Public Sector InvestmentExpenditures and RealizationRate for Nevis Island Administration (inthousandsEC$) 2000 PrimeMinister's Office 233 1,664 550 53 Finance 0 558 3,144 417 CommunicationsandWorks 3,270 6,59 1 16,370 19,712 Agriculture 22 124 166 212 Education,Health andCommunityAffairs 371 0 701 2,189 Tourism 0 0 0 327 Planning 0 0 0 38 Total 3,896 8,936 20,932 22,949 As % St. Kitts andNevis's GDP 0.65 0.25 3.02 3.88 ApprovedEstimates(EC$000) 18,438 49,72 1 39,461 86,115 Actual (EC$ 000) 3,896 8,936 20,932 22,949 RealizationRate(%) 21.1 18.0 53.0 26.6 Source: Nevis, Estimatesfor the Year. 4.29 Since 1995, the Nevis Island Administration (NIA) has carried out a public sector investmentprogramof about 3 percent of GDP. As Table 4.7 shows, over 80 percent of the capital spending was on infrastructure. This investment has included infrastructure required to support the expansion of tourism resultingfrom the new Four Seasons Resort and to rehabilitate infrastructure damaged by hurricanes or protect against future hurricane damage. The next- largest category of public sector investments was in health and education, which accounted for almost 10percent of total public sector investments in2000. 51 4.30 The NIA, like the Federal Government of St. Kitts and Nevis, also has a very low realization rate for public sector investment projects. This is because it follows a similar approach in preparing its capital spending estimates, under which estimates for the budget year do not necessarily reflect expectedspendingfor the year, but are more of a wish list. 4.31 Most of the NIA's expenditures on public sector investments are paid for with its own revenues or by borrowingon commercial terms. For example, in2000,34 percent of its capital expenditure was financed by its own revenues, 63 percent by loans, and only 2.6 percent by grants. 52 5. PUBLIC SECTOREMPLOYMENT AND COMPENSATION 5.1 With a total population of about 45,000 inhabitants, St. Kitts and Nevis has limited resources to carry out the whole range of functions performed by any government. As several studies have shown, small countries tend to have bigger governments in response to the higher cost of supplying public goods and the need to provide a stabilization role to ameliorate the effect of external shocks (Alesina and Wacziarg 1997; Rodrik 1996). However, even in a small country a government canbe too big inseveral ways (Nunberg andLindauer 1994): 0 The public service may be too large in the broad sense that the government sector is overextended, possessing too many agencies and departments charged with too broad a span of responsibilities. 0 The public service may be too large in the narrower sense of employing workers in excess of the requirements of designatedtasks. 0 The public service may be too large andexpensive for soundfiscal management. 5.2 In this chapter, we mainly focus on why the government's wage bill in St. Kitts and Nevis constitutes too high a percentage of GDP, about 15 percent during 1995-2002, for cost- effectiveness and sound fiscal management. The main findings and recommendations can be summarized as follows: 0 The high wage bill of the St. Kitts and Nevis Federal Government is employment driven rather than wage driven. During 1995-2002, the number of established positions increased intotal by 38 percent, and the number of non-established positions in full-time equivalent increased in total by 124 percent. At the same time, government real wages decreased slightly. Downsizing the civil service to the levels of 1995 by cutting 800 established positions and 500 non-established positions (head count) would reduce the wage billby EC$42 million, or 5 percent of GDP. 0 However, to implement a well-thought-out downsizing program, the government will have to introduce and/or upgrade its information systems in order to accurately cost its implications (pension and severance payments) and monitor public employment (established and non-established positions) on a regular basis. 0 Moreover, a medium-term and long-term downsizing program will need to adopt a humanresources management strategy and develop new strategies to carry out essential tasks by leveraging scarce resources. Otherwise, rehiring staff in the medium term can overturnthe fiscal gains obtained by reducing the size of the civil service. For example, a performance appraisal system should be established with pay or bonuses to reward good performance. In addition, the fragmentation and duplication of work among ministries, departments, and agencies, leading to a diffusion of responsibilities, should be addressed, as should the possibility of the government divesting some of its functions to the private sector. 0 The Nevis IslandAdministration (NIA) has reduced the number of positions (established and non-established) since 1995 intotal by 15 percent. Its main challenge for the future is 53 to consolidate these gains and reduce the cost of the administrative ministries such as Finance andthe Premier's Office. I. RECENTTREKDSINEMPLOYMENT AND COMPENSATIONINTHE ST. KITTSAND NEVIS FEDERAL GOVERNMEKT 5.3 There are two types of government positions in the St. Kitts and Nevis Federal Government: established and non-established. Employees in established positions are appointed by the Governor General upon a formal request by the Public Service Commission, with the approval of the Establishment Division in the Prime Minister's Office. Established positions are permanent positions listed in the budget, and their compensation is classified as Personal Emoluments. Employees in non-established positions are hired directly by line ministries, their numbers are not presentedinthe budget, and only their wages are included as a line item.55 5.4 The legislation that governs public sector employment is more protective of established than non-established workers. Established employment is governed by the Constitution, which has been in effect since 1983, and the public service regulations date from the 1970s. If an established worker retired in the "public interest" receives a lump sum immediately and a reduced pension, the lump sum is equal to five times the full annual pension amount. The Pension Variation Act of July 1994 speaks of no pension before 15 years of service, and a pension at age 50 after 15 years of service. The Protection of Employment Act dated June 1986 covers non-established workers. This act states that to dismiss a non-established worker requires only notice of the dismissal. If the notice is given a month in advance no severance payment is necessary, but all benefits must be paid (for example, leave not taken). In the case of dismissal in lieu of notice, a severance payment of three months must be paid, plus the benefits at time of dismissal. Employment 5.5 From 1995 to 2002, the number of established positions increased by a tota1,of 38.3 percent, or 4.7 percent per year. As Table 5.1 shows, the largest increase in established positions occurred in 1997 (12.8 percent). It was mainly the result of the creation of a National Defense Force (about 135 positions) and an Educational Planning Division (19 positions), and the hiring of 60 supernumerary teachers (substitute teachers) in all high schools. One of the largest increaseswas in2000 (an election year). 5.6 In the same period, non-established positions on a full-time-equivalent (FTE) basis increased intotal by 124 percent, or 12.2 percent per year?6 A significant proportion of non- 55The Public Service Commission is also responsible for ratifying promotions and disciplinary actions. Decisions can be appealed to the Public Service Appeals Board. There have been only a couple of dismissals for cause in the `pastfew years. 56A headcount of non-establishedpositions exists only for 2000 and 2002. For all others years the number of non- established positions on an FTE basis was estimated by dividing total wages in the Estimates by the average PersonalEmoluments. 54 established employees were taken on as temporary staff to deal with natural disasters, reportedly to assist inrelief, cleanup, and reconstruction efforts, or in advance of elections; however, their number has not since declined. Furthermore, the head count numbers for non-established employment on a yearly basis shows the extent to which the government has lost control over the level of government employment. As a first step, the government needs to keep track of non- established employment on a regular basis and report it in the estimates similar to the way it reports established employment. Table 5.1 Employment inthe St. Kithand NevisFederal Government, 1995-2002 1995 1996 1997 1998 1999 2000 2001 2002 1995-2002 EstablishedPositions 2,194 2,325 2,623 2,686 2,798 2,940 3,031 3,034 Annual growth rate(%) 6.0 12.8 2.4 4.2 5.1 3.1 0.1 4.7 EstimatedNon- EstablishedPositions (FTE) 411 612 726 767 763 888 932 922 Annual growth rate (%) 48.8 18.7 5.6 -0.5 16.4 5.0 -1.1 12.2 Non-EstablishedPositions HeadCount 1,038 1,362 Note: Non-establishedpositionson a full-time equivalent(RE) basisare estimatedby dividing total wages in the Estimatesby the average PersonalEmolument.Thisyields afigure for 2000 similarto the headcount estimateof 1,038 as of October2000, and 1,362 as of April 2002, and can be usedfor analyzingtrends in non-establishedemploymenton a full-time basis. Source: St. Christopher and Nevis, Estimatesfor the Year 1995-2002. 5.7 Three ministriesaccounted for almost 75 percent of established positions in 2001. Established positions are distributed among 15 ministries. As Table 5.2 shows, the largest ministry was Education, Labor and Social Security with 949 established positions, or 31.3 percent of the total; the second largest was National Security with 771, or 25.4 percent of the total; and the thirdlargest was HealthandEnvironment, with 532, or 17.6 percent of the total. 5.8 The Ministry of Public Works Utilitiesand Transport accounted for 35 percent of all non-establishedpositions in 2002. The Ministry of Health and Environment absorbed 23 percent of non-established workers, and Education, Labor and Social Security 23.2 percent. Unfortunately, information on the distribution of non-established positions across ministries for other years was not available. 55 Table 5.2 Distribution of Established Positions by Ministry in the St. KittsandNevisFederal Government, 2000-01 %T ions GovernorGeneral 5 5 0.2 Parliament 3 3 0.1 Audit 14 I5 0.5 Legal 50 51 1.7 Premier's Office 64 69 2.3 NationalSecurity 758 771 25.4 ForeignAffairs, InternationalTrade, andCARICOM Affairs 47 42 1.4 Finance, Development,andPlanning 192 203 6.7 Community and Social Developmentand Gender Affairs 29 46 I.5 Agriculture, Fisheries, Co-Operatives,Lands and Housing 102 101 3.3 Tourism, Informationand Telecomm, Commerceand Consumer Affairs 13 30 1 .o Communications, Works and Public Utilities 177 176 5.8 Education, Labour and SocialSecurity 949 949 31.3 HealthandEnvironment 511 532 17.6 Culture, Youth and Sports 26 38 1.3 Total 2,940 3,031 100.0 Source: St. Christopher andNevis, Estimatesfor the Year 2001 and 2002. worked for the St. Kitts and Nevis Federal Government as an 3g 14 - I)SL hiiir imiihcui5 established or non-established $ 1 2 - employee. Three in 10 government IO - employees were non-established workers. As Figure 5.1 shows, St. e - Kitts and Nevis's ratio of l5 =5 -f 6 - *Grenada government employment to 4 - ,S+Vlmmtaadthe *S+Ludr Cmnadlna + D o d d e r *Jamaica population (14.2 percent) is at least 3 2 - 5.10 The government has recognized its excess of established and non-established employees. In 2000, the St. Kitts and Nevis government imposed a freeze on new hiring and identifiedthe wage bill as an area where cutbacks were needed as part of a strategy of fiscal consolidation. However, although the freeze has been implemented with some success for established employees, the non-established employment head count has increased since 2000 in total by 31percent, or 3.8 percent infull-time equivalents. 56 5.11 The government, in delaying a decision to downsize its civil service, is hoping that the private sector will providejobs for potential laid-off public servants. The government is hoping that the new Marriott Hotel will be able to absorb some of the potential laid-off public servants when it opens at full capacity in September-October 2003. With 900 rooms and an employee-per-room ratio of 1.5, it can be expected to require some 1,350 workers. This is an increase of about 8 percent in total employment in an economy with an unemployment rate of 4.5 percent, according to the latest labor survey available, dated 1995. The government believes that this employment increase will be enough to downsize public service employment by up to 10 percent, and to absorb some of the workers that will lose their jobs when it closes the sugar company (SSMC). 5.12 Moreover, the downsizing in public service employment required to close the fiscal gap is more than the 10 percent of the current number of employees foreseen by the government. Downsizing the civil service to 1995 levels by cutting 800 established positions and 500 non-established positions (head count) would reduce the wage billby EC$42 million, or 5 percent of GDP. The downsize will have to be across all line ministries, with National Security, Education, Health and Environment taking a larger cut becausethey represent about 70 percent of all government employment. Chapters 6 and 7 present a detailed discussion of options for downsizing in the health and education sector, while maintaining the level and quality of service delivery. On National Security, the ratio of police personnel per 100,000 habitants is 1,133 in St. Kitts and Nevis compared to 516 in Barbados and 605 inDominica.57Although, in St. Kitts and Nevis the figures include immigrationofficers, the numbers are twice those of other small countries inthe regionand call into question the size and composition of the police force in the country. See Box 5.1 for options for and constraints on downsizing in the experience of Guyana. Cottzpetisntion 5.13 In real terms, the average compensation per established worker decreased during 1995-2002. As Table 5.3 shows, personal emoluments from 1995 to 2002 increasedin nominal terms by a total of 62 percent, or an average of 7 percent per year; however, in the same period the number of established positions increased by 38.3 percent, or an average 4.7 percent, and consumer prices increased by 26.2 percent, or an average of 3.4 percent. Thus for 1995-2002, the real compensation per establishedworker decreasedby a total of 7 percent. 57The data for Barbados and Dominica are for 2000, and its source is the Seventh United Survey of Crime Trend and Operations of Criminal Justice Systems, United Nations. The data for St. Kitts and Nevis are from the Budget Estimates, 2000. 57 I Box 5.1 Civil Service DownsizinginGuyana I In Guyana, the Government has implemented a number of actions over the last few years to ensure a substantial downsizing of the civil service and the public sector. The following downsizing methods were implemented with various degrees of success: Termination: As part of the highly-indebted poor country (HIPC)andPoverty Reduction Growth Facility (PRGF) requirements, the government agreed to separate 1,000 core civil service security guards and 1,000 temporary and unqualified workers. The separation issue was put on the agenda because of extremely high wage increasesof 31 and 27 percent, respectively, in 1999and 2000, awardedby an arbitration tribunal. Itwas felt that by reducing the size of the core civil service, the budgetary repercussionsof possible future wage increases would be reduced. By November 1, 2000 the Government had separated about 1,359 security guards, representingover 11 percent of the "core" civil service (including medical workers but excluding teachers, police, and the armed forces). The cost of paying severance and relatedexpenditures totaled G$465 million, or about0.33 percentof current GDP. About 82 percent o f the package consisted of severance benefits, and, nearly 12 percent of the package went to superannuationbenefits for pensionableand non-pensionablestaff. There has not been any termination of employees categorized as "temporary and unqualified" so far. Since the Government still needs security services, it decided to outsource them. In order to ease the transition problems of the severed security guards, the severed security guards formed the company chosen to outsource these services, for the period November 1, 2000 to December 31, 2002. While the outsourcing of security services does not offer any immediate financial benefit to the Government, it could in principle be a right step in a broader approach to civil service reform. The Government has been thinking along similar lines, that is, separationand outsourcing, for some of its other work staff, such as cleaners and clerical services. Again, in principle, we would encourage such thinking, but the benefits wouid only be realized if the outsourcing processes were truly competitive and performancecontinuously monitored. HiringFreeze: SinceOctober 1998, the Governmenthas implemented a temporary employment freeze in the core civil service. The freeze affects promotions, new recruitment, reclassification, and so forth. Any exceptions to this freezehaveto be directly approvedby the Headof the Presidential Secretariat. Decline in Civil Service Employment: Since 1993, the total number of employees incivilian government (including regional administration, police, statutory bodies, and ministries) has dropped from approximately 56,000 to about 31,000, currently. In addition, employment in the public sector corporations has dropped significantly, including a decline on the rolls of the largest domestic employer, Guysuco (Guyana Sugar Company), from 27,000 in1992to lessthan 18,000, currently. Corporatization: Another significant step in public service reform has been the recent corporatization of the Georgetown Public Hospital-the largest in the country and accounting for about 60 percent of health sector staff. Since 1999, this hospital has been converted into a semiautonomous entity that receives Government budgetary support but is managed by its own Board. It has the power to hire and fire employees and to set wages. Since the hospital was taken over by an independentBoard, the number of employeeson its payroll has gone down from 2,000 to 980 (the 2,000 included as many as 900 "ghost" employees who were drawing wages from the hospital but were either working elsewhere, or were working for political parties). The hospital has hired about 100 new employees to fill critical positions and was successful at attracting these new hires with better salaries and conditions of service, since it is no longer bound by civil service rules. Not surprisingly, its current productivity indicators show a marked improvement over those of previous years. Of course, it is important that this institution starts generatingahigher proportion of its revenue than it currently does through fees and charges so that the burden of higher salaries is not eventually borne by governmentbudget. The Guyana RevenueAuthority and other agencies have followed the exampleof the Georgetown Public Hospitaland have recordedpositive results. 5.14 The pattern of increases in average compensation for established positions is more clearly revealed in the profile of salary scale K7, which is at the low end of the salary scale and applies to drivers, messengers, and others comparable positions. The category K7 inthe 58 salary scale applies to various types of relatively unskilled workers, and is not affected by compositional shifts inemployment. If salary scale K7 is compared to increases inthe consumer price index, it can be seen that there was no significant increase inreal terms for the salary scale K7 andthat the averagecompensationfor establishedemployeesbasically keptupwith inflation. 5.15 The ratio of the highest salary (K47) to the lowest (K22) in the St. Kitts and Nevis Federal Government salary scale was about 8.6 during 1995-2002. This compression ratio is close to the one observed inother developing countries, such as Cape Verde, and Chile, but four times higher than the ratio inOECD countries (approximately two to three times). Table 5.3 St. Kitts and Nevis Federal Government Compensation, 1995-2002 1995- 1995 1996 1997 1998 1999 2000 2001 2002 2002 PersonalEmoluments (EC$Millions) 56.59 56.83 62.33 71.60 75.92 78.87 90.68 91.46 Annual Growth Rate(%) 2.9 9.7 14.9 6.0 3.9 15.0 0.90 7.1 Average Compensationfor EstablishedPosition(EC$) 25,793 24,442 23,763 26,657 27,132 26,828 29,917 30,145 Annual GrowthRate(%) -2.9 -2.8 12.2 1.8 -1.1 11.5 0.8 2.3 RealAverage Compensationfor EstablishedPosition(EC$95) 25,169 23,957 21,43 1 23,202 22,846 22,115 24,177 23,884 Annual Growth Rate(%) -4.8 -10.5 8.3 -1.5 -3.2 9.3 -1.2 1.1 Wages (EC$Millions) 10.6 14.95 17.25 20.44 20.69 23.82 27.88 27.79 Annual Growth Rate(%) 41.0 15.4 18.4 1.2 15.1 17.0 -0.3 4.8 Salary Scale K7 (EC$) 10,380 10,380 10,920 11,460 11,460 12,660 12,660 12,660 Annual GrowthRate(%) 0 5.2 4.9 0 10.5 0 0 2.9 AnnualInflation (%) 2.0 8.7 3.6 3.4 2.1 2.0 2.0 3.4 Nore: The figuresfor personalemolumentsand wages inthis table differ from those in Chapter 1 becauseto do an analysisat the ministry level we hadto rely on the figures in the budget.The latterdoes not include"below the line" expendituresandbonuses, honoraria, and overtime. Source: St. Christopher andNevis, Estimatesfor the Year 1995-2002. 5.16 The composition of Personal Emoluments by ministries mainly reflects the composition of established employment. As Table 5.4 shows, Education, National Security, and Health accountedfor about 72 percent of total personal emoluments in2001. Incontrast, the composition of wages reflects the composition of non-established employment. The Ministry that stands out as having the largest share of non-established employment is Communications, Works and Public Utilities, which accounted for 33.2 percent of the wages paid to non-established employees in2001. 59 Table 5.4 Personal Emoluments and Wages by Ministry inthe St. Kimand Nevis Federal Government (EC$) 2000 2001 % Total 2001 Governor General PersonalEmoluments 188,319 212,258 0.2 Wages 112,000 123,200 0.4 Parliament PersonalEmoluments 79,882 92,411 0.1 Wages 23,000 25,300 0.1 Audit PersonalEmoluments 422,899 483,297 0.5 Wages 0 0 0.0 Legal PersonalEmoluments 928,394 1,412,669 1.6 Wages 90,574 131,379 0.5 Office of PrimeMinister PersonalEmoluments 2,125,8 13 2,212,069 2.4 Wages 45 1,268 587,830 2.1 Ministry of National Security PersonalEmoluments 18,600,468 21,787,101 24.0 Wages 922,572 1,014,830 3.6 Ministry of ForeignAffairs, InternationalTrade andCARICOM Affairs PersonalEmoluments 1,018,82 1 1,567,942 1.7 Wages 1,677, I05 1,667,476 6.0 Ministry of Finance, DevelopmentandPlanning PersonalEmoluments 5,814,936 6,669,2 11 7.4 Wages 160,327 275,177 1.o Ministry of Community and SocialDevelopmentand Gender Affairs PersonalEmoluments 1,449,720 1,449,720 1.6 Wages 154,719 324,532 1.2 Ministry of Agriculture, Fisheries, Co-Operatives, Lands and Housing PersonalEmoluments 2,749,002 3,221,435 3.6 Wages 959,669 952,401 3.4 Ministry of Tourism, Information andTelecomm, Commerceand PersonalEmoluments 527,753 1,249,57 1 1.4 Wages 1,720,448 2,182,946 7.8 Ministry of Communications,Works andPublicUtilities PersonalEmoluments 5,398,247 5,883,941 6.5 Wages 8,424,176 9,259,321 33.2 Ministry of Education, Labour and SocialSecurity PersonalEmoluments 25,886,737 28,554,459 31.5 Wages 4,134,056 5,382,904 19.3 Ministry of HealthandEnvironment PersonalEmoluments 13,063,9 14 14,573,092 16.1 Wages 4,838,536 5,716,390 20.5 Ministry of Culture, Youth andSports PersonalEmoluments 738,942 1,308,6 19 1.4 Wages 150,759 235,487 0.8 Total PersonalEmoluments 78,993,847 90,677,795 100.0 Wages 23,819,209 27,879,173 100.0 Note; The figures for personalemoluments and wages inthis table differ from those inChapter 1 becauseto do an analysisat the ministry level we hadto rely on the figures inthe budgetbook, which does not include "below the line" expendituresand bonuses, honoraria, andovertime. Source; St. Christopher and Nevis, Estimates for the Year 2000and 2001. 60 5.17 Government wages in St. Kittsand Nevis are significantly lower than private sector wages in the country across all occupations?' As Table 5.5 shows, this pattern is observed across all occupations, but the difference is more pronounced in managerial-level occupations. This is perhaps one of the reasons why St. Kitts and Nevis suffers from serious shortages of qualified administrators, policy advisors, and technical experts. Below the top level of management inthe government, there is a highturnover among middlemanagement, leaving the executive core responsible for administration, policy formulation, and implementation. The shortage of highly qualified personnel is exacerbatedby the low retirement age (55 years old). Table5.5 PublicandPrivateSector Wages inSt. KittsandNevis, 2001(EC$) Occupation Public Private (%I Driver 14,3 10 18,720 30.82 Junior Accounts Clerk/Officer 19,260 21,288 10.53 Secretary 24,150 25,104 3.95 StatisticalClerk 34,980 44,400 26.93 Computer TechnicianProgrammer 34,980 40,800 16.64 PermanentSecretaryIGeneralMgr.Financial 68,160 93,000 36.44 Source: St. Christopher andNevis, Estimatesfor the Year 2001, andDepartmentof Labor,November 2001. 5.18 In St. KittsandNevis, the Figure6 2 Average Wage as a percentageot GDP per capita,2001 (Saloctsd CanbbaanCountnest average wage for established workers is below other countries 350 in the subregion; however, in B Jamaica terms of GDP per capita it is ; m ~ closer to the ratios observed in tZB + St. Luck OECD countries. The average $ m *Grenada wage for established employees in 2 St. Kitts and Nevis at EC$29,917 g& 150 *St. i s lower than in St. Lucia, but KBsand Nevis higher than in Grenada. However, Cflm as a percent of GDP per capita 50 (Figure 5.2), it is by far the lowest 0 in the subregion, and much lower (1-n ~~~~~~~~~~~~~~~ and Schiao Campo 'Gaemment Ernploymen1and Pay A Globaland Rsgmal Perspectd Policy Rerearch than in Jamaica, and close to the ratio of about 1to 2 observed inOECD countries. 5.19 The personnelinformation system is fairly basic for the needs of St. Kitts andNevis. The information system that usually contains such staff particulars as date of hiring, date of increments, and movements between salary scales, training received, performance evaluation reports, and the date of retirement is manual (paper files) and inadequate for a modern public 58Clearly,jobs inthe government can be more attractivefor other pecuniaryand nonpecuniarybenefits than wages. 61 sector. It does not contain the information required for manpower planning, professional development, and the design and implementationof a civil service reform. 5.20 Moreover, the St. Kittsand Nevis government lacks a humanresources strategy. Job descriptions do not exist even for established positions. No performance appraisal system is in place, which makes it impossible to assess whether public servants are actually doing their job efficiently and effectively and to establish performance pay or bonuses to reward good ~erformance.~~ No analysis of training needs has been carried out. Training in essential skills such as writing, office procedures, financial and personnel management, and basic management is deficient. There is no training plan for individual employees to enable them to advance along their career paths. Training is often offered in an unsystematic way, and dependant on the availability of donor financing. 5.21 In St. Kitts and Nevis there is fragmentation and duplication of work among ministries, departments, and agencies, leading to a diffusion of responsibilities. The survey of public officials, carried out for the OECS Institutional and Organizational Capacity Review (IOCR) in 2001, noted a duplication of functions and that such overlap adversely affected the public service. This suggests that missions, objectives, and functions of the different ministries need to be reviewed to determine the most effective and efficient organizational structure. More broadly, there is concern about the degree of overlap of functions across the various island governments in the subregion. Opportunities to pool resources and realize lower unit costs through greater collaboration extend to almost all areas of public sector activity, including international relations, economic and social policy, and policing. The benefits of common organizational arrangements for tax administration, customs, audit services, police services, magistracy, and environmental protection were discussedinthe OECS IOCR. 5.22 There is considerable scope for the government to divest some of its functions to the private sector to improve quality and efficiency and increase coverage So far, only telecommunications has been privatized and garbage collection contracted out. The electrical utility, water and sewerage, port handling, the cruise ship terminal, and the government printing office could all be considered potential candidates for privatization. Some services where it is easy to specify output can be more efficiently performed by the private sector. These include road maintenance, airport services, laundry and catering services in hospitals, and cleaning and maintenance services in public buildings. The Post Office could be corporatized to improve efficiency. Finally, some public sector activities in the areas of social services, and education, could be transferred to civil society as is beingdone to a limited extent inSt. Lucia. 5.23 A Public Sector Reform Unit was recently established with the mandate to lead the process of public sector reform, which is directed at modernizing the public service. This will include taking an inventory of the human resources capacity of each ministry, identifying the relevant training to address any deficiencies discovered, and taking a critical look at the procedures and systems in use in the public service to make sure they have not become 59Based on the experience in Grenada, even with a good performance appraisal system it may be difficult to get managers to document any deficiencies of their staff given the close nature of personal relationships on the island. Nevertheless, the St. Kitts and Nevis government should implement a system of job descriptions and performance appraisals to facilitate management of the public service. 62 impediments to efficiency. It is encouraging that the government established a Public Sector Reform Unit to modernize the public service and to seek remedies for the deficiencies identified above. However, it is not clear whether the unit will be adequately funded and has the support of the Cabinet to carry out its mandate. 11. RECENTTRENDS PCBLICSECTOREMPLOYMENT COMPENSATIONINTHENEVIS I& AND ISLAND ADMINISTRATION 5.24 The institutions governing public sector employment in the NIA are similar to the ones in the St. Kitts and Nevis Federal Government. The Human Resources Unit in the Premier's Office plays the same role as the Establishment Division inSt. Kitts and Nevis. There i s only one Public Service Commission for both islands, but when making appointments for Nevis, the Public Service Commission is comprised of the chair and one member from St. Kitts and Nevis and two members from Nevis. The same Public Service Appeals Board also functions for both islands. 5.25 The NIA, like the St. Kitts and Nevis Federal Government, does not have a human resources system in place. Job descriptions and performance appraisals, which are required to effectively managepersonnel, are nonexistent. Also, no systematic analysis of training needs has been done. Table5.6 EmploymentinNevisIslandAdministration, 1995-2002 1 1995- 2002 Established and Non- 888 868 852 807 801 800 804 751 established Annual GrowthRate (%) -2.3 -1.8 -5.3 -0.7 -0.1 0.5 6.6 -3.3 Source: HumanResourcesOffice inthe NevisPrime Minister's Office. 5.26 Unlike inSt. Kitts andNevis, overall government employment (established and non- established positions) in Nevis decreased in total by 15.4 percent from 1995 to 2002, or 2.2 percent per year. As Table 5.6 shows, at its peak in 1995, the NIA had 888 positions when additional people were taken on inthe aftermath of Hurricane Luis (up from 779 in 1993). Ithas since fallen to 751 positions and a freeze is in effect. Discussions of the need for further downsizing are underway. 5.27 In Nevis, as in the Federal Government, non-established positions represent an important fraction of public sector employment. But, unlike St. Kitts and Nevis, the NIA maintains a head count of the total number of non-established positions. As of March 2002, NIA reported that it had 751 employees on payroll, of which 518 were established positions and 233 were non-establishedpositions. Moreover, a new personnel information system operational since March2002 identifiedan additional 347 workers paidweekly. 5.28 The main difference in the composition of government employment between the NIA and the federal government of St. Kitts andNevis is the high share inemployment of the administrative ministries. As Table 5.7 shows, the share of administrative ministries, such as the Ministry of Finance and the Premier's Office, is relatively high; both ministries accounted 63 for 16.5 percent of employment compared to about 9 percent in the St. Kitts and Nevis Federal Government.60 Table 5.7 EstablishedandNon-EstablishedPositions,by Ministry,inthe NIA, 2001-02 % Total 2001 2001 Deputy Governor General 4 3 0.4 Legislature 1 1 0.1 Audit 7 7 0.9 Legal 7 8 1.1 Premier's Office 39 41 5.4 Ministry of Finance 81 84 11.1 Ministry of Communications,Works, PublicUtilitiesandPosts 88 75 9.9 Ministry of Agriculture, Lands, Housing, CooperativesandFisheries 55 57 7.5 Ministry of Health, SocialDevelopmentandCommunity andGender Affairs 158 149 19.7 Ministry of TourismandCulture, TelecommunicationsandInformation 27 29 3.8 Ministry of Education 255 252 33.2 HumanResourcesDepartment 7 5 0.7 Ministry of EconomicPlanningandDevelopment,Trade & Industry, Youth andSports 45 31 4.1 Ministry of PhysicalPlanningandInfrastructuralDevelopment,Natural Resources andEnvironment 0 16 2.1 Total 774 758 100.0 Note: The figures for total numbers of positionsdiffer from those presentedinTable 5.6 becausethey are from the budgetbook. Source: NevisIslandAdministration, Estimatesfor the Year 2001and2002. 5.29 Inrealterms, the averagecompensationper establishedand non-establishedworker increased only slightly during 1997-2002. As Table 5.8 shows, personal emoluments and wages increasedinnominal terms by 54 percent, or an average of 9 percent per year. However, over the same period the number of established and non-established positions decreasedby 26.7 percent, or an average of 5 percent per year, and consumer prices increasedby 26 percent, or an average of 2.6 percent per year. Thus, for 1997-2002, the real average compensation per established and non-established worker increasedby a total of 7 percent, or 1.3 percent per year. This increase reflects changes in employment composition rather than an average increase of positions or compensation. It is important inthis comparisonto keep inmindthat becauseof the federalstructureof the government, the NIA does nothave a Ministryof NationalSecurity andForeignAffairs. 64 Table 5.8 PersonalEmolumentsand Wagesinthe NIA, 1997-2002 19 1998 19 2002 1997- 2002 PersonalEmoluments 22.05 25.73 27.66 29.04 33.84 34.03 andWages (EC$MiI) Annual Growth Rate(%) 16.7 7.5 5.0 16.5 0.56 9.1 Average Compensation 36,873 39,89 I 39,913 40,559 43,721 44,894 for EstablishedandNon- EstablishedPosition (EC$) Annual Growth Rate(%) 8.2 0.05 1.6 7.8 2.7 4.0 RealAverage 33,255 34,720 33,601 33,434 35,333 35,570 1.3 Compensationfor EstablishedandNon- EstablishedPosition (EC$95) Annual Growth Rate(96) 4.4 -3.2 -0.5 5.7 0.7 1.3 AnnualInflation(%) 3.6 3.4 2.1 2.0 2.0 2.6 Source: NevisIslandAdministration,Estimatesfor the Year 1997-2002. 65 6. THE HEALTHSECTOR 6.1 The Federation of St. Kitts and Nevis has a health system with a strong primary health care focus. It is dominated by publicly financed and publicly provided health care. There are no private clinics or hospitals, but there are physicians inprivate practice. Public spending on health during 1995-2002 was about 3.3 of GDP, on average. Public health spending inreal terms grew for the same period at about 4.4 percent per year. This growth is unlikely to be maintained over the short to mediumterm giventhe tight fiscal constraints. 6.2 Public facilities appear to provide reasonably good coverage, with a good level of consumer satisfaction. Nevertheless, the health system faces many challenges inthe future. Most immediately, the public system is facing strong budgetary pressures as part of the overall fiscal crisis. The health system also faces the challenge of adapting to a changing disease profile, a steady emigration of nurses, and the type of specialized services that it can provide, and provide most efficiently, given the country's small population size and federation structure. This chapter reviews health outcomes in St. Kitts and Nevis, how services are financed, organized, and delivered, and suggests both short- and medium-term measures to improve health outcomes, equity inthe system, and efficiency. 6.3 The mainfindings and recommendations canbe summarized as follows: e Improving and Sustaining Health Outcomes o Public health and primary health care programs need to find ways to sustain the improvements in infant mortality that were reachedin2001, through promotinghealthier lifestyles, better care duringpregnancy, and quality care inthe neonatal period. o Inaddition, becauseof changes indisease patterns and age structures of the population, St. Kitts and Nevis needs to strengthen public health and primary care in noncommunicable diseases and injuries, inorder to continue to improve health indicators and contain costs over time. This means focusing health promotion and disease control activities as effectively as possible on tobacco control, injury prevention, control of obesity through diet and physical activity, and screening and early treatment of diabetes andhypertension. The government is making strong efforts to strengthen HIV/AIDS prevention through behavior change, communication efforts (especially with high-risk groups), condom distribution, screening, treatment of sexually transmitted diseases, and voluntary counseling and testing. High-level leadership and support needs to continue and to be strengthened. To help in these areas, the government is implementing a World Bank- financed project of US$4.04 million to strengthen its national program. Given the importance of acting now to reverse the HIV/AIDS epidemic, the government needs to ensure adequate implementation and counterpart funding for these efforts. e Improving EfJiciency in the Delivery of Services o As the disease patterns change over time, and with the goal of achieving greater economies of scale, it may be possible to rethink the configuration of health facilities in 67 the federation by, for example, reorganizing and consolidating some of the primary health center services. The government achieves economies of scale in drug procurement by using the Eastern Caribbean DrugProcurement Service, which gets good prices from competitive tendering processes. There may be scope to achieve greater economies of scale by coordinating other services such as specialized laboratory tests and medical supplies. o The parallel health delivery systems inSt. Kitts and Nevis are cause for concern because of duplication inefforts that waste scarceresources.The government could look for ways to rationalize humanresource use and health service facilities by better coordination and referral across St. Kitts and Nevis. o There are user fees on publicly provided services, with broad exemptions (for example, children and the elderly). Collectionrates are low. Given the tight fiscal constraints, there may be some scope for increasing revenue through better collection, especially if facilities are permittedto retain the fees to cover some of their costs. The user fee policy needs periodic review to ensureboththe desiredequity and efficiency impact. e Uses of Health Spending o Wages absorb most of the spending inthe health sector and raise questions as to whether there is sufficient allocation for pharmaceuticals, medical supplies, and maintenance of equipment and buildings for the efficient delivery of services. There may be significant opportunities to generate cost savings without affecting the quality of services by substitutingsomenursingpositions with lower-cost nursing(nurse assistants). o Hospitals are absorbing an increasing share of public spending over time. This is a concern and is likely to continue unless (a) changes are introduced to improve hospital efficiency and contain costs, and (b) noncommunicable disease prevention and early diagnosis and treatment programs are strengthened to keep people healthy and out of hospitals. It may be possible to reduce average length of stay at the Joseph N. France General Hospital by managing more pre- and postoperative tests on an outpatient basis, improving scheduling of the operating room, and other practices. There should be considerable scope to increasingly substitute inpatient surgery for outpatient surgery. Some of these efficiency changes may require capital investments. o Donors are the main source of capital spending in health. As is true in many other countries, the government needs to ensure that donor-funded projects are in line with government priorities and are sustainableover the longrun. I. HEALTH OUTCOMESINST. KITTSAND NEVIS 6.4 St. Kitts and Nevis has good health indicators, although infant mortality-until recently-was higher than it should be for a country at its level of GDP per capita, and life expectancy appears to be lower than expected. Given its stage in the epidemiological transition, noncommunicable diseases and injuries are an increasing share of the burden of disease and a leading cause of mortality. Public health programs and primary care need to be reshaped to prevent injuries andchronic diseases and to provide early diagnosis and treatment. If 68 this i s not done, treatment of these ailments will overburden the health system with high costs from, for example, complications of diabetes and hypertension. Inaddition, HIV/AIDS needs to be addressedas vigorously as possible. Table 6.1 Health Indicators inSt. Kithand Nevis Compared with Barbados, Jamaica, and other OECS Countries, 2000 Total GDPPer InfantMortality T Population Capita Estimate 2001 00OLive Birth 2001 (current$) 2000 St. Kittsand Nevis 41,082 8,345 19* 2.2 71 Antigua and Barbuda 68,487 9,729 15 1.7 75 Barbados 268,189 NA 17 1.8 75 Dominica 73,199 3,563 NA NA 76 Grenada 99,000 4,022 13 3.2 72 Jamaica 2,668,230 2,9 17 20 2.5 75 St. Lucia 158,134 4,356 13 2.0 71 St. Vincent and the *Grenadines andNevis, the comparable IMRnumberin2001 ismuchlower-12.5 per 1,OOO-due to asharpdecline in 115,881 2,9 14 20 2.1 73 For St. Kitts neonatalmortality that year. Source: World Development Indicators. 6.5 Infant and Child Health. Given its level of income, St. Kitts and Nevis has had relatively high infant mortality rates (IMRs), at 19 per 1,000 births in 2000. As shown in Table 6.1, this level of IMR has been on par with Jamaica and St. Vincent and the Grenadines- countries with less than half the per capita income of St. Kitts and Nevis. This is a problemthat the government has recognized and has been trying to address, with some success. In2001, the rate fell to 12.5, the lowest recorded, due to a sharp decline in neonatal mortality in that year (Table 6.2.) It remains to be seen whether this improvement in neonatal health was a one-year change or whether it will be sustained. A problem, linkedto the IMR, is the percentage of births that are low birth weight, at 2,500 grams or less. This is also relatively high, at 12.6 percent in 1998. Low birth weight is a concern bothfor infant health and for long-termhealth. Evidence is growing that low birth weight is also a risk factor for health problems later in life, including obesity andhypertension inadults. 6.6 Total fertility rate is low, at 2.2, but a significant share of births is to teenagers. In 1995, about 17 percent of all births were to teenagers. Teen births are a problem for many reasons. Teenage pregnancy is often associated with school dropout. Teens are rarely ready psychosocially and financially to start families, and teen pregnancy is a risk factor for infant mortality. 69 Table6.2 HealthIndicatorsinSt. Kitts and Nevis, 1995and2001 Life Total Infant Neonatal Post- Proportion Year Expectancy Fertility Mortality Mortality Neonatal atBirth RateBirth Rate Rate Mortality per Per 1,000 Rate Women LiveBirths 1995 69 2.5 25 NA NA NA NA NA 2001 71 2.2% 12.5 7.5 4.9 99.2 2 4.5* "Refers to 2000. Source: World Development Indicators andCore HealthDataInitiative 2002 DataUpdateProcess, HealthInformationUnit, Ministry of HealthandEnvironment, St. Kitts and Nevis. 6.7 Breastfeeding is well accepted, but infant and child health would improve if the numbers initiating breastfeeding and the duration of exclusive breastfeeding were longer. Vaccination levels are excellent-close to 100 percent. InSt. Kitts, 67 percent of babies were exclusively breastfed for one month. By the third month, this falls to only 22 percent. Rates are similar in Nevis. As in neighboring countries, the main health problems in children are respiratory infections, diarrheal disease, asthma, and accidents. Worms are another health problem of children, not surprisingly given the sanitation problems, which the school system is trying to addressthrough mass deworming. 6.8 Noncommunicable Diseuses. Given the aging of the population and the epidemiological transition, noncommunicable diseases are the main causes of death in St. Kitts and Nevis. The major causes of death from noncommunicable diseases are diabetes, stroke, and hypertension (Table 6.3). Public health programs and primary care need to be reshapedto prevent injuries and chronic diseases and to provide early diagnosis and treatment. Table 6.3 LeadingCauseof Mortality, 1996-2001 (number of deaths per year) Disease 1996 1997 1998 1999 2000 2001 Stroke 69 96 56 65 57 62 Diabetes 51 88 _ _ 57 48 66 Hypertension -- _- -- 46 47 56 Myocardial Infarction 41 76 30 35 48 31 Conjunctive Heart Failure 36 71 20 33 28 17 Septicemia 33 55 38 56 45 41 Pneumonia 28 30 11 49 55 41 CardiovascularDisease 28 30 I 1 1 18 16 Senility 21 25 1 0 6 5 Anemia 18 19 18 19 19 17 Cancer 46 49 52 18 RenalFailure 17 35 28 37 Cardiac Arrest 17 18 19 40 Asphyxia 11 11 2 7 Ischemic Heart Disease 2 26 6 7 Natural Causes 7 0 22 11 --AIDS available. 1 7 5 5 8 2 =Not Source: Health InformationUnit, Ministry of Health andEnvironment, St. Kitts and Nevis. 70 6.9 St. Kitts and Nevis is a transit point for the illicit drug trade. Drug abuse is a problem on the islands, although good estimates are lacking on the magnitude of the problem. There is a National Council on Drug Abuse that has received funding from the Canadian International Development Agency (CIDA) for drug abuse programs. Community psychiatric programs report that schizophrenia is the top diagnosis for patients attending those programs, followed by substance abuse. 6.10 HIV/AIDS. While there is underreporting and late reporting of new cases of HIV/AIDS, the epidemic is clearly a problem in St. Kitts and Nevis, as it is for the Caribbean at large. From 1996 to 2000, 26 people died of AIDS-related causes. Of these deaths, 16 were male, and 10 were female, resulting in a male-female ratio of 1to 6. By 2001, the Caribbean Epidemiology Center (CAREC) and the Centers for Disease Control (CDC) estimated there were 347 people living with HIVIAIDS, two of them children. This is an adult prevalence rate of about 1.8 percent, which reflects an epidemic that is still concentrated. Many factors in St. Kitts and Nevis are influencin the shape of the epidemic that the HIV/AIDS program needs to recognize and try to address. Some of these factors are: F, a Sexual activity starts at a young age. a There is a pattern of sex between young girls andolder meninexchangefor food and clothes. a Homosexuals and people living with HIV/AIDS are stigmatized and discriminated against. 0 The Rastafariancommunity does not accept condom use. a The diseasecan be transmitted via migrant labor flows to and from other parts of the Caribbean and tourists. The government has developed an HIV/AIDS project with US$4.04 million in World Bank support (see Box 6.1). 6.11 The under-5 mortality and HIV/AIDS MillenniumDevelopmentGoals (MDGs) are the most applicableto St. KittsandNevis of allthe health-relatedMDGs.This is becausethe other health-related MDGs: malnutrition (defined as the prevalence of underweight children), maternal mortality, and tropical diseases (other than dengue, which is endemic) are not major issues in the country. There is one maternal death every few years in St. Kitts and Nevis; the country has relatively good reproductive health indicators and there are only about 700 births a year. The under-5 mortality rate inSt. Kitts and Nevis dropped from an estimated 36 per 1,000 in 1990 to 24 per 1,000 in2001. This is a decline of about 3.7 percent per year. The MDGtarget-a reduction of two-thirds from 1990 to 2015-would require that the rate fall to about 12 per 1,000 by 2015. On trend projections, St. Kitts and Nevis will not reach but will come close this MDG target, since the present trend would result ina rate of 14 per 1,000 by 2015. The rate of decline is however likely to slow from 2001 to 2015, because as child mortality rates get lower and lower, incremental declines get harder and harder to achieve. It is difficult to assess the likelihood of reaching the HIV/AIDS targets, because baseline data on HIV prevalence and incidence are poor. Adult H I V prevalence is probably somewhere between 1.5 and 2 percent. 6' StrategicPlanfor the National Response to HIV/AIDS St. KittsNevis, 2001-2005. 71 The Government has launched acomprehensiveHIV/AIDSprevention andcontrol programwith the suppott of the World Bank with the objective of reversing its spread. Box6.1 Strengthening the NationalHIV/AIDS Programover 2003 to 2007: The St. Kittsand Nevis HTC'IAIDS Prevention and Control Project In2002, St. KimandNevis met the eligibility criteria for country participation inthe World Bank-financed Multi-Country HIVIAIDS Prevention and Control Adaptable Program Lending for the Caribbean, enabling the country to receivea US$4.04 million loan to strengthen its National HIV/AIDS Program During 2003-07 the project will help control the spread of HIV/AIDS through (a) scaling up prevention programs targeting high-riskgroups and the general population; (b) strengthening treatment, care, and support for people living with HIVIAIDS; and (c) strengthening St. Kitts/Nevis's multisectoral institutional capacity to better respondto the epidemic. With respectto treatment, the project will support capacity strengtheningand upgradingof current services available for introducing antiretroviral therapy, care, and support for people living with AIDS. The total project cost is US$4.5 million, of which the government i s contributing $450,000, or about I O percent. The project is expectedto generate significant net benefits from the slowing and, ultimately, reversal, of the human, social, andeconomic costs that accompany the rise inHIVIAIDS cases. While the project's economic net benefits are expected to be significant, the government needs to realistically plan for the counterpart funding requirements duringthe project period, loan repayments, and the long- term incrementalrecurrent costs generated by the expandedprogram, especially inlight of the current fiscal outlook. The postproject recurrent costs are estimated at US$236,000 per year. Of this, US$184,000 per year is truly incremental, assumingthat the current numbers of governmentstaff dedicated to HIV/AIDS are not increased. 11. THEHEALTHSYSTEM: COVERAGEANDDELIVERY 6.12 The health system in St. Kitts and Nevis, as in the other OECS countries and the United Kingdom (U.K.), is predominantly publicly financed and publicly provided. All hospitals and clinics in St. Kitts and Nevis are public, although about 24 percent of physicians are infull-time private practice (Table 6.4). Health systems can be dominated by public finance, either through general revenue (the U.K.) or social insurance (Germany), by private finance (China), or by a hybridof the two (the United States). Provisioncan be largely public, private, or a mix. There is no ideal model of health systemfinancing and delivery. 6.13 St. Kitts and Nevis's small country size and federation structure pose many challenges to the administration of public health programs and clinical services, such as (a) unexploited potential economies of scale, (b) duplication across islands of planning and monitoring activities, (c) health services mostly publicly provided, (d) health financing mostly through general revenue, and (e) ahighratio of nurses to physicians. 72 Table 6.4 Characteristics of the Health System, St. Kittsand Nevis, 1991-2001 1991 1992 1993 1994 1995 19% 7 1998 1999 2000 2 HospitalBeds 268 268 268 268 268 268 244 150 199 246 178 Doctors 28 31 32 39 46 48 48 46 46 46 49 Full-Time Private 12 12 13 12 13 15 14 15 16 13 12 Temporary 3 5 4 6 7 12 7 5 6 3 5 Dentists 7 9 9 10 11 11 10 12 14 14 15 Nurses 260 262 262 218 200 206 201 203 216 225 249 NursingAssistants 24 24 34 34 42 48 42 46 48 45 45 Community Health 17 17 17 17 17 17 17 32 32 28 48 Workers HospitalAdmissions 4,839 4.004 4,083 NA NA NA NA NA 4,997 4,98 1 4,784 Note: Hospitaladmissionsare shown as NA for 1994-98 because existingdataare for St. Kitts only. Source: HealthInformationUnit, Ministry of Healthand Environment, St. Kitts and Nevis. 6.14 As the disease patterns change over time and with the goal of achieving greater economies of scale, it may be possible to rethink the configuration of health facilities in the federation. For example, it might be possible to reorganize and consolidate some of the primary health center services. There are 17 primary health centers across the Federation, 11in St. Kitts and 6 inNevis. These centers provide general nursing, medical, and some specialist services. The health centers are managed by district nurses/midwives. Secondary care is provided by Mary Charles Hospital (10 beds) and Pogson Hospital (22 beds) in St. Kitts, and Alexandra Hospital (54 beds) in Nevis. The Alexandra Hospital has a 24-bed infirmary for psychiatric and geriatric care, inaddition to its 54 beds. The Pogson Hospital is scheduled for demolition and rebuilding as a polyclinic, at a cost of US$1.7 million. The main referral hospital, Joseph N. France Hospital, located in St. Kitts, has 150 beds. It sustained heavy damages from Hurricane George inSeptember 1998and was recently renovated andexpandedwith assistance from the European Union. Its occupancy rate in 1998 was reported at 65 percent and its average length of stay at 9.5 days. These statistics indicate much room for efficiency improvements. Complicated cases are often sent abroad, to neighboring countries, or to Puerto Rico, with the government sometimes assisting with airfare and some or all of the hospital fees.62The CardinHome inSt. Kittshas 100 beds for the elderly. It also has a ward for the mentally ill.All these facilities are public. 6.15 An excellent example of exploiting economies of scale at the regional level, and that can be extended to other services, is the Eastern Caribbean DrugProcurement Service. The St. Kitts and Nevis government uses a National Formulary to guide purchasing. The government may be able to achieve greater economies of scale by better coordinating other services with its OECS neighbors and other countries in the region, such as specialized laboratory tests and procurement of medical supplies. 6.16 Remarkably, although the federation of St. Kitts and Nevis has a combined population of only about 44,000 inhabitants, each island has largely independent health systems that duplicate each other, only further exacerbating the challenges of running an efficient health service. There i s coordination, but Nevis has its own ministry responsible for 62Reliable information on these expenses was not available. 73 health, and its own chief secretary responsible for the administration of local health services.63 Both St. Kitts and Nevis have separate budgets for health, approved by each island's statutory entity. 6.17 There are no private hospitalsin St. Kitts and Nevis, although there are physicians providing clinical services on a private basis. Table 6.5 indicates that both the poor and the non-poor use a mix of public hospitals and clinics and private doctors when in need of medical care. More non-poor than poor used a private physician when care was needed43 percent compared to 27 percent. Inaddition, the poor turnto public clinics more frequently than the non- p o o r 4 2 percent compared to 20 percent. Both poor and non-poor have about the same reported use of public hospitals, not surprisingly, since there are no private hospitals. Table 6.5 Uses of HealthFacilities inSt. Kitts and Nevis, 1999 (in percentage) PlaceVisited Poor Non-Poor Total Public Hospital 24 26 25 Public HealthCenter 42 20 25 Pharmacy 4 3 Private Clinic 2 1 Private Doctor 27 43 40 Other 6 6 6 Total 100 100 100 N 33 129 162 - =Zero observations. Source: "Poverty Assessment Report: St. Kitts and Nevis" (2000). 6.18 Satisfaction with medical services is very high. The 1999 Poverty Assessment Report documented that about 96 percent of the population in St. Kitts and 86 percent in Nevis was satisfied with the medical services received. Consumer satisfaction is a key component of health service performance. Many health organizations are increasingly using exit interviews and other techniques to provide feedback from consumers on their perception of health service quality. These are tools that St. Kitts andNevis may wish to also adopt ina systematic fashion. 6.19 Public health financing is mainly through the consolidatedfund. Revenue from user fees is minor. User fees at public facilities amount to only about US$6 per capita per year, or 3 percent of health recurrent spendingin2000. Facilities do not retain fees, so have little incentive to put effort into collection. Fees revert to the consolidated account. Furthermore, there are broad exemptions, such as the population aged 16 and under, and over age 62. Fees are modest. In Nevis, for example, the charge for a consultation is about US$15. Daily fees at the Joseph N. France General Hospital are US$7.50 for care in a six-bed unit. Patients are charged US$65 per day for a private room. While fees may seem low on a per service basis, user fees for a catastrophic health problem could put an unacceptably high burden on a poor household's financial situation, because the exemptions are not well targeted. The user fee policies would benefit from a review interms of equity andefficiency. The review should consider whether fees 63The Chief Medical Officer and Director of Community Health Services are responsible for coordinating services throughout St. Kitts and Nevis. 74 should be increased, decreased, or eliminated for certain services. The review should examine whether the indigent are deterred from getting services or needed drugs due to user fees, and if exemptions are working properly or could be improved. If retained, the government should consider permitting facilities to retain fees at the facility level to improve services. The government should also consider how fees might be significantly increased to cover costs for physicians who treat their private patientsinpublic facilities, and for overseas visitors. 6.20 The government is considering adopting a national health insurance system based on payroll taxes. Given that there is a general-revenue-financed system in place, it is not clear why it would make sense to add a payroll-tax system of social insurance. The main motivation may be simply to find an additional source of revenue for the health system. However, payroll tax systems have disadvantages. They are a tax on labor. Inaddition, given that many households are outside the formal sector, it would not cover the full population. Working with the existing financing system, but finding ways to improve efficiency to ensure that the country is getting better value for money, could be a better choice in the long run for such a small country. There appear to be many ways that efficiency could be improved over the short to medium term: reorganizing and consolidating the primary health center services, substituting nurses with nurse assistantswhere possible, and increasing the reliance on well-targeted user fees for financing. 6.21 The physician-to-population ratio is on par with neighboring countries, at about 1 physicians are trained in the University of the West in die^.^^ It also relies on temporary per 1,000. However, the ratio of nurses to physicians is relatively high. Most of the physicians to meet the physician needs in the country. There are about five nurses for every physician (Table 6.4). This high nurse-to-physician ratio can help facilitate the primary health carehealth promotion focus that the government is striving to achieve. However, despite the highratioof nursesto physicians, the government is concerned about nursing emigration andthe many nurse positions that are vacant. Furthermore, St. Kitts and Nevis has a very large proportion of nurses relative to nursing assistants-about 5 to 1. It could consider making efficiency improvements by substitutingnurses with nursingassistantswhere possible. 11. USES HEALTH OF SPENDING INST. KITTSANDNEVIS, 1995-2002 6.22 Public spending on health as a share of GDP in St. Kitts and Nevis is on par with several other countries in the region (3.3 percent). Table 6.6 compares health expenditure estimates in 2000 for several OECS countries, Barbados, and Jamaica. The World Health Organization (WHO) estimates that health spending inSt. Kitts and Nevis amounts to about 5.2 percent of GDP. Public spendingis almost 60 percent of the total, at 3.1 percent of GDP in2000. In2000, total healthspendingfor St. KittsandNeviswas estimated at about $447 per capita. 64There are two small medical schools in the Federation, the International University of Health Services and the Windsor University School of Medicine, both catering to foreign students. 75 Table 6.6 Health Spending inSt. Kithand Nevis Compared to Barbados, Jamaica, and other OECS countries, 2000 (as percentage) PublicHealth Shareof St. Kittsand Nevis 447 5.2 59 41 Antigua and Barbuda 562 5.5 60 40 Barbados 606 6.4 65 35 Dominica 247 6.1 71 29 Grenada 212 4.8 70 30 Jamaica 165 5.5 47 53 St. Lucia 202 4.3 62 38 St. Vincent and the Grenadines 190 6.3 65 35 Note: Estimatesof public spending are more reliablethanestimates of privatespending. Source: World HealthOrganization, PreliminaryInformation. 6.23 Since 1995, the share of recurrent health spending in Figure6.1 Share of RecurrentHealthExpendituresinGDP, S t Kilts andNevis 1995-2000 GDP has been stable at around (percenlage) 3.3 percent of GDP (Figure 6.1). 4.0% In the same period, St. Kitts's 3.5% / health sector received on average 3.0% - 9.8 percent of government 2.5% - recurrent spending. The equivalent 2.0% - figure in Nevis was 13.5 percent. 1.5% These shares seem adequate; however, given the difficult fiscal situation that the government 0.5%0.0%I 0 % l faces, savings canbe obtained. The expected cyclical fluctuations in 1995 1996 I997 1998 1999 2000 2001 2002 Source SI ChnstopherandN e w Esbrnatesfor the Year andNewsIsland Admlnistrabon (NIA) Enmatesforme social expenditures as a response Year Note informmonforNews 1999wasnotavail& 0St.KMr+St.Kittr and Nevir to economic shocks or naturaldisastersare not present inSt. Kitts andNevis health expenditures, mostly because of the large share of personal emoluments and wages in the sector's recurrent expenditures. 6.24 Public recurrent health expenditures in St. Kitts grew during 1995-2002 by 4.7 percent per year in realterms. As Table 6.7 shows, inNevis the growth over the same period was 3.7 percent per year inreal terms. Wages absorb about 80 percent of all health spending in St. Kitts and 73 percent in ne vi^.^^ These figures are on the high side, and raise questions as to whether there is sufficient allocation for pharmaceuticals, medical supplies, and maintenance of equipment and buildings. As discussed in paragraphs 6.15 and 6.22, there is scope to reduce personnel spending by consolidation of health centers and substitution of nurses for nursing assistants. `jThe slightly lower share inNevis may be because no pharmaceuticalspending appears inthe Nevis budget. 76 Table6.7 RecurrentHealthSpendinginSt. KittsandNevis, 1995-2002 (inEC$95) Nevis St. Kitts and evis 1995 14,406,756 5,637,789 20,044,545 1996 14,743,53 1 4,926,805 19,670,336 1997 14,431,226 4,749,325 19,186,55 1 1998 14,833,8 10 5,657,470 20,49 1,280 1999 16,356,606 NA 2000 18,966,38 1 6,585,263 25,55 1,644 2001 19,821,246 7,389,230 27,2 10,477 2002 19,855,460 7,294,585 27,150,045 Growth Rate Per Year 1996-2002 4.7 3.7 4.4 Nore: All numbers referto actualexpenditures,exceptthe numbers for 1995 for Nevis, and 2001-02 for both islands, whichareestimates. Source: St. Christopher andNevis, Estimatesfor the Year and Nevis IslandAdministration, Estimatesfor the Year. 6.25 Public recurrent spending on Figure6.2 Shareof HospitalExpendituresin Told RecurrentHealth hospitals in St. Kitts increased Expenditures.SL ~ m and Nevis 1995-2002 s 90.00% (percentage) dramatically during 1995-2002 (Figure 80.00% 6.2). Its share in total recurrent 70.00% expenditures increased from about 52 60.00% percent in 1995 to close to 59 percent in 60.00% 2002. This finding may be partly due to 40.00% definitional changes. The budget 30.00% categories in use on a year-to-year basis 20.00% hadfrequent changes. However, it is likely 10.00% that hospitals are putting pressure on 0.00% ~ 2002 health spending in St. Kitts, and this is 996 1996 1997 1998 1999 2000 Z M I certain to continue unless (a) changes are introduced to improve hospital efficiency, and (b) chronic disease preventionandearly diagnosis and treatment programs are strengthened to keep people healthy and out of hospitals. InNevis, the pattern is the opposite-the share of hospital spending in recurrent expenditures decreased from 65 percent in 1995to about 62 percent in2002. 6.26 Capital expendituresin health are highly donor dependent, with a very large share from the European Community. The government ultimately needs to start assuming this responsibility with its own revenue base. A complete discussion of the Public Sector Investment Program inthe health sector is presented inChapter 4. 77 7. THE EDUCATIONSECTOR 7.1 Education is fundamental for economic and social development. The knowledge and skills taught in the education system are a highly valued commodity in the labor market. For example, in St. Kitts and Nevis, attainment of postsecondary education, seven to eight years of additional schooling after primary education, allows an individual to increase earnings by more than 70 percent for the entire period of labor market participation, 45 years.66 Furthermore, human capital investment is known to bring better societal outcomes, such as reduced youth delinquency and improved health. 7.2 The close link between education and labor market earnings implies that lack of education is one of the primary sources of poverty in St. Kitts and Nevis. Data from the 1999 Survey of Living Conditions indicate that 27 percent of those with no education or a primary education are poor, while only 18 percent, 12 percent, and 5 percent are poor among individuals with secondary, postsecondary, and university education, respectively. Hence, there are strong economic motives for the individual and the government to invest ineducation. 7.3 Education expenditures in St. Kitts and Nevis, on average, were about 5.2 percent of GDP during 1994-2001, with a rising trend toward the end of the period. However, this allocation inpublic resourceshas not always been accompaniedby an improvement ineducation outcomes, equity, and efficiency. The fiscal crisis the country is facing will require a cut in expenditures for all sectors, but this should not preclude the government from completing the educational agenda by making better use of the available resources and increasing the nongovernmental provision ineducation, for example, intertiary education. 7.4 This chapter reviews education outcomes inSt. Kitts and Nevis, discusseshow education services are financed, organized, and delivered, and suggests interventions to improve outcomes, equity, and efficiency. The mainfindings andrecommendations canbe summarized as follows: Expensive delivery of education at the primary and secondary level. St. Kitts and Nevis has the most expensive delivery system of the OECS countries, mostly due to its high pupil-teacher ratio. Among the initiatives that the government could consider to increase the pupil-teacher ratio are: 0 Merging of primary schools. Inefficient use of teachers takes place predominantly at smaller primary schools. A merging of schools based on school mapping would exploit economies of scale inthe education system and reducecost, while increasing quality. 0 Instituting a curriculum reform of secondary education. A common core curriculum for the first three forms of secondary education would reduce the excessive offering of subjects of a specialized nature, such as typewriting, auto mechanics, andmetalwork, and lead to larger classes. 0 Mandating of a minimum floor on class size at the primary and secondary level. Many schools, in particular at the secondary level, offer subjects with 66Without controlling for other factors influencing labor market earnings, such as age, gender, and ability, these observations suggest a return to investment in human capital of the magnitude of 10 percent, a rate that often exceeds-in both magnitudeand certainty-that of investment inphysical capital. 79 low demand, leading to small class sizes and high costs. Mandating a minimumclass size of, say, 15 students, a rule inmany developed countries, would increaseefficiency. 0 Rewarding wel2-pe~ormingschools. The Ministry of Education could provide incentives for each school to deliver education more efficiently by rewarding schools that commit to increase the pupil-teacher ratio while retaining part of the savings. 0 Low quality of education due to an automatic progression system and very low level of teacher training. The government couldconsider: 0 Introducing an efective monitoring and evaluation system. With automatic progression inprimary school, it is critical to establish a monitoring system to guarantee that all students have learned sufficient to graduate from primary school. 0 Rethinking teacher training and hiring regulations. All graduates from secondary school should be eligible for Teachers College-regardless of prior teaching experience. To address the lack of teaching experience a mentoring programshould be introduced as an element of teacher training. 0 Only certified teachers should qualibfor open positions. This would prevent the quality of the teacher corps from deteriorating further. o A major retraining program of the existing teaching stafs should take place. Provided efficiency gains take place in the system, this would generate time andresourcesfor teachersto attendretraining courses. 0 Reducing public unit costs in tertiary education by: o Raising fees. Currently, nationals are exempted from paying tuition fees, but they have to shoulder other limited fees that amount to 5 percent of total costs. Moreover, many tertiary graduates migrate, which raises the issue of the large subsidy to higher education. Increasing this ratio to 20 percent would generate EC$1million (3 percent of the recurrent education budget). o Abolishing salaried students in Teachers College and nursing education. Currently, around 25 students at the Teachers College and 10 students in nursing education are paid full-time salaries, while students of other disciplines are unpaid. Elimination of this differential treatment of students would provide the government with an estimated saving of EC$650,000 (2 percent of the recurrent education budget). o Increasing regional cooperation among colleges, which in the medium run could pave the way for quality and efficiency improvements through specialization. o Stimulating expansion through negotiating performance contracts with the main provider, which is a medium-term way to increase enrollment at little cost. 0 Reducing administration costs through deeper cooperation within the country and at the regional level with the development of exams, information systems, strategies, and curricula. This would reduce duplication of work and lead to further efficiency gains 80 through a common curriculum, common regional exams, and collective textbook purchases and sharing of teachers. A 25 percent reduction in administration costs would implysavings of EC$940,000 (2.5 percent of the recurrent education budget.) Better targeting of the schoolfeeding and textbookfree-of-charge programs. At least 65 percent of the well off, or half of the student population, receive free meals. Providing school feeding free of charge is therefore generous by the state, which in times of fiscal crisis could be scaled back to benefit only the needy students. The government could therefore recover costs with notable exceptions for the needy pupils. For instance, elimination of free school meals to children from the most affluent half of the population would free up resources amounting to EC$900,000 (2.5 percent of the recurrent education budget) that could be used to give free meals to all poor children. A similar pattern is discussed with free textbooks. I. THEEDUCATIONALSYSTEMINST.KITTSANDNEVIS 7.5 Except for pupils in early childhood, Kittitians and Nevisians receive education predominantly from public institutions. Table 7.1 presents a snapshot of the education system by ownership. The essence of the table is that toddlers predominately attend private institutions-70 percent of early childhood enrollment-whereas pupils in primary and secondary education overwhelmingly attend public establishments. Finally, all students in tertiary education attend public institutions. Table 7.1 EducationInstitutionsinSt. KittsandNevisby Ownership, 2000-01 4,493 (97%) 143 (3%) 4,636 (100%) Note: 1. Early childhood includesenrollment in preschool, daycare, and nursery.2. Training includes youths attending nonformal education (115 students) andadvanced vocational education (26 students). Source: "Education PlanningDivision Statistical Bulletin, 2000-200 1 ." 7.6 Provision of early childhood education is an example where, if correctly managed, private provision can result in substantial savings for the public treasury. The advantages and disadvantagesof the private provision of education continue to be intensely debated inmany countries. However, increasing nongovernmental provision or investment in education could lower the financial burden on the government, while maintaining the nation's overall level of investment. Such policies are discussedbelow inthe case of tertiary education. 7.7 The education system in St. Kitts and Nevis stands out in a Caribbean context in two ways: (a) the dual ministerial structure for one country due to its federal structure, and (b) the automatic progression in basic education, where no child repeats a class unless the parents directly express the wish. The dual ministerial structure increases administrative costs. The 81 automatic progression system leads to impressive educational indicators interms of enrollment, graduation, and repetition. However, it hides quality and inefficiency problems. 11. EDUCATION OCTCOMES INST, KITTSAND NEVIS 7.8 Inearly childhood, the net enrollment rate of children between2 and 5 years of age is estimated to have been around 80 percent in 2001.67 This is a high enrollment rate considering that the state spends a relatively small amount of resources on early childhood education (8 percent of the recurrent education budget). Inthe absence of a private provider of early childhood education, it is highly unlikely that the state would have been able to cater to the same number of toddlers, 2,649, without a substantial increase in allocation. Thus, this provides a telling example of how fee-charging private institutions can alleviate the fiscal burden of educational provision. However, an increased reliance on a fee-charging provision should be accompanied by a careful evaluation of the equity aspects and, if necessary financial aid to those families that cannot afford the fees. 7.9 The Millennium Figwe 7.1 Gross Iu-olmmnt and ConipletionRntes in PrirnnryEducationin Development Goals related to Cwibbenn Countries,1001 primary education have been (in prrrentsgr) attained. As Figure 7.1 shows, 120, gross enrollment and 100 completion rates reached 104 and 110 percent in 2001, 80 respectively. Moreover, both 60 girls andboys present the same gross enrollment rate, 104 40 percent, but boys have a higher 20 gross completion rate (115 percent) than girls (104 0 percent). However, the high completion rate in primary education should be looked at with caution due to the automatic progression system inplace. In this system, the use of an effective monitoring and evaluation system is critical for the system to perform adequately. Hence, in order to assure that all students have sufficient knowledge to graduate from primary school, the introductionof a monitoring system is advisable. 7.10 In secondary education, both absolute enrollment and gross enrollment have increased since 1990.In2001, absolute enrollment was around 4,600 pupils, which corresponds to a gross enrollment rate of 100percent, with no difference between boys and girls. Despite this large enrolled student population, only 67 percent of the school-age cohort completed secondary education, with a completion rate of 76 percent for girls and 59 percent for boys. Nevertheless, completion inSt. Kitts and Nevis is still well above the Caribbean average of 57 percent in2001. 67Basedon the 1999LivingConditionsSurvey. 82 7.11 The large disparity between enrollment and completion rates in secondary education is caused by a large fraction of the school population leaving school during the last two years of the secondary education cycle. Youngsters opt to stop attending school, attracted by higher income and low personal returnto continue schooling, which could be linked to the automatic promotion policy. With automatic promotion in primary education, pupils are allowed to progress in the system, although they might not have learned the necessary skills. Without continual monitoring of the individual pupil's learning outcome-either through testing or monitoringby the teacher-vulnerable pupilsmightnot receive the necessary special attention to learn adequately. Hence, the automatic promotion system could hide insufficient learning among the weakest students until it is too late and the student consequently leaves school at first option, which is at the end of the mandatory schooling age. Extending the mandatory schooling age could reduce the tendency to prematurely leave secondary school; however, deeper education reforms aimed at raising quality and returns to secondary education are needed to successfully reduce the costly low completion rate of secondary education. 7.12 In the CXC exam, a Caribbean-wide exit exam for secondary education, St. Kitts and Nevis compare favorably against other countries in the sub-region. The CXC exam provides an indicator for measuring learning outcomes, and thus quality of instruction, for graduates of secondary education. On average, 59 percent of St. Kitts and Nevis students passed English and math in 2000, while the regional average was 51 percent. However, this result should be interpreted with caution because of a self-selection problem. Not all graduates participate in the CXC exam, and dropouts do not take the exam either; thus, the examination results indicate the quality of teaching receivedby the better half of the school population. 7.13 The country's main tertiary education institution, Clarence Fitzroy Bryant College, grew significantly in the 1990s, from 193 students in 1991to 1,235 in 2001. However, more than half of the student population, 710, or 57 percent, consists of part-time students enrolled in adult continuing education, where secondary education is not a requirement for entry. Thus, the college provides formal tertiary education to only 525 students.68According to the 1999 Survey of Living Conditions, which reports both school attendance and age, less than 10 percent of the 18-to-24-year-olds attend tertiary education. The low domestic formation of advanced human capital is not substituted by foreign formation. The United Nations Education, Scientific and Cultural Organization (UNESCO) estimates that in 2000 only 15 percent of the 18-to-24-year- olds from St. Kitts and Nevis studied abroad. This brings the total enrollment in tertiary education to about 12 percent, a strikingly low coverage, keeping in mind that the Latin American average is 20 percent, and a completion rate of secondary education in St. Kitts and Nevis of 67 percent.69 7.14 In summary, the educational agenda has progressedsteadily in recent years, but is still unfinished. Preschool fails to enroll a fifth of the children, and while children enjoy full access to primary and secondary school, only 7 out of 10 pupils complete secondary education. 68 Further, 141 students are attending part- and full-time vocational and training courses at the Non-formal Education Centreand the Advanced Vocational Education Centre. 69In2001, 246 students started studies at ClarenceFitzroy Bryant College, representinga thirdof the class of 682 students that attended the last form of secondary education in 2000. This implies an estimated transition rate from secondary to tertiary educationof 36 percent, a very low rate comparedto OECDandLatinAmerican standards. 83 Inaddition, there is growing concern about the deterioration inquality of education due to the automatic progression regulation in primary education. Finally, tertiary education is in its infancy. 111. USES OF EDUCATIONEXPENDITURESINST. KITTSAND NEWS, 1994-2002 7.15 On average, St. Kitts and Nevis Figure7 9 Share of EducationExpendituresinGDP. S t KiKs and Nevis1994-2002 spent in recurrent and capital (as %) expenditures on education about 5.2 percent of GDP during 1994-2001, with a rising trend toward the end of the period. As Figure 7.2 shows, the federal government accounts for four fifths of the education expenditures, 4.5 percent of GDP compared to 0.8 percent of GDP for the Nevis Island Administration. In 2000, the education x g I g S g g g g g g g I g " r n 0 " z $ ! z 2 z z R R R $ ! z z z z ? R R sector was allocated a total of 6.1 percent St Kittsand NevlsFederalGovernment N e w Island Administration of GDP from the federal government and the Nevis Island Administration, which places St. Kitts and Nevis slightly below the average Caribbean country, 6.6 percent of GDP, and considerably-above the Latin American average, 4.1 percent of GDP. Ifthis allocation were used efficiently, it would imply a sizable investment into the country's future growth potential. However, as this chapter will show, educational outcomes do not match the highspending, andconsiderable efficiency gains canbemade. 7.16 The largest increase from 1994 to 2000 in the share of education expenditures in GDP belongs to the St. Kitts and Nevis Federal Government and responds mainly to an increase in capital expenditures.The share of education expenditures in GDP increased from 4.6 percent in 1994 to 6.1 percent in 2001. About two-thirds of this increase was the result of capital expenditures. Two basic education projects with the Caribbean Development Bank (CDB) and the World Bank that aim for expansion of secondary schooling, improved quality, and enhancedmanagementcapacity were implemented. 7.17 Expressedin realterms, recurrent Figure7. 3 EducaDionExpenditureinS t K i m andNevis 1994-2002 education expenditures increased by 50 0 (millionsofECf 95) 45.0 about 40 percentbetween 1994 and 2001. 40.0 35.0 As Figure 7. 3 shows, the largest increase 30.0 took place in 1997-about 12 percent- 25 0 20 0 which led to the creation of an Educational 15 0 Planning Division and the hiring of 142 100 5 0 supernumerary teachers for primary and 0 0 secondary education. 1 St. Kitts and Nevis FederalGovernment 1 Nevis IslandAdministration 1 7.18 The Federal government allocated, on average during 1994-2002, Source AUmarf Calculatlans NO@ ~ i g u i tor~anyears e w p t 2002 are actilais in2002 me ngurescornspondto esumatei NOdam avaiia~ietor NR m 1999 and 2002 e 0Recurrent L3Capital 84 about 15 percent of its recurrent Figure 7.4 Share of RecurrentEducation ExpendituresinRecurrent budget to education, while the Nevis Expenditures.St. Ktts and Nevis 1994-2002 (as Xt Island Administration spent about 9 percent of its recurrent budget. As Figure 7.4 shows, since 1998 there has been a positive trend in the share of education in recurrent expenditures of the federal government. In 1994, the share of education was 13.2 percent compared to 15.2 percent in2001. Source AuthorsCalculauans 1994 1995 1996 1997 1998 1999 2000 2001 2002 Note Figuresfor alyears except2002 are actual5 In 2002 the hgurescorrespondto esamates nNensIslandAdminlstrauon -e -Federal government 7.19 Since the mid-l990s, administration costs and tertiary education have taken an increasing share of recurrent education expenditures. As Figure 7.5 shows, the St. Kitts and Nevis Federal Government spent about 10 percent of total recurrent education expenditures on administration in 2002 compared to 8 percent in 1996. For Nevis, the administration takes up four times more resourcesthan the administration inthe average Latin American country. Given the large fixed costs involved in education administration-for instance developing of exams, information systems, strategies, and curricula-it i s expected that smaller states allocate a higher share to administration." However, there are considerable savings to be gained from deeper within and cross-country cooperation, which, moreover, would stimulate quality improvements through international sharing of best practices. The government of St. Kitts and Nevis could therefore consider (a) consolidating educational management within the country, and (b) engaging more actively in the OECS coordination of regulation, administration, and testing of education. A 25 percent reduction in administration costs would imply economies of EC$940,000 (approximately 2.5 percent of the recurrent education budget). 7.20 The share of tertiary education Figure 7.5 St. Kitts and Nevis Recurrent Education Expenditures by in recurrent education expenditures subsector, 1998-2002 (as o/b Total) increased from 14 percent in 1996 to 100% 9056 18 percent in2002. Ina LatinAmerican 80% and Caribbean context, this share is 70% 0Other below average, but so is coverage of E4Teltiary 5ou 60% R Secondary tertiary education in St. Kitts and Nevis. 40% 0Pnmary The combination of limited access to 0 EarlyChildhood 30% Administration tertiary education and a low cost- 20% recovery ratio for tertiary education 10% 0% makes this budget line inequitable. Further, given that a nonnegligible share Source Authorscaiculabons Note Figuresfar 1995 1998and 1999areactuals Figuresfor 2002 areelmates of highly educatedindividuals leaves the country, the public loses a corresponding share of the returns to this public investment. The country should not reduce investment in tertiary education, but rather rethink the financing structure. This level of advancededucation yields the highest private returns on the labor market 'OThe average cost of administration in other education systems inthe Caribbean reaches 10percent, twice the Latin American average of 5 percent. 85 and spurs economic development; consequently, the financing burden could increasingly be shifted toward the beneficiary. With this aim inmind, the government could raise existing fees in tertiary education. Most countries, in particular developing countries, have introduced cost recovery measuresinpublic higher education. Recently, Dominicaestablishedtuition fees for the country's community college that cover 13 percent of the institution's cost. 7.21 Currently in St. Kitts and Nevis, no tuition fee is charged to nationals, but minor fees, such as a graduation fee andtechnology access fee, cover 5 percent of total costs. These could be raised to reflect the true costs of tertiary education; for instance, a 20 percent cost- recovery ratio would generate immediate extra revenue of EC$1 million (3 percent of the recurrent education budget). Further, to achieve higher value for public money, the Ministry of Education could set performance targets for the Clarence Fitzroy Bryant College through a negotiated contract with the institution. The tertiary education budget is also burdened by payment of full-time salaries to 25 students enrolled at the teacher training college and 10 students enrolled in nursing school. This is an exceptional, expensive, and unequal expenditure disproportionately favoring one type of student. Elimination of this spending category would provide the government with an estimated saving of EC$650,000 (2 percent of the recurrent education budget). Box 7.1 presents alternative sources of financing tertiary education in Latin America and the Caribbean. 7.22 The share of secondary education expenditures in total recurrent education expenditures in the St. Kitts and Nevis Federal Government has decreased since the mid- 1990s. Inthe mid-l990s, 4 out of 10education EC dollars went to secondary education. At the start of the new millennium, 3 out of 10 reachedthis subsector. The adjustment reflects a serious squeezing of non-salary expenditures. The share of recurrent expenditures to secondary education should not be allowed to fall further, and a reallocation of recurrent expenditures to non-salary expenditures in secondary education is recommended and in accordance with the educationpriorities that the country is facing. Equity of Spending 7.23 The most important educational barriers for low-income families are limited access to, and inconsistent quality of, the education system. International evidence overwhelmingly demonstrates that such barriers disproportionately hurt the educational chances of children from poor and uneducatedfamilies. Consequently, equitable policies would (a) increase completion of secondary education through quality improvements at the primary and secondary level, and (b) overcome barriers to access to early childhood development and tertiary ed~cation.~' This section further considers two specific budget lines that aim for enhancedequitable outcomes: the School Feeding Program andthe free textbook program. 71Barriers to access and completion of education, regardless of the level, are likely to be o f both social and financial nature. In the case of early childhood education, the survey o f living conditions clearly indicates a relationship between barriers and income. One-fifth o f children from the lowest three income quintiles do not attend early childhood programs, while all children from the highest two quintiles attend. Targeted financial aid could mitigate the disadvantage of children from poor families. 86 Box 7.1 Financingof Tertiary Education in Latin America and the Caribbean Globalized economies depend on productivity increases to compete in the international markets. Technological and educational improvements are key for productivity increases. This creates a major pressure for expansion of tertiary education, which governments are struggling to accommodate through the facilitation of private investments. This facilitation generally takes two forms: cost recovery in tertiary public institutions and growth of nongovernmental institutions. Cost recovery is generally politically sensitiveand necessitatesclear prior communication to society of the reasons for it, such as the high inequity and costs often associated with higher education, and the major monetary gains to the recipients of tertiary education that allow for payment of fees. Normally, fees are introduced under economic crises to reduce public spending or with an understandingthat the extra resources generated will remain with the higher education system. Australia pioneered a successful way of introducing cost recovery. The Higher Education Contributing Scheme (HECS) charges tuition amounting to approximately 25 percent of the costs, and repayment takes place after graduation and is based on the graduate's income. Hence, the student has no risk of defaulting with increased participation. Such contribution schemes are frequently introduced in the context of a generalfinancing reformthat always aims for higher efficiency inthe tertiary education system. Increased investment in private tertiary education is often stimulated by improving the functioning of the market for tertiary education, including (a) no or few regulatory obstacles to establishing and operating of nongovernmental tertiary education institutions; (b) provision of clear information on educational quality, labor market value and relevance of tertiary education; (c) offering of (unsubsidized) student loans to allow a student to study now and pay later; and (d) diversification of tertiary education with developmentof shorter termtechnical and technological degrees, which entail less investment. Latin American and Caribbean countries have used these policies in varying degrees and in different combinations. The two graphs below show the status of private investment in tertiary education by presenting (a) the share of cost borne by students, and (b) the share of nongovernmental provision of higher education for selected Latin American and Caribbean countries. Increasing private investment with appropriate student aid to students from low-income families allows for greater access to higher education and thereby increases a country's humancapital growth. Chile Ecuador El Salvador Colombia Brazil Costa Rica Barbados Colombia Dominica St Kittsand Nsvis Peru Honduras Chile Bolivia Argentina Mexico Brazil St Kittsand Nevis 0% IO%, 20% 3.0% %of costs borneby student inpublicuniversdies Share of nongovernmental enrolment 0% 20% 40% 60% 80% Source: "Brazil Higher EducationSector Study," (World Bank 200 I), updatedwith country-specificsources. Datafrom 1997- 2003. 87 7.24 The School Feeding Program is traditionally perceived as an expenditure line that benefits low-income students. However, inSt. Kittsand Nevis at least 65 percent of the well off, or half of the student population, receives free meals. A school feeding program has at least two purposes: (a) it stimulates demand for schooling, which has proven critical for the attendance and learning of low-income students, whose parents tend to underestimate the value of education for the future of the child, and (b) it provides important nutrition that improves health outcomes and educational learning. For both purposes, the primary target group is children from poor families. For other income groups, school feeding is convenient, but not necessary for either learning or nutrition. Table 7.2 shows that at least 65 percent of the well-off half of the student population receives free meals. Providing school feeding free of charge is therefore generous of the state, but intimes of fiscal crisis, the program could be scaled back to benefit only the needy students. The government could therefore recover costs with notable exceptions for the needy pupils. For instance, eliminating free school meals to children from the most affluent half of the population would free up EC$900,000 (2.5 percent of the recurrent education budget), which could be usedto give free meals to all poor children. 7.25 A similar pattern is observed with free textbooks. Better targeting of subsidies is key to improve equitable outcomes. The lack of targeting for schoolbooks is especially worrisome given that 1out of 5 poor children has no textbook, while only for 1in 12 children from the richest quintile has none. The government could hence either target textbook provision to low-income families only, or introduce book fees for higher-income families. Given the current subdued level of expenditures on learning material (see below), targeted provision of textbooks would not currently yield sizable savings. However, when expenditures for books rise, a targeted provision would assure the best equitable outcome. Table 7.2 Distribution of Education Subsidiesby Quintile, 1999 ReceivedFree Meal? Yes 80 83 64 67 69 75 No 6 9 21 18 14 12 Notstated 14 8 15 15 17 13 Sum 100 100 100 100 100 IO0 a ' Received Free Books? Yes 15 15 7 8 10 11 No 6 1 7 8 8 6 C; Not stated 79 84 86 84 82 83 Sum 100 100 100 100 100 100 ReceivedFree Meal? Yes 100 8 No 50 8 Notstated 50 100 100 84 Sum 100 100 100 100 100 100 ReceivedFree Books? 6 `6 10 9 v1 Yes 18 8 No 6 12 10 6 n" ETd 88 70 90 92 90 85 100 100 100 100 100 100 Source: Author's calculationsbasedon the 1999St. Kitts and NevisLiving Survey Conditions. 88 7.26 Salaries are, by far, the largest spending item, taking up about 90 percent of all recurrent education expenditures in 2000?* At the secondary level, there is especially reason for concern because the wage share in total secondary education spending increased from an already high92 percent in 1996 to 99 percent in2000. This leaves exceptionally little money for other crucially important learning inputs, such as writing paper, textbooks for poor students, and school maintenance. This skewed allocation of expenditure toward salaries unnecessarily reduces the effectiveness of public resources and results in inefficiently low learning. Hence, there is a severe need to increase the expenditures on learning materials at the expense of salaries. The government recognizes this need in the 2003 St. Kitts and Nevis Estimates Book; however, politicalwill and changing priorities are necessary to remedy the situation. 7.27 The main reason for the large share of salaries in recurrent education expenditures is the low pupil-teacher ratio. This indicator i s closely linked to unit costs of education, since emoluments to teachers represent about 90 percent of total recurrent costs. As Figure 7.6 shows, St. Kitts and Nevis spends the most per student per year in primary and secondary education among the Caribbean countries--US$860 and US$l,OSO, respectively. Importantly, higher spending does not necessarily imply better outcomes. As Jayasuriya and Wodon (2003) demonstrate, countries differ intheir efficiency inturninginvestments into outcomes. Figure 7.6 Pupil-Teacher Ratio and Costs of Education inSt. Kithand Nevis, 2000 (in US$) Pisisiq~ ~ I i m , ~ ~ s o r i Secondary Elluaition 1000 , 700 - ; 2 'sL 600 - DO *BE `SL ,BE ,TT I DO 5 0 0 - ePVG 400- GR ,JAS" p: 100- atB8 3 0 0 - JA GR 2 0 0 - *DR *DR 0 ' 0 15 20 25 30 Seulse Mm%gEdunnadP&m~icr rnhr Canbbcan.WddBakt&en& 20 25 30 35 0 5 10 40 studentratio Pupilstudent ratio This i s also the case for educational investment in the Caribbean. A forthcoming World Bank report, "Monitoring Educational Performance in the Caribbean," shows that educational outcomes in the Caribbean dependjust as much on how much you spend as on how you spend the money. For St. Kitts and Nevis, the data clearly indicate that with the current level of investments flowing into education, outcomes could be improved substantially through efficiency gains. 7'Wages include personnelemoluments, wages, and allowances. 89 7.28 Overall, the government of St. Kitts and Nevis manages teacher staff in an inefficient manner. Inprimary and secondary education, there are, on average, 18 and 13 pupils per teacher, respectively. This teacher deployment is tremendously inefficient. If these indicators were raised to the average Caribbean level, 25 and 18 pupils per teacher in primary and secondary education, respectively, the government would save about EC$2.3 million (6 percent of the recurrent education budget). This saving presupposes a shedding of 117 teacher positions, through either attrition or layoffs, which would become redundant following reforms aimed at a more effective provision of education.73One important element in the rationalization of teacher staff is an evaluation of the advisability for the highnumber of so-called supernumerary teachers, 142 positions in 2002. The supernumerary teachers make up close to half the teacher force in primaryeducation and a quarter insecondary education. According to the Ministry of Education, the majority of these teachers are substituting for other teachers on study leave, which explains the finding of a regionallow pupil-teacher ratio, and suggests major savings are possible through strengthening lax teacher management. 7.29 Among the initiatives that the government could consider to increase the pupil- teacher ratio to lead to efficiency and quality improvements inthe teacher deployment are: e Merging of primary schools. Inefficient use of teachers takes place predominantly at smaller primary schools. A merging of schools based on school mapping exercises to examine bussing options for children from small villages and future demand for schooling, would allow the system to exploit the economies of scale in the education system andreduce costs, while increasing quality. e Instituting a curriculum reform of secondary education. A common core curriculum for the first three forms in secondary would reduce the excessive offering of subjects of a specialized nature, such as typewriting, auto mechanics, and metalwork, and would lead to larger classes. Further, such a reform could be combined with a modernization of offered courses that equip pupils with more skills that are relevant for today's labor market. Such a reform is underway in an OECS context, but implementation has been lagging.74 e Mandating of a minimum floor on class size at the primary and secondary level. Many schools, in particular at the secondary level, offer subjects with low demand, leading to small class sizes and highcosts. Mandating a minimumclass size of, for instance, 15-a rule that is inplace inmanydeveloped countries-would increase efficiency. 73 The savings in primary education would arise from a reduction in established teacher positions from the 2002 level of 296 to 232, the level needed with a pupil-teacher ratio of 1 to 25. The average wage of a primary or secondary education teacher remunerated in the wage categories (K10-2 1) is EC$I9,260. From secondary education, the saving arises from a reduction of established positions from 309, the 2002 level, to 256, the level required at apupil-teacher ratio of 18. This saving does not take into account the cost of retrenchment. 74 Besides improving efficiency through larger classes, a common curriculum within OECS could lead to further efficiency gains through common regional exams during the course of secondary education, collective textbook purchases, and sharingof teachers. 90 a Rewarding well-per$orming schools. The Ministry of Education could provide incentives for each school to deliver education more efficiently, possibly through establishing a contract with each school management, in which the school commits itself to increasing the pupil-teacher ratio and as a reward retain part of the savings. For such a decentralization of budget management to function effectively, international experience shows that local and national school management often needs strengthening. 7.30 In addition, the exceptionally highunit costs inSt. KittsandNevis are explained by the large number of hours of instruction that the students receive compared to the Caribbean average. In St. Kitts and Nevis, students receive 1,400 and 1,480 hours per year of instruction in primary and secondary schools, respectively. This exceeds the Caribbean average by 400 hours (40 percent). Theoretically, this should lead to significantly higher learning outcomes. Measured by the passing rate in math and Englishon the Caribbean Exam Council's exam, which is the only international measurement in which the country participates, St. Kitts and Nevis is second best. Dominica fares best and provides only 975 hours of instruction per year. Consequently, the 40 percenthigher quantity of instruction seems to have an observable yet limited payoff. This report does not recommend reducing the number of school hours per year, but rather recommends focusing on efficiency and quality improvements inorder for childrento learnmore at a lower cost. 7.3 1 Quality of instruction, and especially teacher qualification, is key for increasing effectiveness of learning. Teacher qualifications are inadequate in St. Kitts and Nevis. Figure 7.7 shows that only 56 percent and 29 percent of the teacher corps in primary and secondary education, respectively, are adequately trained to perform their task. The lack of qualifications is a severe and long-term barrier for increased efficiency in the education sector. The inefficiency is partially a result of the rule of admission at the teachers college that stipulates two years of teaching experience for eligibility. Hence, teachers are barred from training before the commencement of teaching. The damage to children's learning outcome due to instruction from untrained and inexperienced young teachers is serious. Figure 7.7 Share of Trained Teachers inPrimary and Secondary School, St. Kithand Nevis, 1999 (inpercentage) PrimaryEducation SecondaryEducation 100 80 70 n BO 50 40 30 20 i o 0 91 7.32 To improve and strengthen teacher qualifications, the government could consider: Rethinking teacher training and hiring regulations. All graduatesfrom secondary schools should be eligible for the teacher training college-regardless of prior teaching experience. To address the lack of teaching experience, a mentoring program could be introduced as an element of teacher training. a Additionally, the government could stipulate that in thefiture only certified teachers will qualify for open positions. This would prevent the quality of the teacher corps from deteriorating further. Toseriously address the shortfall of training, a major retrainingprogram of the existing teaching stafs could take place. Provided efficiency gains take place in the system, this would generate time for teachers to attend retraining courses. 92 8. SOCIAL PROTECTION PROGRAMS 8.1 Social protection programs are public interventions to assist the most vulnerable groups of society with the ob'ective of ensuring a minimumlevel of welfare for individuals, households, and c ~ m m u n i t i e s ? ~ ~ ~mosthcommon social protection tools to address the social risks of ~ T e vulnerable groups are social insuranceand social assistance. e Social insurance consists of mechanisms that pool social risks across populationgroups, such as pension, unemployment, or health insuranceprograms. 0 Social assistance programs are designed to help the poor and vulnerable individuals and households cope with temporary or chronic poverty by providing income support and access to basic services. Social assistance programs aim to address the specific needs of groups at risk, and include a variety of programs such as cash and in-kind transfers, subsidies, workfare, public works, housing programs, and social funds. 8.2 Social protection programs in St. Kitts and Nevis emphasize risk-coping strategies; that is, programs to help households cope with the effects of risk (especially poverty) rather than prevention and mitigation strategies. Moreover, the existing system is not well poised to respond to either economic downturns or natural disasters. Targeting and distribution systems that can be gearedup quickly during crisis periods are required. 8.3 Several programs, includingthe school lunch program for primary students and a school bus service for secondary students, are not targeted and are provided cost free irrespective of income. In St. Kitts, health care subsidies are targeted based on categorical criteria (children, elderly, disabled) rather than income. 8.4 Since the mid-l990s, social protection expenditures in St. Kitts and Nevis have been on average approximately 3.5 percent of GDP. Although below the average for the Latin America and the Caribbean Region, any effort to strengthen social protection cannot include costly new programs giventhe difficult fiscal positionthat the St. Kitts and Nevis government is facing. 8.5 This chapter identifies vulnerable groups, describes the social protection strategy and programs of the Federation, reviews social assistance and social insurance spending, and suggests policy recommendations for strengthening social protectioninSt. Kitts and Nevis. 75 It isimportant to note at the outset that very little social protection sector work had been undertaken in St. Kitts and Nevis before this report. Further, there are over 20 social protection programs implemented by five different ministries and agencies in St. Kim, and a parallel set of programsoperated with separate administrative and budget systems inNevis. This made programs hard to identify, and collection of complete times series data on expenditures and beneficiaries impossible. Despite these constraints, considerable progress was made in identifying social protection programs, expenditures, andkey issues. 76This chapter examines current spendingon social protection programs.Except for Upgrading of DaycareCenters with BNTF funding (EC$22,000 in 1999 and EC$700,000 in 2000), capital expenditures on social protection are limitedto low-income housing programs. It was decided not to include capital expendituresfor low-income housing programs because, as reported by program directors, these differ enormously from the reported PSIP expenditures and, as implementedcurrently, are not targetedto the poor or most vulnerable. 93 8.6 The main recommendations to strengthen social protection programs can be summarizedas follows: a On Social Protection Programs o Strengthen targeting of systems for social assistance, in particular, for income transfers and the Student Education and Learning Funds (SELF), and introduce cost sharing for programs that are not currently means tested, including health care services in St. Kitts and the School Feeding Program, with cost recovery from those who can afford it and targeted fee waivers for those who cannot.77 o Identify strategies for using social protection programs to promote human capital development-possibilities include introduction of conditional cash transfer and/or conditionalities into SELF. o Reduce administrative costs by merging or eliminatingcategoriesof social assistancethat are a duplication of programs already inplace (for example, skills training programs) and by exploring alternative administrative strategies, including private sector provision of skills training. o Assess the government post-hurricane public employment programs to ensure that they support Government's social protection strategies. o Expand coverage of the Social Security Scheme (SSS) and undertake an actuarial review to determine the fiscal viability of the SSS and the SeveranceFund. a On Institutional Reforms o Develop a social protection strategy. This would require sectorwide planning; that is, all ministries and agencies responsible for implementing social protection programs would come together to identify social protection priorities and coordinated strategies to address these priorities. o Develop a more effective and efficient system to identify beneficiaries, including mechanismsto target households and communities (poverty maps).78 o Implement regular household surveys to assess poverty levels, incidence, and coverage of social safety net programs that could be used to assist in the development of targeting systems. o Improve labor market data to make labor market programs more responsive to the requirements of the labor market. o Enhance fiscal accountability through introduction of program budgeting for social protection programs not currently identified as programs in the budget, such as social public assistanceand skills training programs. 77A means test is an appraisal of household income and assets to determine whether a household qualifies for program benefits. 78Despite the small size o f St. Kitts and Nevis, there are considerable geographic differences inpoverty rates. InSt. Kitts, three of the nine parishes account for 70 percent of all the poor. InNevis, St. George parish accounts for 43 percent o f the poor. 94 8.7 Reducing program leakage (from SELF and other targeted programs) and introducing targeting of school feeding and allowing few waivers for health services to the poor, would considerably ease budgetary pressure. Divestment of programs could also result incost savings. However, capacity buildingandinstitutionstrengthening, including development of targeting, management information systems (MIS), and survey capacity, have their own development and administrative costs. Expanding program coverage with provision of benefits conditional on activities that promote human capital also comes with its own investment costs. The net fiscal impact of the above reforms would needto be carefully assessed. I. ST.KITTSANDNEVIS SOCIAL PROTECTIONSTRATEGY ANDPROGRAMS Social Protection Srrcrtegy in St. Kitt,saiid Nevis 8.8 The government of St. Kitts and Nevis has adopted a social protection strategy designed to ensure access to health care and primary education and to expand access to secondary and tertiary education." Health care is heavily subsidized and there is no requirement for cost sharing in education. The government also attempts to ensure that other programs and services (including income support, food, clothing, housing, and utilities) are available for the poor. A Social Security Board provides retirement, disability, illness, and maternity benefits to people who have contributed to the social security scheme, and administers a noncontributory old age and invalidpension program. Training programs for disadvantagedjob seekers are provided. Inaddition, the government identifies tax holiday measuresfor the tourism industryas an element of its overall social protection strategy inthat tax holidays are viewed as an employment-generation mechanism. 8.9 Social protection programs are directed primarily at children, the elderly (62 years old and over), and the chronically ill or disabled. Single-parent households, unemployed youth, or people who face economic downturns-especially those inthe sugar industryand those who have suffered the effects of natural disasters-are also targeted. 8.10 This stated, the existing social protection strategy is not clearly articulated or well coordinated. There is an array of programs in both St. Kitts and Nevis with little coordination among programs within or between islands. With the exception of the Social Security Program, programs are financed and administered separately by the two islands. There is recognition of the need to better coordinate social protection programs. Responsibility for coordination has been assigned to the Ministry of Social Development, Community, and Gender Affairs through the Council for Human and Social Development (COHSOD). However, the primary objective of COHSOD is to address priorities identifiedby CARICOM, and the committee is not specifically mandated to coordinate social programs in St. Kitts and Nevis. Although the government has expressed its intent to better coordinate its social protection programs, a mechanism to oversee the development of an integrated social protection strategy and a rationalized set of programs is required. ''Bothasa federation and as separate island states. 95 8.11 The result is that social protection priorities are not clearly identified and fiscal prioritization of programs is lacking. This is reflected in the budgetary process during which budgets are essentially determined by previous budget allocations and not in response to changing needs and priorities. Social Protection Programs and Expenditures in St. Kitts and Nevis, 2000 8.12 St. Kitts andNevis implements an array of social protectionprograms including: 0 Social insurance programs. Social Security, a contributory pension, disability and sickness fund; and the Severance Fund, which provides severance benefits to workers retrenched from the private sector. Social assistanceprograms. Social and Public Assistance, means-testedincome transfer programs, School Feeding Programs, and the SELF for needy students, skills training and adult education programs, microenterprise development programs, and fee waivers for medical care. The key social protection programs are summarized inTable 8.1. Appendix C presents a detailed description of social protection programs inSt. Kitts and Nevis. Social Iiuurunce Program 8.13 The Social Security Scheme accounts for the largest share in total social protection expenditures (68 percent). The Social Security Scheme (based on contributions) covers employed and self-employed people, including public employees. It provides benefits for retirement, temporary illness, disability, maternity, employment-related injuries, medical and funeral expenses, and funds a noncontributory pension and invalid assistance program with eligibility subject to a means test. The contributory scheme with total expenditures equal to US$8.3 million (or EC$22.5 million) in 2000 is by far the largest social protection program in terms of both expenditures and beneficiaries. The program provided benefits to approximately 8,700 people and absorbed over two-thirds of the social protection budget in 2000. Approximately 30 percent of the elderly population of St. Kitts and Nevis received retirement benefits. In comparison, expenditures on the targeted noncontributory pension program were equal to US$0.7 million (EC$ 1.8 million). 8.14 The Severance Fund is targeted to employees who have been retrenched from the private sector and provides a one-off benefit. Its objective is to mitigate shocks caused by private sector downsizing. The fund is funded from a 1percent employer-paid tax on the payroll of registered enterprises. The tax is collected by the Social Security Board and transferred to the consolidated fund for administration by the Department of Labor. Information on expenditures and number of claimants was not available. Moreover, reviews or evaluations of the program have not been undertaken. SocialAssistance Progrnms 8.15 Income Support Programs represent the second-largest share in total social protection expenditures (16 percent). Community, social, youth, and development programs 96 represent 10percent of total expenditures, or US$1.3 million(EC$3.4 million).80These programs include targeted cash transfers, school uniforms, food vouchers, community assistance for the elderly, and gender- and community-based training programs. The other 6 percent of total expenditures belongs to the noncontributory pension program. 8.16 Education-based welfare programs account for the third-largest share intotal social protection expenditures (8.3 percent). The main education-based programs implemented in both St. Kitts and in Nevis are the School Feeding Programs and SELF. Both islands also implement a variety of other education-based social protection programs, including nontargeted school bus service for secondary students, targetedscholarships, anduniforms. 8.17 The St. Kitts School Feeding Program provides a centrally prepared free lunch to all primary school students, and to needy secondary school students. InNevis, the program provides a meal or snack at all government preschools and primary schools. In 2000, combined recurrent expenditures on the School Feeding Program were estimated at US$0.7 million (EC$1.8 million), with approximately 4,500 students receiving meals in St. Kitts and 585 in Nevis. 8.18 The Student Education and Learning Fund (SELF), implemented and financed separately in St. Kitts and in Nevis, provides assistance with textbooks, examination fees, and uniforms for needy secondary and tertiary students.8' Total expenditures were equal to US$0.2 million infiscal 2000 (EC$0.6 million). The program is estimated to serve less than one- thirdof needy students. Preliminary data from the 1999Poverty Assessment Report indicate that the program i s poorly targeted, with 15 percent of students in St. Kitts in the first quintile (the poorest) and 10percent of students inthe fifth quintile (the richest) reporting that they received free schoolbooks. Both islands implement a variety of other education-based social protection programs, including target scholarship programs, school bus services, and uniform and book programs. 8.19 Health-based welfare programs represent only a small share of social protection expenditures (1percent), not including the value of health subsidies in St. Kitts. St. Kitts and Nevis has adopted very different approaches to the provisionof health subsidies. InSt. Kitts, fee waivers for medical care and pharmaceuticals are provided to all children, pregnant and lactating women, people aged 62 and over, and people with chronic illnesses. Incontrast, Nevis provides means-tested fee waivers and transportation allowances for medical care. The expenditure implications of these different approaches are dramatic. Although it is impossible to estimate the full extent of the subsidies in St. Kitts, they are considerable. A proposal to means test the fee waivers in St. Kitts (using the lists of beneficiaries under the Social Assistance Program) has been approved by the Ministers of Health and Finance, and is currently before the Cabinet. Ministry-level expendituredataare reportedbecause program-level expenditure data were not available. However, changing portfolio responsibility for programs made it necessary to report the programs together because in some years they fell under the administrative and budget responsibility of the ministry that implements the cash transfer programs. ''Before 2001, the SELF program inNevis was financed from the central budget of St. Kitts and Nevis. SELF was introduced as a separateprogram inthe Nevis budget in2001, with an estimated allocation of EC$25,000. 97 Table 8.1 Social ProtectionPrograms and ExpendituresinSt. Kitts and Nevis, 2000 Recurrent SocialProtection Expenditures % of Total (inEC$) Recurrent % St. Kitts Nevis Total Expenditures GDP SocialInsurance 22,468,000 67.63 2.52 SocialSecurity (ContributoryPension and Administration) 22,468,000 22,468,000 67.63 2.52 SeveranceFund SocialAssistance 10,754,076 32.37 1.21 IncomeSupportPrograms 5,253,879 15.81 0.59 Community, Social, Youth Development (IncludingPublic/SocialAssistance) 2,670,18 1 766,698 3,436,879 10.35 0.39 SocialSecurity(Noncontributory Pension) 1,817,000 1,8 17,000 5.47 0.20 Education-BasedWelfare Programs 2,752,160 8.28 0.31 StudentEducationLearningFund 628,139 628,139 1.89 0.07 SchoolFeedingProgram 1,768,379 75,000 1,843,379 5.55 0.21 SchoolBus Service 260,642 20,000 280,642 0.84 0.03 GovernmentDay Care Health-Based Welfare Programs 379,900 1.14 0.04 SubsidizedHealthCare 168,700 168,700 0.51 0.02 Off-IslandMedicalCare 135,000 70,000 205,000 0.62 0.02 BurialAssistance 2,700 3,500 6,200 0.02 0.00 LaborMarket Interventions 1,576,639 4.75 0.18 YouthSkills Training 464,054 81,000 545,054 1.64 0.06 RuralEducationDevelopment ProjectStrong Adult Education 28,554 28,554 0.09 0.00 SmallEnterpriseDevelopment 118,472 229,875 348,347 1.05 0.04 OtherLabor Department 493,120 161,564 654,684 1.97 0.07 FeedingPrograms 125,406 0.38 0.01 Maternal andChildFeeding 125,406 125,406 0.38 0.01 FeedingProgramfor the Elderly FoodVouchersfor the Poor Housing Programs 640,352 1.93 0.07 Homefor theElderly 500,352 140,000 640,352 1.93 0.07 Other Public Assistance 25,740 25,740 0.08 0.00 Total 31,447,185 1,744,891 33,222,076 100.00 3.73 Note: All figuresare estimates. Source: Author's calculations. 8.20 Although expenditure information is incomplete, labor market intervention programs represent a small share (5 percent) of the total social protection budget, with considerable overlap in the skills development programs. St. Kitts implements the Rural Education Development Program (a community-based skills training program) and two youth 98 skills programs-the Youth Skills Program and Project Strong Program. Nevis implements an Adult Education and Skills Training Program and the Woman's Training Program. Both islands support microenterprise development programs through their Small Enterprise Development Units. 8.21 Both islands identify low-income housing programs among their menu of social protection programs; however, these programs exclude low-income people by virtue of their income and mortgage requirements. Other housing programs for the poor include housing repair programs and old age homes for the elderly poor and disabled. Expenditure and beneficiary data were not available for the housing repair programs. Expenditures on homes for the elderly, admission to which is based on a means test, were equal to US$0.2 million (EC$0.6 million) infiscal 2000, with 100residentsinSt. Kitts and 30 inNevis. Plmiiing and Excmting Cupacily~fiwSocial Protectioii Progruins 8.22 The large number of programs, each with separate administrative systems and procedures, strain an already overburdened public sector. The lack of coordinated efforts across ministries and departments has resulted in overlaps, duplication, inefficiencies, and administrative waste. To the degree that they serve the elderly, Social and Public Assistance and noncontributory Social Security target the same group. Moreover, Social Security, the Social and Public Assistance programs, SELF, and the Ministry of Health all have different systems for identifying beneficiaries. This increases administrative and time costs for beneficiaries. There is a need to strengthen the planning and executing capacity of these programs and to put inplace mechanismsfor their fiscal prioritization. 8.23 Effective planning will require updated poverty and labor market information at regularintervals.Timely andreliable informationonpoor and vulnerable groups is essentialfor the design and implementation of social safety net programs. The CDB-financed Poverty Assessment that was conducted in 1999 provided useful information; however, data on poverty and labor markets needto be collected on a regular basis. 8.24 Budgeting and management informationsystems need to be strengthened. Several programs, including cash transfer programs and labor market programs, are not identified inthe government's budgets as separate programs. This limits effective planning and weakens fiscal accountability. Elevating social protection programs to program status would mean that expenditures, including expenditures by object code, could be easily tracked over time and evaluated with respect to program inputs and outcomes. In addition, management information systems are not computerized. 8.25 Regional initiatives to build information systems, data collection, and analysis capacity could address the human resource constraints and reduce the overall costs of development and implementationof monitoring and informationsystems. Inadequate MIS systems and capacity constraints with respect to data collection, monitoring, and evaluation are common across the OECS. Regional approaches to MIS development and monitoring and evaluation capacity building would address the human resource constraints in individual countries and would be more cost-effective than developing systems on a country-by-country basis. 99 I Box 8.1 A Closer Look at the Social Assistance PrograminSt. Kitts The Social Assistance Program is administered by the Ministry of Social Development, Community, and Gender Affairs and is fully funded from the consolidated fund. The Social Development Assistance Act of 1998 (which covers both St. Kitts and Nevis) provides the framework for provision of social assistance for poor and indigent people; that is, for people who are poor or needy or who because of infancy, old age, illness, disease, infirmity, or mental incapacity are unableto maintain themselves. Social Assistance, as defined by the act, includes homehelp, housing repairsor replacement, in-kindassistance, or any other assistancethat may be identified. The Social Assistanceprogramimplements several subprogramstargeted to the poor. These include: 0 Cash transfers 0 School uniforms for needy students 0 Food vouchers for needy families 0 Community assistancefor the elderly 0 Low-income housing assistance 0 Foster care 0 Grants for glasses, hearingaides, and prescriptions 0 Funeral grants 0 Travel allowance for off-island medical care. Applications for assistance are made at the central ministry offices. People may be self-referred or may be referred by churches, the police, or schools. Under a decentralization program initiated in January 2002, applicants can apply to Social AssistanceOfficers to be located incommunity centers. Eligibility is subject to a means test and home visit. Criteria for all componentsof the Social Assistance Program have not been established, and eligibility and determination of the modality of assistance is subject to the discretion of the Officer. Social Assistance Officers are responsiblefor the distributionof benefits. Income transfers are paid incash every two weeks. Food vouchers are also distributed every two weeks. Pay stations are located in villages 01 sometimesinprivate homes. There are approximately 20 to 25 pay stations across the island. Cash transfers of $30 per person every two weeks are given to single individuals, the elderly, and the disabled. The value of the cash transfer is equal to approximately 25 percentof the adult equivalent poverty line. The FoodVoucher Programprovides food vouchersto families inneed. Vouchers are distributed every twc weeks and may be redeemed at preselectedgrocery stores. Grocery stores submit the vouchers and are reimbursed by the Ministry. Allotments of food are determined based on family size. The items specified on a typical food voucher for a mother and two children are: 5 pounds of rice, 5 pounds of sugar, 5 pounds of chicken, 12 tins oi evaporatedmilk, 3 tins of corned beef, 3 tins of sardines, 1bottle of orange concentrate, 1 box of cream of wheat, 2 packagesof pulses, 4 packagesof macaroni, 1 box of soap powder, and 2 bars of soap. The SchoolUniformProgram is targeted to low-income householdswith children. Families with incomes less thanEC$250 per week (irrespectiveof family size) are eligible for assistance. Preselectedsuppliers provide the uniforms. A child receivesone uniform if they are returning to a school andtwo uniforms ifthey are new to a I school. Depending on the size of the family, a child may receive only the uniformor they may also receive shoes andsocks. 100 11 COMPOSITIoNAND EVOLUTION SOCIAL PROTECTIONSPENDINGINST. KITTS OF AND NEVIS, 1996-2001 8.26 The task of determining the amount allocated to social protection in the budget is full of twists and turns.** Social protection programs cut across ministries and are integrated into a wide variety of programs and projects. Several of the programs traditionally defined as social protection programs (such as school feeding programs, scholarship programs for needy students, and subsidized health care for people who are poor) may also be represented in the budgets of the health and education sectors. In 2000, public expenditures on social protection programs (including social security) were equal to EC$33.2 million, and accounted for approximately 3.7 percent of GDP and 11percent of Central Government expenditures. 8.27 During 1996-2001, social protection expenditures fluctuated Figure8.1 SocialProtectionExpendituresinSt KittsandNevis (percentage at GDP) between 3.1 percent of GDP in 1996 6 O i - and 3.8 percent of GDP in 1999. As Figure 8.1 shows, on average social 503 - protection expenditures represented 3.5 4m. percent of GDP. The highest increases were in 1998 and 1999-hurricane years-but the increases seem to be of a permanent rather than transitory nature. This calls into question the use of social 1M protection expenditures as a OM countercyclical policy tool for either 189wYmots ealcui&2 1888 1888 2000 2001 economic shocks or natural disasters. ElSocialAssistance 0SocialInsurance(SocialSecurity) 8.28 Average expenditures on social protection (including social security) are somewhat lower in St. Kitts and Nevis (3.5 percent of GDP) than in other countries in the LAC Region (4.7 percent of GDP). This compares with expenditures on social protectionof 11to 12 percent of GDP in Western Europe, North America, and East and Central Europe, and 1 to 2 percent inSub-SaharanAfrica and South Asia. 82I t should be pointed out that the expenditures for the individual programs for fiscal 2000 (Table S.l), and the expenditures for the time series analysis, are slightly different. That is because, when looking at programs for a single year, we were able to get more program-specific data, while for the time series analysis we had to rely on spendingdata for ministerial departmentsthat traditionally house social protection programs.Various ministries had portfolio responsibility for these departments in different years. Efforts have been made to track expenditures for departmentsno matter what ministry the department was housedin.Expenditure data do not include central ministry administration, because it was impossible to reliably apportion central ministry administration costs to the different departments. Program-level expenditure data on programs housed inthe Ministry of Health (feeding programs, fee waivers, and so forth) and the Ministry of Education (school feeding and student welfare programs) are also reported. Spendingon the microenterprise development, low-income housing, and contributory and noncontributory social security are reported. Given the number of agencies and programs involved, time series data could only be provided for 1996-2001. 101 8.29 Social protection Figure 8.2 Social Protection Expendituresin St. Kitts and News expenditures (including social (in millions ECF95) 30.0 security) increased in real terms on average by about 5 percent 25.0- per year between 1996 and 2001. As Figure 8.2 shows, this increase 20.0- i s mostly driven by the increase in 150- Soclal Protection social insurance expendituressocial (5.8 percent) rather than 100- assistance (3.4 percent). However, 5.0- social assistance accounts for only a small share of recurrent 0.0 - expenditure (3 to 4 percent). Source Auihols sdculmons 1-SocialAssistance Social Insurance 111 SOCIALPROTECTIONAND RISKMANAGEMENTKITTSANDNEVIS INST. 8.30 National 1999 survey data indicate that approximately 1 in 3 individuals is poor. Further, approximately 11percent of St. Kitts's population and 17 percent of Nevis's population are indigent; that is, they do not have sufficient resources to meet their basic dietary needs.83 Poverty and indigence rates for St. Kitts and for Nevis are shown inTable 8.2. Table 8.2 Incidence of Poverty, Indigence, and Other Related Indicators, 1999 (percentage) St. Kitts Nevis Poor Individuals 30.5 32.0 IndigentIndividuals 11.0 17.0 PoorHouseholds 16.0 NA IndigentHouseholds NA NA PoorUnder 25 67.8 58.0 Poor60+ 5.0 5.O Femalesas a Percentof All Poor 56.0 62.6 Incidenceof PovertyAmongFemales 32.0 36.0 Incidenceof PovertyAmongMales 29.0 26.0 UnemploymentRateAmongPoorFemales 9.1 11.1 UnemploymentRateAmongPoorMales 0.0 0.0 HouseholdHeads inPrimaryOccupations 70.0 94.0 With NoEducationalCertification 57.1 37.0 InI11Health 6.4 13.0 Poor inWoodenDwellings 17.5 12.5 Poor with OutdoorKitchens 21.5 4.2 Poor with PitLatrines 30.5 30.5 Poor with OutdoorBaths 35.6 20.8 Source: "St. Kittsand Nevis Poverty Assessment Report," Caribbean Development Bank, draft 2000. 83 In 1999, the adult equivalentpoverty and indigencelines were equal to, respectively, EC$3,361 andEC$2,135 in St. Kitts, andEC$3,941andEC$2,453 inNevis. 102 8.31 Poverty is not the only source of vulnerability in St. Kitts and Nevis. We examine other sources of vulnerability by identifyingrisk indicators throughout the life cycle (Table 8.3). Information for this table was gathered through interviews with central ministry and project staff and a review of secondary documents. The matrix also identifies indicators not related to age, and existing social protection programs and requirements. This exercise should be treated as preliminary. Statistical analysis of these risk indicators to determine prevalence, incidence, correlations with income, and other analytic work on household vulnerability could provide useful insights to help identify and prioritize appropriate social protection interventions, but will require updated and improved survey data. Table 8.3 St. Kitts and Nevis Risk Indicators by Age Cohort, Existing Programs, and Suggested Interventions Suggested11 erventions Age RiskPrevention/ Group RiskIndicator St. Kitts Nevis Mitigation Risk Coping Birth Low birth weight; Maternal and Child Ministry of Increasepre- and post- Early stimulation to 4 mildto moderate Health Programs. Health programs. natal attendance. programs to ensure malnutrition. Maternal and child cognitive and healthprograms. physical development. Not participating Government Government Increaseaccess to inearly Centersand private Centers and early childhood Financial support childhood provision. private provision. development (SELF or Public development programs. Assistance) program. conditional on school attendance. Abuse, Child Welfare, Child Welfare, ParentingEducation abandonmentand Parenting Parenting Care andProtection neglect. Education. Education. Services. 5-9 Obesity. MoHprograms and IIMoH programs Maternal and child Nutritionandprimary Respiratory services. and services. healthprograms. health care disease. interventions. Abuse, Child Welfare, ChildWelfare, Multipleprevention Care andProtection abandonmentand Parenting Parenting strategies required. Services. neglect. Education. Education. 10-14 Not attaining Universal primary Universal primary Improvequality of Remedial education. basic literacy and /secondary /secondary primarylsecondary numeracy. enrollment. enrollment. education. Not having SELF. SELF. Textbook Rental Financialsupport textbooks. Programs. (SELF or Public Assistance) conditional on school attendance. Abuse, ChildWelfare, Child Welfare, abandonment, Parenting Parenting Care andProtection and neglect. Services. 103 Existing 'rograms Suggested :erventions Age RiskPrevention/ Group RiskIndicator St. Kitts Nevis Mitigation RiskCoping 15-24 Limited human Rural Education; Adult Educationand Labor Intensive capital Youth Skills; Youth Skills. Growth Life Long development, STEP; SEDU; job Women's Training; Learning responsive unemployment; search assistance. SEDU. to market demands. reen mother, underemployment. HIV/AIDS and STD md substanceabuse Teenage Healthy lifestyle Healthy lifestyle Healthy lifestyle reatment programs. pregnancy/STDs ; promotion promotion promotion substance abuse. programs; HFLE. programs; HFLE. programs; HFLE. 25-59 Limited human Rural Education; Adult Education; Evaluate Severance Workfare- capital SEDU; job search Women's Training; Fundandoptions for :specially inpost- development. assistance. SEDU. expansion. nurricane periods. Unemployment. Severance Fund. Social Security. Expanded coverage rargeted fee waivers. of social security N o pension or Social Security. Ministry of Health. system. disability coverage. Healthy lifestyle Chronic diseases. Ministryof Health. promotion programs. 60 and No pension Social Security. Social Security. Social Security. Public Assistance and over coverage. Noncontributory pension. Chronic diseases; Ministry o f Health; Ministry of Health; sickness, Social welfare and Social welfare and Ministryof Health; disability, and feeding programs, feeding programs, Social welfare and social isolation. homes for the homes for the feeding programs, elderly. elderly. homes for the elderlv. Risks HIV/AIDS. HIV/AIDS HIV/AIDS HIV/AIDS Treatment programs. Not Education Education Education Programs. Related Hurricanes. Programs. Programs. Workfare and to Age expanded social Poor housing Post-hurricane Housing Repair Disaster assistance inpost- material, no public employment Program. management. hurricane periods; kitchen or programs. temporary food, bathroom, no Housing water, and shelter piped water. improvement provision; program. reconstruction Disability. Special Education. Special Education. programs. Early stimulation programs, expanded access to education Social Assistance. and training. Source: sed on author's discu: ins with central ministry and project staff and areview of secondarydocuments. 8.32 Despite the diversity of programs currently in place, coverage of poor and vulnerable groups is inadequate. Improvedtargeting would reduce costs and promote fiscal accountability. Existingprograms are not well targeted, coverageof the target population is low, and program design and implementation is not guided by adequate data-including household 104 and labor market information. Programs that could be targeted, such as School Feeding for primary students, are provided free of cost to all students. 8.33 Programs could do better at promoting human capital development initiatives. International experience suggests that a conditional cash transfer with benefits linked to health clinics or school attendance would be ideal, but given current capacity and financial constraints and the time required to establish these systems, it is not at all clear that this should be the short- term priority. Alternative strategies, particularly the SELF program (the means-testedprogramto helpto defray the out-of-pocket educational expenditures of needy students) couldbe redesigned to better promote human capital development initiatives. Although it is too early to draw conclusions, the St. Lucia model, in which a grant from Taiwan. China was used to capitalize a fund, with the interest earned used to finance student welfare programs, may also serve as a modelto bereplicated. 105 BIBLIOGRAPHY Alesina, A., R. Hausman, R. Hommes, and E. Stein. 1996. "Bugget Institutions and Fiscal Performance inLatin America." NationalBureau of Economic Research, Working Paper No. 5586. Cambridge, MA, May. Alesina, A., and R. Perotti. 1996. "Fiscal Adjustments in OECD Countries: Composition and Macroeconomic Effects.'' National Bureau of Economic Research, Working Paper No. 5730. Cambridge, MA, August. Alesina, A., and R. Wacziarg. 1997. "Openness, Country Size and the Government." National Bureauof Economic Research, Working Paper No. 6024. Cambridge, MA, April. Baxter, M., and R. G. King. 1999. "Measuring Business Cycle: Appropriate Band-Pass Filters for Economic Time Series." Review of Economic Studies,November, :575-93. Beveridge, Stephen, and Charles R. Nelson. 1981. "A New Approach to the Decomposition of Economic Time Series into Permanent and Transitory Components with Particular Attention to Measurement of the `Business Cycle' ."Journal of Monetary Economics, March, :151-74. Blanchard, 0. 1990. "Suggestions for a New Set of Fiscal Indicators." Organization for Economic Cooperation and Development, Department of Economics and Statistics, Working PapersNo. 79. Paris, April. Blanchard, O., J. C. Chouraqui, and R. Hagemann. 1990. "Sustainability of Fiscal Policy: New Answers to anOld Question." OECDEconomic Studies, (15). Organization for Economic Cooperation and Development. Paris, Autumn. Brown, E. C. 1956. "Fiscal Policy inthe Thirties: A Reappraisal.'' American Economic Review, December, :857-79. Buiter, Willem H. 1993. "Measurement of the Public Sector Deficit and Its Implications for Policy Evaluation and Design." In Mario I.Blejer and Adrienne Cheasty, eds., How to Measure the Fisca2 Deficit. Washington, D.C.: International Monetary Fund. Chouraqui, J. C., R. Hagemann, and N. Sartor. 1990. "Indicators of Fiscal Policy: A Reexamination." Organization for Economic Cooperation and Development, Department of Economics and Statistics, Working Paper No. 78. Paris. Crucini, M. 1997. "Country Size and Economic Fluctuations." Review of International Economics, May, :204-20. D e Ferranti, D., G. Perry, 1. Gill, and L. Serven. 2000. Securing Our Future in a Global Economy. Washington, D.C.: World Bank. Eastern Caribbean Central Bank. 2001. "Developing SustainableDebt Indicators for the ECCB." Mimeo, July. ."The Targeted Public Sector Savings inthe Member Countries of the ECCB." Mimeo, undated. Head, A. 1995. "Country Size, Aggregate Fluctuations, and International Risk Sharing." Canadian Journal of Economics, November, :1096-1 119. 107 Hodrick, Robert J., andEdwardC. Prescott. 1997."Postwar U.S. BusinessCycles: An Empirical Investigation." Journal of Money, Credit, and Banking, 29 :1-16. Jayasuriya, R., and Q. Wodon. 2003. "Efficiency in Reaching the Millenium Development Goal." Washington, D.C.: World Bank. King, J., and R. Fulford. 2001. "St. Kitts and Nevis: Reform of the Consumption Tax." InternationalMonetaryFund, FiscalAffairs Department. Washington, D.C., March. Loayza, N., P. Fajnzylber, and C. Calder6n. 2002. "Economic Growth inLatin America and the Caribbean." Washington. D.C.: World Bank. Marcel, M., M.Tokman, R. ValdCs, and P. Benavides. 2001. "Balance Estructural del Gobierno Central. Metodologia y Estimaciones para Chile: 1997-2000." Estudio de Finanzas Pliblicas, Ministerio de Hacienda, Direcci6n de Presupuestos, Septiembre. Nonegaard, J., and J. Isaac. 1996. "St. Kitts and Nevis: Formulation of a Strategy for Tax Policy." International Monetary Fund, Fiscal Affairs Department. Washington, D.C., June. Nunberg, B., and D. Lindauer (Eds.). 1994. "Rehabilitating Government: Pay and Employment Reform inAfrica." Washington. D.C.: World Bank. OECS Secretariat. 2002. "OECS HumanDevelopmentReport 2002." St. Lucia. Perotti, R. 1996. "Fiscal Consolidation in Europe: Composition Matters." American Economic Review, Papers and Proceedings, May, :105-10. Perry, G., and L. ServCn. 2002. "Can Fiscal Rules Help Reduce Macroeconomic Volatility in LAC?" Washington, D.C.: World Bank. Poterba. 1994. "State Responses to Fiscal Crises: The Effects of Budgetary Institutions and Politics." Journal of Political Economy, August, :779-821. Price, W. R., and P. Muller. 1994. "Structural Budget Indicators andthe Interpretation of Fiscal Policy Stance inOECD Economies." OECD Economic Studies. Paris, Autumn. Ravn, M., and H. Uhlig. 2001. "On Adjusting the HP-Filter for Frequency of Observations.'' CEPR, DiscussionPaper No. 2858, June. Rodrik, D. 1996. "Why Do More Open Economies Have Bigger Governments?" National Bureauof Economic Research, Working Paper No. 5537. Cambridge, MA, April. Talvi, E., and C. VCgh. 2000. "Tax Base Variability and Procyclical Fiscal Policy.'' National Bureauof EconomicResearch, Working Paper No. 7499. Cambridge, MA, January. Von Hagen, J. 1992. "Budgeting Procedures and Fiscal Performance in the European Communities." EuropeanCommunities Commission, Directorate General for Economic andFinancialAffairs, EconomicPapers No. 96, October. Von Hagen, J., and I.Harden. 1995. "Budget Processes and Commitment to Fiscal Discipline." European EconomicReview, April, :771-79. World Bank. 2001. "Brazil Higher EducationSector Study." Washington, D.C. .2002. "OECS Country FinancialAccountability Assessment.'' Washington, D.C. 108 .2002. "OECS InstitutionalOrganizationalCapacityReview." Washington, D.C. . 2003. "Caribbean Economic Overview 2002: Macroeconomic Volatility, Household Vulnerability, andInstitutional andPolicy Responses." WashingtonD.C. . "Monitoring Educational Performance in the Caribbean." Washington, D.C., forthcoming. Zee, H., J. Stotsky, and E. Ley. 2002. "Tax Incentives for Business Investment: A Primer for PolicymakersinDevelopingCountries." WorldDevelopment, September, :1497-15 16. 109 APPENDICES Appendix A The Hodrick-Prescott DetrendingProcedure There are several methods to describe the cyclical properties of any economic variable. The literature on business cycles usually defines a cycle, for example of output, as Y, = Y,* +Y,', where Y, is actual output, q' is the cyclical component of output, and Y,*is some output benchmark (trend). A number of benchmarks canbe the basis of a measureof the business cycle: 0 The levelof potential output 0 The trend inoutput definedby a linear trend inits logarithm 0 The trend inoutput as definedby Hodrick and Prescott (1997) 0 The permanent component in output as defined by the Beveridge and Nelson (1981) decomposition 0 The Baxter-Kingfilter 0 The trend inoutput as defined by a peak-to-peak trend line. Each of these decomposition methods entails some relatively arbitrary decisions and has advantages and disadvantage^.^^ Chapter 2 of this report uses the Hodrick and Prescott filter (HP filter) to detrend GDP and revenue and expenditure categories. The HP filter was selected because of its simplicity and transparency. Moreover, it is the most commonly used filter in empirical studies and policy analysis to identify trend components inmacroeconomic series. The HP filter proposes the following minimizationprocedure to calculate the trend yr*: T T-1 This procedure implies extracting the trend from the original series by minimizing the sumof the deviation of yt from its trend (the first term) andthe variability of the trend itself (the second term). The relative weight of the second goal (variability) is represented by the smoothing parameter A The higher the preference for a smoother trend, the higher the value one e should assign to A . In empirical applications of the HP filter, different values for Ahave been used. For quarterly data, a value of 1,600, recommended by Hodrick and Prescott in their original work, is commonly employed. For annual data, current empirical practice is to use A= 400 or 100. However, using frequency domain analysis, Ravn and Uhlig (2001) have shown that a value of 1,600 for quarterly data corresponds to a value of 6 to 8 for annual data. Inthis report we use a value of A = 8. 84 See Baxter and King (1 999) for a discussion of how alternative filtering methods affect the measurement of business cycles. 111 Appendix R Structural FiscalBalance Methodology The literature on adjusted budget surplus measures can be traced back to the paper by Brown (1956), where he argued that to measure the stance of fiscal policy correctly one had to distinguishbetween "automatic" and "discretionary" policies. Brown's paper did not propose an adjusted measure of the budget surplus, because he explicitly argued in favor of the differential treatment of the various components of revenue and expenditure with reference to an explicit Keynesianmodel of the economy. Since the publication of Brown's paper, economists have sought a single indicator of the stance of fiscal policy, similar to the budget surplus as a percentageof GDP, but adjustedfor the business cycle. A number of government and international agencies produce these sorts of measures, including the OECD, the World Bank, the IMF, the European Community, and their various member governments. A number of indicators have been suggested. Chouraqui, Hagemann, and Sartor (1990) and Price and Muller (1994) present good discussions of the various indicators. Blanchard (1990) and Buiter (1993) provide arguments against using single indicators. Cyclical adjustment of the budget usually begins with the decomposition of output into some trend, or potential, component and its cyclical component. Appendix A describes several methods to detrend GDP, but the method used here is the same as the one the European Community uses. It simply uses the Hp filter-based trend in GDP as the measure of potential output. To compute cyclically adjusted surplus measures the European Community, IMF, and OECD estimate the elasticities of various components of revenue andexpenditure with respectto output. They use the estimated elasticities to make cyclical adjustments to these components of the budget. At this stage an important set of assumptions must be made: one must decide which revenue and expenditure components fall into the automatic category and which fall into the discretionary category. Because the assumption is that the business cycle causes those that fall into the automatic category, while those inthe discretionary category potentially cause the cycle, only those components that fall into the automatic category should be adjusted. To illustrate the methodof adjustment, take as an example income and profit tax revenue, T", typically thought of as an automatic category. Its elasticity with respect to output, eTy i s givenby The elasticities are sometimes estimated using purely statistical models of the relationship between income tax revenue and GDP. In other cases they are obtained with reference to statutory tax rates, and a statistical model of the relationship between personal income and GDP, as in the method employed by the Bureau of Economic Analysis of the U.S. Department of Commerce. Givenan estimate of the elasticity, revenue is adjustedby the amount 112 where yf is the cyclical component of output in logarithmic percentage terms. If the cyclical component is zero, clearly no adjustment to tax revenue is made. If the cyclical component is positive and the elasticity is positive, then the adjustment will be negative. This makes intuitive sense: during a cyclical upturntax revenues rise simply because the economy is expanding. To adjust for this effect, tax revenue should be adjusteddownward. Ingeneral, with a method such as this the adjustedbudgetsurplus is easy to compute. In any standard budget surplus measure, A, is defined as the difference between revenue, R, ,and expenditure, X, . To adjust the budget surplus for the business cycle and create a new budget surplus measure denoted At, one uses data on the cyclical component of output, y: ,along with estimates of the revenue and expenditure elasticities. Suppose there are N revenue categories, {R,t,., .,R, },and Mexpenditure categories, {Xi,,.. ,X,} . ,tobeadjusted. Supposetheelasticity of Rjr with respect to output is given by eRj ,while the elasticity of Xj, with respect to output is givenby exj The adjusted surplus measureis givenby . A; =A, +adjustment (3) For the purposes of Chapter 2 of this report, the following revenue and expenditure categories inSt. Kitts andNevis data were considered for adjustment: 0 Income and Profit Taxes, R,, 0 Domestic Goods and Services Taxes, R,, 0 International Trade Taxes, R,, 0 Transfers, R,, . Estimates of the elasticities of these budgetary items with respect to the output gap were estimated using the following statistical model, illustrated inthe case of income taxes: (4) qf =eRlyf +1 3 where 5; and yrc represent the cyclical components of income and profit taxes and output, respectively, as measuredusingthe HI' filter. Estimates of the elasticities are found inAB.Table 1.Since the elasticity transfer payments were not statistically significant, no adjustmentsto these items were made. Adjustments to the three revenue categories were made. Thus, the adjusted surplus measureis givenby 3 A; = R, - X , -CR,[exp(e,yf)-l]. j=1 113 AB. Table 1 Estimatesof Revenueand Expenditure Elasticities Elasticity RevenueSource Income andProfit Taxes 2.04* Domestic Goods and Services Taxes 2.3 1* International Trade Taxes 3.60"' Expenditure Category Transfers 0.54 ** Significantat 1percent. * Significantat 10percent. Notes: The estimates were computedusingordinary least squares. 114 % a P W " J $d ai q a li " f d N 3i d I E d 1 :C c c 0 P 0 a Br J ; @ 8 : : cI d 1 Q 1% E ! yd'il 1E ; 2 % 9' h I- v o r$ i i O - ; o oo 5? f cccc 4s5 e $ C 0 09 0 C 2 0 2 i 0 ? 3 0 3 d 9 m i Y c1 2 2 3 m 0 - 2 % 3 " 3 "0 d "" - w 0 2 ri "13 I : 2 x 0 0 t 0 0 a l- 3 3 c1 0 2 0 a 0 t - 0 3 m 2 0 c! 00 9 c m Q\ . J \o Y 0 m 2 00 2 x [I 3 0 0 m 2 2 2 t': N 0 a 3 \o 8 rn 3 0 W Y d m 2 m 2 2 2 00 c 0 W c! 0 00 ": 3 o e m m m m w m B g .I 5 8 w 3 vi 2 a - :? vi3 x 0 0 2 !2 0 0 2 B 2B n Q .I U u 4 m 3 r? 3 0 ? - W B 3 v) 4 m 00 *%- s", 3 3 m \4 9 9 9 9 9 1 ~ ~ 0 0 0 0 s 3 THIS PAGEINTENTIONALLY LEFT BLANK THIS PAGEINTENTIONALLYLEFTBLANK