Report No. 18791-KH Cambodia Public Expenditure Review Enhancing the Effectiveness of Public Expenditures (In Two Volumes) Volume II Main Report January 8, 1999 Poverty Reduction and Economic Management Sector Unit East Asia and Pacific Region Document of the World Bank CURRENCY EVALUATIONS Currency Unit = Cambodian Riel AVERAGE VALUE OF US$1.00 IN 1995 1996 1997 December 1998 2451 2,624 2,946 3,700 WEIGHTS AND MEASURES Metric System GOVERNMENT'S FISCAL YEAR January I - December 31 ABBREVIATIONS AND ACRONYMS ADB - Asian Development Bank ADD - Accelerated District Development Program BIS - Bank for International Settlements ASEAN - Association of South-East Asian Nations CCC - Cooperation Committee for Cambodia CDC Council for the Development of Cambodia CIF - Cost, Insurance and Freight CG - Consultative Group CR - Cambodian Riel CRDB - Cambodia Reconstruction and Development Board CSES - Cambodia Socioeconomic Survey CVAP - Cambodia Veterans Assistance Program DHC - District Health Center DFW - Department of Forestry and Wildlife EPI - Expanded Program of Immunization FAO - Food and Agriculture Organization of the United Nations FDI - Foreign Direct Investment GDP - Gross Domestic Product GSP - General System of Preferences MEF - Ministry of Economy and Finance MFN - Most Favored Nation MOEYS - Ministry of Education, Youth, and Sports MOH - Ministry of Health MOP - Ministry of Planning MTEF - Medium-Term Expenditure Framework NBC - National Bank of Cambodia NGO - Non-Governmental Organization O & M - Operations and Maintenance ODA - Official Development Assistance PER - Public Expenditure Review PIMS - Public Investment Management System PIP - Public Investment Program PSI - Pre-shipment Inspections RCAF - Royal Cambodian Armed Forces SEDP - Socioeconomic Development Plan SWAP - Sector-Wide Approach UNTAC - United Nations Transitional Authority in Cambodia TA - Technical Assistance TVET - Technical and Vocational Education and Training UNDP - United Nations Development Programme UNESCO - United Nations Educational, Scientific and Cultural Organization UNICEF - United Nations Children's Fund USAID - United States Agency for International Development VAT - Value Added Tax WHO - World Health Organization WTO - World Trade Organization Vice President: Jean-Michel Severino, EAP Chief Economist: Masahiro Kawai, EAP Country Director: Ngozi Okonjo-Iweala, EACSM Acting Sector Manager: Kyle Peters, EASPR Task Manager: Su-Yong Song, EASPR TABLE OF CONTENTS Page No. INTRODUCTION...................................................... i 1. RECENT ECONOMIC DEVELOPMENTS AND STRUCTURAL REFORMS.............1 A. OVERVIEW...............................................................1 B. REAL SECTOR DEVELOPMENTS AND POVERTY SITUATION........................3 C. FISCAL DEVELOPMENTS AND POLICIES .......................................5 D MONETARY DEVELOPMENTS, INFLATION, EXCHANGE RATES, AND FINANCIAL SECTOR REFORM .......................................................... .......6 E. EXTERNAL SECTOR DEVELOPMENTS AND POLICIES .......... ...................7 F. FORESTRY MANAGEMENT AND OTHER STRUCTURAL REFORMS ....................8 G. CONCLUSION...............................................................8 2. PROSPECTS FOR DOMESTIC REVENUE MOBILIZATION .................9 A. CURRENT SITUATION ....................9............ ....................9 B. IMPLEMENTING THE LAW ON TAXATION ......................................9 C. IMPLEMENTING THE VAT ..................................................1. 0 D. LIMITING TAX AND DUTY EXEMPTIONS ......................................10 E. PROPER CAPTURE OF FORESTRY REVENUE..............................12 F. PROPER TAXATION ON OTHER NATURAL RESOURCES .......................... 15 G. STRENGTHENING TAX ADMINISTRATION /GOVERNANCE ................... .....15 H. PROPER TRANSFER OF NoN-TAX REVENUES FROM LINE MINISTRIES ..............16 I. CONCLUSION............................................................ 16 3. ANALYSIS OF LEVEL AND COMPOSITION OF PUBLIC EXPENDITURES.........17 A. INTRODUCTION .......................................................... 17 B. LEVEL OF SPENDING ON PUBLIC SERVICES ...................................19 C. ROLE OF GOVERNMENT AND AID AGENCIES IN PUBLIC SERVICE PROVISION.......21 D. FUNCTIONAL COMPOSITION OF PUBLIC SPENDING ............ .................22 E. INTRASECTORAL ALLOCATION OF PUBLIC SPENDING: HEALTH AND EDUCATION SECTORS ...............................................................24 F ECONOMIC COMPOSITION OF PUBLIC SPENDING.............................28 G. REGIONAL COMPOSITION OF PUBLIC SPENDING .........................30 H. CONCLUSION ............................................................31 4. ASSESSMENT OF PUBLIC EXPENDITURE MANAGEMENT: INSTITUTIONAL DIMENSION ................................. ......32 A. OVERVIEW ..............................................................32 B. LINKING PLANNING, POLICY, AND BUDGETING ...................................................... 33 C. BUDGET PREPARATION ........................................................................................... 34 D. BUDGET EXECUTION.....................................................35 E. BUDGET MONITORING AND REPORTING ................................38 F. RECOMMENDATIONS .....................................................39 G. CONCLUSION............................................................46 - 11 - 5. MACROECONOMIC FRAMEWORK, RESOURCE ENVELOPE, AND PUBLIC EXPENDITURES ............................................. .....47 A. INTRODUCTION .................................................. ........47 B. THE FULL REFORM SCENARIO ..............................................47 C. THE STATUS QUO SCENARIO ...............................................50 D. DOWNSIDE RISK.................... ............................. .......52 E. CONCLUSION.. ............................................................52 ANNEXES: ANNEX 1: FORESTRY SECTOR ISSUES ...................... ......................53 ANNEX 2: PEACE DIVIDEND ANALYSIS OF MILITARY DEMOBILIZATION ... ............63 ANNEX 3: PUBLIC EXPENDITURE ANALYSIS OF THE HEALTH SECTOR .................71 ANNEX 4: PUBLIC EXPENDITURE ANALYSIS OF THE EDUCATION SECTOR .............86 STATISTICAL APPPENDIX ...................................... .....1 03 TABLES 1.1 MACRO ECONOMIC INDICATORS, 1991-98 ................. ................... 1.2 MACROECONOMIC INDICATORS: BEFORE AND AFTER THE SHOCKS .....................2 1.3 MACROEcoNoMic INDICATORS: TARGET VS. OUTCOME, 1997-98...................3 1.4 CHANGE IN POVERTY SITUATION ..................................................4 1.5 BUDGET IMPLEMENTATION, 1997-98 ................5......... .............5 2.1 FORGONE REVENUE BASED ON KNOWN INFORMATION ............................... 10 2.2 DIFFERENT REVENUE POTENTIAL FROM FORESTRY.................................. 13 3.1 ALTERNATIVE CONCEPTS OF EXPENDITURES ON PUBLIC SERVICES, 1996 ...........................20 3.2 TOTAL SPENDING ON PUBLIC SERVICES FROM ALL SOURCES, 1994-97.......... .....20 3.3 PUBLIC EXPENDITURE IN CAMBODIA: AN INTERNATIONAL PERSPECTIVE .. ................21 3.4 COMPOSITION OF SPENDING ON PUBLIC SERVICES IN CAMBODIA, 1996 ...... ................22 3.5 ECONOMIC AND FUNCTIONAL CLASSIFICATION OF PUBLIC EXPENDITURES FINANCED FROM ALL SOURCES, 1996.................................................23 3.6 OUTCOME Vs. BUDGETED NON-WAGE O&M EXPENDITURES ......................30 3.7 ACTUAL SPENDING PATTERN AND PLAN ALLOCATION ................................30 3.8 REGIONAL COMPOSITION OF PUBLIC SPENDING, 1996 ..........................31 5.1 MACROECONOMIC FRAMEWORK - THE FULL REFORM SCENARIO ......................48 5.2 EXTERNAL FINANCING REQUIREMENTS - THE FULL REFORM SCENARIO ...................49 5.3 REQUIRED NEW COMMITMENTS OF OFFICIAL DEVELOPMENT ASSISTANCE ................49 5.4 MACROECONOMIC FRAMEWORK - THE STATUS QUO SCENARIO .......................51 FIGURES 1.1 MONEY, INFLATION, AND EXCHANGE RATE, 1993-98 .............. .............6 3.1 THREE CONCEPTS OF PUBLIC EXPENDITURES....................................... 19 3.2 SOURCES OF FINANCING PUBLIC SERVICES, 1994-97.............. .............21 4.1 MONTHLY DISTRIBUTION OF REVENUES AND EXPENDITURES IN CAMBODIA, 1997 ............36 BOXES- 111 - BOXES 1.1 ECONOMIC REFORMS SINCE 1993 .......................... ............2 2.1 EVIDENCE OF UNOFFICIAL PAYMENTS IN THE FORESTRY SECTOR............ ........... 13 4.1 THE BUDGET PREPARATION PROCESS IN CAMBODIA ............. ...................34 4.2 FINANCIAL CONTROLLER SYSTEM................................................36 4.3 ANECDOTAL EVIDENCE OF LEAKAGE OF FUNDS...............................38 4.4 ACCELERATED DISTRICT DEVELOPMENT SYSTEM (ADD)..........................41 4.5 PROVINCIAL BUDGET MANAGEMENT LAW ................................... ......43 ACKNOWLEDGEMENTS This report was prepared on the basis of an economic mission that visited Cambodia in March 1998. The report was prepared by a team led by Su-Yong Song (task manager), and composed of Bill McCleary (advisor), Tom Hart (revenue mobilization), Govinda Rao (public expenditure analysis), Geoff Dixon (public expenditure management), Wijaya Wickrema (accounting and audit), Marc Quintyn (macroeconomic framework, IMF), Bill Magrath (forestry issues), Markus Kostner (military demobilization), Anil Deolalikar (health and education), Vin McNamara (education), and Ky Tran (statistics and projections). Contributions were also made by Rama Lakshminarayanan, Peter Moock, and Christopher Thomas. The peer reviewers were Vinaya Swaroop, Malcolm Holmes, Jeffrey Hammer, Pascale Kervyn De Lettenhove. It was prepared under the guidance of Masahiro Kawai (Chief Economist, East Asia and Pacific Region), Ngozi Okonjo-Iweala (Country Director, EACSM), Kyle Peters (Acting Sector Manager, EASPR), and Richard Newfarmer (Lead Specialist, EASPR). The manuscript was edited by Emily Evershed and typed and formatted by Lily Tsang. The team would like to express its gratitude to various ministries and agencies of the Government of Cambodia throughout the preparation of this report, in particular the Ministry of Economy and Finance, the Council for the Development of Cambodia, the Ministry of Health, and the Ministry of Education, Youth, and Sports. The team would also like to express its gratitude to various donor and NGO agencies, especially WHO, UNESCO, UNICEF, and MEDICAM. The report was discussed with the Government of Cambodia in December 1998 and revised on the basis of these discussions. Introduction 1. This Public Expenditure Review (PER) undertakes, for the first time, a systematic review of the adequacy and effectiveness of public expenditures in Cambodia. By making concrete recommendations, the PER aims to help the government strengthen its capacity to manage public expenditures. The PER, which is to be presented to the Consultative Group (CG) meeting scheduled for February 25-26, 1999, makes the following main contributions: * The PER tackles the issues of governance, in particular through estimating the extent of revenue loss resulting from weak governance, especially from illegal logging, as weak governance in public resource management is a critical impediment to sustainable development in Cambodia. * The PER estimates foregone revenues with credible accuracy based on known information and assesses the potential for enhancing revenue mobilization as revenue mobilization is a prerequisite for effective expenditure policy in Cambodia. * The PER constructs a consolidated public expenditure database--encompassing expenditures incurred by the government, donors, and NGOs--and analyzes the level and composition of more complete public sector expenditures in collaboration with the government as it is difficult to form a reliable picture of the total expenditure of a public nature in Cambodia because a significant portion of aid-financed expenditures are not included in the government's official budget. * The PER assesses the implications of the institutional arrangements and management practices in the budgetary process on determining budgetary outcomes and makes specific recommendations for reforming institutional procedures in view of Cambodia's weak institutional capacity. 2. The coverage of issues and sectors is selective. First, while the report analyzes sector-specific issues for various sectors within the overall framework, it provides a more in-depth analysis of the health and education sectors because of their direct relevance to poverty reduction strategy through human resource development. Second, a full assessment of issues such as civil service reform, public enterprises, and fiscal decentralization is outside the scope of the PER. Third, more specific aspects, such as benefit incidence analysis of public expenditure, are left to the forthcoming report Cambodia: Poverty and Social Services. 3. This volume is the main report and is composed of five chapters: * Chapter I assesses recent economic developments and structural reforms. A focus is given to the interrelated fundamental and governance problems and also to the impact of the July 1997 events and the regional financial crisis. * Chapter 2 examines the current situation in domestic resource mobilization from various sources, estimates the extent of foregone revenue resulting from weak governance in public resource management, and assesses the prospects for revenue enhancement over the medium term. -il- * Chapter 3 examines the scope of public expenditures, the level of spending on public services, the role of the government and aid agencies in public service provision, and the functional, economic, and regional composition of public expenditures. * Chapter 4 assesses the effectiveness of institutional arrangements and management practices in determining budgetary outcomes and makes specific recommendations for reforming institutional procedures. * Chapter 5 presents two alternative macroeconomic scenarios (i.e., with or without reform) over the medium term by pulling together the discussions of the preceding chapters, and also highlights the implications for the level and reorientation of public expenditures. In addition, this volume has four annexes and statistical appendix: * Annex I examines issues related to forestry management. * Annex 2 presents a peace dividend analysis of military demobilization. * Annex 3 provides in-depth analysis of public expenditure issues in the health sector. * Annex 4 provides in-depth analysis of public expenditure issues in the education sector. 1. RECENT ECONOMIC DEVELOPMENTS AND STRUCTURAL REFORMS A. Overview 1.1 Political Background, The Royal Government--a coalition government formed in 1993 after the two decades of conflict--performed well during its initial years and made remarkable progress in stabilizing the economy, in restoring economic growth, and in undertaking policy reforms to transform the economy into a market-oriented one (see Table 1.1 and Box 1.1). In particular, inflation which averaged 140 percent per annum during 1990-92 was reduced to single digits during 1995-97. Continued Khmer Rouge insurgencies, however, led to pressures on defense expenditures. Since 1996, tensions between the two main coalition partners over power sharing escalated and culminated in the July 1997 events. Partly reflecting this, the government's reform efforts weakened considerably, which adversely affected economic performance. More recently, the political and security situation improved significantly with the formation of the new coalition government resulting from the national elections held on July 26, 1998 and the marked weakening of the Khmer Rouge elements. Table 1.1: Macroeconomic Indicators, 1991-98 1991 1992 1993 1994 1995 1996 1997 1998 Est. Real GDP Growth (%) 7.6 7.0 4.1 4.0 7.6 7.0 1.0 0.0 Inflation (final quarter basis, %) 150.4 112.5 41.0 17.9 3.5 9.0 9.1 12.0 Broad Money Growth (%) 28.6 209.0 40.0 29.4 44.3 40.4 16.0 5.4 Budget Revenue (% of GDP) 4.4 6.2 5.4 9.6 8.9 9.1 9.7 8.1 Budget Expenditure (% of GDP) 7.8 9.8 11.2 16.5 16.7 16.3 13.9 11.7 Current Budget Balance (% of GDP) -1.2 -4.3 -1.4 -1.4 -0.6 -0.8 0.6 -0.5 Export of Goods (US$ Mn., excluding re-export) 81 101 168 234 269 298 404 469 Import of Goods (US$ Mn.. retained import) 113 160 361 509 673 749 707 775 External Current Account Deficit (% ofGDP) -1.5 -2.5 -9.4 -13.7 -16.1 -15.5 -11.4 -11.6 Official Exchange Rate (annual avg., Riels/USS) 703 1,253 2.470 2,543 2,462 2,624 2,989 3,700 Source: Data provided by the Cambodian authorities. 1.2 Fundamental Fiscal and Governance Problems. While macroeconomic stability has largely been maintained, despite the government's piecemeal efforts little progress has been made in implementing the measures to address the interrelated fundamental fiscal and governance problems. Revenue mobilization is a critical problem, owing especially to a plethora of tax exemptions. Inadequate budgetary expenditures for social and economic infrastructure, owing to weak revenue mobilization and overruns in defense expenditure, have hampered rapid and broad-based economic growth. Recent progress is also threatened by insufficient accountability and transparency in economic management. In particular, the lack of action in implementing a sustainable and transparent forestry management policy and in channeling forestry revenue into the budget has posed serious concerns for sustainable development. Concerns on weak domestic revenue mobilization, inadequate and inefficient public expenditure management, and lack of governance in economic management--forestry management in particular--were strongly raised at the Consultative Group (CG) meeting during July 1-2, 1997. -2- Box 1.1: Economic Reforms Since 1993 Cambodia's economic reform program began in earnest after the elections of May 1993. Reforms have moved the system substantially toward a market economy, reduced the size of public sector, and brought about substantial stabilization. * Pricing: Most prices were liberalized except those for utilities. * Exchange Rate: The exchange rate was liberalized and the spread between the official and the parallel market rates has been narrowed to the minimum level. * Tax Reform: Increasing emphasis has been put on income, turnover, and excise taxation and improvements in tax administration. The resulting rise in revenues has helped in holding deficits to reasonable levels and virtually eliminating monetary finance. * Expenditure Policy: An Organic Budget Law has been passed which specifies content, procedures, and responsibilities for the budget. Three-year rolling Public Investment Programs have also been prepared. * Fiscal Policy: Improved controls over expenditure and improved revenue mobilization, combined with significant foreign aid inflow, have reduced the need for monetary finance and significantly contributed to stabilization success. * Banking System: Substantial progress has been made toward a two-tiered banking system. The system opened up to entry by foreign and domestic commercial banks. * State Enterprises: The number of state enterprises was substantially reduced through privatization and long- term leases to the private sector. A Public Enterprise Law was adopted which lays out the responsibilities of enterprises and the state and puts public enterprises under the same legal framework as the private sector. * Foreign Investment: A liberalized investment law was passed which has contributed to substantial increases in FDI flows. * Trade Regime: Quantitative restrictions on imports have been eliminated and the bulk of import categories are governed by five tariff bands ranging from zero to 50 percent. The government is preparing for membership in ASEAN and WTO. 1.3 Two Shocks. In addition, since Table 1.2: Macroeconomic Indicators: mid-1997 the economy has suffered two Before and After the Shocks shocks. One was the political events of First-Half Second-Haff Jan.-Oct. early July 1997 and the ensuing suspension 199a 19 15 of a substantial part of donor assistance. Exchange Rate Depreciation 2 20 It The other was the regional financial increase in Tourist Arrivals (%) a! 24 -50 -23 criis Alhogh coomi prfomace Private Sector Credit Growth (in US$) 17 -2 -2 crisis. Although economic performance II I I_LJ deteriorated significantly in the second half a/ Compared with the same period in previous year. of 1997, the economic outcome for the Source:Data provided by the Government entire year was somewhat better than expected immediately after the onset of the two shocks, thanks to the governments continued fiscal discipline. Little tangible progress, however, has been made in implementing key structural reforms, especially in the area of improving governance in public resource management. 1.4 Although it is difficult to separate the respective impact of the two shocks, the impact of the political events appears to have been initially more severe than that of the regional financial crisis. While the July 1997 events and the ensuing political uncertainty substantially affected growth, particularly through undermining investors confidence, the impact of the regional financial crisis to date has been mitigated by the highly dollartized nature of the economy. For example, the acceleration of inflation in the second half of 1997 was lessened by significant declines in dollar-denominated prices of imports from neighboring countries. The adverse impact of the regional financial crisis is manifested mainly in the cancellation/delay of committed investment projects and in the drop in tourism. The impact of the two shocks is demonstrated in Table 1.2 and Table 1.3. -3- Table 1.3: Macroeconomic Indicators: Target vs. Outcome, 1997-98 Indicator 1997 1998 Target a/ Outcome Target a/ Estimated GDP Growth (%) 6.5 1.0 6.5 0.0 CPI Inflation (%) 5.0 9.1 5.0 12.0 Exchange Rate (end of period, riels/USS) 2,800 3.450 2,900 3,700 Budget Revenue (% of GDP) 9.7 9.7 11.1 8.1 Budget Expenditure (% of GDP) 16.0 13.9 16.9 11.7 Current Budget Balance (% of GDP) 1.1 0.6 2.1 -0.5 Current Account Deficit (USS mil.) 536 346 561 329 Current Account Deficit (% of GDP) 16.0 11.4 15.2 11.6 Foreign Direct Investment (USS mil.) 252 135 269 120 ODA (US$ mil.) 399 208 372 203 Gross Foreign Reserves (US$ mil.) 290 264 310 386 b/ Gross Foreign Reserves (months of imports) 2.5 2.5 2.7 3.5 a/ Policy Framework Paper, 1997-99. b/ Including government's gold holding of US$ 117 million (that had been blocked since the Pol Pot regime) recently released by BIS. 1.5 As the 1996 and 1997 Economic Reports analyzed in depth the macroeconomic and structural reform issues for the period 1991-96, this chapter assesses the economic developments and structural reforms since 1997. A particular focus is set on the interrelated fundamental fiscal and governance problems and also on the impact of the two shocks. B. Real Sector Developments and Poverty Situation 1.6 Real Sector Developments. After registering an average rate of 6 percent during the previous five years, GDP growth decelerated to 1 percent in 1997, mainly reflecting the July events and the financial turmoil in the region. Tourism and construction were the hardest hit. For example, despite a strong increase during the first half of the year, tourist arrivals decreased by 16 percent for the whole of 1997 compared with 1996. Agriculture growth was stagnant as a result of bad rice harvest. Strong growth was recorded in manufacturing (garments in particular), utilities, and transportation and communications. On the demand side, a squeeze in government expenditures (non-wage civilian O&M in particular) owing to revenue shortfalls, and a substantial slowdown in investment due mainly to the suspension or postponement of foreign financed projects (both ODA and FDI funded) in the aftermath of the political conflict and the regional financial crisis, led to a slack in aggregate demand. Stagnation deepened in 1998. GDP growth in 1998 is estimated at zero percent due to the draught and continued regional financial crisis and political uncertainty in the runup to the elections in July. In particular, tourist arrivals during the first ten months in 1998 decreased further by 23 percent compared with the same period in 1997. 1.7 Poverty Situation. According to the Bank's forthcoming report, Cambodia: Poverty and Social Services, the poverty situation seems to have improved modestly between 1993/94 and 1997.2 The population below the poverty line declined from 39 percent to 36 percent between 1993/94 and This deceleration of GDP growth also reflects long-run (i.e., productivity enhancing) factors which have been adversely affected by the recent slow progress in the government's efforts to provide a better enabling environment for private sector development. 2 The report updates the baseline poverty profile estimated for 1993/94 in A Poverty Profile of Cambodia, World Bank Discussion Paper No. 373, October, 1997, utilizing the 1997 Cambodia Socio-Economic Survey. It should be -4- 1997 for the country as a whole3 (Table 1.4). On a Table 1.4: Change in Poverty Situation a/ regional basis, poverty incidence declined 993/9 1997 significantly in other urban areas (from 37 to 30 Phnom Penh 11.4 11. percent), modestly in rural areas (from 43 to 40 Other Urban 36.6 29.9 percent), and not at all in Phnom Penh (remaining at Rural 43.1 40.1 11 percent). While the poverty level is still high TOt 39.0 36.1 compared with that of other East Asian countries, it is encouraging that poverty incidence declined in areas Soure coder outside of Phnom Penh, where the poverty situation WorldBank,forthcoming is much worse. 1.8 The main characteristics of the poverty incidence and implied policy directions are the following: " The highest incidence of poverty is found in rural areas, with 40 percent estimated as poor, four times higher than the II percent in Phnom Penh. Rural households continue to account for almost 90 percent of Cambodia's poor. Government policies and associated donor support to reduce poverty must, therefore, mainly target the rural population. * The highest poverty rate--44 percent--is found among people living in households headed by farmers. By contrast, households headed by some who work in the government have a poverty rate of only 18 percent. Clearly, policies aimed at reducing poverty through enhancing income generating capacities should be targeted toward the agricultural sector. * Because three-fourths of the poor continue to be self-employed, policies designed to make markets more efficient (particularly markets for agricultural commodities, inputs, and rural credit) will be more useful as poverty alleviation policies than policies targeted to employers and employees, such as minimum wage legislation, social security, and mandatory fringe benefits. * Poverty rates are high for households in which the head of the household has had either no schooling (42 percent) or only some primary schooling (40 percent). These two groups account for four-fifths of all poverty. The prevalence of poverty among households in which the head had completed higher and lower secondary education falls to 12 percent and 24 percent, respectively. Raising educational attainment (secondary as well as primary) is clearly an effective poverty alleviation intervention. * The incidence of poverty averages 33 percent in female-headed households, compared with 37 percent in male-headed households. The gender and poverty patterns by household head are similar to those observed in other East Asian countries such as Vietnam and Indonesia. * Rural poverty in Cambodia (40 percent) is lower than in its Indochina neighbors--Vietnam (47 percent) and Lao PDR (53 percent)--but considerably higher than elsewhere in East Asia. It shows the magnitude of the development gap which remains between the economies of Indochina and the rest of East Asia, while suggesting that Cambodia's starting point is not very different from its neighbors in Indochina. cautioned that the updated poverty estimates for 1997 are not strictly comparable with those of the baseline poverty profile as there are differences in coverages. In particular, only 56 percent of rural villages were covered in the 1993/94 survey owing to the security situation whereas 87 percent were covered in the 1997 survey. However, the fact that ownership of many consumer durables has increase in all quintiles from 1993/94 to 1997 seems to support the observed decline in poverty. 3 Among various poverty measures, this assessment is based on the head count index--the most commonly used index of poverty--which is simply the proportion of the population whose expenditure levels fall below the poverty line. -5- C. Fiscal Developments and Policies 1.9 Fiscal Performance. Fiscal performance in 1997 largely repeated the Table 1.5 Budget Implementation. 1997-98 pattern of the past few years: civilian non- wage operating expenditures were compressed Outcome/Budget Est. Outcome/Budget significantly to sustain macroeconomic 1997(/) 1998(%) stability in response to a revenue shortfall and Total Revenue 98 93 an overrun in defense and security outlays Comties 79 94 (Table 1.5.) While the squeeze in civilian Non-tax Revenue a/ 128 86 operating expenditures enabled the Total Expenditure 86 88 government to avoid recourse to bank Current Expenditure 94 100 financing of the budget deficit, the current Wages 106 111 pattern of fiscal adjustment raises serious CivilAdministration 101 111 concerns about medium term sustainability. Defense and Security 108 112 Total government revenue amounted to 9.7 Non-wages percent of GDP. It is noteworthy that total Civil Administration 76 75 revenue as a share of GDP has been hovering CaptalExpnditurs 74 69 around 9 percent during the last four years. Domestic Financing 91 57 The main culprit in the revenue shortfall was Exteral Financing 70 73 trade taxes. Despite the depreciation of the a/Including capital revenue. riel, only 79 percent of budgeted trade taxes Sour: Data provided by the government were collected, mainly because of the widespread granting of ad hoc exemptions on garment imports and exports of sawn timber and rubber. While domestic tax revenue exceeded the target, it should have been higher had the new measures under the Law on Taxation been fully implemented without delay. In contrast to an overrun in defense and security expenditures (by 8 percent), civilian non-wage operating expenditures were compressed significantly (by 24 percent). Capital expenditure was lower than planned (by 26 percent), owing to a slowdown in aid inflows since July 1997. .10 Fiscal performance in 1998 deteriorated further--worse than envisaged even though the budget was not ambitious in terms of mobilizing revenue and reorienting expenditure. Budget revenue is estimated at 8.1 percent of GDP compared with the budgetary target of 8.9 percent. Defense expenditure significantly overran the budgetary target due partly to the integration of the Khmer Rouge defectors. (It should be noted that these expenditures included some spending of social and economic nature such as allocations for the families of defected Khmer Rouge soldiers and infrastructure rehabilitation for them.) Worrisome development was that the Treasury has accumulated salary arrears for three months. Thus, the government has borrowed CR 82 billion (or 0. 8 percent of GDP) from the National Bank of Cambodia (NBC--for paying salary arrears, integrating Khmer Rouge defectors, and also financing a part of election costs--after having avoided bankfinancing of the budget deficit since 1994. Furthermore, as of December the government accumulated salary arrears for another two months. 1.11 Fiscal Reform As reflected in the fiscal performance, progress in implementing fiscal reform measures was not satisfactory and in some cases took a backward step (for details, see Chapter 2). Full implementation of the Law on Taxation has been delayed because of technical difficulties and a lack of consensus regarding the pace of implementation. The initial introduction of Value-added Tax (VAT) for about 800 large taxpayers, which was to be effective beginning in 1998, has been postponed until January 1, 1999. In addition to granting ad hoc tax and customs exemptions, the government extended the exemption from pre-shipment inspections (PSI) to garments and cigarettes after the July 1997 events. -6- Combined with the existing exemptions on imports for the projects approved under the Law on Investment, the PSI-exempted imports amount to some 60 percent of total customs imports, causing serious concerns about potential under-invoicing of imports. While the Implementing Regulations for the Law on Investment were adopted in January 1998 with a view to strictly limiting the scope of ad hoc tax and duty exemptions, the recommended areas of investment cover virtually every sector in the economy, including, in particular, natural resource exploitation (see Chapter 2). 1.12 The 1999 budget aims at significant improvement in fiscal performance with the revenue target of 10.4 percent of GDP, current expenditure of 9.0 percent of GDP, and defense and security expenditure of 3.7 percent of GDP. For the envisaged revenue target to be realized, effective implementation of the VAT as scheduled will be critical. Improving governance in tax and non-tax collection will be also important. Reigning and rationalizing defense and security expenditure will be key to make the expenditure target feasible. D. Monetary Developments, Inflation, Exchange Rates, and Financial Sector Reform 1.13 Monetary Developments. The Cambodian economy is highly dollarized and is largely cash based. The most significant monetary development during 1997-98 was a substantial withdrawal of foreign currency deposits (10 percent in U.S. dollar terms in 1997), indicating waning confidence in the Cambodian economy in the aftermath of the July 1997 events. This withdrawal led to a substantial slowdown in broad money growth to 16 percent as compared with 40 percent in 1996 (Figure 1.1). While the level of foreign currency deposits was stable during the first six months of 1998, it dropped by 14 percent in U.S. dollar terms as of end October compared with the beginning of the year, largely reflecting the uncertainty in the runup to the elections. This led to the increase in broad money supply by only 7 percent as of end October 1998. Following a strong increase in the first half of 1997, the downturn in economic activity resulted in a decrease in extending private sector credit (see Table 1.2). Several commercial banks had problems with customers that were unable to service their loans since the second half of 1997. During 1998, one commercial bank has been closed and two downsized their operation due mainly to the continued regional financial crisis. Figure 1.1: Money, Inflation, and Exchange Rate 1993-98 95 75 5 5 15 1993 1994 1995 1996 1997 Oct-98 -0--Inflation -----Exchange Rate Depreciation -A--M2 Growth 1.14 Inflation and Exchange Rates. Consumer prices increased at a moderate rate in the first half of 1997, but the rate of increase accelerated during the third quarter. Afterward, this acceleration was mitigated by significant declines in dollar-denominated prices of imports from neighboring countries that were undergoing financial crises, such as Thailand and Malaysia. The inflation rate for the whole year was 9 percent. To date in 1998, inflation has been on the rise (12 percent as of October on a 12-month basis), owing mainly to the rice price increase resulting from a rice harvest shortfall and also from the -7- hoarding of some commodities ahead of the July election. After a rapid rise during May-July in the runup to the elections, inflation decelerated. The exchange rate remained largely stable during the first half of 1997, but depreciated by 20 percent during the second half. As of end October 1998, the riel had depreciated by some 11 percent during the year despite the National Bank's occasional dollar auctions in modest amounts. The exchange rate stabilized after a rapid rise during May-July owing to the uncertainty in the runup to the July election. 1.15 Financial Sector Reform. Some progress has been made in the areas of banking supervision and prudential regulations. Eight banks were audited by international audit firms on the basis of on-site inspections in 1997. These inspections revealed problems in one bank which was given six months to implement a recovery plan. The NBC strengthened reporting requirements by asking the commercial banks to submit a list of loans exceeding US$100,000 according to adequate loan classification criteria, on a monthly basis. The NBC raised the minimum capital requirement for new banks to US$15 million, and increased the reserve requirement from 5 to 8 percent and the capital guarantee deposit from 5 to 10 percent effective January 1997, to ensure the liquidity and solvency of the banking system. However, the adoption of the Commercial Bank Law, which is to provide a transparent framework for banking operations, has been delayed by more than a year. E. External Sector Developments and Policies 1.16 Balance of Payments. The most significant development in 1997 was a strong growth in nontraditional exports (mainly garments). Garment exports in 1997 tripled to US$202 million from the 1996 level, benefiting from the Most Favored Nation (MFN) and General System of Preferences (GSP) status. Although the rate of depreciation of the riel has been substantially lower than that of the currencies of other countries in the region, and therefore Cambodia has lost some of its export competitiveness in terms of prices, as yet this has had little effect on exports as Cambodia has no quota on garment exports (its main export item) whereas most regional competitors have already filled their quotas. Thus, domestic exports (excluding re-exports) increased by 35 percent. In turn, retained imports (excluding imports for re-exports) declined by 6 percent. 1.17 Although aid and FDI related imports dropped substantially, this was partially offset by a sharp increase in government imports. Even though the amount of approved FDI projects did not drop by very much since the July events, the impact of the regional financial crisis began to be felt in terms of actual FDI inflows from the last quarter of 1997, as investors from such countries as Korea failed to follow through with approved investment projects. Combined with the impact of the July 1997 events, FD1 inflows plummeted to US$135 million in 1997, about half of the 1996 level. Official transfers and concessional loans contracted sharply (as can be seen in Table 1.3) in the aftermath of the July events. While no detailed balance of payments data for 1998 are available, one significant development was that gross international reserves increased significantly in 1998 as the BIS released the Cambodian government's gold holdings (estimated at US$117 million) that had been blocked since the Pol Pot regime. 1.18 Trade Policy. Cambodia is to become a member of ASEAN (and hence in the ASEAN Free Trade Area and Common Effective Preferential Tariff) shortly. Progress is being made toward joining the WTO. In the aftermath of the regional financial crisis, prices of imports from main trading partners such as Thailand have declined substantially in U.S. dollar terms. The government was considering raising import taxes on some products such as cement to protect local industries from a flood of cheap imports. -8- 1.19 External Debt. Cambodia's external debt at the end of 1997 amounted to US$2 billion (orabout 65 percent of GDP). The Russian Federation is Cambodia's main creditor with US$1.3 billion (if valued at the historical exchange rate used by the Russian Federation). Among the four Paris Club creditors that had agreed on granting highly concessional terms (Naples terms), a bilateral agreement has yet to be signed with the United States. Cambodia also needs to make a bilateral agreement on debt reduction/rescheduling with the Russian Federation within the framework of the Paris Club, as the latter recently became a member. F. Forestry Management and Other Structural Reforms 1.20 Forestry Management. Conditions in the forestry sector continued to decline and illegal logging and log smuggling have reached alarming levels (for details, see Chapter 2 and Annex 1). It was estimated that illegal logging in 1997 totaled between 3 and 4 million cubic meters. This is three-eight times levels that are sustainable. If the current anarchic situation continues, the commercial potential of Cambodia's forest resource would disappear in less than five years. 1.21 Despite the government's piecemeal initiatives, no significant improvement in addressing the fundamental problems in forestry management has been made. While the first months of 1997 saw some improvement following the imposition of a new export ban effective January 1, 1997, the situation deteriorated rapidly after the July 1997 events, with illegal logging and exports intensifying. In addition, provincial governors and military commanders issued quotas, permits, and other authorizations for logging and log transport outside of the concession contracts. One positive development is that four studies on forestry, commissioned by the government under the Bank Technical Assistance project (log control and monitoring, forest policy, forest concession management, and a legal study), are essentially completed and have produced implementable policy recommendations (see Chapter 2 and Annex 1). 1.22 The regional financial crisis has exacerbated the forestry situation. The average stumpage value of Cambodian timber plummeted from more than US$70 per cubic meter to less than US$14, and this contributed to the acceleration of illegal logging to the level of 4 million cubic meters mentioned above. The following results are the cumulative impact of the crisis working through several paths: (i) a general weakening of prices owing to contraction in key export markets (Japan, Korea, and Thailand); (ii) crisis- driven efforts to increase exports by Cambodia's more established competitors (primarily Malaysia and Indonesia); and (iii) because of the overall flattening of economic activity in the domestic economy, greater pressure on weak sector governance to allow the expansion of fellings beyond controlled or sustainable levels. 1.23 Other Structural Reforms. The envisaged civil service reform and military demobilization have been put on hold owing mainly to political events. Only limited progress has been made in privatization. While about 20 medium-size enterprises were divested, little progress was made toward the divestiture of six major rubber plantations because of lack of consensus within the government on the legal statute of the future entity or entities in charge of rubber cultivation. G. Conclusion 1.24 To reduce poverty significantly and enhance economic and social well-being, Cambodia must sustain rapid and broad-based economic growth over the long term. Toward this end, Cambodia must resolve the interrelated fundamental fiscal and governance problems that pose a serious threat to medium-term sustainability, which are discussed at length in the subsequent chapters. -9- 2. PROSPECTS FOR DOMESTIC REVENUE MOBILIZATION A. Current Situation 2.1 In Cambodia, weak domestic revenue efforts severely constrain the effectiveness of public expenditure policy as an instrument for accelerating growth and reducing poverty. While domestic revenue as a share of GDP doubled between 1991 and 1994, it was flat at about 9 percent during 1995-97, and worsened to 8 percent in 1998. Cambodia's current revenue-to-GDP ratio is very low by international standards, amounting to about half the average for low income countries. Domestic revenues barely cover current expenditures; capital expenditures are financed almost entirely from external sources. 2.2 The main reasons for the country's poor revenue performance are: * Inadequate implementation of the Law on Taxation and delay in introducing a VAT * Generous tax exemptions granted under the Law on Investment * Ad hoc tax and customs duty exemptions, often associated with unwarranted political interference * Weak capture of forestry revenues * Limited capacity in tax and customs administration. 2.3 The current low revenue effort is extremely detrimental to achieving sustainable development over the medium term. Nevertheless, there exists substantial revenue enhancing potential. As shown in Table 2.1, foregone revenues in 1996 and 1997 estimated by the PER based on known information were large, amounting to 5-6 percent of GDP each year (equivalent to two-thirds of actual budget revenue). In addition, owing to the delay in the implementation of a VAT until January 1, 1999, another 1.7 percent of GDP was foregone during 1998. The basis for each estimate is provided in the following sections. This chapter examines the current situation in domestic revenue mobilization from various sources, estimates the potential for revenue enhancement over the medium term, and makes specific recommendations for realizing the potential. B. Implementing the Law on Taxation 2.4 In sharp contrast to stagnant customs revenues, which grew by only 0.9 percent, domestic tax revenues grew by 31.5 percent in 1997, owing mainly to the new measures of the Law on Taxation which was adopted with the 1997 budget law. It should be noted, however, that tax revenue collection from the new measures of the Law on Taxation was significantly lower than initially anticipated. The two new major revenue measures for 1997 were the tax on turnover charged on the first sale after import and the excise tax on automobiles. Foregone revenue from these two sources amounted to 0.3 percent of GDP. This was due mainly to the late signing of the Law on Taxation, the delay in issuing implementing regulations, and the July 1997 events. Several other new measures in the Law on Taxation have yet to be implemented, including the base expansion of the tax on salary, the fringe benefits tax, the new tax accounting rules for the tax on profits, and the introduction of withholding taxes under the tax on profits. In terms of revenue generation, about 80 percent of the new tax measures under the Law on Taxation are currently being implemented. The full year implementation of the tax on turnover should have generated an additional CR 9 billion and the excise tax on specific goods another CR 6 billion for 1998 (taken together, about 0.2 percent of GDP). -10- Table 2.1: Foregone Revenue Based on Known Information a/ 1996 1997 Mn. S GDP% Mn. $ GDP% Law on Taxation - - 9 0.3 Law on Investment Tax on Profit 9 0.3 12 0.4 Custom Duties 39 1.2 37 1.2 Forestry b/ Tax on Profit 22 0.7 22 0.7 Royalties 60 1.9 60 2.0 Government Imports Duty Exemptions 10 0.3 13 0.4 Administation/Governance Tax Administration 4 0.1 3 0.1 Customs Administration 13 0.4 12 0.4 Non-tax Revenue 5 0.2 6 0.2 Total 163 5.2 174 5.7 a/ As the estimates are based on known information only, these should be considered as the minimum estimates. b/ Foregone revenue in 1996 is assumed to be the same level as in 1997. Source. Bank staff estimates based on data provided by the government. C. Implementing the VAT 2.5 The major revenue enhancing measure of the Law on Taxation for 1998 was to be the implementation of the VAT initially for about 800 large taxpayers registered with the Large Taxpayer Unit of the Tax Department, beginning January 1, 1998. It has been postponed until January 1, 1999 due to the initial slow progress in preparation. Since then, however, preparation for the implementation has been largely on track. The Tax Department selected 50 staff to be assigned to VAT-related activities, and several training sessions have been held for the selected personnel. All ministerial regulations (prakas) and Council of Ministers regulations (sub-decrees) have been prepared and are ready for submission. Extensive work has been carried out to develop taxpayer forms, administrative forms, registration procedures, and taxpayer publications. After some delay since the elections in July 1998, the preparation for the VAT implementation--publicity campaign, taxpayer registration, education, and advisory visit-- accelerated during the last few months. For the effective implementation, the Government needs to swiftly issue necessary sub-decrees and prakas, and continue and intensify the training at the Tax Department, education of taxpayers, and publicity campaign. 2.6 The whole year implementation of the VAT would add 1.7 percent of GDP to revenue collection, because a significant portion of turnover tax--currently collected at the border in the form of a consumption tax at a standard rate of 4 percent--would be transformed into the VAT at a single rate of 10 percent. As it would take some time before becoming fully operational, the expected revenue enhancement from the VAT implementation in 1999 would be somewhat lower than the full effect. D. Limiting Tax and Duty Exemptions 2.7 The Law on Investment is one of the most critical impediments to improved revenue mobilization. The Law provides for tax holidays of up to eight years, a permanent 9 percent corporate tax rate after the end of the holiday period, duty free importation of capital goods and other fixed and movable assets, duty free importation of raw materials for export enterprises, and restrictions on the -11- taxation of remissions abroad. After a long delay, the Implementing Regulations for the Law on Investment were adopted in January 1998 with a view to strictly limiting the scope of ad hoc tax and duty exemptions. The implementation of these Regulations was a significant step in improving investor accountability for taxes. The recommended areas of investment, however, cover virtually every sector in the economy, including natural resource exploitation, in sharp contrast to the initial proposal of limiting the recommended activities strictly to medium and large scale manufacturing activities. The combination of the Law and the Regulations has eliminated any room for the business income tax to be a policy instrument in the revenue mobilization effort. 2.8 Assuming a modest return on the capital base, the lost revenue under the tax on profits would have been US$12 million, or 0.4 percent of GDP in 1997.1 Customs exemptions under the Law on Investment in 1997 amounted to US$249 million of imported goods. Assuming a duty rate of 15 percent, the foregone customs revenue in 1997 was US$37 million or 1.2 percent of GDP. 2.9 With the existence of the Law on Investment in its present form, tax revenues would be limited to domestic taxes on goods and services (sales taxes) and trade taxes. Once Cambodia becomes a member of ASEAN shortly, revenue from trade taxes will be limited. Membership will place downward pressure on both import and export tariffs and tax revenues. With these limitations, the only real revenue potential is limited to domestic consumption taxes. Thus, the Law on Investment would need to be significantly rewritten if the government is to realistically improve the overall revenue effort of the tax system. There is little justification for the tax incentives provided under the Law on Investment. The corporate tax rate (20 percent) is 10-15 percentage points lower than those of other ASEAN countries. This is a significant reduction in the level of taxation, and thus would support the elimination of the permanent 9 percent rate under the Law. In addition, the tax on profits provides other incentives which include rapid depreciation (most assets except structures can be written off in five years or less) and a five-year operating loss carry forward. The combination of these two provisions would eliminate the need to grant tax holidays. Moreover, the experience of other countries indicates that tax incentives rank low on the list of the major determinants of investment flows as compared to political and economic stability, a sound legal environment, and an adequate quality of physical infrastructure. Furthermore, the exemption of certain earnings, dividends, and interest, paid abroad, is hard to justify on the grounds that they in fact represent a subsidy to foreign governments, as most of the payments exempted in Cambodia are taxable in the home country of the investor. Almost all of these countries provide relief for Cambodian taxes through the foreign tax credit or deduction mechanism. 2.10 The government is urged to revise the Law on Investment and the Implementing Regulations by: * Removing all the incentives currently granted to resource and resource processing firms from the Implementing Regulations (see paras 2.15 and 2.20). * Removing the production of consumption goods from the preferred sectors in the Implementing Regulations. (However, the revision of the Implementing Regulations to this effect should not be retroactive.) * Repealing the 9 percent tax rate, the tax holiday period, and the profits reinvestment rule in the Law on Investment, with grandfathering clauses. Those activities other than resource activities could maintain their holiday status but not to exceed three years. All the holiday As actual amounts of fixed investment are not available, the amounts of FDI are used. Information suggests that most FDIs have short gestation periods (one year or less). Therefore, investment in place in 1997 would be US$606 million. All of these should be under tax holidays. Hence, foregone revenue assuming a 10 percent annual return and a 20 percent tax rate for 1997 would be US$12 million. -12- period would be rescinded after this period. Those firms currently paying taxes at the rate of 9 percent would be given a grace period of up to three years. * Amending the Law on Investment to make all repatriations and payments of interest subject to border withholding tax. * Reducing the duty exemption under the Law on Investment to 50 percent in duty in all provisions and put the reduction on a drawback scheme. While the implementation of the above recommendations (except for the first two which can be quickly done) could be carried out over the medium term, the legislative preparation needs to be carried out in next few months. 2.11 In addition to the exemptions granted under the Law on Investment, various ad hoc tax exemptions, such as on rubber exports, play a significant role in undermining the revenue potential of the tax system. It is, however, difficult to estimate the loss of revenue resulting from these ad hoc exemptions. The authorities need to amend the Law on Taxation to require that any other exemptions not specified in the Law on Taxation should be void. Even more serious, the government has granted a quasi-legal character to the exemptions as in the case of logging and cement production. This has been accomplished through the Implementing Regulations which now extend its scope beyond the initial intent to cover only manufacturing activity. In addition, the government received ad hoc duty exemptions on government imports which should have been paid, and the loss of revenue from this was 0.4 percent of GDP in 1997. E. Proper Capture of Forestry Revenue 2.12 Forests are Cambodia's most developmentally important natural resource and could bring in significant additional revenues. The current forestry situation, however, poses serious governance problems and shows evidence of pervasive corruption. The revenue loss arising from illegal logging to the government in 1997 is estimated as well in excess of US$60 million, or 2 percent of GDP (as compared with the actual collection of one-fifth of that amount in the budget). It was reported that private and unofficial rent amounted to US$14 per cubic meter, in addition to the official royalties2 and the unofficial charges associated with the original felling operation, as seen in Box 2.1. These levels of side-payments and extrabudgetary charges are indicative of the lack of governance in the forestry sector. The stresses in the Cambodia economy as a result of the regional financial crisis and the political events of July 1997 and the ensuing losses in financial support from various donors have exacerbated the situation. Thus, in the short term, much of the potential for increasing government revenue would come from re-channeling private and unofficial rent capture. 2.13 With a view to contrasting the revenue potential from forestry under different policy implementations, a simple financial model of the Cambodian forestry sector is developed and applied to three alternative scenarios over the period of the next 30 years (for a full description, see Annex 1): * Scenario 1: continuation of the current levels of illegal logging and minimal revenue collections from authorized concessions and illegal log "collections" * Scenario 2: full implementation of the concession system under the current set of contracts and prices 2 The government has decided to increase the official royalty from US$14 per cubic meter to US$54 beginning January 1, 1999. Realization of this remains to be seen as it could involve re-negotiating concession contracts. -13- * Scenario 3: adoption of a full rent recovery policy which would involve open trade, market- based royalties, and harvests held to a sustainable level as recommended by the Bank, UNDP, and FAO (1996). Box 2.1: Evidence of Unofficial Payments in the Forestry Sector Table I illustrates the problems associated with unofficial payments, and the reach of these practices. For a shipment of logs from the forest landing in east central Cambodia to processing facilities near Phnom Penh, at least eight supplementary payments totaling US$14 per cubic meter (or II percent of the CIF value of the shipment) are needed beyond official royalties and unofficial charges associated with the original felling operation. The largest single charge, US$7 per cubic meter (or 5.3 percent of the CIF value), is paid to provincial authorities and is widely recognized as a significant source of resources for regular provincial operations. The second single largest charge is to the local military, $1.70 (1.3 percent). In total, extrabudgetary charges by forestry personnel amount to somewhat more than 4 percent. Total additional charges (often called facilitation fees), including those directly associated with the felling operation, amount to approximately US$50 per cubic meter. Logging operators in Cambodia claim that these charges, on top of official royalties and other taxes, currently make profitable operation impossible. Table 1: Selected Costs Associated with an Internal Log Shipment (4.000 m3 from an East Central Province to Phnom Penh, 1997-98) Per Cubic Shipment Percent of Meter Total Value Us$ US$ Estimated Value (delivered) 130.00 520.000 100 Provincial Authorities 7.00 28,000 5 Military Authorities 1.70 6,800 1 Provincial Forestry Officers 0.88 3.500 1 Foresters at Checkpoints 2.75 11,000 2 Foresters at Intervening Provinces 0.50 2,000 0 Forestry Officials at Final Destination 1.38 5,500 1 Total Extrabudgetary Charge 14.20 56.800 11 Source: ARD report. 2.14 The full rent recovery scenario would generate a net present value return to the budget of US$631 million compared to only US$75 million under the continuation of the current situation(see Table 2.2). It should be noted that revenue enhancement under the full recovery scenario would be gradual, as this scenario assumes that there will be environmentally sustainable harvesting as well as step- by-step implementation of the medium-term recommendations (see para. 2.17). Table 2.2: Different Revenue Potential from Forestry Avg. Annual Average Annual Average Av.nuN Present Value of Avrg nul Aeae Increase as % of . Rent Recovery Government Extrabudgetary Total a994 Ratio (%) Revenue(30 Revenue (Mn. Flows Until 2003 years, Mn/ US$) (Mn. US$) Govemment Revenue 1/ 1: Continuation of Current Situation 9 55 2.3 151 1 2: Full Implementation of Existing System 39 245 18.0 - 9 3: Full Rent Recovery Policy 100 631 82.4 - 41 1/ Share of timber revenues was 17% in 1994 which was an all time high. 2.15 In addition to the proper capture of royalties on forestry exploitation, tax revenue collection from -14- forestry also needs to be strengthened. The prepayment of the tax on profits in 1997 for log exporters was CR 182 million and for sawmill and board factories was CR 709 million. As all companies were declaring a loss for 1997, the prepayment would be transferred to the minimum tax and the total amount of CR 891 million would be the extent of the tax revenue collections from log exporters and sawmill operators in 1997. This amount represents less than 0.3 percent of the total value of exported logs and sawn timber. Timber has escaped taxation through the manipulation of the Law on Investment. In order to avoid taxation, timber companies set up sawmills, which allowed them to export rough sawn timber, and had the logging activity classified as manufacturing and therefore enjoying the provisions of the Law on Investment and, most important, escaping from the ban on log exports. A conservative estimate of the foregone revenue from the tax on profits for both legal and illegal exports for 1997 is US$22 million or CR 66 billion as compared with the actual collection of CR I billion. 2.16 For the tax and non-tax revenue potential to be realized, the government must take the following "emergency measures" immediately to arrest the deteriorating situation: * Stop granting new concessions * Freeze approving all new investments in wood processing * Establish mechanisms to monitor and prevent illegal log felling and exports * Enforce cancellation of all permits for "collection" of logs and stop issuing new permits * Strengthen the Forestry Department, with clear mandates and responsibilities. 2.17 Furthermore, over the medium term the government must implement the recommendations of the four studies it has conducted with the aid of Bank Technical Assistance--which are essentially completed and embody the framework suggested by the Bank/UNDP/FAO and also by government's own forest policy consultants: * Log Control and Monitoring. The current forest law enforcement crisis justifies a special effort to control illegal logging, increase revenue collection, and facilitate movement toward a more efficient log trade regime. A set of protocols for assessing the scope, spatial distribution, and character of illegal logging has been tested and needs to be developed for transfer to the government. * Legal Study. Concession contracts are badly biased against the government, but they do specify performance obligations which are likely to form the basis for a structured and legally defensible challenge to individual contracts on a case-by-case basis. A draft model contract has been prepared, comments on the draft forest law have been submitted, and a training program has been conducted. Forest legislation needs to be resubmitted to the National Assembly, following revision. An earlier draft did not provide an effective platform for forest exploitation, risks perpetuating the current system of arbitrary allocation of forest resources, and does not resolve key institutional issues. * Forest Concession Management. Concessions can provide the basis for sustainable sectoral development provided a satisfactory policy framework is put in place, regulations are strengthened, and security and technical guidelines are enforced. A framework of planning, harvesting, and revenue systems will be provided and, following review and consultation, will need to be formalized through appropriate legislation or regulation and implemented through pilot projects. * Forest Policy. Elements of the government's forest policy are at odds. Pressures for immediate revenues are now swamping the concerns about sustainability; strong measures and commitments will be needed to redirect privately appropriated rent to the public This estimate is obtained by applying the log f.o.b. value US$190/cm and sawn wood f.o.b. value US$310/cm to the total amount of log and sawn timber exports and assuming a net profit rate of 7 percent. -15- treasury; land allocation is chaotic and arbitrary; community involvement in forest development is severely limited; and public statements are inconsistent. 2.18 The government recognizes the importance of decisively implementing the recommended actions of the four studies, and has indicated its willingness to embark on these challenging tasks. To achieve these medium term objectives, the government must take the following actions for the next few months: prepare a proposal for illegal logging control; prepare the illegal logging monitoring program, undertake a performance review of concessions, and prepare the Cambodia Code of Forest Practice (CCPF) which is essentially an operational manual that lays out the rules for concession operations. In the subsequent few months, these actions should be followed by: the preparation of the revised forest law; decisions on the termination or renegotiation of concession contracts; the issuance of a sub-decree establishing the operational manual for concession management; the strengthening of the implementation capacity of forestry management; and the strengthening of law enforcement capacity on illegal logging. F. Proper Taxation of Other Natural Resources 2.19 Natural resources represent a major revenue potential for Cambodia. Cambodia has a natural resource base similar to its ASEAN neighbors and most ASEAN countries generate between 2 and 4 percent of GDP from the taxation of resources, in addition to non-tax revenues that they may generate. Cambodia has yet to effectively utilize this revenue source. Timber is a prime example, as discussed above. Gem mining makes no definable contribution to tax revenues. A portion of the gem exports from Thailand (US$2 billion in 1997) can be attributed to uncut gemstones mined in Cambodia. The potential exists to derive revenue from the exploitation of petroleum resources. Three petroleum contracts held by two contractors were carried into 1997. A fourth agreement was entered into in 1997. This agreement is a conditional arrangement for the exploration of the overlapping claims area with Thailand. If natural gas is found in viable quantities, revenues could be from I to 2 percent of GDP, given the proper contracts such as the 1991 draft production sharing contract and proper negotiation. 2.20 Other potential tax revenue sources continue to evade taxation either through a liberal interpretation of the Law on Investment or through political arrangements (for example, cement production, which is normally classified as a mining activity, has escaped taxation in the same manner as timber). It should be borne in mind that the Implementing Regulations for the Law on Investment include natural resource extraction as a recommended area of investment. If the same broad rules are applied to natural resource extraction as are applied to logging and cement production, the revenue potential from these activities in the medium term will be nonexistent. G. Strengthening Tax Administration/Governance 2.21 Although tax administration has improved considerably, it still remains very weak. The first detrimental factor is the low level of audit capability in the Tax Department. Most audit activity in the Tax Department is confined to desk audits of information supplied by the taxpayer. The only attempts to use information other than that supplied by the taxpayer has been the use of SGS's pre-shipment inspection data on imports. Most staff assigned to the audit function lack the skills to be able to successfully audit the financial records of an international company. The second detrimental factor is insufficient political support and law enforcement to collect taxes. Thus, it is difficult for personnel from the Tax Department to audit, re-assess, or challenge a taxpayer. The Tax Department identified tax arrears of CR 14 billion during 1996-97, but it has collected only 10 percent of the total assessments. As it takes on average 6 to 12 months to collect an assessment, with the aging of total assessments foregone revenue is estimated at US$4 million. -16- 2.22 The Law on Taxation was the first serious attempt to consolidate tax legislation into a code. Legislation passed prior to the Law (such as the tax on turnover, the tax on specific merchandise and services, the tax on rent, and the tax on unused land), however, remains outside the Law on Taxation. The intent was that as these separate tax laws were redone they would be included as separate chapters in the Law on Taxation. In addition, there are several draft laws in circulation that also contain tax provisions, such as the draft commercial code, the draft law on casinos, and the draft law on NGOs. All issues relating to taxation should be included in the Law on Taxation with the Ministry of Economy and Finance responsible for legislative drafts, implementation, and administration. 2.23 One related issue is that a part of the Law on Taxation has not been implemented and that has the potential to seriously erode revenue if it is not implemented with the proper rules and controls. As part of the administrative protest procedure, the final step in the process involves the Committee of Arbitration. Although composition of the Committee is at the discretion of the Minister of Economy and Finance, it appears that the Council of Ministers demands a substantial part of the membership. Constraints must be placed on this Committee so that it does not legitimize the process of appealing to the Prime Minister's Office or the Council of Ministers for special relief or ad hoc exemptions. The Prime Minister's Office and the Council of Ministers must cease to be viewed as the major route for tax relief The authorities need to amend the Law to require that any other exemptions not specified in the Law should be void. 2.24 Regarding customs, the government extended the exemption from pre-shipment inspections (PSI) to imports for the projects approved under the Law on Investment, garments, and cigarettes. To protect the import tax revenue base and to generate the revenues that should be collected at import, the government must place more reliance on the PSI. The PSI should be required on all shipments over some minimal (US$3,000) c.i.f. value. The government should require that the official PSI valuation be used as the customs value for duty and domestic tax purposes. If customs administration had good governance, customs revenue elasticity with respect to GDP should have been around 1.0, as international experience shows. Under this assumption, foregone revenue resulting from weak Customs administration--through under-invoicing and false declaration--would have been 0.4 percent of GDP. There has been some concern expressed by the government about the cost of extending PSI to all imported goods and materials as the government currently pays the cost of PSI (0.75 percent of c.i.f. value). This cost would be more than offset through increased revenues collected at the time of import attributable to proper valuation and to the expanded tax base when the VAT is applied to all imported goods. H. Proper Transfer of Non-tax Revenues from Line Ministries 2.25 Additional revenue can be generated by requiring that all rents and fees collected by some of the line ministries should be deposited in the Treasury. At present, revenues from items such as passport and visa fees, profits from enterprises run by various ministries, and rents accruing from letting out buildings belonging to individual ministries, are collected by the line ministries and are partly spent by them. Although revenue retention is an important incentive for those providing services, the current situation is such that rents and fees are often collected and used in a non-transparent manner. Transferring these revenues to the MEF would increase the revenues by 0.2 percent at a conservative estimate. I. Conclusion 2.26 A substantial potential for revenue enhancement does exist. With the decisive implementation of the recommendations made in this chapter (the core of which is improving governance), the government should be able to substantially increase the revenue-to-GDP ratio from the current 8-9 percent to about 13-14 percent by 2002 (see Chapter 5). Without such implementation, the effectiveness of public expenditure policy will continue to be severely constrained and the transition to a sustainable development wifl not be passible. -17- 3. ANALYSIS OF THE LEVEL AND COMPOSITION OF PUBLIC EXPENDITURES A. Introduction 3.1 As a result of two decades of conflict, the weak physical infrastructure and inadequate capacity of human resources in Cambodia are hindering the optimal use of resources and constraining the capacity to adapt to the changing needs and requirements of a growing economy. The majority of people lack access to potable water and the schools and health clinics are yet to be rebuilt. Rural areas suffer from lack of serviceable roads, irrigation systems, and agricultural extension facilities. Industry and trade are hampered by high transport costs, the low quality and unstable supply of electric power, and poor telecommunications facilities. The effect of conflict has been more devastating on human resources. Almost 35 percent of the total population and 37 percent of the rural population are illiterate and only about 10 percent of the population is engaged in skilled occupation. About a third of the population aged five years and above has had no education, only 20 percent have had schooling beyond the primary level, and only 4 percent have completed lower secondary education. The infant mortality rate, at 90 per 1,000 live births is among the highest among the countries in the region and about 50 percent of children suffer from malnutrition. Moreover, lack of adequate physical and social infrastructure limits the access of the poorer households to public services. Most of the poor households, particularly in rural areas, have lower access to such social and economic services as basic education, health facilities, piped water, roads, and electricity. These factors place severe limitations on labor productivity. 3.2 Thus, an integral element of the developmental strategy in Cambodia consists of rebuilding physical infrastructure, developing human resources, and improving the accessibility of public services to the poor. Public expenditure policy plays an essential role in this task. First, the level of public expenditures and the methods employed to finance these expenditures are important from the viewpoint of macroeconomic stability. In an economy where economic revival requires substantial investments in social and physical infrastructure but where public savings are very low, the government has to pursue a careful path in the trade-off between macroeconomic stability and the large investment needed for reconstruction of physical infrastructure and human resource development. Second, spending on social and economic services directly contributes to poverty alleviation and broad-based economic growth. Provision of economic infrastructure and support services help to extend market economy to rural areas and enhance income-earning opportunities for the poor. Expenditures on social services such as basic education, health, and family planning enhance human capital, raise labor productivity, and contribute to a shared basis of economic growth. Third, the impact of public expenditure policy depends not only on the level of expenditure but also on its allocative and technical efficiency. The pattern of allocation of public expenditure among various activities determines its efficiency and the ability of the economy to deliver social and economic services. Fourth, the assignment of fiscal responsibilities between different levels of government determines the responsiveness and incentives in the provision of public services according to the diversified needs in different regions. The regional distribution of public expenditures determines the accessibility of public services to the poor. Fifth, the institutional mechanisms for formulating and implementing expenditure policy contribute to accountability and incentives, and also determines the efficiency with which expenditures are translated into public services. 3.3 This chapter examines the important allocative and technical efficiency and equity issues pertaining to the level of spending on public services, the role of government and aid agencies in public service provision, and the functional, economic, and regional composition of public expenditures. While this chapter briefly summarizes the main findings of the analysis of the efficiency and equity of public -18- expenditure within sectors (intrasectoral allocation) for the health and education sectors, more in-depth analysis is presented in Annexes 3 and 4. 3.4 Public expenditure allocation needs to be evaluated and selected using a three-step analysis or criteria.' First, it needs to be assessed as to whether there is a rationale for government intervention in general and public expenditures, in particular in the area underlying market failures. Public expenditures should be concentrated first on goods and services that the private market will not provide or will not provide sufficiently, rather than merely substituting for, or even marginally improving upon, the private market outcome. Second, if there is an underlying market failure, it is necessary to assess how large a discrepancy between social and private values this imposes and therefore to what extent alternative expenditure allocations can improve on the private market. In other words, the social cost-benefits of alternative expenditures need to be computed in order to select those that maximize the net contribution to social welfare. Third, while the cost-benefit analysis would help meet the criteria for economic efficiency, the impact of alternative expenditure allocations on the poor needs to be taken into account to ensure that the composition of public spending helps meetpoverty alleviation objectives. This Chapter and Annexes 3 and 4 apply this framework to the extent feasible. 3.5 Consolidated Database. In Cambodia, various donors and NGOs provide assistance of a public expenditure nature: they significantly supplement the government's budgetary expenditure, especially in the provinces. It is, however, difficult to obtain a reliable picture of the total expenditure of a public nature on the various sectors in the Cambodian economy because a significant portion of aid-financed expenditures are not included in the government's official budget. This omission of a major part of the extrabudgetary aid flows results in a distortion of the disaggregated breakdowns by function, economic classification, and region. Thus, with a view to presenting a more complete picture, in collaboration with the government, the PER constructed a consolidated database covering from 1994 (when a budgeting in a proper format started) to 1997, combining the budget data at the Ministry of Economy and Finance (MEF) and the donor aid database at the Council for the Development of Cambodia (CDC) in a consistent manner by function, by economic classification, and by region. As the 1997 data are still preliminary and also of an aberrant nature (as a result of the suspension of a substantial part of donor assistance), the analysis mainly focuses on the 1994-96 data. 3.6 Three Concepts of Public Spending. As a result, three different concepts of spending on public services are evaluated in this chapter (see Figure 3.1). All three concepts include the budgets of provincial governments but exclude the budgets of public sector enterprises. The first concept is public expenditures financed from domestic revenues raised by the government through its tax and non-tax revenue sources (about 8.4 percent of GDP in 1996). These are not readily available and have to be estimated by deducting the donor assistance from the official government budget. A part of the donor assistance, however, is given for budget support, and while it is possible to exclude this part to arrive at expenditures financed from domestic revenue sources on the overall level, it is not possible to separate this part from sectoral expenditure details. Thus, in analyzing sectoral allocations, public expenditures financed from domestic revenues inclusive of budget support is used. 3.7 The second concept is expenditures captured in the government's official budget (about 17 percent of GDP in 1996). These include all expenditures financed from domestic revenues by the government as well as a part of the expenditures financed from donors and NGOs. The coverage of the expenditures financed from donors and NGOs is not complete because most of the aid-financed projects Sanjay Pradhan, "Evaluating Public Spending -- A Framework for Public Expenditure Reviews," World Bank Discussion Paper 323, May 1996. -19- are contracted and executed outside of the ministries. The MEF has tried to collect information separately on these projects and to include them in the budget. Although the coverage has become more comprehensive over the years, it remains incomplete.2 Figure 3.1: Three Concepts of Public Expenditures A: First concept A+B: Second concept A+B+C: Third concept C A: Domestically-financed expenditures B: ODA & NGO financed expenditures captured in government budget OC. ODA & NGO financed expenditures not captured in government budget 3.8 The third concept is the total spending on public services (about 25 percent of GDP in 1996). This is obtained by combining public expenditures financed from domestic revenues from the government budget with expenditures financed from ODA and NGOs compiled by the Council for the Development of Cambodia (CDC) in a consistent manner. This consolidated database gives a comprehensive picture of spending on public services without double counting. A significant portion of external assistance flows outside the budgetary process and is implemented directly by the donors or through the NGOs. Although such assistance is not part of the government budget, it is nevertheless used to augment public services. A comprehensive picture of spending on public services, therefore, should include the budgetary expenditures as well as donor financed expenditures on public services even when they are incurred outside the budgetary process. The consolidated database also helps to compare the relative roles of the government, donors, and NGOs in providing public expenditures in specific sectors. B. Level of Spending on Public Services 3.9 Table 3.1 presents expenditures on public services for 1996, according to the three concepts discussed above. Expenditures financed from the government's own revenues formed only 8.4 percent of GDP (and if budget support is included, 10.3 percent). Expenditure estimates included in the official budget were 17.2 percent of GDP. On the other hand, aggregate expenditures on public services from all sources constituted 24.8 percent of GDP. This means that expenditures financed from the government's own revenues contributed only about one-third of total spending on public services. It also implies that about one-third of total spending on public services would be lost if the analysis is limited to the official government budget. 3.10 Total spending on public services from all sources as a share of GDP was more or less constant at about one-quarter during 1994-96 as seen in Table 3.2. (A preliminary estimate for 1997 shows that it dropped to 22 percent owing mainly to the partial suspension of external assistance.) Given the low level of per capita GDP in the country, the level of per capita spending on public services in 1996 was only CR 191,000, or about US$73. 2 Among other areas, donor assistance in the form of technical assistance and assistance through NGOs is not included, although it is rather substantial, especially in the social sectors. -20- Table 3.1: Alternative Concepts of Expenditures on Public Services, 1996 (in percent of GDP) Concept I a/ Concept 2 Concept 3 1. Administrative Services 1.9 2.9 3.4 2. Defense and Security 4.9 4.9 4.9 (a) Defense 3.5 3.5 3.5 (b) National Security 1.4 1.3 1.4 3. Social Services 2.3 3.9 6.7 (a) Education 1.0 1.7 2.7 (b) Health 0.5 1.4 2.3 (c) Social Action, Veterans' Affairs 0.5 0.5 1.3 (d) Other 0.2 0.3 0.3 4. Economic Services 1.2 5.3 9.8 (a) Agriculture 0.2 1.7 2.4 (b) Rural Development 0.1 0.2 3.1 (c) Transport and Civil Aviation 0.3 1.8 2.3 (d) Water Supply and Sanitation b/ .. .. 0.3 (e) Communications 0.4 09 1.0 (f) Industry, Energy, and Mining 0.2 0.5 0 6 (g) Other 0.1 0.1 0.1 Total Expenditures 10.3 17.2 24.8 a/ Includes external budget support (1.9 percent of GDP) as this part cannot be separated into sectors. b/ Water supply and sanitation in concepts I and 2 are included under Industries. Energy. and Mining. Source Staff estimates based on government data. Table 3.2: Total Spending on Public Services from All Sources, 1994-97 (in percent of GDP) 1994 1995 a/ 1996 1997 b/ 1. Administrative Services 3.0 3.0 3.4 2.8 2. Defense and Security 6.8 6.1 4.9 4.8 (a) Defense 5.4 4.5 3.5 3.3 (b) National Security I 4 1.5 1.4 1.5 3. Social Services 6.5 5.6 6.7 7.2 (a) Education 2.3 3 2 2.7 3.2 (b) Health 1.4 1.1 2.3 2.0 (c) Social Action 2.2 0.9 1.3 1.6 (d) Other 0.6 0.5 0.3 0.4 4. Economic Services 7.8 9.4 9.8 7.3 (a) Agriculture 1.2 1.3 2.4 1.0 (b) Rural Development 1.2 2.5 3.1 2.6 (c) Transport (including air) 3.1 3.0 2.3 1.6 (d) Water Supply, Sanitation 0.8 0.9 0.3 0.0 (e) Communication 0.3 0.3 1.0 1.1 (f) Industry. Energy. & Mining 1.1 1.4 0.6 0.6 (g) Other 0.1 0.1 0.1 0.3 Total 24.1 25.3 24.8 22.1 a/ Numbers do not add up to total as NGO contribution (1.2% of GDP) cannot be broken into sectors. b/ Provisional. Source: Staff estimates based on government data. -21- 3.11 Table 3.3 compares the public expenditure/GDP Table 3.3: Public Expenditure in Cambodia: ratio in Cambodia with East Asian economies as well as An International Perspective with other regional averages for developing countries. In Country Public Expenditure principle, the relevant expenditure concept to be analyzed as%ofGDP would be budgetary expenditures (Concept 2). In 1996 it Cambodia (1996) was much lower than the average expenditure levels Concept 1 prevailing in East Asian economies and most of the Concept 2 17.2 regional averages. In fact, the budgetary expenditure level Concept 3 in Cambodia was much lower than that in neighboring Vietnam (1996) 24.7 Vietnam and Lao PDR with a similar level of development. East Asia a! 22.5 When the total spending on public services (Concept 3) is South Asia a/ 21.3 compared, Cambodia's level would be in line with the Africa a/ 23.2 levels in other countries. But this comparison is based on a Latin America a/ 17.3 rather doubtful assumption that the official budgets of most All Developing Countries a/ 24.4 other countries largely capture externally financed capital a/ Average of 1985-1989. expenditures. Even if this is the case, the problem remains Source: Staff estimates based on government data. with too much dependence on externally financed expenditures, which is not sustainable over the longer term. C. Role of Government and Aid Agencies in Public Service Provision 3.12 From the viewpoint of expenditure maneuverability, the sources of financing expenditures are important. Figure 3.2 presents spending on public services financed from different sources during 1994- 97. Financing from government's own revenue sources formed only about 42 percent in 1994 and steadily declined to 34 percent in 1996 (although it jumped back to the 1994 level in 1997 owing to the partial suspension of donor assistance). Expenditures financed by NGOs (from their own budgets, excluding donor assistance provided through them) were around 5 percent. Thus, a major proportion of financing for public services (almost three-fifths) was from external donor assistance. Figure 3.2: Sources of Financing Public Services, C994-97 70 6 60 5556 52 Conep 22 1721 Conp 3 40 7 10 ~ , 1994 1995 1996 1997 UVGovt. Financed DODA ENGa 3.13 Table 3.4 shows the shares of government's own revenues, ODA, and NGOs in financing different public services as ratios of both GDP and total expenditures in 1996. The table shows the overwhelming dependence on external assistance to finance almost all social and economic services. Government financing of social services was about 34 percent, and for economic services it was a mere -22- 12 percent. The ODA contribution amounted to 53 percent of expenditures on social services, and its share in financing economic services was 83 percent. The NGO contribution was relatively pronounced in the social sectors. It is noteworthy that while expenditures on economic services financed externally dropped sharply to 5.8 percent of GDP from 8.6 percent of GDP in 1996. expenditures on the social sectors financed externally increased to 5.0 percent of GDP in 1997 from 4.4 percent in 1996. This reflects the fact that donors suspended their assistance mainly in the economic sectors in the aftermath of the July 1997 events, but continued and strengthened their assistance in the humanitarian/social sectors. 3.14 The predominance of donor financing of public services limits the ability of the government to determine expenditure allocations according to its own priorities. The issue of sources of financing, however, is immaterial if resources are fungible. While there exist some empirical evidence in other countries that aid may be partially fungible, it is difficult to assess the extent of fungibility in Cambodia. Table 3.4: Composition of Spending on Public Services in Cambodia, 1996 Percent of GDP Percent of Each Sector Total Expenditure Govt ODA NGO Total Govt ODA NGO Total Financed a/ Financed a/ 1. Administrative Services 1.9 1.4 0.0 3.4 57.0 42.1 1.0 100.0 2. Defense and Security 4.9 0.0 0.0 4.9 99.7 0.3 0.0 100.0 (a) Defense 3.5 0.0 0.0 3.5 100.0 0.0 0.0 100.0 (b) National Security 1.4 0.0 0.0 1.4 98.8 1.2 0.0 100.0 3. Social Services 2.3 3.5 0.9 6.7 33.9 52.6 13.5 100.0 (a) Education 1.0 1.3 0.4 2.7 36.4 48.3 15.3 100.0 (b) Health 0.5 1.5 0.3 2.3 22.5 64.8 12.7 100.0 (c) Social Action 0.5 0.6 0.2 1.3 42.5 45.0 12.5 100.0 (d) Other 0.2 0.1 0.0 0.3 59.2 40.8 0.0 100.0 4. Economic Services 1.2 8.2 0.4 9.8 12.3 83.4 4.3 100.0 (a) Agriculture 0.2 2.0 0.2 2.4 9.4 82.9 7.7 100.0 (b) Rural Development 0.1 2.8 0.2 3.1 1.9 90.2 7.8 100.0 (c) Transport (including air) 0.3 1.9 0.0 2.3 14.8 85.2 0.0 100.0 1 (d) Water Supply and Sanitation b/ .. 0.3 0.0 0.3 .. 100.0 0.0 100.0 (e) Communication 0.4 0.7 0.0 1.0 35.4 64.6 0.0 100.0 (f) Industry, Energy, and Mining 0.2 0.4 0.0 0.6 26.0 74.0 0.0 100.0 (g) Other 0.1 0.1 0.0 0.1 50.4 49.6 0.0 100.0 Total 10.3 13.1 1.4 24.8 41.6 53.0 5.5 100.0 a/ Includes external budget support (1.9 percent of GDP) as it cannot be allocated among sectors. b/ Water Supply and Sanitation in government budget are included in Industry. Energy and Mining. Source: Staff estimates based on government data. D. Functional Composition of Public Spending 3.15 Severely Compressed Development Expenditures. The most conspicuous feature of the functional composition of expenditures incurred from government's own revenues is that outlays on defense and security have crowded out resources for spending on development activities. (It should be noted, however, that defense and security expenditures included some spending of social and economic nature such as allocations for the families of defected Khmer Rouge soldiers and infrastructure rehabilitation for them.) This has been due largely to the need to fight against Khmer Rouge insurgencies. It has been due also to the competing military buildup by the two coalition partners, which eventually led to the armed conflict in July 1997. In 1996, for example, the preemption of almost a half -23- of government revenues for defense and security left the government with only about one-third for spending on the reconstruction of physical and social infrastructure (Table 3.5). Table 3.5: Economic and Functional Classification of Public Expenditures Financed from All Sources, 1996 (in percent of GDP) Financed from Government Revenue a/ Non-wage Subsidies & Locally Counterpart Externally Wages O&M eS ses Financed Cuntrn Subtotal Financed Total OM Transfers Funding Investment Investment 1. Administrative Services 0.4 0.8 0.4 0.3 0.0 1.9 1.5 3.4 2. Defense and Security 2.7 2.2 0.1 0.0 0.0 4.9 0.0 4.9 (a) Defense 22 1 3 0.1 00 00 35 00 3.5 (b) National Security 0 5 0 9 0.0 0 0 00 1.4 0.0 1.4 3. Social Services 0.9 0.8 0.5 0.1 0.0 2.3 4.4 6.7 (a) Education 0.7 02 0.0 00 0.0 1 0 1 7 2.7 (b) Health 0.1 04 00 00 00 0.5 1 8 2.3 (c) Social Action 0.0 00 05 00 0.0 0.5 07 1 3 (d) Other 0.1 01 00 0.1 00 0.2 0.1 0.3 4. Economic Services 0.2 0.5 0.1 0.3 0.1 1.2 8.6 9.8 (a) Agriculture 0 1 0 1 00 0.0 00 02 2.2 24 (b) Rural Development 0.0 0.0 00 0.0 00 01 3.0 3 1 (c) Transport and Civil Aviation 0 1 0.0 00 0.2 0.0 0 3 1.9 2 3 (d) Water Supply and Sanitation b/ . . . . . 0.3 0 3 (e) Communication 00 03 00 00 0.0 04 0.7 1.0 (f) Industry, Energy and Mining 00 00 0.1 00 00 02 0.4 06 (g) Other 0.0 0.0 0.0 00 00 0 1 0.1 0 1 Total 4.2 4.3 1.1 0.6 0.1 10.3 14.5 24.8 a/ Includes external budget supports (1.9 percent of GDP). b/ Water Supply and Sanitation in government budget are included in Industry. Energy and Mining. Source: Staff estimates based on government data. 3.16 The squeeze on development expenditures is the combined result of revenue shortfalls and an overrun in defense and security expenditures. Even when there was a significant shortfall in actual collection of revenues from the budget targets, the actual outlay on defense and security consistently exceeded the budget estimates during 1994-97. For example, even actual revenues from domestic sources were higher than the budget target by about 20 percent in 1994, and actual expenditures on social services and economic services were less than the budgeted estimates by about 9 percent and 13 percent, respectively. In contrast, the defense and security sectors received 64 percent more than the original budget allocation. From 1995 to 1997, domestic revenue collection fell short of the budget targets. Nevertheless, expenditures on defense and security continued to receive significantly more than the budgeted allocation at the expense of those sectors that would have a greater impact on poverty reduction, such as education, health, agriculture, and rural development. This highlights three critical weaknesses in budget management: (i) unreliability of the macro framework (revenue shortfall); (ii) the lack of legitimacy of the budget (reallocation during budget implementation); and (iii) the residual nature of expenditures on social and economic services (see Chapter 4). 3.17 In Cambodia, as elsewhere, the poor do not have adequate access to land or capital, and investment in human resources is a critical policy component of the development strategy to accelerate growth and reduce poverty. The development experiences of successful countries in the region underline the importance of investment in education and health facilities in improving labor productivity and increasing a shared basis of economic growth. However, the outlays on human resource development in Cambodia from the government's own resources have been extremely low. In 1996, for example, expenditures on education and health from government resources constituted only about 1 percent and 0.5 -24- percent of GDP, respectively as seen in Table 3.5. What is more striking is that the primary education outlay financed from government's own resources showed a decline during 1994-96. Similarly, for the health sector, expenditure incurred on primary health care was very low at 0.5 percent of GDP. Unless the authorities fully implement the revenue mobilization and expenditure rationalization measures recommended in the PER, reaching the SEDP targets of 3 percent of GDP for education and 2 percent for health (financed from government's own resources) by the year 2000 would not be realized. Even when the total (including externally aided) expenditures on public services are considered, the education and health sectors secured only 5 percent of GDP or about one-fifth of total expenditures incurred. The current total public expenditure for health and education is inadequate in view of the massive human resource development needs. Similarly, the expenditure available for physical infrastructure (transport, communications, and industry) was squeezed (less than 4 percent of GDP). In particular, in a country in which 85 percent of the population and 88 percent of the poor are concentrated in rural areas, the agriculture and rural development sectors received only 22 percent of the total expenditure allocation. It is of critical importance to allocate adequate expenditures to irrigation, rural roads, and agricultural extension service to help extend a market economy to rural areas and enhance income earning opportunities for the poor. 3.18 Military Demobilization. To make expenditure reorientation toward development spending possible, it is critical to significantly reduce military expenditures. Defense expenditure during the period 1994-97 varied from 3.3 percent to 5 percent of GDP, and this level of spending compares unfavorably with the average of 2 percent in 1990 for the East Asian countries and 2.3 percent for the developing countries taken together. The government urgently needs to reduce and rationalize military expenditure to make savings for development expenditure. In this regard, decisive implementation of military demobilization--which has been put on hold owing mainly to the political events--will be essential. In 1995-96, the government prepared the comprehensive Cambodia Veterans Assistance Program (CVAP) which also foresaw the reintegration into civilian life of several thousand Khmer Rouge defectors. Owing to the political turbulence in July 1997, the CVAP has yet to be implemented. Given the markedly improved political and security situation, prospects seem to be bright. With a view to highlighting the importance of reducing military expenditures, a "peace dividend"'analysis of the military demobilization program is described in Annex 2. According to the analysis, with the effective implementation of the demobilization program (demobilizing 43,000 soldiers over the next three years), the share of defense expenditure in GDP could be reduced from 3.3 percent to 1.9 percent by the year 2002. Toward this end, the government needs to take the following preparatory steps for next few month: (i) the updating of the CVAP (especially in view of Khmer Rouge defectors); (ii) the identification and registration of special target groups; and (iii) the pilot demobilization of special target groups. E. Intrasectoral Allocation of Public Spending: Health and Education Sectors 3.19 To better understand the appropriateness of sectoral strategies and expenditure priorities, an analysis of the efficiency and equity of public expenditure allocation within sectors (intrasectoral allocation) is needed. Among various sectors, Annexes 3 and 4 provide an in-depth analysis of the health and education sectors in view of their direct relevance to poverty reduction strategy through human resource development, which is urgently needed to overcome the dissipated human resource situation. Each sector is analyzed in the areas of: (i) sector overview; (ii) government policies; (iii) funding mechanisms and cost recovery; (iv) expenditure trends; (v) efficiency and equity of public spending; (vi) evidence of expenditure leakage/ineffective targeting; (vii) scope for private sector participation; (viii) A number of issues--externalities, public/private mix, private versus social benefits, etc--need to be considered before the appropriateness of the target numbers can be assessed. -25- institutional constraints; (ix) evidence of the impact of public expenditure; (x) evaluation of the PIP; and (xi) recommendations. Emphasis is given to the marked differences in several aspects of budgetary management in these two social sectors, especially in cost recovery strategy. A brief summary of the main findings is presented below. 1. Health Sector 3.20 Funding Mechanisms. In 1996, households financed an overwhelming 82 percent of all health expenditure (i.e., both public and private) in the country. Donors were the next important funders, contributing 11.5 percent of total health spending. The MoH health budget accounted for merely 4 percent of total health expenditure, and NGOs accounted for the remaining 2.4 percent. There are two important points here: first, Cambodian households are paying far more than they can afford for health care (US$33 per capita per annum), and, second, government-financed health spending is woefully inadequate (US$7 per capita per annum). Thus, there is a need to increase government spending on health significantly, particularly through greater cost recovery. Until recently, there were no official fees at any government health facilities, and, in principle, health services were available without charge to everyone. While cost recovery is an established policy of MoH, cost recovery is not widespread, and is currently limited to a few selected national and central hospitals, most of which are in Phnom Penh. The MoH is piloting programs of cost recovery, including user charges, community-based prepaid insurance schemes, and the contracting out of specific services in public facilities to the private sector, on a trial basis. Household survey data from the Cambodia Socioeconomic Survey (CSES) 1997 indicate that there is substantial scope for increased cost recovery via user fees for health services. It is important to ensure that the user fee collection process be transparent and that the collected revenues be kept at the facility level to improve the quality of services. 3.21 Expenditure Trends. In 1996, there was a major increase in non-private spending on health, more than doubling from 1995. This increase in public spending means that in 1996 total spending on health in Cambodia was approximately US$41 per capita. Total health spending (by households, government, and donors) constituted about 13.2 percent of GDP in 1996, while total government- financed health spending constituted only 0.5 percent of GDP. The MoH budget indicates that operations and maintenance (O&M), including drug supply, accounts for a major part of the budget. Indeed, the proportion of the budget allocated to O&M increased from 63.5 percent in 1995 to 76.8 percent in 1996. These are desirable trends. 3.22 Efficiency and Equity of Public Spending. Both the efficiency and equity of total public spending on health appear to have improved in that the share of spending on tertiary curative care (hospitals) has fallen, while that on primary health care has increased (albeit much of it owing to changes in donor spending). First, during 1995-96 the share of hospitals in government health expenditures fell to a mere 8.2 percent from 25 percent, while that of primary health care increased to 26.6 percent from 16.1 percent.4 This is a favorable trend because of its implications for both efficiency and equity. To the extent that secondary and tertiary based facilities are disproportionately used by the better-off, while primary care facilities such as health centers are used mostly by the poor, government health policy needs 4 Because of the inability to separate primary from other levels of care and the pooling of government expenditures on all hospitals, these data have to be treated with caution. First referrals to district or provincial hospitals are very much part of an integrated primary care system. Hence, only national or central hospitals, which are typically based in Phnom Penh, can be considered as offering a tertiary level of care. The referral system functions poorly in Cambodia, and roughly more than one-half of the expenditure on hospitals is on hospital administration and on provincial and central hospitals. -26- to shift its focus from secondary and tertiary care to primary care. Second, the share of drug supply in total public health spending also increased dramatically--from a virtually negligible level in 1994 to 30.8 percent in 1996. This, too, is a desirable trend. In contrast, the provincial distribution of per capita public health spending appears to be inequitable, with richer provinces having higher levels of government health expenditures per capita than poorer provinces. This is not due to a bias in the formula for allocating public health spending across provinces, but instead is the relative ease of implementing health projects in more affluent provinces that have higher capacity and better infrastructure, facilities, and administrative resources. In addition, the inequality in health spending per capita may occur because the scarcity of local resources in poor provinces makes reallocation from health spending more likely. Whatever the reasons, the inequitable distribution of health spending across provinces needs to be reversed. 3.23 Impact of Government/Donfr Spending. There is evidence that donor assistance and government health expenditures, especially on communicable disease control programs such as the Expanded Program of Immunization (EPI) and malaria control, have been effective not only in increasing the utilization of health services (e.g., immunization coverage) but also in reducing the morbidity and mortality rates. In addition, provincial data indicate a strong positive correlation between public health spending per capita and reported morbidity rates, suggesting that higher levels of government health expenditure per capita in a province do indeed contribute to lower levels of reported morbidity. 3.24 Scope for Private Sector Participation. Compared with other sectors, private sector participation in health services delivery is widespread. National household survey data for 1997 show that nearly two- thirds of all individuals reporting an illness visited non-government (private, donor, NGO) facilities for their outpatient care. Since the private sector (other than pharmacies) caters disproportionately to the more affluent groups in Cambodian society, mainly in the urban areas, this implies that government health facilities need to target their services increasingly to the poor individuals and underserved provinces. 3.25 Institutional Constraints. There is impressionistic evidence that, until recently, it was common for a provincial health department to receive authorization from the provincial governor for only 70-80 percent of the budgeted amount. Recently, the MoH and the MEF designed and released a new cash release process called the Accelerated District Development Program (ADD) for 22 districts (see Chapter 4). By reducing the number of administrative layers in the budget disbursement system, the ADD program is expected to speed up disbursements and reduce the discrepancy between budgeted and actual health expenditures at the district !evel. Household survey data indicate that the level of utilization of health services at health centers and referral hospitals is significantly higher in ADD than in non-ADD districts. Expanding of the coverage of the ADD system will raise the efficiency of health expenditures. 2. Education Sector 3.26 Funding Mechanisms. In 1996 nearly 46 percent of all education expenditure (i.e., both private and public) in the country was funded by international donors and NGOs. The MoEYS and the private household sector contributed about 27 percent each. Officially, schooling is free all the way from the pre- primary to the university level in Cambodia. Household survey data from the CSES 1997 indicate that there is substantial scope for increased cost recovery in the education sector through charging official school fees. Unlike the MoH, however, the MoEYS has yet to establish a cost recovery policy to benefit from regularizing or formalizing the currently widespread unofficial fees and contributions. The MoEYS needs to consider introducing user fees especially in higher education institutions, with transparent collection mechanism at the facility level. -27- 3.27 Expenditure Trends. Total public spending on education increased nearly 60 percent from 1994 to 1996. The increase in public spending meant that in 1996 total national spending on education (i.e., including spending by households) amounted to about 3.7 percent of GDP. In turn, total public spending constituted 2.7 percent of GDP. If public spending financed from government is considered, however, it was only 0.9 percent of GDP. Comparative data suggest that Cambodia's spending on education financed from its own source in relation to its GNP is very low in comparison with other countries in the region. 3.28 Efficiency and Equity of Public Spending. Two aspects stand out in the current allocation of public education spending in Cambodia. First, it is too heavily concentrated in wages. The MoEYS budget indicates that wages (mostly teacher salaries) account for nearly three-quarters. The dominance of teacher salaries in the recurrent education budget means that complementary inputs, such as textbooks are underfunded. And communities and households are held responsible for providing these complementary inputs to students. This means that communities and households, especially those at the bottom of the income distribution, cannot afford to provide books and other learning materials to students. Inadequate provision of textbooks and other learning materials reduces the effectiveness of teachers. To raise the effectiveness of public spending on education, the allocation to non-teacher recurrent inputs will need to be raised substantially. In turn, as teacher salaries are so low, the MoEYS is under enormous pressure from teachers to raise their wages and has raised teacher salaries in 1994 and 1998. An unfortunate feature of both pay raises is that they have been determined on a single rate for all teachers. It would have been more prudent, and certainly more efficient, if the pay raises had been given out as incentive payments for better training and performance, longer teaching hours, and service in the rural areas. 3.29 Second, tertiary and technical education (including language, technical, and management training) takes up a disproportionately large share of the total public spending (i.e., including donor aid) on education. While tertiary and technical students account for merely 0.5 percent of all students in Cambodia, public spending on tertiary and technical education took up 30 percent of total public spending on education in 1996. In contrast, primary education, which accounts for 85 percent of all students in the country, received only 40 percent of the total allocation. At this stage of development, such an allocation is not only inequitable but also inefficient. Shifting of expenditures from tertiary and technical education to primary and lower secondary education will increase the efficiency and equity of education spending.5 3.30 Impact of Government/Donor Expenditure. Between 1994 and 1997, student enrollments grew by over 15 percent as public expenditure on education increased by 35 percent. While the growth in enrollments could have occurred because of other factors such as income growth, it is likely that increased public spending on education has had a positive impact on school enrollments. Another evidence is that provinces that have higher levels of public spending on education per capita tend to have significantly higher gross enrollment ratios. In addition, various donor-driven initiatives have made significant contributions. It could be argued that R&D from higher education would have big externalities and that there would be a capital market constraint (market failures) since higher education is expensive and human capital is hard to borrow against. But Cambodia is still far away from any basic and applied research and almost all of the people obtaining higher education come from the richest income quintile who would not be credit-constrained. 6 One such example is UNICEF's cluster school projects. The analysis of statistics from three sample clusters indicate significant improvements in average cluster enrollment (from 73 percent to 91 percent in rural cluster), repetition (from 38 percent to 20 percent), and dropout rates (from 8 percent to 4 percent) during 1993/94 - 1996/97. -28- 3.31 Scope for Private Sector Participation. Although there has been some growth in the private sector in education, the private sector's role in the provision of education is still relatively small. The CSES data indicate that overall less than 1 percent of students are enrolled in non-government schools. The only level at which the private sector accounts for more than a trivial share of students is the tertiary level (20 percent). Data show a generally strong positive relationship between private schooling and per capita expenditure. Thus, although the proportion of students enrolled in private schools is very small at this time, it can be expected to grow very significantly with economic growth and rising incomes in Cambodia. As there exists considerable scope for growth in the private provision of education, the government should look for ways to form partnerships with the private sector in order to maximize the amount of high quality education to the people. To this end, the government needs to focus spending more of its own scarce resources on the provision of basic education and lower secondary education. At the same time, it needs to put in place policies which encourage private sector involvement in upper secondary, tertiary, and technical education. Such policies need to include a regulatory and quality control framework, public information systems, and policies which ensure access for the poor. 3.32 Institutional Constraints. One unique feature of the budgeting process is that numerous capital expenditures are not reflected in the official MoEYS budget. There were large capital outlays on the construction and renovation of schools from several sources as well as smaller school construction schemes provided by donors (international and national, political and private), expatriate Khmers, and collective community efforts. While these school construction and renovation efforts are commendable, enough thought may not be paid to their long-term recurrent cost implications and obligations, which necessarily fall on the MoEYS. The MoEYS needs to exercise formal powers of approval over capital expenditures in education in the context of careful school mapping and planning exercises. F. Economic Composition of Public Spending 3.33 Government revenues barely covered the current expenditures; they could provide hardly any significant counterpart funding to externally aided projects and their contribution to locally financed investment was negligible. Thus, almost all capital expenditures had to be financed from external assistance. Within current expenditures, the government's own resources were hardly adequate for salaries and maintenance expenditures for the social and economic services. 3.34 Salaries. Salary expenditures in 1996 constituted 4.2 percent of GDP or about 43 percent of current expenditures, which is lower than the average for the developing countries taken together at 6.1 percent of GDP or 52 percent of current expenditures. This is mainly due to the very low salary levels even with substantial over-employment. In the 1997 budget, about 3.2 percent of the population or about 358,000 persons were employed in permanent government jobs (including defense and security). It is noteworthy that there were no systematic differences in salaries in different employee categories and that various categories received broadly similar salary levels, as items such as family and children's allowances, or special pay for extra hours of services to lower categories of employees, appear to offset the differences in basic pay scales. Thus, all employee categories, irrespective of seniority or skills, earn less than one dollar per day. The emoluments rece:ved by the defense forces are somewhat higher. 3.35 Low salary levels tend to reduce productivity levels in the public sector. As the differential in wages between public and private sectors is large, persons with better quality and skill levels tend to take up employment with the private sector. Low salary levels also encourage a high degree of absenteeism on the one hand and of rent seeking on the other. Many government employees, particularly at lower levels of employment, are known to take second jobs or to run petty businesses to earn extra incomes. These practices tend to adversely affect the standards of services; in particular, absenteeism among school teachers and health sector employees tends to lower the quality of education and health services. -29- 3.36 Civil Service Reform. In this regard, effective implementation of public sector reform including civil service reform as envisaged is critical. The government has prepared a comprehensive civil service renewal program, but progress has been stalled owing to the lack of consensus and political uncertainty. A component of the reforms was to reduce the size of the civil service by 20 percent. Even a 20 percent reduction is unlikely to save much money, as the rationalization of the pay structure and the payment of reasonable salaries is likely to absorb the money saved from civil service reduction. While there may not be much scope for expenditure reduction, the scope for rationalizing expenditures to make them effective, efficient, and productive is immense (i.e., significantly improving the efficiency and effectiveness of programs and service delivery, see Chapter 4). 3.37 Toward this end, it is imperative that work on civil service reforms accelerates, with emphasis on institutional change for greater effectiveness rather than a narrow focus on employment and pay. First, efforts are needed to strictly limit further hiring and reduce the number of staff through eliminating ghost workers and through normal attrition. Second, the civil service census must be carried out at the earliest opportunity; this will provide a complete record of numbers and grades, education attainments, experience, and skills which will facilitate personnel planning. Third, work must accelerate on the preparation of options for government and civil service reform. These options should contain analyses of and recommendations for the appropriate role of government, the appropriate structures for that role, the connections between financial and personnel functions, the needed skills mix for the reformed government, and the pay, training, and recruitment policies to support that mix. For this work to progress expeditiously, a strong commitment to civil service reform will need to be expressed at the highest levels of government. 3.38 Non-wage O&M. As seen in Table 3.6. expenditures for civilian non-wage O&M were squeezed in the course of budget implementation--an average of 23 percent shortfall as compared to the budgeted allocation during 1994-97--owing to revenue shortfalls and overruns in defense non-wage O&M. Non- wage O&M for social and economic services amounted to a mere 1.3 percent of GDP (as compared to 2.2 percent of GDP for defense and security). Underfunding of non-wage O&M has resulted in the underuse, or inefficient use, of capital investment and in ineffective service delivery. It is, thus, of utmost priority to allocate adequate amount of non-wage O&M to ensure the efficiency of capital investment and the quality of service delivery. 3.39 Externally Financed Investment. The government adopted Cambodia's Five Year SEDP (1996- 2000) which sets out the development objectives and strategies for the medium term and makes an indicative allocation of investment patterns in the economy. The primary concern of the Plan's development strategy is to reduce poverty by developing human resources in the country and to improve rural infrastructure. To achieve these objectives, the Plan projects a public investment of US$2.2 billion during 1996-2000, of which 65 percent is to be channeled into rural areas. Although some progress has been made to link the Plan more closely with the investment allocations included in the annual budget through the PIP process, the linkage between the medium term Plan and the annual budget remains weak (see Chapter 4). Table 3.7 examines the extent to which the level and composition of public investment envisaged in the SEDP have materialized. The actual public investment (donor-funded) in 1996 was in line with the average annual investment level in the SEDP. In terms of composition, the actual public investment focused more on agriculture and rural development (34 percent) than did the SEDP's indicative allocations (10 percent). In contrast, public expenditure financed from government revenues during 1996-97 (which is essential to the effectiveness of public investment since it provides O&M support) allocated only 5 percent to agriculture and rural development, which raises concerns as to how the target can be met of allocating 65 percent of public investment benefits into rural areas where 88 percent of the poor live. -30- Table 3.6: Outcome vs. Budgeted Non-wage O&M Expenditures (outcome as % of budget) 1994 1995 1996 1997 a/ Defense and Security 204 177 123 132 Defense 278 210 123 166 National Security 109 135 122 79 Administrative Services 86 102 91 88 Social Services 69 94 81 42 Education 84 155 99 33 Health 61 66 72 46 Social Action 115 112 86 75 Other 88 118 83 41 Economic Services 73 55 88 57 Agriculture 90 102 81 39 Rural Development 132 90 77 46 Transport (incl. Civil Aviation) 86 77 71 129 Communications 63 35 96 63 Industry. Energy. and Mining 85 92 72 49 Other 89 93 80 24 Total Civilian Expenditures 77 84 86 61 Total 126 127 102 91 a/ Provisional. Source Staff estrmates based on government data Table 3.7: Actual Spending Pattern and Plan Allocation Category SEDP 1996-1997 PIP 1998-2000 Donor Funded Govemment Financed Investment in 1996 Expenditure in 1996-97 Avg. Annual Avg. Annual ODA & Annual Percent Percent Percent Preto Investment of Total Investment NGO Avg. (Mn. Pcto ofTtlof Total of Total Total (Mn. US$) (Mn. US$) (Mn US$) US$) Agriculture and Rural Development 44 1o 56 12 151 34 9 5 Manufacturing and Mining 18 4 11 2 3 1 5 3 Transport and Communications 101 23 110 24 82 19 26 a/ 14 a/ Electricity 35 8 33 7 13 3 Water Supply and Sanitation 35 8 19 4 11 2 Education and Training 48 11 48 10 55 12 30 16 Health 44 10 80 17 57 13 17 10 Social and Community Services 44 29 6 22 5 19 10 and Religious Affairs Administration/Special Programs 48 11 81 17 50 11 73 40 Unallocated 22 5 0 0 4 2 Total 440 100 467 100 442 100 182 100 a/ Includes electricity and water. Source: Staff estimates based on government data G. Regional Composition of Public Spending 3.40 Cambodia has a highly centralized system of public finances (see Chapter 4). Thus, it is not surprising that the provincial share of expenditures financed from the government's own revenues in 1996 was only about 22 percent. If the total public expenditures (including ODA/NGO) are considered, -31- the provincial share works out to only 14 percent. Even this was possible because of the relatively high share of expenditures on primary education (37 percent) and war veterans' pensions (23 percent). Almost 97 percent of provincial budget expenditures were of current spending. Thus, the provinces have scarcely any role in building physical infrastructure. As they lack the powers to raise revenues or vary the level and composition of expenditures, the provinces have neither incentives nor accountability to provide public services efficiently to match the preferences and requirements of different regions. However, the government recently approved the Provincial Budget Management Law, which devolves certain limited revenue and expenditure powers to provinces (see Chapter 4). 3.41 Spending by the aid agencies is concentrated in the better-off regions, particularly Phnom Penh. In 1994, for example, almost 66 percent of the expenditures incurred by the aid agencies was concentrated in Phnom Penh alone. In spite of government efforts to bring about a more balanced distribution of aid funding, Phnom Penh continued to claim a predominant share of 51 percent, even in 1996 (Table 3.8). This could be explained by the fact that the bilateral and multilateral donors tend to dispense funds in areas which are more accessible, have facilities, and can be monitored more effectively. Even the regional distribution of expenditure from government resources seems to favor relatively better off provinces. Although they were more balanced than the ODA/NGO financed expenditures, per capita budgetary expenditures were positively related to per capita consumer expenditures. Phnom Penh, with a population share of only 7.5 percent, received the highest budgetary expenditure share of 12.3 percent. The consequence of these two factors has been the concentration of total public expenditures in Phnom Penh: total expenditure on public services per capita in Phnot Penh was ten times higher than that in other provinces together in 1996. Since poverty is heavily concentrated in rural areas, this pattern does not serve the equity objectives of the government well. Table 3.8: Regional Composition of Public Spending, 1996 1/ All Sources Government ODA/NGO I Exp. per Capita Share of Exp. per Capita Share of Exp. per Capita Share of (Thousand Total (%) (Thousand Riels) Total (%) (Thousand Riels) Total (%)Riels) Phnom Penh 42 404 12 28 51 377 Other Provinces 58 45 88 16 49 29 1/ Excludes countrywide government budget and donor programs. Source: Staff estimates based on government data. H. Conclusion 3.42 Much room exists for improving the efficiency and equity pertaining to the level and allocation of public expenditures in Cambodia. The government needs to decisively carry out the recommended actions. It should be emphasized that simply increasing spending and improving allocations does not guarantee better results. These efforts need to be complemented by improving budgetary institutions and processes, as discussed in Chapter 4. -32- 4. ASSESSMENT OF PUBLIC EXPENDITURE MANAGEMENT: INSTITUTIONAL DIMENSION A. Overview 4.1 The effective translation of government expenditures into expenditure outcomes depends critically on institutional arrangements through affecting incentives that govern the size, allocation, and use of budgetary resources. Institutional arrangements influence the quality of the outcomes on three levels: (i) aggregate fiscal discipline; (ii) resource allocation and use based on strategic priorities (allocative efficiency); and (iii) efficiency and effectiveness of programs and service delivery (technical or operational efficiency).' Institutional arrangements alone, however, may not necessarily have any effect. For such arrangements to be binding, the mechanisms that make adherence to these rules transparent and that hold the government and its ministries accountable for bad performance are necessary. Transparency and accountability mechanisms impose implicit costs on politicians and bureaucrats for violating rules and thus can make their commitment to the rules credible. This chapter examines the effectiveness of public expenditure management applying this analytical framework. 4.2 To better assess the budgetary process in Cambodia, it is essential to understand the current fiscal system which was adopted with the Organic Budget Law in December 1993. Under the current system, provincial governments do not have revenue-expenditure decision making authority. They merely act as agents or implementing bodies of the central government. Prior to the introduction of the 1994 budgetary reforms, although tax and expenditure policies were determined by the central government, the provinces were empowered to collect revenues and the center depended on the transfers from the provinces to meet its expenditure requirements. The unification of the budget in 1994 swung the pendulum completely in the opposite direction, from a rather anarchic to a highly centralized system of public finance. 4.3 In the current system, the revenue collection was completely centralized and the provinces were transformed into mere administrative units disbursing expenditures as delegated by the relevant line ministries of the central government. The Organic Budget Law stipulates that the ministers and heads of relevant public agencies are the principal fiscal officers of the state budget, responsible for their relevant line ministries and departments both at the center and in the jurisdiction of the provinces; the ministers confer on the provincial governors the function of deputy financial officers of the state budget and also delegate the responsibility for executing some expenditure functions. The recently approved Provincial Budget Management Law, however, recognizes provinces as independent budgeting entities and devolves some limited revenue and expenditure powers (see Box 4.5, later in this chapter). 4.4 Budget institutions and processes in Cambodia have made a remarkable contribution to macroeconomic control. In the late 1980s, unsustainable large budget deficits funded by new money creation contributed to galloping inflation. This peaked at 150 percent in 1991 and continued at high levels until 1994, before decreasing markedly to single digits during 1995-97 with the adoption of the Organic Budget Law and the subsequent reforms aimed at improving budget management. Under the Law, the government's annual fiscal program is detailed in the national budget, which must be prepared by the Ministry of Economy and Finance (MEF) in association with other line ministries and passed by the National Assembly. The Law lays down the timetable for completing the various steps involved in the budget preparation to ensure that the budget for the year is approved prior to the beginning of every Ed Campos and Sanjay Pradhan, "Budgetary Institutions and Expenditure Outcomes," September 1996; "The Public Expenditure Management Handbook", June 1998. -33- financial year. The Law provides for a single unified national budget consolidating all the fiscal functions of both central and provincial governments, establishing the authority of the central Treasury over all revenue collected, and ensuring a priori as well as ex post control over government expenditure by the Treasury. By adhering to fiscal discipline underpinned by the Law and the subsequent efforts, authorities have succeeded in reining in the current budget deficit. Thus, from the perspective of expenditure control, Cambodian budget processes have been performing effectively. 4.5 Budget institutions and processes in Cambodia, however, have been very detrimental to allocative and technical efficiency, which is discussed in the following sections. The fundamental problem is that the Cambodian budgetary process has sacrificed too much in terms of allocative and technical efficiency to maintain aggregate fiscal discipline. Thus, the challenge for budgeting in Cambodia is to enhance the allocative and technical efficiency of the budget process without compromising the high level of aggregate expenditure control which is being achieved by current highly centralized budgeting practices. This involves improving the current unnecessarily large trade-off between the aggregate fiscal discipline on the one hand and the allocative and technical efficiencyon the other. In fact, in the longer run there is evidence that improving allocation and operational performance can support improved fiscal discipline. The issues relating to improving the unnecessarily large trade-off are the central concern of this chapter. B. Linking Planning, Policy, and Budgeting 4.6 Closely linking planning, policy, and budgeting is one of the most important factors contributing to desirable budgetary outcomes. The Five-Year Socio-economic Development Plan (SEDP) of Cambodia (1996-2000) sets out the developmental objectives and strategies for the medium term and makes an indicative allocation of investment pattern in the economy. This has to be translated into concrete expenditure programs in the annual budgets. Over the last four years, some important progress has been made in linking the plan with the annual budget through three-year rolling PIPs; nevertheless, the linkage between the medium term plan and the annual budget remains weak. 4.7 The Ministry of Planning (MOP) prepared the First Five-Year SEDP and subsequently three PIPs. The third PIP (1998-2000) makes a significant advance in strengthening the institutional capacity to prioritize inter-sectoral and intra-sectoral investments to meet the sectoral targets indicated in the SEDP. It also links the three-year rolling plan on investment with the investment outlay in the annual budget for 1998 in some important respects. First, greater coordination between the MEF and the MOP in the preparation of the PIP and the budget has helped to forge a better technical linkage between the planning and the budgeting of investment outlay. In particular, the PIP process was initiated by a joint circular issued by the MEF and the MOP which set out the objectives, guidelines, and timetable for the preparation of the PIP and which stated the policy that henceforth only those projects included in the PIP would be financed by the national budget. Second, the setting up of a formal inter-ministerial steering committee--consisting of high ranking officials from the MOP, MEF, CDC, and NBC, and chaired by the MOP--aids considerably in coordinating the functions of policy and investment planning, budgeting, and 2 The MEF has made significant efforts despite difficult political situation. In particular, the MEF have issued numerous reports on the occasions of the preparation of the budget and the reporting of its implementation. These reports have made a broad analysis of the fiscal situation focusing on the lack of discipline in fiscal management and the failure in implementing proper management process. These reports are usually sent to the Council of Ministers and the National Assembly. They are regularly published in the Budget Book and the MEF Economic and Financial Review (quarterly). Their recommendations are translated into laws, regulations, decisions of MEF and/or administrative circulars. For details of fiscal reform efforts during the last four years, see 45 Months at the Ministry of Economy and Finance by H.E. Keat Chhon, June 1998. -34- aid coordination. The steering committee is responsible for setting the overall investment ceiling and sectoral allocation priorities in keeping with the objectives of the SEDP. Third, there has been improvement in the appraisal and prioritization of projects. Prioritization in the PIP is to be based on sectoral priority, absorptive capacity, domestic resource availability, and implementation capacity. 4.8 In spite of efforts to improve the efficacy of the institutional and procedural framework, the government's ability to determine inter-sectoral and intra-sectoral resource allocation remains weak. Consequently, linkages among the indicative sectoral allocations in the SEDP, programmed allocation in the PIP, and the annual budget allocation remain tenuous. The most important reason for this is that, with virtually no public savings, the government is not in a position to balance its investments so as to steer them to meet priorities according to the medium term plan. Thus, the government can hardly afford to give up externally funded investments, irrespective of sectoral priority or the impact on poverty reduction. 4.9 A critical institutional weakness is the absence of a proper system of monitoring. There is no system in place to see whether the actual disbursement of assistance to these sectors and sub-sectors has, in fact, occurred. When actual investment allocation is not monitored, it is difficult to direct future expenditures or the disbursement of external assistance to meet the plan targets. The task of monitoring is not confined to realizing investment targets. It also encompasses examining the achievement of the physical targets laid down. The assessment of additional investment requirements can be made only after the achievements of both financial and physical targets have been monitored. The government recognizes this problem and a system for monitoring the PIP is to be established under the Public Investment Management System (PIMS), which is intended to ensure the coherence and complementaritv of the separate capital budgeting activities. C. Budget Preparation 4.10 Budget preparation in Cambodia takes place in several elaborate stages (see Box 4.1). Despite this elaborate system, although some progress has been made in involving other ministries and agencies (especially the Ministry of Health), in practice the MEF largely dominates the entire budget preparation process (with the notable exception of the Ministries of Defense and Interior). The involvement of strategic agencies such as the CDC, the MOP, and the NBC is still limited. Although every line ministry has to prepare its budget allocation on the basis of the expenditure ceiling provided, the financial officers- -officials of the MEF attached to line ministries--play the lead role. The involvement of the provinces is even smaller. The preparation of the budget by different provincial departments is carried out under the guidance of the financial officials deputized to the province by the MEF. While this arrangement is due largely to the weak capacity of other agencies as well as the need to exercise strict fiscal discipline, it brings out problems of incentives and accountability. Box 4.1: The Budget Preparation Process in Cambodia Thefirst stage (beginning in June of the preceding year) is the establishment of the macroeconomic framework of the budget. This involves an MEF review of the macroeconomic constraints on the broad budget aggregates and their implications for overall budget balances. Forecasts of GDP growth and associated tax and other revenues (based on outcomes for the current year to date) are prepared by the MEF. Revenue forecasts determine the room for domestically financed budget spending. The second stage of budget preparation begins the policymaking process at the MEF. Initially, this focuses on the revenue side. Available funds are reviewed against broad spending objectives, and the scope for revenue augmentation is examined. Any options for reducing spending on major items to free up funds for new priorities is also discussed in general terms at this stage. -35- Once the availability of resources for domestically funded spending is determined, the third stage in budget preparation is to begin the prioritizing of these funds between the key sectors. This involves a first draft of the composition of the budget by broad strategic sectors, designed to maximize allocative efficiency. This draft sets ceilings for recurrent and investment expenditure by ministry and by major expenditure chapters of the budget. These indicative shares are based on the priority themes identified by the government for that particular budget (for example, the 1997 budget increased sector shares for education and health). The proposed broad sector shares are then approved by the Council of Ministers. With the sectoral architecture of the budget established, the fourth stage of the budget process begins, which involves translating the sector shares set by the MEF and Council of Ministers into the detailed spending programs for each line ministry. This process begins with a technical circular from the MEF to line ministries in July of the year preceding the budget which outlines the broad strategy and priorities for the coming budget. Against this background, the budget circular asks each ministry to present detailed bids for funding for both the central administration and the provincial departments of the ministry. The preparation of detailed estimates for the receipts side of the budget proceeds in parallel with this process. In parallel, line ministry bids for funding are prepared during July and the first part of August for an August 15 deadline at the MEF. Ministry bids for different budget lines are heavily based on historical budgeting. The outlays in the last budget become the bids in the next budget. However, the process is more complex for ministries such as Education and Health which deliver programs through provincial and district offices. The MEF receives line ministry spending proposals by August 15 in the year preceding the budget. Then, the reconciliation of the top down and bottom up phases follow. Thefifth stage, taking place in the second half of August, involves the technical analysis of line ministry budget proposals by the MEF, which checks compliance with priorities in the medium term plan, accuracy of costings, and consistency with initial ministry ceilings. The starting point for this MEF examination of line ministry bids is the budget currently under way. This is adjusted for technical parameters -- the actual use of budgeted funds by agencies and movements in wages and prices. The result is the budget for 1996 (for example) on the parameter base of 1997. Line ministers' budget proposals for 1997 are then compared with this benchmark to identify adjustments, extensions, and new programs incorporated in their bids for 1997. The sixth stage, during September, involves the resolution of conflicts between funds earmarked by the MEF for each ministry and funds sought by the ministry. This involves meetings between the MEF and line ministry officials. Where a disagreement cannot be resolved by officials, it may be raised to the level of the Minister of Economy and Finance and the line ministers. Upward and downward adjustments of ministry funding can also be made to ensure that proposals that can be fitted within one ministry's envelope are not less cost effective than proposals which cannot be squeezed into the envelope of other ministries. In the seventh stage, final issues and disputes are resolved during October at the prime ministerial level. The draft budget that emerges is then submitted to the Council of Ministers for approval. With the details of the budget determined, the eighth stage is to prepare the budget bill. The bill is submitted to the National Assembly, and is subject to committee discussion and a plenary session. Following signature by the President of the National Assembly, the budget law is promulgated by the King. Finally, in the ninth stage the MEF issues a Praka to line ministries outlining the funds budgeted for them. Line ministries in turn distribute funds internally to their operational units on various decision rules. 4.11 Treatment of the capital component of the budget is different from recurrent expenditures. Most of capital budget is donor funded. Proposals for capital investments which appear in the budget originate from line ministry agreements with individual donors about funding for specific projects in their respective sector. However, the prioritization of capital spending options for inclusion in the budget takes place separately from the prioritization process applying to recurrent expenditures. It operates through the preparation of a PIP. While some progress has been made through the PIP process, the linkage between decision making on the investment and recurrent budgets is weak. This linkage needs to be strengthened through a more integrated approach. D. Budget Execution 4.12 Highly Centralized Budget Execution. Budget execution is very highly centralized, with the MEF approving the individual spending decisions of line ministries to an extent that has few parallels. In some cases, three authorizations are required from the Minister of Economy and Finance--for entering obligations, for accepting tender, and for releasing cash. The running costs of line agencies are -36- prescribed in Chapter II of the budget by the MEF in fine detail. Virement processes involving MEF approval are required to switch funds between, for example, electricity and telephone, inhibiting cost effective program management. Expenditure approval processes in the provinces are even more complex. A single expenditure action can involve approvals from the responsible line ministry, the Department of Financial Affairs at the MEF, the provincial governor, and the provincial treasury. Expenditures are subject to complex central controls intended to tightly control the evolution of the budget deficit through the course of the year. This highly centralized approach to budget execution ensures a high level of expenditure discipline. However, these controls make proactive management of programs by line agencies extremely difficult and also significantly reduce the allocative and technical efficiency of the budget process, as these complex processes absorb large amounts of resources, reduce budget flexibility, and focus senior officials on processing paper. 4.13 Monthly Cash Allocation of the Bulget. Figure 4.1: Monthly Distribution of Owing to the uncertainty of government receipts from Revenues and Expenditures in Cambodia, month to month and the ad hoc (and often unwarranted) 200 injection of new spending priorities (in particular in defense and security) into the budget process, budget C execution is dominated by a monthly cash allocation 'o0 process (Figure 4.1). At the beginning of each month, expenditure allocation to different spending ministries 50 and agencies is carried out by a committee composed 0 of the Budget, Tax, Financial Affairs, and Treasury Departments of the MEF, on the basis of the stock of Treasury balances and the expected revenues during the month. On this basis, a Praka of the monthly expenditure plan is issued by the MEF. This prioritizes the access each line ministry will have to the cash expected to be available during the month. 4.14 Financial Controller System The process of authorization, however, is a long and complex one. Financial controllers employed by the MIET and assigned to line ministries, provinces, and municipalities play a central role in this process (see Box 4.2). Financial controllers undertake a pre- audit of each expenditure proposal by line ministries. This audit is based on expense, consistency with budget allocation, accuracy of costing, availability of funds, compliance with budget laws and regulations, consistency with budget intent, and quality of supporting documentation. Box 4.2: Financial Controller System The process of centralization was strengthened with the introduction of a system of financial control through a sub-decree in 1995, implemented beginning in mid-1996. Following this sub-decree, 50 financial controllers were appointed by the MEF and trained under a USAID program. Under this system, there is a chief financial controller for each line ministry and several deputy and assistant controllers, depending on the size of the bud get of the line ministry. The functions of the chief financial controller could be similar to that of a chief financial officer of a line ministry. However, the controllers are located in the MEF and report to the Director of Financial Affairs. The system remains largely a function that pre-audits all expenditure commitments made by the line ministries. The sub-decree allows the financial controller to authorize all expenditures up to CR 20 million and certain recurrent expenditures such as salaries on an annual basis. Following the July 1997 events and the financial crisis faced by the government, however, all expenditure commitments other than salaries and wages now require the approval of the Minister of Economy and Finance. As a result, the financial controllers are being used to process individual expenditure commitments for approval by the Minister of Economy and Finance. The financial controllers remain largely ineffective in the financial managemrent of the ministerial expenditure programs for which -37- they are responsible given their preoccupation with processing commitment requests for the approval of the Minister of Economy and Finance, as well as their lack involvement in the work of the finance and accounting departments of line ministries. If these officers could be integrated with the line ministries within a decentralized accounting and financial management framework, they could play a significant role in the management of the line ministries. Proper control of expenditure commitment is important, but financial management does not stop there. Strategic financial management requires allocating resources to priority needs, evaluating the cost effectiveness of services, and obtaining value for money. These activities become even more important in view of the resource constraints currently faced by the government. The financial controllers, supported by adequate financial management systems, should be able to exercise better management of revenues, recurrent expenditures, and capital expenditures. 4.15 Since virtually all line agency expenditures must be approved by the relevant MEF financial controllers, these officials play a key role in ensuring that line agencies observe the letter of the budget. They also provide a control point for monitoring the level of spending commitments incurred by agencies. Other roles include reporting on budget implementation and providing advice on line agency spending proposals for the next year's budget. Payment orders issued by the financial controllers in response to a purchasing activity which they approve are cleared through the MEF. This is a highly centralized process which currently involves approval by the Minister of Economy and Finance himself of relatively small and routine outlays proposals. Notification of MEF approval is then given to the national or provincial Treasury, authorizing the Treasury to make a direct payment to the external supplier. Substantial delays are involved as well as uncertainty about whether funding will be approved. 4.16 If cash flow problems arise during the course of the month, three types of adjustments are made in order to prevent an unplanned increase in the budget deficit. First, the monthly cash allocation by the MEF to individual ministries is reduced in relation to that contained in the budget appropriation. Second, payment authorizations are withheld by financial controllers, which restricts line agencies from spending their full monthly cash allocation (a pre-audit control). Third, where payment has been authorized by the financial controller and the transaction has been undertaken by the line agency, funds may not be immediately released by the national or provincial Treasury to the external supplier of goods and services. 4.17 "Serial spending" occurs in Cambodia owing to the introduction of new spending decisions in each monthly cash allocation. Ten percent of each month's cash allocation is set aside for new political decisions in regard to spending. These decisions relate mostly to defense and security. The monthly cash allocation process does take account, to some extent, of the claims of the competing ministries, and involves a form of ad hoc reprioritizing in the face of unforeseen cashflow constraints. However, reprioritizing is undertaken without the same transparency and review mechanisms that characterize the budget preparation phase, thus reducing the allocative efficiency of the budget process. More serious is the effect on the technical efficiency of budget execution of ad hoc monthly revision of the budget. The operational inefficiencies created by the highly centralized system of cash allocation to line ministries relate particularly to the tendency for cash release to be irregular and unpredictable. Funding release tends to be concentrated toward the end of the year. Thus, operational units of the line ministries are uncertain about how long each cash allocation will last, which distorts the manner in which they determine their own spending priorities. For example, while the Ministry of Public Works and Transport would like to spend on road repairs when the road conditions are bad after the monsoon, they cannot undertake these repairs until the end of the year. Lack of resources to finance the projects on a continuous basis can also result in time and cost overruns. In addition, there exists a practice of informally borrowing from local money lenders at exorbitant rates of interest to meet immediate expenditure obligations and repay the money when the liquidity position improves. -38- 4.18 The monthly allocation practices during budget execution, often resulting from unwarranted political intervention, raises questions about the value of the elaborate processes associated with budget preparation. It is clearly demonstrated by Table 3.6 in Chapter 3 which shows significant discrepancies between the outcome and budgeted non-wage O&M expenditures, hugely in favor of defense and security at the expense of economic and social services. It also points to the importance of developing capacity for reliable and honest monthly revenue projection. 4.19 Leakage of Funds. A further reason why ministerial and provincial departments effectively fail to receive their budgeted allocation is associated with the informal diversion of funds released to individual agencies for uses unrelated to the agency program ("leakages" of funds). In the course of accessing their monthly cash allocations, ministries and provinces encounter informal arrangements for deductions from authorized funding at various approval points. These leakages fall into two categories: the diverted funds are used for other programs, informally reprioritizing budget outlays at the margin, and compromising the allocative efficiency of the budget, and the diverted funds supplement the personal income of those processing the payments (outright corruption), resulting in technical inefficiency owing to the higher than necessary budget cost of achieving given program deliverables. Box 4.3 presents the anecdotal evidence of such leakages. Box 4.3: Anecdotal Evidence of Leakage of Funds Anecdotal evidence suggests that "leakage" is particularly high during the course of processing cash allocations by the provincial treasuries (possibly as much as 15 percent of funds being disbursed to an individual ministry leak to other uses), with further leakages of possibly 5 percent at the provincial finance department office (in the course of issuing payment approvals) and possibly 1 percent at the provincial office.' Additional leakages may occur further down the line in the form of "commissions" by district officials. Further leakages arise owing to the delegation held by the provincial governor for the release of line ministry funds to ministry operatives in the province. This results in the practice by provincial governors of siphoning off funds intended for the regional activities of a particular ministry for their own purposes and programs, with the provincial ministry officials being obliged to sign for funds which they do not in fact receive. The propensity of provincial governors to skim funds from programs in this way reflects, in part, the lack of provincial revenue measures available to the governors. Based on information collected from provincial health departments in 1997. E. Budget Monitoring and Reporting 4.20 The government has taken several steps to improve financial management and public accountability. A series of laws, decrees, directives, and accounting circulars has been put in place to provide the regulatory framework for sound financial management. In addition to the financial controller system described in Box 4.2, specialized units have been established in each line ministry to strengthen procurement. A budget reporting system, called TOFE, provides monthly macro-level information. These measures have improved budget management and accountability for public funds compared with the pre- 1993 era. 4.21 Nevertheless, budget monitoring and reporting remain weak The government's efforts to increase transparency and accountability in the public sector have yet to bring about the expected outcomes, owing to weaknesses in implementation and lack of a proper accounting and auditing system and sound financial management at the line agencies. The severe limitations of the accounting and auditing systems are detrimental to technical efficiency, because they create an environment that is conducive to widespread leakages. There is an urgent need to establish a standard accounting system for line agencies, which would provide greater transparency of transactions and would help curb leakages. Linked with this is the need to establish audit capability, both internal and external, for line agencies. -39- 4.22 Accounting. The current accounting system is modeled on the pre-War French government system. The Treasury maintains accounts under double-entry accounting principles, with a modified cash basis accounting. This system requires that accounts be maintained on a cash basis but that they include certain commitments at year end under certain rules. Owing to the cash shortage, however, even the cash basis accounting system is not properly maintained. The expenditure commitments and the corresponding cash payments are occasionally in arrears. The Treasury then issues payment orders to creditors which are not immediately cashable by contractors and suppliers. The creditors in turn discount these payment orders with taxpayers who can use them to settle tax dues. It is not clear how these transactions are reflected in Treasury account books. 4.23 The Treasury accounts are maintained by chapters of expenditure such as wage expenditure, non- wage expenditure, and capital expenditure. Revenues are better analyzed by tax and non-tax revenues and their sub-components. This system of accounting facilitates producing the monthly financial statement (TOFE). This is used for month-to-month budgeting versus reporting actual revenue, expenditure, and financing sources. The functions of line ministry accounting departments primarily relate to processing applications for approval of commitments to the MEF financial controllers. 4.24 Procurement. Following the issue of the Decree on Public Procurement in 1995, procurement units were established in all line ministries and provinces to undertake public bidding to procure goods, services, and civil works. In addition, a central Procurement Unit was established in the MEF to undertake all procurement contracts in excess of CR 20 million (approximately US$5,100). Notwithstanding this law and the very low threshold for public bidding, only about 14 percent of total non-wage expenditure in 1997 was procured through the MEF Public Bidding Office. This low volume of public bidding results from the inability of the MEF Public Bidding Unit to impose the legal requirements, owing to pressure from the ministers and the circumvention of bidding regulations through the breakdown of bids into smaller parcels. The law remains largely ineffective and the transparency of the procurement process is questionable. At times, after the announcement of public bids, contracts have been awarded outside of the bidding process through political interference. 4.25 One reason for the ineffective implementation of the procurement procedures could be the very low threshold set for centralized procurement. As centralized bidding takes time and the line ministries lose control, there is no incentive for line ministries to comply strictly with the rules. It would be better to give more flexibility to ministry procurement units and strictly enforce rules through post audits by the MEF procurement staff. 4.26 Auditing. There is no proper internal or external audit in place. As discussed above, financial controllers perform the pre-audit function. The current audit requirements are based on a Decree issued in 1984. The Decree explicitly exempts the Defense and Interior Ministries from audit unless it is authorized by the Prime Minister. A draft Audit Law is under consideration to establish an independent Auditor General's Office reporting to the National Assembly. When the Audit Law and the relevant decrees under discussion are approved in current form, the laws would establish the basis for a modern auditing function. Both external and internal auditing can be useful only if audit findings are properly addressed and reported deficiencies are corrected. In this context, establishing a modem accounting function and training staff is critical. The Law on Accountability of Public Managers under consideration would also contribute to enhancing the accountability of budget with fines, disciplinary actions, and charges under the penal code for the mismanagement of public managers. F. Recommendations 4.27 Given the history of fiscal and macroeconomic instability, it is clearly undesirable to introduce -hanges to budget processes which might weaken the current effective management of the budget deficit. However, considerable scope exists for improved allocative and technical efficiency of the budget -40- process consistent with the present high levels of fiscal discipline. This relates particularly to the institutions and processes involved in budget execution. Because of the weak technical capacity at the central and provincial levels, this section focuses on the pragmatic removal of impediments to effective budget operation. Toward this end, it provides a strategy for reform in the form of six key directions and, under each direction, makes concrete recommendations with indications of sequencing. The six key directions for reform are: * Ensuring political commitment to an authoritative budget process * Establishing responsibility for performance at the level of spending units * Rationalizing the roles of the central and provincial administrations o Developing financial participation schemes (user charging) * Improving the allocative and technical efficiency of donor funding * Strengthening the Public Investment Management System. 4.28 The top priority is for the government to respect the budgetary process, because if the government does not abide by its own budget rules and procedures (which has often been the case in Cambodia), there is little point in making efforts in designing budget reforms. In this regard, reducing unwarranted political interference in the budgetary process should be considered as the prerequisite for any budget reforms. Another prerequisite is to improve the managerial and accounting and reporting capacities at the spending units. This is especially so if responsibility for performance at the level of spending units is to be established--which is the key to improving the trade-off between fiscal discipline and allocative and operational efficiency--without unacceptable loss of financial control. 1. Ensuring Political Commitment to an Authoritative Budget Process 4.29 Ensuring predictability of the budget through political commitment to an authoritative budget process is the top priority. This reform would be a necessary condition for credible efforts towards improving management of expenditures at the sectoral level. 4.30 Recommendations: * Ad hoc spending decisions taken outside the budget preparation context--which occur because of newly identified political priorities-- should be minimized. * Where new spending measures are proposed during the course of the year, they should not exceed the reserve budgeted for this purpose. * Where this is not achievable, the request for new spending from the Prime Minister's Office should also specify the savings option from which the funding will be drawn, and the affected line ministries should be informed of the revision of their budget estimates. 2. Establishing Responsibility for Performance at the Level of Spending Units 4.31 Devolving Management and Financial Freedoms. The foremost medium-term direction for change is to increase responsibility for effective program performance at the program agency level. This would entail empowering program agencies and functional units to manage their programs with minimal interference rather than processing unpredictable levels of funding through multiple external approval processes. The aim is to reduce the pre-audit of individual disbursement activities by the MEF and replace it with a post-audit of expenditures from MEF approved expenditure plans. To give greater freedom to line agencies to determine the way in which they spend money in response to their knowledge of program environment, the government needs to extend the scope for lump sum program allocations rather than detailed line allocations. This would be a gradual process, as developing a program basis for budgeting requires good financial management skills within line agencies, so that envelopes of funds for particular programs are not diverted to less productive uses. -41- 4.32 Therefore, the devolution of management freedoms through program-based budget allocations should be conditional upon progress in capacity building in the spending units and hence confined to pilot programs rather than to broad-based changes for some time to come. A pilot approach was introduced in 1996 in the form of the Accelerated District Development (ADD) System (Box 4.4).' This system provides more efficient and flexible funding for hospitals and health centers in a small number of districts on the basis of: appropriating of a separate envelope of funds (on a program basis) for each district selected for participation, rather than having the district's funding drawn from separate ministry- wide line items for numerous individual categories of running costs (to which individual districts have uncertain access); and providing more streamlined funding arrangements that bypass the provincial governor and associated leakages. Box 4.4: Accelerated District Development System (ADD) The ADD is a cash advance system (introduced in June 1996) designed to give program managers greater certainty in regard to the level of budget funds available to them and greater flexibility in the use of these funds. It provides funds for referral hospitals and health centers in (currently) 22 ADD districts. The ADD's main feature is that operating costs (excluding wages) are budgeted as a single envelope (Chapter 13). This gives program managers discretion in the use of funds for some 24 purposes, such as patient food, emergency patient transport, furniture, maintenance of buildings, and vehicle running costs. Part or all of the operating costs in ADD districts comes from the Accounting and Finance Department of the Ministry of Health (MoH) rather than from the provincial Treasury. This eliminates pre-audit and provides greater certainty and stability of funding. However, the release of funds through the year is still subject to Treasury restrictions and does not ensure regular cash release and budget certainty. The scheme operates as follows: (i) the MoH receives a cash advance from the MEF, which is charged against the MoH Chapter 13 allocation and becomes an imprest account; (ii) this is distributed to ADD districts on the basis of their budget allocation; and (iii) allocations are reimbursed by the MEF via a payment order signed by the MEF and given to the Treasury, reimbursing the imprest account. To ensure accountability, the imprest account is reimbursed by the MEF only after ADD districts submit invoices to the MoH indicating proof of expenditure, together with a summary of spending against each budget line by the ADD district. These invoices are checked by the MoH and the financial controller for the MoH at the MEF. If everything is correct, a payment order to the Treasury is signed by the MEF and the imprest account is reimbursed and charged against the MoH Chapter 13 allocation. The MoH can then make a further distribution to the ADD districts. The districts have the freedom to spend the Chapter 13 budget according to their district's needs. However, the MEF is encouraging them to give priority to health-specific expenses--such as patient food and repair of medical equipment--over administrative expenses that are less directly related to health, such as furniture and maintenance. 4.33 The piloting of program budgeting through the ADD is an important step toward a better trade- off between fiscal control and the allocative/technical efficiency of the budget process. It sacrifices little in the way of fiscal control while offering a more streamlined and cost-effective use of funds by functional units. While strong central control of the total monthly cash release ensures continued fiscal discipline, spending decisions made closer to the program delivery level can respond more sensitively to priorities than can funding that is determined prescriptively from the center. Preliminary indications to date are that this is operating sticcessfully. 4.34 Recommendations: * An evaluation of the effectiveness of the ADD in its 22 districts is needed. ADD is not a program budget in a strict sense, among others, as it does not include wages. It is more of cost center budgeting, with a program component. -42- * Depending on the findings, a unit needs to be established in the MEF for: addressing problems in introducing the ADD in the existing 22 districts; planning for the extension of ADD to other districts; identifying other agencies in which program budgeting might also be piloted. * Depending on the results of the evaluation, a trial of a similar program-based scheme should be planned in other agencies (with MoEYS an early priority). 4.35 Improving the Financial Management of Program Staff Before introducing more devolved budgeting, good financial management skills within line agencies needs to be established. There are at least three pre-conditions for this: (i) the establishment of basic financial reporting systems in line ministries which indicate the purposes for which funds are used within each functional unit within the agency; (ii) a commitment by functional units to spend non-mandated funds so as to maximize program outcomes rather than the personal or managerial interests of the unit; and (iii) a commitment on the part of senior line ministry management to actively monitor the use of funds by functional units under their control and a willingness to redirect funds between units in response to ineffective use of funds. 4.36 An initial step is to improve transparency and accountability through establishing standardized accounting procedures for line agencies. For this purpose, the accounting systems of the major line ministries should be reviewed to highlight the improvements needed before program budgets are used more widely. Following this, the MEF should develop accounting standards for line ministries so that an integrated financial information system could interface Treasury general ledger accounts with the line ministry general ledger accounts. Establishing a proper audit system with an independent Auditor General's Office is also critical. 4.37 Recommendation: The MEF needs to establish a financial management improvement team to help agencies introduce ministry-wide accounting systems for reporting, analyzing, and benchmarking resource use by each of their functional units. The unit would work initially with the most innovative of the large agencies (MoH). 4.38 Accounting for Deliverables. If financial decision making is to be devolved, the responsibility for program performance needs to be clearly assigned to the program units. This requires the identification of the outcomes for which they are responsible. The government needs to ensure that budget estimates of payments to line ministries are accompanied by statements of the deliverables expected from the ministry. The Health and (to some extent) Education Ministries are moving toward clearer identification of program outputs, because their budget bids are based on bottom up estimates of service that provide functional units and unit cost formulas for their operation. Over the medium term, rigorous activity-based budget bids along these lines need to be encouraged, as they would pinpoint areas in which operating costs are unacceptably high in relation to outputs and would reveal areas in which some flexibility is desirable in budget allocations and increase the cost effectiveness of budget outlays as a whole. 4.39 Recommendation: The MEF needs to prepare a timetable for introducing activity-based budget bids by line agencies over a five-year period. This could involve penalties where funding bids are inadequately supported by costing information. Assistance needs to be made available to line agencies from the proposed MEF financial management improvement team for the development of activity-based budget bids. 4.40 Improving Technical Skills in Line Agencies. A critical issue for devolving responsibility is whether line agencies have the technical capacity to undertake financial and management planning functions. The current situation is such that the need to build a financial management skill base in -43- program ministries is urgent. One possible solution to this problem would be to create a technical specialist grade within the public sector for financial and policy analysts, with a wage structure closer to private sector payments for similar skills. Formal training in financial and program management skills would also help develop a suitable environment for greater prioritization of resource use within agencies. Equally important is that capacity building at the project management level needs to be more systematic and programmed. 4.41 Recommendation: Consideration needs to be given to the establishment of a technical specialist grade in the government service with wages more closely related to comparable technical positions in the private sector. 3. Rationalizing the Roles of the Central and Provincial Administrations 4.42 Program delivery at the provincial level involves too many players. And spending actions can be modified or delayed at many points, which can result in excessive uncertainty, delay, and leakage of funds. A particular issue is that the large number of government agencies involved increases the siphoning off of funds because of the payment of commissions at each stage of authorization. In addition, provincial governors redirect funds intended for line ministry functional units. The number of individuals and agencies involved in the approval process also results in an inability to assign responsibility for shortcomings in the operation of line ministry functional units. 4.43 Thus, a priority in budget process reform is the separation of the funding processes for provincial units of line ministries from clearance mechanisms involving the provincial governor. This is currently being tried successfully with ADD. A parallel reform, the Provincial Budget Management Law, gradually in phases would give provincial governors their own sources of revenue for undertaking the local programs which they identify while also gradually terminating the provincial governors' role in channeling line ministry funds to their regional operations (Box 4.5). This would increase the control over cash allocations that can be exercised by the provincial departments of the line ministries, district offices, and managers of operational units, and would reduce the power of provincial governors in regard to cash disbursement. Box 4.5: Provincial Budget Management Law For administrative purposes, Cambodia is divided into provinces, districts, communes, and villages. There are 21 provinces in Cambodia. The country follows a unified budgetary concept in which the tax and expenditure powers were centralized with the adoption of the Organic Budget Law in December 1993 . The provinces do not have powers to raise taxes, fees, or use charges and they execute expenditure programs on behalf of the central government merely as its agencies. Under this Law, provincial governors and city mayors must prepare and implement their budgets and are responsible for the maintenance and upkeep of the public assets transferred to them. The Law assigns certain local expenditure functions, such as maintenance of street lighting, fire fighting, public parks, water supply, drainage, sewerage, garbage disposal, roads, and parks to the provinces/cities, which are also required to meet their own employee payrolls and to devolve grants to districts, communes, and villages. To meet these expenses, the Law assigns certain minor taxes, for example, land taxes, registration fees and patent fees, vehicle taxes, and non-tax revenues such as user charges on electricity and water supplied to local residents and fees for the occupation/use of public property (markets, parking lots, etc.). These revenues will be supplemented by central grants to ensure that no province or city has a deficit budget. The Law is expected to be implemented in phases. Initially, the Law is to be applied to Phnom Penh, Siem Reap, and Battambang. It would be extended to other provinces as the administrative capacity develops. Full -44- application of the Law would require detailed working out of the framework for the local levy of taxes and non-tax revenues as well as a methodology for making transfers which is objective, equitable, and transparent and at the same time contains incentives for better fiscal management at the provincial level. The degree of decentralization advanced by the new law, however, would be limited. In the proposed arrangement, the expenditure share of the provinces remains broadly the same. Nor do provinces have appreciably greater flexibility in regard to revenues, since the revenues assigned to them contribute less than 2 percent of total tax revenues and about 1.5 percent of total revenues. Nevertheless, the proposal would provide better incentives to provincial governments for providing services. Enabling the provinces to raise revenues to meet a part of their expenditures (although marginal) enhances flexibility, provides greater incentives, and improves accountability. 4.44 It should be borne in mind, however, that the development of a transparent and objective transfer system devoid of disincentives to fiscal management is a challenge. Moreover, as the expenditure control system at the provincial level is weak, developing the provincial financial management capacity, including an independent audit system to ensure that funds are utilized efficiently and for their intended purposes, is critical to the decentralization of expenditures. 4.45 Recommendations: * A priority should be given to separating the roles of the line ministries and provincial governors in the disbursement of budget funds. * In parallel, the financial management skills of the regional units of line ministries should be upgraded within each line ministry. 4. Developing Financial Participation Schemes (User Charging) 4.46 The Cambodian budget system is characterized by a variety of arrangements for user charging by service delivery agencies. Some arrangements are officially sanctioned, such as fees for logging and fishing rights, while others are unofficial, such as fees for school examinations or health services. Ministries vary greatly in the degree to which they acknowledge the existence of user charges. For example, the MoH is open about their existence but the MoEYS is reluctant to acknowledge them. At face value, extrabudgetary funding is inconsistent with the budget law and prevents the overall prioritization of the use of funds available to the government. Informal user charging also places part of the financing of public sector goods and services outside the system of financial transparency and accountability. In addition, there are likely to be adverse income distribution effects where the less well- off are unable to purchase a service. However, user charging does contribute to greater allocative efficiency between the public and private sectors, since the level of consumption of publicly provided goods and services is connected to the resource cost of their provision. User charging also helps overcome deficiencies in central revenue collection. However, the schemes should meet minimum standards of efficiency, equity, and transparency. 4.47 Recommendations: * The existence of financial participation (user charging) schemes should be acknowledged as a significant budget tool, and the incidence and terms of such arrangements should be progressively documented over time * A set of guidelines should be developed by the MEF on the circumstances in which user charging is appropriate, the appropriate basis for determining charges, and the handling of associated equity issues. 5. Improving the Allocative and Technical Efficiency of Donor Funding 4.48 Much room exists for improving the allocative and technical efficiency of donor funding. The current situation is such that much donor assistance bypasses the budget and involves direct funding of -45- project contractors. This reflects lack of donor confidence in line ministry capacity to identify, design, and implement projects. Magnifying the problem is the weak coordination at the sectoral level. There are no formal groups involving multilateral and bilateral aid agencies in any sector. On the other hand, NGOs are better organized, with the Cooperation Committee for Cambodia (CCC) facilitating consultation among its members and with the government. 4.49 While such an approach to aid may have been to some extent unavoidable, it has resulted in a fractionalized public sector effort with little central organization and control and with a number of adverse side effects such as: (i) the absence of strong government ownership of many projects; (ii) piecemeal efforts by aid agencies toward sectoral issues and institution building; (iii) weak coordination among donor programs; (iv) a proliferation of different procurement, disbursement, auditing, and progress monitoring activities among agencies; and (v) the creation of special project units, staffed by expatriates or by nationals with "topped up" salaries, with adverse impacts on institution building. 4.50 Capacity Building. To improve the efficiency of donor funding, the government needs to improve technical, financial, and managerial capacities in line ministries to achieve better project implementation. Improved project implementation would encourage donors to increasingly work through the existing public administration. The government also needs to improve project identification, design, and costing in line ministries. Tighter quality standards for project proposals from line ministries could then be introduced by the MoP and the CDC when admitting line ministry proposals to the PIP. The government would then be in a better position to negotiate with donors by presenting better prepared options. 4.51 Aid Coordination. In parallel, the government needs to move toward better aid coordination. A promising way would be the Sector Wide Approach (SWAP). SWAPs would involve the creation of sector working groups with representation from the MEF, CDC, MOP, relevant line ministries, donor agencies, and NGOs. Such groups would provide the forum for exchanges between aid agencies and the central and line ministries and could help plan sectoral strategies, coordinate donor activities, monitor sectoral progress, and address implementation problems. Under SWAPs, the responsibility for sector development is shared collectively by donors and the government. While SWAPs will involve considerable efforts on the part of the government and donors, this approach would link donor funded capital spending more closely with the priorities identified in the SEDP. It would be important that SWAPs should not create a new central clearance mechanism for investment proposals, which would increase bureaucracy and lead to delays in getting individual projects under way. 4.52 Donors' Role. To effectively complement the government's efforts, donors need to play a significant role. Donors need to provide well-coordinated technical assistance and training to strengthen the line ministry's capacity in a more systematic and programmed manner. Ad hoc approaches to capacity building would make the proposed budget reforms more difficult to implement. Donors also need to move toward better aid coordination, such as through a SWAP. The effective subordination of donor activities to mutually negotiated sector-wide strategies would be not be easy and there would be problems of agreement among individual donors regarding responsibility for specific inputs or outcomes. However, while SWAPs give donors less freedom in selecting the individual projects they wish to fund, it give greater involvement in developing (along with the Cambodian government) the overall priorities and strategy for each of the key sectors. Furthermore, a SWAP would move toward common management arrangements and could lead to more coordinated donor involvement in sector development which would help strengthen the weak institutional capacity in line ministries. -46- 4.53 Recommendations: The preparation for a SWAP needs to be initiated in the health sector, as the MoH has a better capacity for this; in addition, coordination between the MoH and donors is relatively close (e.g., regular monthly meetings). * A number of system-wide issues, such as managerial decentralization, reforms in accounting and budgeting systems, and changes in staffing incentives should be addressed for a SWAP to be adopted effectively. 6. Strengthening the Public Investment Management System 4.54 Capital budgeting is composed of five stages: (i) identification of the development framework; (ii) project identification; (iii) investment programming; (iv) investment budgeting; and (v) investment monitoring. These stages involve a number of agencies such as the CDC, MOP, MEF, and line ministries. The Public Investment Management System (PIMS), introduced in March 1998 by the CDC, is intended to ensure the coherence and complementarity of the separate capital budgeting activities of each of these agencies. In particular, PIMS draws the activities undertaken by the MOP and MEF during the preparation of the PIP and the annual budget into a coordinated management framework. The major weaknesses in the chain of PIMS sub-systems include inadequate investment appraisal at the line ministry level, insufficiently robust cost estimates in the PIP for transmission directly into the annual capital budget, and shortcomings in the monitoring of project execution by line ministries. The major benefit from PIMS would be the focusing of resources on these three weak links in the capital budgeting chain and the acceleration of their repair. 4.55 The PIMS master plan would work effectively only when these shortcomings are resolved. In particular, a more formal interface between the PIP and the budget requires better and more detailed programming by the MOP of cash flows associated with each project. This would enable the MEF to include in the budget a more accurate schedule of counterpart payments than at present. Preparation in the PIP of budget quality splits for annual capital spending by project has a further important advantage. It is closely linked to the use of the "out-year" estimates of spending for each project as a nascent set of forward estimates or a starting point for an eventual medium term expenditure framework (MTEF), which would help link policy, planning, and budgeting more closely and consistently. 4.56 Recommendations: * An urgent need is for the line ministry's capacity in monitoring the implementation of the PIP to be strengthened. * Over the medium term, the development of PIMS needs to be extended beyond a broad framework for linking capital budgeting institutions, to become a vehicle for developing concrete strategies aimed at overcoming the current weak links in the operation of the capital budgeting cycle G. Conclusion 4.57 There exists significant potential for enhancing the allocative and operational efficiency without undermining the current high level of aggregate fiscal discipline. To realize this potential, the foremost priority is to respect the budgetary process, minimizing unwarranted political interventions. Another priority is to improve the managerial and accounting and reporting capacities at the spending units, in particular because responsibility for performance at the level of spending units must to be established without losing financial control. -47- 5. MACROECONOMIC FRAMEWORK, RESOURCE ENVELOPE, AND PUBLIC EXPENDITURES A. Introduction 5.1 As discussed in the previous chapters, Cambodia is at the crossroads of a transition toward sustainable development. To make this transition successful, Cambodia must resolve its interrelated fundamental fiscal and governance problems. This chapter brings together the findings and recommendations of the previous chapters--in terms of revenue enhancement, improvement in governance, and expenditure rationalization--and presents two alternative scenarios that highlight the implications of implementing policy recommendations on the medium-term prospects, in particular on the level and reorientation of public expenditures. 5.2 The next section discusses a macroeconomic scenario in which the government is determined to resume structural reforms following the slippages since 1996, and, as part of these policies, is in a position to substantially increase revenue collection (the full reform scenario); under this scenario, the likely outcomes for public expenditure financed by the government's official budget (the second concept in Chapter 3) are presented. In section C, comparison of the full reform scenario with the other scenario, which assumes no significant change in government policy (the status quo scenario), clearly demonstrates that policy changes are needed in order to achieve the public expenditure goals and improvements set out in this report. Section D discusses the downside risk, focusing on a further deepening of the regional financial crisis. Section E provides a brief conclusion. B. The Full Reform Scenario 5.3 The full reform scenario is built on the assumption that the government will take decisive actions to reinforce macroeconomic stability, mainly through prudent fiscal policies, and, more important, to implement structural reforms at an accelerated pace. Key areas are forestry, revenue collection, civil service reform, and military demobilization, but reforms in other areas (such as legal) should also be addressed to create an environment conducive to private sector development. Under this scenario, it is expected that domestic investment would pick up and foreign investment would accelerate, provided the regional crisis would not deepen further. If the government is decisive in pursuing the reforms, the international donor community would continue its financial support (through grants and concessional lending), which was partially interrupted in 1997 following the July events. As a result of this favorable combination, economic growth would start to accelerate from 2000 and would reach 6-7 percent by 2002 (Table 5.1). Inflation would gradually decrease to 5 percent by 2002. 5.4 One of the key areas in this full reform scenario is improvement in revenue collection. It is assumed that revenue would increase from the current 8-9 percent of GDP in 1998 to 13-14 percent by 2002. The largest increase would be realized in 1999 with the introduction of VAT (the equivalent of about 1.3 percent of GDP), and about 1 percent of the increase would come from improved governance.' I The full year impact of the implementation of the VAT is estimated at 1.7 percent of GDP (see Chapter 2). As it would take some time before becoming fully operational, the expected revenue enhancement from the VAT implementation in 1999 would be somewhat lower than the full effect. -48- Table 5.1: Macroeconomic Framework - The Full Reform Scenario (in percent of GDP) 1996 1997 1998 1999 2000 2001 2002 GDP Growth(%) 7.0 1.0 0.0 4.5 5.5 6.0 6.5 Inflation (%) 9.0 9.1 12.0 10.0 6.0 5.5 5.0 Revenue 9.1 9.7 8.1 10.9 12.1 12.9 13.7 Tax 6.5 6.6 5.9 8.0 9.0 9.5 9.9 Non-tax 2.6 3.1 2.1 2.9 3.1 3.4 3.8 Grants 5.5 3.2 2.2 4.1 4.1 3.8 3.5 External borrowing 1.4 1.7 0.6 2.9 3.1 3.2 3.1 Total Expenditure 16.3 13.9 11.7 17.1 18.1 18.9 19.5 Current Expenditure 9.9 9.0 8.3 10.2 10.8 11.4 11.6 Capital Expenditure 6.4 5.0 3.4 6.9 7.3 7.5 7.9 Current Balance -0.8 0.6 -0.5 0.7 1.3 1.5 2.1 Overall Balance (before grants) -7.2 -4.3 -3.7 -6.2 -6.0 -6.0 -5.8 Overall Balance (after grants) -1.7 -1.1 -1.5 -2.1 -1.9 -2.2 -2.3 Domestic financing 0.2 -0.6 0.8 -0.8 -1.2 -1.0 -0.8 Expenditure By Functional Classification Defense and Security 4.8 4.6 4.3 3.7 3.2 2.7 2.6 Non-defense, of which 11.5 9.3 7.5 13.4 14.9 16.2 16.9 Social sectors Health (financed by government revenue) 0.5 0.5 0.5 1.1 1.4 1.7 2.0 Education (financed by government revenue) 1.0 0.9 1.0 1.8 2.2 2.6 2.8 Source: Staff estimates and projections. Better governance would involve: (i) eliminating illegal logging and increasing the revenue from timber royalties and the profit tax from logging companies; (ii) reducing ad hoc import duty exemptions and thereby increasing revenue from import duties; and (iii) improving tax administration, the collection of non-tax revenue by other ministries, and the transfer of these revenues to the Treasury. If governance issues are seriously addressed, the full impact on revenue would be greater than the 1 percent of GDP assumed here (probably in the range of 2.0 - 2.5 percent), but it would be more realistic to assume gradual improvements. The full implementation of the Law on Taxation would also increase revenue by about 0.5 percent of GDP. Further improvements beyond 1999 (estimated at about 0.8 percent of GDP per year) would come mainly from further gradual revenue enhancement through the establishment of sustainable forestry management, through further improvements in governance, and through changes to the Law on Investment to reduce its high degree of generosity (as recommended in Chapter 2). It should be noted that the full benefits of implementing these measures would materialize over the longer term beyond the period under review. A revenue level of 13-14 percent of GDP by 2002 would put revenue in Cambodia at the same level as that of several other Asian countries when they were at Cambodia's level of development. 5.5 In addition to benefiting from growing tax and non-tax revenue, the government's full implementation of structural reforms would engender higher foreign inflows. The current account deficit would widen (in U.S. dollar terms) from US$346 million in 1997 to US$497 million in 2002 (Table 5.2). Cambodia's growing external financing requirements, which also reflect reserve accumulation and debt amortization, would be met by increased borrowing on concessional terms, foreign direct investment inflows, and grants. Thus, there would remain a substant- need for the disbursement of official -49- assistance in the form of grants and concessional loans. The projected increase in these inflows is justified because, under the full reform scenario, multilateral institutions and bilateral donors would be willing to lend their continued support to the country's rehabilitation. In terms of percent of GDP, grants would first increase, and early in the next century would decrease in line with the expected worldwide trends. External borrowing, at concessional terms, would increase to around 3 percent of GDP. In particular, multilateral organizations would support the implementation of structural reforms with budget/balance of payments support loans. Thus, new commitments of official assistance of about US$405 million (excluding IMF assistance which is for international reserve accumulation) per year during 1999-2001 would be required, of which about US$50 million would be budget/balance of payments support to cover some recurrent expenditures--especially for civil service reform and military demobilization. (Table 5.3.) Foreign direct investment inflows would increase steadily at a moderate rate, reflecting the effect of the regional financial crisis. Table 5.2: External Financing Requirements (The Full Reform Scenario) (in million of US dollars) 1997 1998 1999 2000 2001 2002 Est. Projection Financing Needs Current Account Deficit 346 329 397 440 475 497 Reserve Accumulation of NBC 30 101 80 86 86 96 Medium & Long-term Debt Amortization 0 1 4 9 15 16 Total 376 431 480 534 576 609 Financing Resources Official Transfers 171 154 208 223 226 220 Medium & Long-term Borrowing 38 50 96 119 123 132 IMF (net) 0 0 37 33 27 27 Foreign Direct Investment 135 120 140 160 200 230 Other Short-term 32 -10 0 0 0 0 Gold a/ 117 Total 376 431 480 534 576 609 a/ Including government's gold holding of US$117 million (that had been blocked since the Pol Pot regime) recently released by BIS. Source: Staff estimates and projections. Table 5.3: Required New Commitments of Official Development Assistance a/ (in millions of US$) Existing Commitments Required New Commitments CG 1997 1999 2000 2001 BOP/Budget Support 54 55 50 45 Project Assistance 390 355 355 355 Total 444 410 405 400 a/ Excluding IMF assistance which is for international reserve accumulation. Source: Staffestimates and projections. 5.6 The growing resource envelope would allow the government to increase spending, and to bring about the required reorientation in spending, along the lines presented in this PER. Current expenditure, which has been further compressed in 1997 and 1998 because of revenue shortfalls, could be expanded to -50- about 10 percent of GDP in 1999. In later years, it would stabilize at around 11-12 percent. The room for increase in current expenditure would allow for some long-awaited wage increases for the civil servants-- which should go hand in hand with the civil service reform that stalled in 1996-98--and for increases in operations and maintenance, which would prevent the deterioration of capital stock and lead to improvement in the programs and service delivery needed to reduce poverty. Despite these increases in current expenditure, the government would be able to improve the current budget balance from a deficit of 0.5 percent of GDP in 1998 to a surplus of 2 percent of GDP by 2002, thereby contributing to national savings. 5.7 Capital expenditure, currently at 3-4 percent of GDP, could be increased to 7-8 percent, partly as a result of higher foreign inflows and higher domestic revenue, which would make possible a larger locally financed part of investments. The overall deficit would increase to around 6 percent of GDP and would stabilize at that level, but during the period under review foreign inflows would be greater than the deficit, allowing the government to build up its deposit base with the banking system by around 1 percent of GDP per year in 1999-2002. 5.8 Expenditure reorientation would be facilitated not only by the larger resource envelope but also by the peace dividend of military demobilization. Spending on defense and security would decrease slightly in 1998, but greater gains would be possible in the years to come with the effective implementation of military demobilization (see Annex 2). Defense and security outlays would be reduced to 2.6 percent of GDP (of which 1.9 percent of GDP for defense outlays) from this peace dividend of military demobilization and increased revenue, expenditure for the non-defense/security sectors could be doubled in GDP terms by 2002. Thus, spending for health and education (those parts financed by government revenue), taken together, could grow from the current 1.5 percent of GDP to about nearly 5 percent. It should be emphasized, however, that increasing spending and improving allocation alone would not bring about better program and service delivery. To deliver better program and service delivery, the authorities should also address the full range of institutional issues affecting performance as discussed in Chapter 4 and Annexes 3 and 4. 5.9 In sum, this scenario underlines the critical need for enhanced revenue performance, improved governance, and expenditure reorientation as the core of the government's macroeconomic stabilization and structural reform policies. Credible policies, aimed at stabilizing the economy and providing the framework for durable economic growth and at increasing government revenue through consistent tax policies, would also increase foreign aid inflows and enable the government to pursue its expenditure plans. Given Cambodia's current level of economic development, this is the only possible way to reduce poverty and rebuild the country. C. The Status Quo Scenario 5.10 Contrasting with the full reform scenario is the status quo scenario, which assumes that the government would not demonstrate decisive political will to implement a package of far-reaching structural measures and to reinforce macroeconomic stability. The assumed lack of government commitment to structural reform would be reflected in minimal increases in budgetary revenue and only a slight reduction in defense and security expenditures. Moreover, the lack of political will to implement credible reforms would also reflect on the business climate, leading to a lack of growth in domestic, and in particular foreign, investment. As a result, GDP growth would be 2-2.5 percent and inflationary pressures--stemming from central bank financing of the budget--would intensify and keep inflation in the 16-20 percent range (Table 5.4). -51- Table 5.4: Macroeconomic Framework - The Status Quo Scenario (in percent of GDP) 1996 1997 1998 1999 2000 2001 2002 GDP Growth (%) 7.0 1.0 0.0 2.5 2.3 2.2 2.0 Inflation (%) 9.0 9.1 12.0 16.0 18.0 19.0 20.0 Revenue 9.1 9.7 8.1 8.9 9.1 9.2 9.3 Tax 6.5 6.6 5.9 6.5 6.7 6.8 6.9 Non-tax 2.2 3.1 2.1 2.4 2.4 2.4 2.4 Grants 5.5 3.2 2.2 3.0 2.8 2.7 2.5 External borrowing 1.4 1.7 0.6 1.5 1.4 1.2 1.1 Total Expenditure 16.3 13.9 11.7 13.6 13.6 13.6 13.7 Current Expenditure 9.9 9.0 8.3 9.2 9.3 9.6 9.8 Capital Expenditure 6.4 5.0 3.4 4.4 4.3 4.0 3.9 Current Balance -0.8 0.6 -0.5 -0.3 -0.2 -0.4 -0.5 Overall Balance (before grants) -7.2 -4.3 -3.7 -4.7 -4.5 -4.4 -4.4 Overall Balance (after grants) -1.7 -1.1 -1.5 -1.7 -1.7 -1.7 -1.9 Domestic financing 0.2 -0.6 0.8 0.2 0.3 0.5 0.8 Expenditure By Functional Classification Defense and security 4.8 4.6 4.3 4.1 4.0 3.8 3.7 Non-defense, of which 11.5 9.3 7.5 9.5 9.6 9.8 10.0 Social sectors Health (financed by government revenue) 0.5 0.5 0.5 0.6 0.6 0.7 0.7 Education(financed by government revenue 1.0 0.9 1.0 1.0 1.0 1.1 1.1 Source: Staffestimates and projections. 5.11 Under these circumstances, an increase in government revenue would come only from implementation of the VAT in 1999. It is assumed, however, that not all of the expected proceeds from VAT implementation would accrue to the budget, owing to a partial implementation of the VAT, given the government's weak capacity. The increase in tax revenue is therefore assumed to be only about 0.6 percent of GDP. For the subsequent years, tax revenue would broadly stagnate at around 7 percent of GDP, as no additional revenue from enhanced governance and the changes to Law on Investment would be expected. Non-tax revenue would also stagnate, since no policies are assumed to address illegal logging and to enhance revenue from royalties. Thus, revenue would still be about 9 percent of GDP by 2002. It is also expected that, without major improvements in the country's economic situation, several donors would reduce their grants and loans, so that foreign finance would decrease in GDP terms. In particular, multilateral organizations would not provide budget/balance of payments assistance pending the implementation of structural reforms. The total inflow of ODA would gradually decrease from 5 percent of GDP in 1997 to about 3-4 percent by 2002. 5.12 This insignificant revenue growth in GDP terms would prevent the government from improving expenditure patterns. The current budget balance would remain in deficit. As the government would not be able to finance the overall deficit entirely with foreign inflows, it would become necessary to resort to domestic bank financing, which would threaten macroeconomic stability, fueling inflation. Under this scenario, no improvements could be expected in sectoral allocation. Outlays for defense and security would remain high without the effective implementation of military demobilization. This, in conjunction with low revenue, would limit the growth of spending on key sectors such as health and education. -52- 5.13 In sum, this scenario clearly illustrates the shortcomings of the policies during the last two years or so, in particular in terms of revenue collection and forestry management. A lack of credible changes in the government's policies and failure to address the main governance problems in revenue collection and forestry management would prevent domestic revenue from growing while, at the same time, major donors would reduce their assistance to the country. Such a chain of events would render the fiscal situation unsustainable. 5.14. Intermediate Scenario. Between the two extreme scenarios, a host of intermediate cases and outcomes are possible, depending on the extent of the efforts made to address the issues discussed in this PER. One probable scenario is that the government would only implement reform measures minimally affecting powerful entrenched interests. Any effort to raise revenues above the path envisaged in the status quo scenario, as well as any effort to reduce military expenditure, would open up the possibility of increasing and/or reorienting some civilian expenditure categories. Such a policy might lead some bilateral donors to lend more funds or give more grants, taking away some pressures for domestic financing of the budget deficit. A likely outcome would be higher GDP growth and lower inflation as compared with the status quo scenario. D. Downside Risk 5.15 One distinct downside risk is a further deepening of the regional financial crisis. A further deepening of the regional financial crisis would threaten macroeconomic stability and limit the scope for growth over the medium term. Under the full reform scenario, the effects of such a risk could be reduced with prudent macroeconomic policies. In addition, the high degree of dollarization and the fact that the country is still largely agrarian would also help mitigate some of the effects of the regional crisis. Nevertheless, it is likely that foreign direct investment inflows from the region would slow down, which would have a negative impact on growth and employment. A slowdown in tourism would also have a negative impact. Withdrawals of foreign currency deposits could further weaken the already fragile financial system, and a further depreciation of currencies in neighboring countries would erode Cambodia's export competitiveness. In addition, a further slowdown in the economy would have a negative impact on government revenue, which would in turn slow down the reorientation of government expenditure. It is obvious that the impact of the above effects would be all the more dramatic under the status quo scenario. Indeed, a prolonged regional crisis under this scenario is likely to render the economic situation deeply unsustainable over the medium term. E. Conclusion 5.16 The current economic situation poses a serious threat to medium-term sustainability. In order to make a transition toward sustainable development with significant poverty reduction and broad-based economic growth, Cambodia must substantially upgrade the current weak physical infrastructure and the poor quality of human resources. Toward this end, the government must decisively implement the recommendations made in the PER to resolve the interrelated fundamental fiscal and governance problems. As the political and security situations improved significantly, the new government is now in a position to demonstrate clearly strong political will to renew its efforts to reinforce macroeconomic stability and implement structural reforms. The Prime Minister's speech on October 22, 1998 and the subsequent Royal Government of Cambodia's Platform on Second Term, 1998-2003 reaffirm the new government's commitment to implementing the necessary fiscal and structural reforms. While it is encouraging that they are broadly consistent with the recommendations made in the PER, the challenge for the government is to fully follow through with its renewed commitments. -53- Annex 1 FORESTRY SECTOR ISSUES A. Introduction 1. As a potentially major source of public revenues and as a possible model for the fiscal management of other publicly owned natural resources (such as fossil fuels, gems, minerals), forestry merits special attention. Conditions in the forestry sector in Cambodia, however, continue to decline. In 1997, long-standing problems in the sector were compounded by the impact of the regional financial crisis as well as by a political climate that helped undermine progress toward improved governance. While data on illegal logging are inherently uncertain, the severity of illegal logging has worsened over time. The best available estimate is that illegal logging in 1997 totaled between 3 and 4 million cubic meters with an estimated loss to the government, even with depressed wood markets, well in excess of US$60 million.' Moreover, because illegal logging takes no account of forest management standards or environmental safeguards, the true damage from illegal logging is even greater and, if left unchecked, it could devastate the commercial potential of Cambodia's forest resource in less than five years. 2. To examine the leakage of public resources from the sector and also the impact of the regional financial crisis, a simple financial model of the forestry sector, similar to that developed for the World Bank/UNDP/FAO Forest Policy Assessment (1996), is applied. The model, which is calibrated to accord closely with actual 1997 harvest levels, prices and timber theft, suggests that the government stands to lose, in present value terms, over US$576 million over the next 30 years ifpresent trends are allowed to continue. This reflects the projected decimation of Cambodia's commercial forest in five years and diversions from the public budget of an average of over US$150 million per year while commercial operations can continue. If through concerted actions, the government can control illegal logging and revive its efforts to introduce sustainable concession management, annual revenues over a 30-year horizon could average over US$80 million, even starting from a low base for the initial few years because of step-by step development of the concession system and slow recovery from depressed prices. 3. Actions that the government could take to provide stability to the sector have been identified through studies undertaken with World Bank, UNDP, FAO and government of the Netherlands support. These actions include: legislative reform; improved partnerships with the private sector through better concession terms and regulations; the introduction of emergency measures to control illegal logging; and, eventually, liberalization of the log trade. Reforms would be facilitated through fiscal decentralization measures that would give provincial governments a greater stake in sustainable forest management and civil service reform, including salary reform and the restructuring of public forestry agencies, so as to minimize incentives for corruption in forest administration and the transport of forest products. 4. This Annex summarizes data on selected issues facing forestry,' discusses its potential as a source of public revenues and the impact of the regional financial crisis on the sector, and summarizes investment and policy decisions now available to the government. I This is based on a rough estimate, provided by consultants (DAI) to RGC, of a merchantable volume in the commercial forest area of approximately 20 million cubic meters. 2 This is not an exhaustive treatment of the forest sector. For greater depth on conservation of biodiversity and on protected areas management, see the National Environmental Action Plan. For background on the industrial sector, see World Bank/UNDP/FAO, Cambodia: Forest Policy Assessment, 1996. A study is under preparation by the government, with support from the World Bank. Much of the material presented in this paper is based on output related to this work. -54- B. Issues and Progress to Date 5. Despite its significant potential, forestry continues to make a very limited contribution to sustainable development and poverty alleviation in Cambodia. Following on the World Bank/FAO/UNDP Forest Policy Assessment, the Royal Government of Cambodia (RGC) commissioned four technical studies aimed at producing detailed proposals on forest law enforcement, assessment of concession contracts, forest concession management, and ways of improving the overall development performance of forest policy. This work has identified specific policy and implementation measures which the RGC could now begin to undertake. 6. Forest Revenues and Illegal Logging. Work to date continues to support the contention in the Assessment that the long-term solution to illegal logging is full scale development of a properly regulated concession system and a managed network of parks and protected areas. However, in view of the urgency of the situation, special measures are needed and could be managed under a special interagency arrangement. This would include implementation of a routine monitoring and surveillance system, deployment of specialized suppression measures, and institution of public education and information campaigns. Detailed recommendations and proposals are available from the government's consultants. 7. Forest Law and Concession Contracts. In 1996, the Council of Ministers submitted to the National Assembly draft legislation on forestry. A Bank-financed review of this legislation by the government's advisers found that the draft (i) did not resolve the confusion that results from the patchwork of existing laws and regulations concerning forests; (ii) was not suited to the practical aspects of forest management in Cambodia; (iii) did not provide procedures for the award of harvest rights: and (iv) appeared to contemplate regulatory controls that are beyond the scope of RGC. Following these suggestions, the government withdrew the draft with a view to resubmitting a revised draft withmajor revisions that would link the legislation to a defined policy, clarify responsibilities of specific RGC agencies, and lay the basis for consistent implementation. Detailed redrafting will require assistance from international expertise. 8. A critical issue is the disposition of Investment Agreements and long-term Timber Licenses (concession contracts) issued by the RGC, committing nearly all of the country's commercial forest area to approximately 30 concessionaires. Legal advisers have concluded that although the contracts are badly drafted and seriously biased against the RGC's interests, the government could best move forward with a case-by-case evaluation of concessionaire performance against contractual obligations. These technical and financial evaluations could lead to individually tailored, bilateral renegotiation of contract terms and, in some seriously flawed cases, termination of the contracts. Legal advisers have provided a draft "model concession contract" and will provide basic training to staff on contract law applied to forestry. International technical assistance in the form of specialized business representatives would be required to assist the RGC in contract dispute resolution (draft terms of reference have been supplied by the legal consultants). 9. Forest Concession Management. The RGC, assisted by forest concession management consultants, is well advanced in designing detailed revisions to the entire system of concession planning, operation, and regulation. Proposals are being developed in consultation with key concessionaires and Department of Forestry and Wildlife (DFW) staff, and are building on inventory methods already established in the DFW under the UNDP-financed, FAO-executed Forest Inventory Project. A draft Code of Forest Management Practice has been prepared and arrangements have been made for DFW staff to work with concessionaires on management plan revision as a step toward bringing operations into accordance with the new requirements. -55- 10. Full adoption of the revised concession management system would require issuance of appropriate regulatory instruments as well as investment in equipment and vehicles, staff training, and technical assistance (a portion of concession supervision could be contracted to third party monitors). Implementation would begin on concessions already targeted with expansion to new areas depending on the rate at which contract disputes can be resolved and the possible pace of capacity building and contracting for additional services. I1. Forest Policy. All these recommendations relate to the need for fundamental changes in how forest policy is made and implemented in Cambodia. Beyond the specific measures discussed above, analysis has reinforced the need to improve governance in the sector and has also assessed the ways in which stakeholders are involved in forest management and forest policy (see Box 1) and ways to improve compensation and working conditions for government foresters. Analysis has also highlighted the inadequacy of land allocation procedures, indicated the need to find ways to employ forest resources to support the demobilization and reintegration of military forces, and suggested the scope for using forest revenues to support decentralization. Box I. Village Livelihood Impacts of Industrial Forest Concessions--The Need for Reform A study by DFW staff in Kampong Thom province dramatically illustrates the need for changes in the way commercial forests are utilized in Cambodia. Properly introduced, these reforms can be beneficial to local communities, concessionaires, and the national Treasury. In a concession operated by an international investor, the study documents how the concession's operations have deprived villagers of road access, of the traditional use of the forest for resin and gum collection, and of the harvest of small, non-commercial trees for firewood and other local uses. The concessionaire even bulldozed the area around the village pagoda and used the area as a log dump. As relationships deteriorated, the company seized ox-carts and animals being used by villagers to collect logs in the concession area, and in another incident ox-carts were burned and livestock killed. The deputy chief of the commune is quoted as saying, "I was unable to help my people and my guess is perhaps that the burning of the company trucks is related." In a neighboring concession, while relations are not perfect, a greater effort has been made by the concessionaire to give villagers access to and use non-timber forest products. As a result, violence has been avoided and commercial activities are proceeding unimpeded. These observations provide the basis for specific recommendations for forest management regulations and demonstrate to DFW staff the need for proper regulation of concessionaires and the involvement of local people in land allocation and forest management. Source: Executive Secretariat of the National Committee on Forest Policy in association with ARD, "The Study of Effects of Forests on Rural Livelihoods in Kampong Thom Province, Cambodia, " February 1998. C. Weak Governance and the Effects of the Regional Financial Crisis 12. Cambodia's position as a relative latecomer to the international timber market and its relative lack of a coherent long-term strategy have left it susceptible to shocks arising from financial dislocation in East Asia. Thus, the onset of regional financial crisis in the second half of 1997 temporarily erode the average stumpage value of Cambodian timber from more than US$70 per cubic meter to less than US$14 per cubic meter, and contributed to the acceleration of illegal logging to the level of 3-4 million cubic meters. These results are the cumulative impact of the crisis working through several pathways: (i) a general weakening of prices owing to contraction in key export markets (particularly Japan, Korea, and Thailand); (ii) crisis-driven efforts to increase exports by Cambodia's more established competitors (primarily Malaysia and Indonesia); and (iii) the effect of the overall flattening of economic activity in the domestic economy, which put greater pressure on weak sector governance to allow fellings to expand beyond controlled or sustainable levels. -56- 13. Timber and Sawnwood Prices. World markets for timber, sawnwood, and wood based panels are increasingly integrated and are highly responsive to developments in the global economy and to policy actions by major market players. Prices for tropical logs from Asia reached their peak in 1993 and started to decline in part because of the relaxation of restrictions that Malaysia had introduced on log exports. This downward trend, from a peak of nearly US$400 per cubic meter,' brought prices down to less than US$300 per cubic meter in 1995 (in nominal terms). This trend accelerated in 1997 because of currency depreciations in the region and the collapse of demand. Malaysian log prices fell 11 percent between the third and fourth quarters, and by December 1997 prices were below their December 1996 levels. 14. Prior to the onset of the financial crisis in the second half of 1997, recession in Japan had already begun to lower the demand for timber from Southeast Asia. Housing starts in Japan were down by more than 20 percent in the last quarter of 1997 compared with the previous year. The currency depreciations in Southeast Asia have also had a further strong negative impact on timber demand in the region. Recorded imports of timber products by Thailand and the Philippines, the two largest net importers in the region, have nearly come to a standstill. 15. While the weakness of forestry sector administration means that detailed and highly reliable data are not available on wood prices in Cambodia, RGC studies provide some indication of the development of prices and markets. In 1996 the World Bank estimated that the average stumpage value (the economic value of standing timber calculated from the border price of wood less all costs of felling, processing, and transport) was approximately US$70 per cubic meter. By the 1997-98 logging season, however, prices had deteriorated in line with the global and regional trends discussed above. Based on the best information available, by early 1998 Cambodian stumpage values had deteriorated to below US$40 per cubic meter. 16. Market Responses. While comprehensive data are not available, industry observers report that log exporters in Indonesia and Malaysia have responded to the drop in prices by attempting to accelerate exports. This has been accentuated in the case of Indonesia by the lowering of export duties on logs from the prohibitive level of taxation that had replaced the Indonesian log export ban approximately one year ago. At the same time, the construction demand for wood in Thailand appears to have stalled in 1998. The Vietnamese demand appears to continue to be strong owing partly to the need for reconstruction following storms in 1997. 17. Impacts in Cambodia. In Cambodia, these price trends have combined to place strong downward pressures on wood prices. The estimated residual stumpage value of Cambodian logs given above is based on conventionally measured costs of production. However, much of potential for increasing government revenue lies in re-channeling private and unofficial rent capture. Ruzicka (1998) offers a number of estimates of stumpage prices in Cambodia over the 1997-98 period. His estimates are differentiated mainly by the extent to which explicit attention is given to unofficial charges (bribes and other charges) and to the possibility that extra costs might be associated with achieving sustainable forest management. While the latter is open to question because of strong evidence from many settings that sustainable practices are actually cost reducing, the entrenched practice of extrabudgetary charges is an important consideration with respect to the behavior of logging interests and to the prospects for increasing government receipts from forestry, especially in the short term. For meranti logs delivered to Japan. 4 Taxation and Policy Reform in Cambodian Forestry, Working Paper no.1, ARD Forestry Policy Reform Program. -57- 18. Evidence of Unofficial Charges. Table I illustrates the problems associated with unofficial payments and the reach of these practices. For a shipment of logs from the forest landing in east central Cambodia to processing facilities near Phnom Penh, at least eight supplementary payments totaling US$14 per cubic meter (or 11 percent of the CIF value of the shipment) are needed on top of official royalties5 and unofficial charges associated with the original felling operation. The largest single charge, US$7 per cubic meter (or 5.3 percent of the CIF value) is paid to provincial authorities and is widely recognized as a significant source of resources for regular provincial operations. The second single largest charge is to the local military US$1.70 (or 1.3 percent). In total, extrabudgetary charges by forestry personnel amount to somewhat more than 4 percent. Total additional charges (often called facilitation fees), including those directly associated with the felling operation, amount to approximately US$50 per cubic meter. Logging operators in Cambodia claim that these charges, on top of official royalties and other taxes, currently make profitable operation impossible. These levels of side payments and extrabudgetary charges are indicative of weak governance in public resource management which has been exacerbated by the stresses in the Cambodian economy as a result of the financial crisis, the political events of July 1997, and the ensuing losses of financial support from various donors. Table 1: Selected Costs Associated with an Internal Log Shipment of 4,000 Cubic Meters from a Province in East Central Cambodia to Phnom Penh, 1997-98 (in US dollars) Per Cubic Meter Shipment Total Percent of Value Estimated Value (delivered) 130.00 520.000 100 Provincial Authorities 7.00 28.000 5 Military Authorities 1.70 6,800 I Provincial Forestry Officers 0.88 3.500 1 Foresters at Checkpoints 2.75 11.000 2 Foresters at Intervening Provinces 0.50 2.000 0 Forestry Officials at Final Destination 1.38 5.500 1 Total Extrabudgetary Charge 14.20 56.800 11 Source: ARD report. 19. Evidence of Illegal Timber Trade. Even more serious is the general breakdown of law and order that appears to be taking place in the forestry sector. Table 2, based on work conducted for the National Steering Committee on Forest Policy, illustrates the estimated distribution of illegal activity in the forestry sector by province and by type of infraction. While the illegal activity is spread across the country, it is interesting to note that nearly half of all illegal activity arises in two provinces (Koh Kong and Kratie) and nearly three-quarters in only five (Koh Kong and Kratie plus Battambang, Rattanakiri and Kampong Speu). In Koh Kong alone nearly 95 percent of illegal activity is the processing of illegally felled logs for exports.' Data and analyses of this kind are first step toward the design of an effective and targeted law enforcement campaign. The government has decided to increase the official royalty from US$14 per cubic meter to US$54 beginning January 1, 1999. Realization of this remains to be seen as it could involve re-negotiating concession contracts. This suggests that the government's log export ban is having an effect, but not an overall impact, on illicit fellings. -58- Table 2: Estimated Timber Trade by Province of Origin and Category, 1997-98 (m3 Roundwood Equivalent) Province Concession Collection of Illegal Domestic Illegal Sawnwood Total Total Roundwood Operations illegal Fellings and Export Domestic and Roundwood Equivalent (%) in Log Form Exports Equivalent (Roundwood Equivalent) Rattanakiri 447 300.000 12.800 313.247 7 Stung Treng 100 110.000 96.000 206.100 5 Kratie 105.900 183.800 140.000 192.000 621.700 15 Mondulkiri 14.500 19.200 33.700 1 Kampong Thom 68,400 25.200 7S,000 57.556 226.156 5 Kampong Cham 45000 66.667 111 L667 3 Kampong Speu 10.000 240.000 250.000 6 Kampong Som 15,400 128.000 143.400 3 Kampot 102.222 102.222 2 Koh Kong 58.300 300 15.000 1.392.000 1.465.600 35 Battambang 250.000 200.000 450.000 Bantev M. C. 88.889 88.889 2 Pursat 70.000 96.000 166.000 4 Preah Vihear 65.000 0 65000 2 Siem Reap 0 0 0 Others 3.600 1 0 3.600 0 Total 248,447 213.000 1.094.500 2.691.333 4,247,280 i00 Percent of Total 6 5 26 63 100 D. Revenue Potential Under Alternative Scenarios 20. With a view to contrasting the revenue potential from forestry under different policy implementations, a simple financial model of the Cambodian forestry sector is developed. The model provides a way to track sectoral performance and compare different scenarios according to several key indicators, including harvest level, operated area, and government revenues. Initial data on the forest resource, prices, price trends and costs, and the rate of harvesting are extrapolated by the model to yield, for different policy scenarios, internally consistent predictions of the various performance measures. In other words, for example, harvest levels are fed back through the model to ensure that future harvests are consistent with future timber availability. The model is not an optimization model and is adjusted in several respects by assumptions on the pace of the growth of harvesting. 21. The model is applied for three alternative scenarios over the period of the next 30 years: (i) a continuation of the current levels of illegal logging and minimal revenue collections from authorized concessions and illegal log "collections"; (ii) full implementation of the currently conceived concession system (based on essentially fixed royalties of US$14 per cubic meter); and (iii) an improved concession policy framework, with phased introduction and royalties based on border price equivalents (see Table 3). The model calculates revenues on the basis of price projections, the level of logging activity generated by government policy, and initial estimates of the area under management and the possible pace of expansion. It uses data on the Cambodian forest resource, the current levels of illegal logging, and recent prices and their prospects for recovery. The model employs price estimates based on current international market wood prices adjusted to stumpage values by subtracting transport, harvest, and other costs. Prices are projected forward into time on the basis of World Bank projections for tropical logs. Bank projections forecast recovery of the timber market from the effect of the regional financial crisis and the resumption of a long term upward trend in real prices. Because the Bank -59- Table 3: Assumptions in the Forestry Revenue Projection Model Parameter Base Case Value Basis of Estimate Annual Allowable Cut 500,000 cubic meters See text Royalty (existing contract rate) US$14 per cubic meter Concession contracts Royalty (economic value) USS40 Ruzicka (1998) Source: Staff estimates. 22. For 1997, model results closely parallel actual results in the sector, with total fellings at approximately 4.3 million cubic meters, of which a total of around 11 percent is officially or quasi- officially sanctioned. Government royalty revenues come to approximately US$10 million (somewhat more than the US$7 million now estimated for the 1997-98 season). Extrabudgetary flows, which derive more closely from international prices than do official royalties, amount to approximately US$70 million. 23. To compare a continuation of this situation, the model projects the Cambodian forest into the future requiring consistency over time between extractions and future harvests, and compares the results to two alternatives: full development of the concession system under (i) the current set of contracts and prices and under (ii) a full rent recovery policy. The full rent recovery policy would involve open trade, market-based royalties, and harvests held to a sustainable level as recommended by the World Bank, the UNDP, and the FAO (1996). Under current conditions, overexploitation has lowered the value of the forest and is projected to remove some 25 million cubic meters, eliminating the commercial potential of the forest in about five years (DAI report, 1998). Under the concession as planned scenario, production is assumed grow to a peak level of extractions equal to approximately 3.6 million cubic meters per year in 2001 and to tail off to a depressed level of extractions of approximately 350,000 cubic meters for the balance of the 30-year planning horizon. It is likely that the impact of the high level of logging implicitly proposed for the current concession program would, by "high-grading" the forest, result in even greater reductions in the long term potential of the forest, making the model results very conservative in terms of its projected benefits to forest policy reform. 24. For the full rent recovery scenario, the single most critical parameter in the model is the estimated annual allowable cut (AAC). The AAC is intended to be the maximum level of extractions that can be sustained into the future. In forests with even age and size distributions, the calculation of AAC is relatively simple, provided that data are available on forest areas and stocking. In Cambodia, where data are scarce and limited, and where the structure of the forest is badly skewed because of the legacy of bad management, there may not be a single or unique harvest level that can be maintained over time. It is more likely that the desirable harvest level will vary in response to the gradual evolution of the uneven structure of the forest. This is partly subject to modification through the application of silvicultural interventions (for example, enrichment planting, thinnings) and needs to be defined operationally on the basis of detailed plans based on inventories and site assessment. On a national basis, the key variables affecting the approximately desirable harvest level in the near future are the areas and conditions of the various forest types shown in Table 4. projects especially strong recovery in timber markets, the rate at which prices are projected to recover has been reduced by 75 percent. In the model, this translates into recovery from stumpage of US$30 per cubic meter at present to pre-crisis levels in four years increasing to above USS200 per cubic meter by 2010. To permit comparison of revenue streams over time with quite different compositions, financial values are discounted at a rate of 10 percent. -60- Table 4: Forest Cover in Cambodia Forest Type Area Percent of Percent of Forest Area Land Area Evergreen Closed 634.869 6 3.5 Evergreen Disturbed 3.323.527 3 18.3 Evergreen Mosaic 135.632 1 0 8 Semi-Evergreen Closed 111.842 1 0.6 Semi-Evergreen Disturbed 1.184,689 11 6.5 Semi-Evergreen Mosaic 97.905 1 0 5 Deciduous 3.777.678 36 20.8 Deciduous Mosaic 274.553 3 1 5 Regrowth 544.778 5 3 0 Swamp Regrowth 23.591 0 0.1 Swamp 222.150 2 1.2 Mangrove 77.260 1 0.4 Bamboo 21.823 0 0.1 Swamp Mosaic 105.465 I 0 6 Total Forest 10.535.762 100 58.1 Note: Totals may not add owing to rounding. Source: Department of Forestry and Wildlife. Inventory Office, data from Mtekong River Commission Forest Cover Monitoring Project. 25. With respect to the determination of an approximately sustainable harvest level, two forest types, Evergreen and Deciduous, account for more than half of the total forest area, and the Deciduous forest has limited commercial potential. Of the Evergreen forests, the great bulk falls into the Disturbed category and it is here that the worst excesses of illegal and poorly control logging have taken place. Evergreen Closed forests, which are believed to be well stocked, however, are in total only 6 percent of total forest area and all of this area may not be suitable, for environmental and other reasons, for commercial exploitation. The allowable cut depends critically, as well, on the condition of the Disturbed Evergreen area, which could range from severely overlogged to areas with adequate regeneration. Assuming the area of Evergreen forest could be reasonably well distributed and its regeneration potential protected, the total area of approximately 4 million hectares with a growth of approximately 0.3 cubic meters per hectare per year would yield an allowable cut in the range of 1-1.2 million cubic meters. If, however, as appears to be the case, uncontrolled logging has significantly disrupted the structure of the forest and delayed regeneration, a sustainable yield could be significantly lower. Consequently, for this calculation it has been conservatively assumed that the annual allowable cut will be on the order of 500,000 cubic meters per year (under the assumption that controlled harvest would rise from 0 to 500,000 cubic meters over six years). This is equal to only 13 million cubic meters over the entire planning horizon, allowing what should be adequate regeneration to ensure sustainability. 26. The full rent recovery scenario generates a present net value return to the budget of US$631 million compared with only USS75 million under the continuation of the current situation. It should be noted that the revenue enhancement under the full recovery scenario would be gradual as this scenario assumes environmentally sustainable harvesting and step-by-step implementation of medium-term recommendations to be made by the four studies. -61- Table 5: Revenue Potential from Forestry under Alternative Scenarios Rent Present Value Average Average Average Recovery of Revenue Annual Extrabudgetary Annual as % Scenario Ratio (30 years) Government Flows of Total 1994 1/ (Percentage) Revenues (until 2003) Government Mn. $ Mn. $ Mn $ Revenues 1: Continuation of Current Situation 12 75 3.2 151 2 2: Full Implementation of Existing System 39 245 180 - 9 3: Full Rent Recovery Policy 100 631 824 - 41 I/ Share of timber revenues was 17 percent in 1994 which was an all time high. Source: Staff estimates. E. Recommendations and Assistance Needs 27. The government must take the following "emergency measures" immediately to arrest the deteriorating situation: * Stop granting new concessions. * Freeze granting approval for all new investments in wood processing. * Establish mechanisms for the monitoring and prevention of illegal log felling and exports. * Enforce cancellation of all permits for "collection" of logs and stop issuing new permits. * Strengthen the Forestry Department, with a clear mandate and responsibilities. 28. Over the medium term, the government must implement the recommendations of the studies assisted by the Bank technical assistance, which are essentially completed and which embody the framework suggested by the Bank/UNDP/FAO and by the government's own forest policy consultants. * Log Control and Monitoring. The current forest law enforcement crisis justifies a special effort to control illegal logging, increase revenue collection, and move toward a more efficient log trade regime. A set of protocols for assessing the scope, spatial distribution, and character of illegal logging has been tested and is being developed for transfer to the RGC. Government has recently issued orders calling for cooperation by the military and police in controlling illegal logging and several seizures of equipment of logs have been published. These need to be reinforced with budgeted program. Proposals for log and logging control are being developed as are recommendations for public awareness and education and government staff training. * Legal Study. Concession contracts are badly biased against the government, but do specify performance obligations which are likely to form a basis for a structured and legally defensible challenge to individual contracts on a case-by-case basis. A draft model contract has been prepared, comments on the draft forest law have been submitted, and a training program has been conducted. Forest legislation needs to be resubmitted to the National Assembly following revision. An earlier draft did not provide an effective platform for forest exploitation and risks perpetuating the current system of arbitrary allocation of forest resources and also does not resolve key institutional issues (for example, provincial-central lines of authority). -62- * Forest Concession Management. Concessions can provide the basis for sustainable sectoral development provided a satisfactory policy framework is put in place, regulations are strengthened, and security and technical guidelines are enforced. A system of planning, harvesting, and revenue systems will be provided and, following review and consultation, will need to be formalized through appropriate legislation or regulation and implemented through pilot projects. * Forest Policy. Elements of the government's forest policy are at odds. Pressures for immediate revenues are now overwhelming concerns about sustainability, strong measures and commitment will be needed to redirect privately appropriated rent to the public treasury, land allocation is chaotic and arbitrary, community involvement in forest development is severely limited, and public statements are inconsistent. 29. To reach these medium-term objectives, the government must take the following actions over the next few months; prepare a proposal for illegal logging control; prepare the illegal logging monitoring program; undertake a performance review of concessions, and prepare the Cambodia Code of Forest Practice (CCFP) which is essentially an operational manual that lays out the rules for concession operations. In the subsequent few months, these actions should be followed by: the preparation of revised forest law; decisions on the termination or renegotiation of concession contracts; the issuance of a sub-decree establishing the operational manual for concession management; the strengthening of the implementation capacity of forestry management; and the strengthening of the law enforcement capacity regarding illegal logging. 30. Assistance Needs. The immediate assistance needs consist of four main activities: (i) developing improved forest legislation; (ii) establishing special emergency measures to control illegal logging; (iii) facilitating the resolution of contract disputes; and (iv) strengthening RGC supervision of forest management on selected concessions. Basic guidelines for these activities are already available from study results, although some additional effort is needed to cost out specific inputs. The DFW Secretariat, which is now managing the technical assistance studies on behalf of the RGC would need to be strengthened to serve as a project management unit for the implementation phase. 31. Preconditions for assistance could include requiring commitments by the government to eliminate logging authorizations outside of concession areas, to use the draft model concession contract, to revise the draft forest legislation, and to consider replacing the log export ban with a temporary, presumptive tax on log and sawnwood exports.' Commitments would also be sought from select concessionaires to undertake to manage in accordance with draft management plans prepared under the ongoing studies to formalize the proposed system for forest concession management through an appropriate government order. 32. Further development of this policy framework would also create a need for, and would allow for the effective development of, substantial incentives in the regulatory and operational capacity of the DFW. This would have to involve substantial investments in physical infrastructure (new and rehabilitated field offices), equipment (transport, communications, field equipment), staff and forest worker training, and planning and management systems. Agreement in principle has already been indicated by RGC on most of these points. -63- Annex 2 PEACE DIVIDEND ANALYSIS OF MILITARY DEMOBILIZATION A. Introduction 1. Background Cambodia's two recent attempts at demobilization have not materialized. In 1993., the four fighting factions were to unite and demobilize from a unified army under UNTAC supervision. The eventual refusal of the Khmer Rouge to participate in this program led to incomplete demilitarization and in fact prolonged the armed conflict. In 1995-96, the government prepared the comprehensive Cambodia Veterans Assistance Program (CVAP) which also foresaw the reintegration into civilian life of several thousand Khmer Rouge defectors. Owing to the political turbulence that led the July 1997 events. the CVAP has yet to be implemented. 2. As a result of these two aborted demobilization attempts, defense continues to absorb a large portion of government resources, thereby preempting developmental expenditures. Table 1 clearly demonstrates the very high portion of defense expenditures compared with other countries. Table 1: Central Government Budget - International Comparison 1/ Country Defense (%) Social (%) Def./soc. Country Defense (%) Social (%) Def/soc. Cambodia 44.4 21.7 2.04 Cameroon 9.4 29.0 0.32 India 14.5 11.9 1.22 Mongolia 11.5 38.8 0.30 Syria 28.2 26.8 1.05 Malaysia 12.7 48.0 0.26 Yemen 30.3 30.8 0.98 Kenya 6.2 27.1 0.23 Turkey 15.8 21.6 0.73 Albania 7.1 40.2 0.18 Philippines 10.6 26.3 0.40 Ghana 4.9 42.3 0.12 Guatemala 15.2 38.4 0.40 Indonesia 6.2 70.4 0.09 Croatia 21.1 60.9 0.35 Sri Lanka 2.6 46.2 0.06 1/ For other countries, data are for 1995 or the most recent year: for Cambodia. data are for 1995. Source: World Bank World Development Report 1 997. 3. Objective. The objective of this Chapter is twofold. First, defense expenditure is analyzed in relation to key economic and financial data to assess the importance of reining and rationalizing defense expenditure in economic development, thereby identifying a financial rationale for demobilization. Second, the dimensions and financial benefits of implementing the CVAP are investigated under different demobilization scenarios, from a comparative static and a dynamic vantage point. The aim is to assess the long-term sustainability of military restructuring relative to economic development. 4. Data Sources. Accuracy in military expenditure data in developing countries is almost always an elusive goal. Military investment often bypasses the official budget. The size of the army is frequently overstated by the military so that certain individuals can benefit from payments to "ghost soldiers". Where average army wages are low, such overstatement may be a deliberate decision by unit commanders to enable them to more properly feed and house their troops. In Cambodia, the government budget does not include military investment data, and no attempt is made here to trace any such possible expenses. This chapter deals with recurrent data only. 5. Three main sources are used for analysis: (i) staff estimates and projections of GDP and other macroeconomic data based on data provided by the government; (ii) the government's official budget data; and (iii) data on army size and remuneration by rank from the Royal Cambodian Armed Forces -64- (RCAF). Not surprisingly, deviations--sometimes substantial--arise during the comparison of data analyzed from these three sources. 6. Particular uncertainty pertains to the size of the army, which is officially estimated in early 1998 to have been over 138,000. However, the identification and registration process planned under the CVAP has not yet been initiated. Hence, actual numbers may be higher (in particular when local militia have been included) or lower. As a result of such data limitations, this discussion is not intended to be precise in As estimations. Rather, it attempts to give indications of the order of magnitude as much as possible. B. Analysis of Defense Expenditure 7/ Throughout the 1994-97 period, recurrent defense expenditure decreased in real terms, from CR 392 bilion in 1994 to CR 305 billion in 1997 (in 1997 prices). This decrease is almost entirely due to the reduction in non-wages, whereas wages essentially remained constant at an average of around CR 200 billion. Between 1994 and 1997, non-wages declined from CR 194 billion to CR 96 billion. In relative terms, the share of wages and non-wages in the recurrent defense budget changed from 49.5 percent each in 1994, to 66.5 percent for wages and 31.4 percent for non-wages in 1997. This reduction reflects the RCAP's intention to maintain, to the extent possible, the troops while suspending many administrative functions (Table 2). Table 2: Recurrent Defense Expenditure (millions of riel in 1997 prices) 1994 1995 1996 1997 Salaries and indemnities 194,103 223,143 197,196 202.567 Administration 194,008 153,763 114,489 95,556 Public interventions 3,728 4,446 6.545 6.620 Total 391,8391 38l.352 318,230 304,743 Source: Data provided by the government. 8. The share of national defense in the government's recurrent expenditure has been decreasing fairly steadily. National defense accounted for 37.3 percent of total recurrent expenditure in 1997, down from 46.4 percent in 1994. Combined, public security and national defense reached a staggering 58.6 percent in 1994, and remained above 50 percent throughout the period. Over 1994-97, the share of national defense in GDP decreased from 5.0 percent in 1994 to 3.3 percent in 1997. 9. in 1997 the share of social and economic services in the government's recurrent budget reached 23.5 percent and 7.5 percent, respectively. Combined, they reached only about 80 percent of defense spending that year, up from almost two-thirds in 1994. Similarly, defense expenditure in real terms per person decreased from around CR 40,000 in 1994 to CR 29,000 in 1997 (Table 3). Yet, defense's importance remains striking. Comparing recurrent defense expenditure to the government's fiscal situation, recurrent defense expenditure represented, on average, over 60 percent of the overall deficit during the 1995-97 period. (It should be noted that these expenditures included some spending of social and economic nature such as allocations for the families of defected Khmer Rouge soldiers and infrastructure rehabilitation for them.) Overall, total recurrent expenditure decreased from CR 861 billion in 1996 to CR 816 billion in 1997. Noteworthy in this respect is the nominal increase of 4.4 percent in 1997 in defense expenditure when social services and economic services decreased by 3.8 percent and -65- 16.9 percent, respectively, reflecting the events of July 1997 and the subsequent increase in the officially estimated size of the RCAF from 125,000 in 1996 to 138,000 in 1998.1 Table 3: Defense and Social/Economic Services 1994 1995 1996 1997 Defense/GDP 5.0% 4.5% 3.6% 3.3% Social Services/GDP 2.3% 2.2% 2.4% 2.1% Economic Services/GDP 0.8% 0.7% 0.9% 0.7% Defense/Social Services 2.2 2.0 1.5 1.6 Defense/Economic Services 6.2 6.2 4.0 5.0 Defense/Social and Economic Services 1.6 1.5 1.1 1.2 Real Per Capita Defense Expenditure (riel) 39,984 38.135 31.199 29,023 Source: Staff estimates. 10. Defense wages andNon-wages. Army wages include a salary and a variety of indemnities: (i) a 20 percent bonus on the base wage, which differs by rank; (ii) monthly child and spouse allowances; (iii) a daily food allowance; (iv) a monthly rice ration; and (v) an annual clothing allowance. All allowances and rations are independent of rank. A socioeconomic profile developed in 1995-96 in the preparation of the CVAP determined that a soldier has on average 2.8 children and that two-thirds of soldiers are married. With these assumptions and using RCAF data, per capita allowances and rations, monetized, amount to CR 805,500 annually. 11. This amount compares to median salaries (including bonus) of CR 820,800 for generals and CR 300,960 for privates. Because of the substantial allowances and rations, the official wage structure is fairly egalitarian, with median annual wages ranging from CR 1.6 million for generals to CR 1.1 million for privates. The total wage bill amounts to an estimated CR 163 billion, leading to an average wage of 1.2 million annually.2 This compares with an estimated CR 867,000 per capita GDP in 1997. The average family size as determined by the socioeconomic profile was 4.5. If a soldier was the sole breadwinner in a family, his wage would be abysmally low and completely insufficient to cover daily needs. I An estimated 20,000 Khmer Rouge fighters have joined the RCAF since the short civil war in 1997. The number of FUNCINPEC soldiers currently in exile or in the bush as well as the remainder of the Khmer Rouge is not known. Hence, the figure of 138,000 may be more an indication of the potential maximum size of the RCAF with all groups incorporated, rather than its actual strength. 2 Comparing budget and RCAF data leads to an interesting observation. Using RCAF data, the wage bill amounts to CR 163 billion whereas salaries and indemnities as reported in the budget amounted to 203 billion in 1997. The budgeted average wage is thus CR 1.5 million, or one-quarter higher. Possible explanations for this difference, which cannot be clarified from either source, may include: the inclusion in the budget (but not in the RCAF) of local militia, a change in the socioeconomic profile of RCAF soldiers, an underestimate of the official RCAF size (up to a maximum of 34,100), and/or the diversion of funds to other uses (up to a maximum of CR 40 billion). Using budget data it is also interesting to note that the average wage decreased from CR 1.6 million in 1996 to approximately 1.4 million in 1997, given the absorption of defected Khmer Rouge fighters. -66- C. Cambodia Veterans Assistance Program 12. The CVAP foresaw the demobilization in three phases and the subsequent reintegration 'to civilian life of 40,000 RCAF soldiers as well as the reintegration of 3,000 additional Khner Rouge defectors. The program included the provision of a transitional safety net ("departure allowance") fo- 'he most critical reinsertion phase as well as support for their social and economic reintegration, with a ocus on the disabled and chronically ill. The total departure allowance was calculated at CR 2.25 mi.iion per soldier (US$900 at 1996 exchange rates), irrespective of rank or years of service. On a per capita basis, these US$900 would have placed a veteran's family in the upper fourth decile of the country s Income distribution scale. Total program costs were estimated at US$72.7 million. The government remains committed to the CVAP as a sign of both demilitarizing society and restructuring public expenditure toward productive purposes. 13. Scenarios. Using the parameters of the CVAP, the program would reduce the RCAF by 43,000, from 138,500 to 95,500 (Scenario 1). The current rank structure is biased in favor of higher ranks; hence, demobilization would reduce the number of generals and commissioned officers by almost 29,000. In comparison, the ranks of non-commissioned officers and privates would be reduced by only 14,000, increasing their share in the post-demobilization RCAF to almost three-quarters. However, the events of 1997/98 led to the defection of Khmer Rouge fighters and their subsequent incorporation into the armed forces. Hence, the CVAP requires updating, especially regarding the number of soldiers (and defectors) to be demobilized. For illustrative purposes, therefore, an additional analysis was undertaken, assuming the demobilization of 78,300 soldiers, to reach a force of 60,200 (Scenario 2). The calculations for both scenarios and methods assume that wage and non-wage costs per soldier do not increase in real term as a result of demobilization. 14. Methods. The budgetary impact of both scenarios is calculated using two methods: (i) a static analysis and (ii) a dynamic analysis. The static analvsis presents a hypothetical case which assumes that other government expenditures remain constant. It determines the effect of military downsizing if the CVAP were implemented within one year. Though unrealistic, the static analysis provides an illustration of the budgetary savings had demobilization occurred in 1997. The dynamic analysis traces the impact over time to assess the long-term sustainability of military restructuring.As planned in the CVAP, both scenarios are calculated using three phases of demobilization over a three-year period. The assumed demobilization calendars are as follows: (i) 15,000 in 1999, 15,000 in 2000, and 13,000 in 2001; (ii) 25,000 in 1999, 25,000 in 2000, and 28,300 in 2001. In both cases, it is assumed that the lower ranks would be demobilized first, and that generals and commissioned officers would not be demobiEzed before 2001. Furthermore, each phase of demobilization is assumed to be undertaken at the beginning of the respective year. D. Static Impact Analysis of Military Demobilization 15. Under Scenario 1, other things being equal, after completion of the demobilization exercise. annual savings in the government's wage bill wvould amount to CR 52 billion, using RCAF data. Reducing non-wage costs proportionaily (i.e., keeping total per capita costs constant at ore- demobilization levels) would reduce expenses by another CR 44 billion, resulting in an overall savings ("peace dividend") of CR 96 billion, or 11.8 percent of the total 1997 recurrent budget (1.1 percent of GDP). Post-demobilization defense expenditure would amount to CR 208 billion. Using the 1997 budget and assuming the savings from demobilization are used to increase spending on general (1/4), social (1/2) and economic (1/4) services, the share of the recurrent defense budget in total government expenditure would be reduced to 25.5 percent. Other things being equal, the recurrent defense budget's share ;n t P -67- (base 1997) would decrease to 2.3 percent. The ratio of defense expenditure over social and economic services combined would decline to 0.6 (Table 4). Table 4: Static Impact Analysis (millions of riel) 1997 Share Scenario I Scenario 2 _____ (%) (%) (%) General Services 1/ 136,088 16.7 19.6 22.0 Defense and Security 419.336 51.4 39.6 30.1 National Defense 2/ 304.744 37.3 25.5 16.1 Public Security 114.592 14.0 14.0 14.0 Social Services 1/ 189.700 23.2 29.2 33.9 Economic Services 1/ 61,310 7.5 10.5 12.8 Other 9,576 1.2 1.2 1.2 Total 816.010 100.0 100.0 100.0 Defense/Social & Economic Service (ratio) 1.2 0.6 0.3 Defense/GDP (base 1998) 3.3 2.3 1.4 1/ Savings from demobilization are used to increase services. 2/ Post-demobilization per capita costs are constant. Scenario 1 (2): demobilization of 43.000 (78.300) soldiers. Source: Staffestimates. 16. Under Scenario 2, other things being equal, reducing non-wage costs proportionally, the peace dividend would amount to approximately CR 173 billion, and the share of national defense in the total recurrent budget would decrease to 16.1 percent, a reduction of 21.2 percent compared with pre- demobilization recurrent expenses. The post-demobilization share of national defense in GDP (base 1997) would decline to 1.4 percent, and the ratio of defense expenditure over social and economic services combined would decline to 0.3. E. Dynamic Impact Analysis of Military Demobilization3 17. Scenario 1: 43,000 soldiers. With per capita costs constant, the share of defense in total recurrent expenditure would decrease from 37.3 percent in 1997 to 16.0 percent in 2002. Defense expenditure would decrease by over 10 percent annually during the period 1999-2001, finally reaching CR 210 billion in 2002. The non-defense budget would increase on average by almost one-third during the same period and would amount to CR 1,104 billion in 2002. The demobilization of 43,000 soldiers in three phases would reduce the recurrent deficit by 36.6 percent in 2000 and 49.1 percent in 2002, with per capita costs constant. (Table 5). 18. Scenario 2: 78,300 soldiers. With per capita costs constant, the share of defense in total recurrent expenditure would decrease from 37.3 percent in 1997 to 10.1 percent in 2002. Defense expenditure would decrease by on average almost one-quarter annually during the period 1999-2001, finally reaching CR 132 billion in 2002. The non-defense budget would increase by on average over one- 3 The dynamic impact analysis is based on the high case scenario, which presupposes the implementation of the CVAP. Preliminary figures for 1998 indicate a stagnant real GDP and a stark decline in real government expenditure. Compared to the budgeted amounts, defense spending already crowds out other civilian expenditure. It is assumed throughout the dynamic impact analysis that defense expenditure would ultimately reach the level of 1997, whereas non-defense expenditure would bear the necessary adjustment within a fixed budget of CR 896 billion in nominal terms. -68- third during the same period and would amount to CR 1,181 billion in 2002. The demobilization of 78,300 soldiers in three phases would reduce the recurrent deficit by 60.9 percent in 2000, with per capita costs constant. (Table 6). Table 5: Dynamic Impact Analysis - Scenario I (in millions of riel in 1997 prices) 1997 1998 1999 2000 2001 2002 Total Current Expenditure 816,009 755,300 969.969 1.083.512 1.212.330 1,313.783 Defense 1/ 304,743 304.743 271.740 238.737 210.136 210.136 Defense/Total Current Expenditure 37.3% 40.3% 28.0% 22.0% 17.3% 16.0% Defense/GDP 3.3% 3.3% 2.9% 2.4% 2.0% 1.9% RCAF Strength 138.506 138,506 123,506 108.506 95,507 95,507 Per Capita Wages (Riel) 1.173,907 1.173,907 1,173,907 1.173,907 1.173.907 1,173,907 I/ Per capita wage and non-wage costs constant in real terms. Source: Staffestimates Table 6: Dynamic Impact Analysis - Scenario 2 (in millions of riel in 1997 prices) 1997 1998 1999 2000 2001 2002 Total Current Expenditure 816,009 755.300 1.103,215 1.216.759 1,356,585 1.363,779 Defense 1/ 304.743 304,743 249,738 194,732 132,453 132,453 Defense/Total Current Expenditure 37.3% 40.3% 25.7% 18.0% 10.9% 10.1% Defense/GDP 3.3% 3.3% 2.6% 1.9% 1.2% 1.2% RCAF Strength 138.506 138,506 113,506 88,506 60.200 60,200 Per Capita Wages (Riel) 1,173.907 1.173.907 1.173,907 1,173,907 1.173,907 1,173,907 1/ Per capita wage and non-wage costs constant in real term. Source: Staff estimates. 19. Scenario without Demobilization. Assuming that (i) the government would not embark on downsizing the military and that (ii) per capita wages and non-wages would remain constant in real terms, the share of defense in total recurrent expenditure would decrease from 37.3 percent in 1997 to 23.2 percent in 2002. Defense's share in GDP would decline from 3.3 percent in 1997 to 2.7 percent in 2002. 20. Summary of the Costs and Benefits of Demobilization. Table 7 presents a summary of the impact of demobilization in relation to total recurrent expenditure and GDP. Without demobilization, the share of defense in GDP would decrease from 3.3 percent to 2.7 percent. The demobilization of 43,000 and 78,300 soldiers would reduce this share to 1.9 percent and 1.2 percent, respectively. Table 8 shows the costs and benefits in financial resources. The demobilization of 43,000 and 78,300 ex-combatants would lead to annual savings of US$25 million and US$45 million, respectively, starting from 2001. In contrast, total costs for implementing demobilization and reintegration of 43,000 and 78,000 soldiers would amount to US$73 million and US$125 million. -69- Table 7: Impact of Demobilization (in percent) Without CVAP Scenario I Scenario 2 1997 2002 1997 2002 1997 2002 Defense/Total Current Expenditure 1/ 37.3 23.2 37.3 16.0 37.3 10.1 Defense/GDP 1/ 3.3 2.7 3.3 1.9 3.3 1.2 I/ Assuming per capita wage and non-wage costs remain constant in real terms. Sources: Staff estimates. Table 8: Costs and Benefits of Demobilization (in millions of US$) Scenario 1 Scenario 2 Defense Expenditure 1997 80.7 80.7 Defense Expenditure 2002 (w/ demobilization) 1/ 55.7 35.1 Savings (w/ demobilization) 1/ 25.1 45.6 CVAP costs (1999-2002) 72.7 124.5 1/ Per capita wage and non-wage costs constant in real terms. Source: Government budget data, staff estimates. 21. Comparison of Financial Returns to Demobilization. Demobilization and reintegration programs have been undertaken in increasing numbers, especially on the African continent. A recent Bank study on the experiences with and lessons from such programs attempted to quantify the financial returns to demobilization so as to make them comparable across countries. Table 9 presents a comparison between the two scenarios in Cambodia and the results in Ethiopia, Namibia, and Uganda on the basis of three-year averages before and after demobilization.4 In Ethiopia, for example, demobilization was undertaken in the immediate aftermath of war. Hence, budgetary savings cannot be attributed to military downsizing alone. Uganda, on the other hand, implemented a peacetime program. Cambodia's demobilization and reintegration program would rank well in terms of financial returns compared with similar programs undertaken in other countries. F. Conclusion 22. If the government were to proceed with the demobilization of 43,000 soldiers as envisaged in the CVAP, national defense would still account for around one-fifth of total recurrent expenditure if per capita costs remained constant. However, in view of the improved security situation and the increased defection rate of Khmer Rouge fighters, the number of soldiers to be demobilized is expected to be higher, and a reduction of 78,300 soldiers would result in more impressive savings. A revision of the CVAP is therefore necessary to determine the target number for demobilization and reintegration as well as the ensuing program costs. 4 The calculations for Cambodia underestimate the long-term savings because the three-year averages are calculated from 2000-2002, namely, while demobilization is ongoing. A similar limitation applies to Uganda. -70- Table 9: Financial Returns to Demobilization Cambodia I Cambodia 2 Ethiopia Namibia Uganda Demobilization 43,000 78.300 475.800 49.500 36,400 Defense Exp./Recurrent Exp. Pre-Demobilization 1/ 36.7% 36.7% 46.6% 12.4% 34.9% Defense Exp./Recurrent Exp. Post-Demobilization 1/ 18.5% 13.0% 16.4% 6.3% 25.0% Budget Savings/Recurrent Expenditure 2/ 18.3% 23 7% 30.2% 6.1% 9.9% Defense/Social Exp. Pre-Demobilization (ratio) 3/ 1.58 1.58 2.50 0.36 1 43 Defense/Social Exp. Post-Demobilization (ratio) 3/ 0.88 0.48 0.71 0.14 0.91 Budget savings/DRP costs (ratio) 4/ 0.36 0.38 1.74 0.95 0.24 1/ Three-year averages before and after demobilization Assumption for Cambodia: per capita costs constant in real terms. 2/ Budget savings refer to annual government savings from reduced military expenditure using the three-year averages. 3/ Three-year averages before and after demobilization. Assumption for Cambodia: per capita wage and non-wage costs remain constant in real term. 4/ Annual post-demobilization savings compared to program costs. The inverse indicates the number of years until DRP costs have been 'earned'. Note: Only Uganda, as a peace-time demobilization. is strictly comparable to Cambodia. Source: Colletta et a/. 1996b Case Studies in War-to-Peace Transition: staff estimates. 23. To have a lasting impact on development policy, the analysis suggests a more drastic reduction in the size of the RCAF and an improvement in living conditions (that is, an increase in wages and non- wages) for the remaining soldiers, to improve motivation and discipline. Such a policy, however, requires higher short- to medium-term investment that is beyond the government's financial capability. The currently estimated CVAP budget for 43,000 veterans amounts to US$72.8 million. The demobilization of another 35,300 soldiers would increase the budget to at least US$125 million. Without full donor support, neither program is feasible. 24. It should be noted that the analysis is undertaken for the period up to 2002, and that the full benefit of military restructuring would be realized over the longer term. Moreover, in the long run, the peace dividend of army demobilization cannot be seen in financial terms (budgetary savings) only. A well-trained professional army can enhance security and trust which in turn would improve the investment climate at both the local and the national levels. An efficiently implemented CVAP would undoubtedly contribute to such an improved economic environment and to the strengthening of civil society. In addition, if, after having received reinsertion and reintegration support, veterans' incomes would correspond to the (low) national average, veterans would contribute CR 37 billion to national production in 1997 prices (CR 68 billion under the second scenario). The total resources created annually after completion of the demobilization program, through (i) budgetary savings (assuming costs per soldier remain constant in real terms) and (ii) additional personal income, would then amount to CR 132 billion and CR 240 billion, respectively, corresponding to 1.4 percent and 2.6 percent of GDP. Clearly, demobilization would be an important investment in sustainable development. -71- Annex 3 PUBLIC EXPENDITURE ANALYSIS OF THE HEALTH SECTOR A. Sector Overview 1. Health conditions in Cambodia are among the poorest in the world. Cambodia has an infant mortality rate of 90-115 per 1,000 live births, the highest in the region (the regional average being 42 per 1,000 live births). Average life expectancy at birth has been estimated at 54.4 years, much lower than that in other countries in the region. For example, the average life expectancy at birth in 1995 was 69.5 years in Thailand, 66 years in Vietnam, 61.3 years in India, 56.4 years in Bangladesh, and 55.3 years in Nepal. The maternal mortality rate has been estimated at 900 per 100,000 live births, making it the highest in the region. For example, maternal mortality has been estimated at 650 per 100,000 live births in Laos, 160 in Vietnam, and 200 in Thailand. Child malnutrition rates are also extraordinarily high in Cambodia.' Malaria is the major cause of deaths reported, and is responsible for a large proportion of the inpatient caseload: it is estimated that 500,000 cases and 5,000 deaths occur annually. Cambodia probably has one of the most serious HIV epidemics in Asia. It is estimated that there are 100,000- 120,000 HIV infected people in Cambodia and that the number of AIDS cases by the year 2000 could range from 6,000 to 12,000. The current rate of natural increase in population is estimated to be 2.6 percent. Fertility is estimated to be 5.2 children, which means that Cambodia has the second highest fertility rate (next to Lao PDR) in Asia. 2. Despite recent improvements, access to good quality health care remains problematic because of the low level of socioeconomic development, literacy, and awareness of basic hygiene and sanitation combined with poor knowledge of the possibilities and limitations of the public health system. The number of new outpatient consultations in all health facilities by the total population was 0.35 per capita in 1996, below the WHO international standard of 0.60 for rural areas and well below the contact rates observed in the region. However, reported child immunization coverage is wide, thanks to vertical immunization programs. National coverage for 1996 is estimated at 90 percent for BCG, 76 percent for polio, 75 percent for DPT, and 72 percent for measles. In part, the low rate of utilization of health services reflects the total inadequacy of the health infrastructure and personnel in the country. Cambodia has an approximate ratio of 6,400 persons per physician, whereas the corresponding ratios are 4,450 in Lao PDR, 4,420 in Thailand, 5,220 in Bangladesh, and 2,300 in Vietnam. B. Government Policies 3. The Ministry of Health (MoH) has taken concrete steps to reconstruct and revitalize the public health system and address disease control priorities through innovative strategies. Health program priorities include HIV/AIDS, malaria, tuberculosis, and maternal and child health. The MoH has been instrumental in developing and expanding infectious diseases programs, such as the National Tuberculosis Program and the National Malarial Program. 4. In 1995, the MoH launched the health coverage plan, which calls for the establishment of a network of health centers and referral hospitals grouped into operational districts, so that the entire Recent data suggest that, among children under the age of 5, the percentages of moderately and severely malnourished children are 49.3 percent for weight-for-age, 56.1 percent for height-for-age, and 12.2 percent for weight for height (Cambodia Human Development Report 1997, UNDP). Demographic Survey of Cambodia 1996. -72- population has a rational and equitable access to basic health and referral services. Under the plan, 914 health centers and 68 referral hospitals will be established in 71 operational districts. On average, a health center will serve 14 villages and a referral hospital will serve 13 health centers. The Health Coverage Plan seeks to redress Cambodia's main shortcoming, namely, that unlike other developing countries, it does not have a countrywide network of health centers and referral units. To date, 145 health centers have been built and 21 renovated or extended under the Health Coverage Plan. In addition, 18 referral (district) hospitals and 16 referral (provincial) hospitals have been extended or renovated. In addition, the MoH has begun to implement innovative approaches in cost recovery and a new cash release process to some selected districts. C. Funding Mechanisms 5. Sources of Funding. Currently, the overwhelming inaority of health expenditures Figure 1: Funding Sources of Health are financed hy households. In 1996, Expenditures, 1996 households financed 82 percent of all health Govemment NGOs Bud2ct expenditures (i.e., both public and private) in 2 ol the country (Figure 1). Donors were the next important funders, contributing 11 .5 percent Donors of total health spending. The MoH health 12% budget accounted for merely 4 percent of total health expenditure, and NGOs accounted for the remaining 2.4 percent. This pattern of expenditure is extremely different from that in Households the education sector, where households and s2% government each finance about a quarter of total education expenditure (with donors and NGOs accounting for the remaining half). 6. Cost Recoveiy. Until recently, there were no official fees at any government health facilities, and, in principle, health services were available without charge to everyone. (In practice, however, as discussed below, Cambodians were charged informal fees in public health facilities.) Cost recovery is an established policy of MoH. However, cost recovery is not widespread, and is currently limited to a few selected national and central hospitals, most of which are in Phnom Penh. The MoH is piloting programs of cost recovery, including user charges, community-based prepaid insurance schemes, and the contracting out of specific services in public facilities to the private sector, on a trial basis. Anecdotal evidence from one of the newest hospitals in Phnom Penh suggests that cost recovery is working well. For the foreseeable future, however, overall contributions from cost sharing and cost recovery to the national health budget are unlikely to be anything more than modest. 7. Household survey data from the Cambodia Socioeconomic Survey (CSES) 1997 indicate that there is substantial scope for increased cost recovery via user fees for health services. The Survey shows that, on average, Cambodians spend CR 33,122 on consultation fees and an approximately equal amount on medicines and supplies. Even in public facilities that presumably offer free services (i.e., the district health center or DHC), individuals report paying CR 22,633 in consultation fees and CR 23,052 in medicine costs. Obviously, since DHCs do not have official fees, these amounts represent informal payments by individuals to providers at the DHCs. 8. In addition, the data in Table I show that individuals are willing to pay substantially more in consultation fees for perceived better quality care. For example, at private clinics (which presumably -73- provide better quality care), individuals are willing to spend nearly CR 35,000 in consultation fees and CR 24,000 in medicines. Even khru khmers (traditional healers) charge nearly CR 45,000 in consultation fees and CR 50,000 in medicines for a single outpatient treatment visit. Thus, if government health facilities were to improve the quality of care they offer and make official the informal fees currently paid by patients at public facilities, there could be a significant increase in the budgetary resources available for health. Table 1: Average Expenditure on an Outpatient Treatment Visit, by Type of Provider Consulted, Cambodia, 1997 Cost of a single visit Provider visited Consultation Drugs supplies Transport Total cost Khum clinic 37.960 33.390 2.508 73,858 District health center 22.663 23.052 2.634 48.349 Provincial hospital 28.663 56.603 6.013 91,279 Central hospital 84.639 59.096 5.891 149.626 Pharmacy 14.851 14.715 8.153 37,720 Private doctor/nurse 38.933 39.673 6.138 84.743 Private clinic 34.900 23.717 5.382 63,999 Private hospital 52.482 71,250 5.021 128.753 Khru Khmer 44,899 50.492 3,662 99,053 Other traditional 18,930 26.395 2.820 48.144 Other provider 34.965 25.677 10.682 71.323 All providers 33.122 33.102 5.322 71.546 Note Numbers are averaged overall nonzero values of payments. Source: Cambodia Socioeconomic Survey (CSES) 1997. 9. Table 2 also reports average expenditures on an outpatient visit, but disaggregates the data by economic group. What is interesting is that even the poorest 20 percent of the Cambodian population are willing to pay (and, in fact, are already paying) a total of nearly CR 28,000 (for consultations and medicines) for a single outpatient treatment visit. This again indicates a significant possibility for expanding the scope of cost recovery and cost sharing in the health sector in Cambodia. Table 2: Average Expenditure on an Outpatient Visit, by per Capita Expenditure Quintile, Cambodia, 1997 Per capita Expenditure on a single health visit Per capita quintile monthly expenditure Consultation Drugs supplies Transport Total cost Poorest 12.871 14,899 5.125 32.896 7,565 Second 20.650 14.514 2,875 38,039 12.203 Third 27.757 22.338 5.268 55,362 17,466 Fourth 30.039 31.094 6.766 67.899 25,957 Richest 60,941 57.543 5,230 123,714 91,524 Source: Cambodia Socioeconomic Survey (CSES) 1997. D. Expenditure Trends 10. Overall Health Expenditures. There was a major increase in non-private spending on health in 1996, more than doubling from 1995. The increase in health spending occurred across all sources: the MoH budget, donor assistance, and NGOs (Figure 2). The increase in public spending means that, in 1996, total spending on health in Cambodia was US$40.52 per capita - out of which US$33.36 was household spending, US$4.64 was spending by donors, US$0.91 was spending by NGOs, and US$1.61 -74- was spending by the government. Total health spending (by the households, the government, and donors) constituted about 13.2 percent of GDP in 1996, while total government financed health spending constituted only 0.5 percent of GDP and 5.1 percent of government expenditure. There are two important points here: first, Cambodian households are paying far more than they can afford for health care, and, second, government-financed health spending per capita is woefully inadequate. Indeed, government- financed health spending per capita of US$7.16 is significantly less than the US$12 needed to finance an essential package of health services, including public health, for a low income countries.3 Figure 2: Non-private health spending in Cambodia, 1994- 96 (million rie Is) 200,000 180 000 160 000 At 809 000 80 000 40000 20 000 0 1994 1995 1996 Alsources _ 'A6Hbudget - Donorassistance I1. Pattern of Expenditures. Table 3, which reports the MoH budget by line item in 1995 and 1996, indicates that operations and maintenance (O&M), including drug supply, accounts for a major part of the MoH budget. Indeed, the proportion of the budget allocated to O&M increased from 63.5 percent in 1995 to 76.8 percent in 1996. Expenditure on drugs constituted 43.2 percent of total MoH health spending in 1996, which was a sharp increase from 1995. Salaries and allowances accounted for only 21.2 percent of total MoH spending -- down from 32.9 percent in 1995. These are desirable trends, as the health budgets of many developing countries are dominated by salaries and allowances, with consequently little resources being left for drugs and other essential operating expenditures. In Lao PDR, for example, salaries and allowances accounted for 66 percent of total MoH expenditures in 1994-95, with drugs accounting for only 11 percent. 12. Center-Province Distribution. Nearly two-thirds of the government health budget is distributed at the central level, and the remaining one-third to the provinces. Central spending consists primarily of budgets for hospitals, training centers, and research institutions for vertical disease programs. In addition, it includes expenditure on drugs, most of which are distributed by the central medical store to provinces and districts. If central drug expenditures are removed, approximately one-third of the health budget is allocated to the central level and the remaining two-thirds for the provincial and district levels. This is very similar to the situation in neighboring Lao PDR. 3 Obviously, the US$12 figure is only a rough guide to determining whether the health sector is sufficiently funded or not. It would depend on a number of country-specific factors, such as the efficiency of public spending on health, the public-private mix in heath spending, and so on. -75- Table 3: MoH Budgetary Expenditures, by Line Item, 1995-96 1995 1996 Line Item Exp. mn. riels % of total Exp. mn. riels % of total Total MoH budgetary expenditure 26.322 100.00 43,177.5 100.00 Wages 8.645.9 32.85 9.139.7 21.17 O& M 16705.2 63.46 33,150.0 76.78 Social subsidies 134.0 0.51 205 0.47 International subsidies 640.2 2.43 109.1 0.25 Domestically financed construction & equipment 196.7 0.75 387.0 0.90 Domestically financed local counterparts - - 185.8 0.43 Source. Data provided bY the government 13. Provincial Distribution. Three observations can be made about the distribution of public health expenditures across provinces. First, actual health expenditures (namely, cash released from the Treasury) are significantly different from budgeted health expenditures in all provinces (see Table 4). The amount not spent as a percentage of the budgeted amount varies from a low of 4.3 percent in Stung Treng to a high of 50.1 percent in Prey Veng. There appears to be a relationship between the percentage of budgeted health expenditures that are never spent and the per capita GDP of a province (as proxied by the per capita consumption expenditure of households), which suggests that poor provinces experience greater bottlenecks and constraints to full utilization of budgeted health funds. The bottlenecks might be the result of poor human resources, including administrative skills, and poor infrastructure. Second, there is substantial variation in per capita government health expenditures across provinces. Low population provinces such as Mondulkiri and Ratanakiri have significantly higher per capita government health expenditures (CR 6,099 and CR 3,944, respectively) than large provinces such as Phnom Penh (CR 974). In part, this reflects the very difficult terrain and scattered populations in Mondulkiri and Ratanakiri, which makes health services delivery more expensive. Third, there appears to be a positive relationship between government health expenditures per capita and per capita provincial GDP. This is observed in Figure 3, which plots the two variables against each other. This suggests that, even if the formula for planned allocations of government health expenditure across provinces is equitable, the actual allocations favor the richer provinces. One reason for this may be that it is easier for the MoH to implement health projects in more affluent provinces, which have a higher absorptive capacity that comes from better infrastructure, facilities, and administrative resources. Another reason may be that in provinces such as Prey Veng the scarcity of local resources makes reallocation from health more likely. E. Efficiency and Equity of Public Spending 14. One of the key messages in health economics is that preventive care interventions should have the first claim on public resources, because of the substantial externalities they create for the society and also because of under-spending on them by private agents. In most developing countries, however, the bulk of government health spending goes into providing curative services, especially at the tertiary level, with very few resources allocated to preventive care and primary health programs. Such a distribution is not only inefficient but is also highly inequitable, as household survey data typically show that a disproportionately large share of hospital users tend to be drawn from the affluent sections of the population. -76- Table 4: Government Health Budget and Expenditures and Per Capita Household Consumption Expenditures, by Province, Cambodia, 1996-97 Payment Actual Per capita order Cash released per capita household. Budget approved from Treasury govt health cons. exp. Province 1996 1996 1996 exp. 1996 1997 Phnom Penh 1,271.3 975.8 799.2 974 109.172 Kandal 1.709.8 1 540.2 1.139.7 1.218 25.449 Kg Cham 2.245.4 1,891.4 1.891.4 1.304 19,484 Battambang 1.843.3 1.525 1 L.525.1 2.311 27.859 Prey Veng 1.708.0 L.482.0 852 0 874 14.649 Siem Reap 1.468.8 1.008 1 963.0 1,462 13.694 Kg Thom 1,173.5 911.6 855.5 1.503 21.668 Takeo 1.335.9 1.128.9 1.128.9 1,548 21.880 Svay Rieng 948.6 833.8 833.8 1.816 21,564 Pursat 823.0 639.1 607.4 1.932 18,403 Kg Chhnang 979.3 841 6 840.4 2.420 21,953 Kg Speu 1,014.4 531 5 527.4 990 23,472 Kampot 1.022.0 910.9 857.6 1.822 20,728 Kg Som 475.8 412.4 409.5 3.393 20.436 Koh Kong 212.0 186.6 186.6 1,894 129.272 Preah Vihear 332.9 262.7 262.7 2.410 Kratie 637.9 546.3 478.2 2,165 24.006 Ratanakiri 368.1 308.1 304.5 3,944 43.825 Mondolkiri 216.8 188.7 188.7 6.099 Banteay Meanchey 1.345 7 1096.0 991.0 1,982 20,398 Stung Treng 411.9 394.1 394.1 4.796 24.918 Kep 39.6 22.5 22.5 870 34.838 Source: Ainistry of Health, 1996 Expenditure Book Phnom Penh, 1997. Figure 3: Relationship between government health expenditures and household consumption expenditure per capita across Cambodian provinces, 1996 5.000 4,500 4.000 3.500 3,000 2.500 2.D00OO .,500- 1000 - 500 (2 0. 10000 15.000 20,000 25,000 30,000 35.000 40,000 45,000 Household coonsumption expenditure per capita in province (rtels) 15. Table 5 attempts a reclassification of MoH spending from all sources (i.e., budget, overseas development assistance, and NGOs) for the period 1994 to 1996. This reclassification is not perfect, as some items of expenditure, such as drug supply, cannot be allocated to primary, secondary, and tertiary health facilities. Nevertheless, examination of the data can provide some idea about the efficiency of public spending on health. -77- Table 5: Public Health Spending by Function, 1994-96 (From All Sources - Budget, ODA and NGOs) 1994 1995 1996 a/ Function Mn. riel % of total Mn. riel % of total Mn. riel % of total Total 87,792 100.00 81.865 100.00 192.074 100.00 Sector policies and planning 39.310 44.78 39,414 48.15 33.151 17.26 Primary health care 14.101 16.06 9.025 11.02 51.116 26.61 Disease control 7,030 8.01 4.877 5.96 5.147 2.68 Family planning 1.383 1.58 2.127 2.60 2.690 1.40 Hospitals 21.952 25.00 18.871 23.05 15.741 8.20 Training 32 0.04 28 0.03 45 0.02 Drug supply 0 0.00 7.326 8.95 59.198 30.82 Construction, equipment 3.984 4.54 197 0.24 592 0.31 a! Numbers do not add up to total since NGOs contribution (24 billion riels) cannot be broken into sub-categories. Source: Data provided by the Government. 16. The following observations stand out from the data reported in Table 5. First, as in many developing countries, hospitals took up a disproportionately large share of government health expenditures in 1994, accounting for a full 25 percent of total government spending on health. On the other hand, primary health care accounted for merely 16.1 percent of public spending on health.4 Second, these trends appear to have been completely reversed in 1996. Data for 1996 show that the share of hospitals in government health expenditures fell to merely 8.2 percent in 1996, while that of primary health care increased to 26.6 percent. This is an extremely favorable trend because of its implications for both efficiency and equity. Third, the share of drug supply in total public health spending also increased dramatically - from a virtually negligible level in 1994 to 30.8 percent in 1996. This, too, is a desirable trend, and suggests that government health facilities are better stocked with drugs and medical supplies than they used to be. Fourth, vertical disease-control programs, which are generally preventive health interventions, saw a decline in their share of total government health spending. 17. A more disaggregated look at the functional spending data by source (i.e., budget, donors, and NGOs), however, shows that the shift away from hospitals to primary health care was primarily attributable to changes in international donor spending. In fact, spending on hospitals from the MoH budget increased from CR 1.7 billion in 1995 to CR 2.8 billion in 1996. However, the shift of government health expenditures into drug supply came about because of spending shifts among donors as well as in the MoH budget. 18. In summary, both the efficiency and equity of public spending on health appear to have improved in that the share of spending on tertiary curative care (hospitals) has fallen, while that on primary health care has increased (albeit much of it because of changes in donor spending). One unfortunate trend is that the relative share of disease-control programs has fallen; however, it is not clear from these data 4 Because of the inability to separate primary from other levels of care and the pooling of government expenditures on all hospitals, these data have to be treated with caution. First referrals to district or provincial hospitals are very much part of an integrated primary care system. Hence, only national or central hospitals, which are typically based in Phnom Penh, can be considered as offering a tertiary level of care. The referral system functions poorly in Cambodia, and roughly more than one-half of the expenditure on hospitals is on hospital administration and on provincial and central hospitals. -78- whether this reflects a real cutback in preventive health programs or a shifting of many elements of disease-control programs to primary health care. F. Evidence of Expenditure Leakage/Ineffective Targeting 19. UNICEF data on the availability of drugs show little evidence of leakage of public sector drug supplies into the private market for drugs. However, this evidence pertains only to the flow of drugs within the public drug supply system (i.e., from the center to provincial/district levels and from the provincial/district levels to the health centers). It is not clear what happens to the drugs at the level of the facility (i.e., whether the facility staff provide it to patients or sell it to patients, or sell the drugs in the market). UNICEF is planning a study to look at this issue more seriously following discussions with the government and donors. 20. There is another form of expenditure leakage (see Section H on Institutional Constraints). The vertical budgetary system employed in Cambodia provides too many opportunities for "leakage" of authorized funds at intervening administrative levels. Since the provincial health department is typically ignorant of the health budget authorized for its province by the Ministry of Economy and Finance (MEF) and has to depend on approvals for expenditure by the provincial governors, there is an incentive for provincial governors to disburse only a portion of the health funds they receive from the MEF. 21. There is a third form of leakage, and this concerns the leakage of health subsidies and benefits intended for the poor to the non-poor. There is some evidence from the Cambodia Socioeconomic Survey 1997 that indicates that these leakages are significant. Table 6, which reports the percentage of individuals in each per capita expenditure quintile who reported paying nothing for their visit to government hospitals, indicates that about 12 percent of individuals visiting provincial hospitals and 14 percent of individuals visiting central hospitals were fully waived from all fees. However, there is no pattern of granting fee exemptions to the poor. Indeed, 17 percent of individuals in the fourth quintile reported paying nothing for a visit to provincial hospitals, as compared to only 10 percent of individuals in the poorest quintile. This indicates a very high degree of leakage of benefits intended for the poor and thus points to substantial inefficiency and waste in the system of granting fee waivers at government hospitals. Table 6: Percentage of Individuals Obtaining Treatment Who Reported Paying No Consultation Fee for Theii Outpatient Visit, by Type of Provider Visited and by per Capita Expenditure Quintile, Cambodia, 1997 Per capita expenditure quintile All Provider individuals Poorest Second Third Fourth Richest Provincial hospital 952 11.54 10.25 1737 8.11 12.33 Central hospital 18.07 0.00 6.07 12.07 18.95 13.60 Source: Cambodia Socioeconomic Survey (CSES). 1997. G. Scope for Private Sector Participation 22. Compared with the education sector, private sector participation in health services delivery is widespread. National household survey data for 1997 show that the public sector is not the provider of choice for most individuals in the country. Nearly two-thirds of all individuals reporting an illness in a four-week reference period visited nongovernment facilities (i.e., pharmacies, private doctors/nurses, private clinics, and private hospitals) for their outpatient care (see Table 7). In contrast, fewer than 20 percent of individuals visited a khum clinic or district health center. With the exception of pharmacies, -79- which the poor use relatively more than the rich (primarily on account of their being low cost), there is a positive association between the utilization of private medical facilities and economic status. Table 7: Distribution of Outpatient Visits by Individuals Reporting III in the Last 4 Weeks, by Provider Visited and by per Capita Expenditure Quintile, Cambodia, 1997 Per capita expenditure quintile Provider Poorest Second Third Fourth Richest Total Khum clinic 10.80 13.41 11.26 8.32 7.97 10.01 District health center 1206 12.09 8.52 9.56 3.98 8.70 Provincial hospital 2.24 2.96 4.02 4.53 2.89 3.49 Central hospital 0.76 1.11 4.92 4.85 9.32 4.85 Pharmacy 44.16 24.63 25.54 2049 18.22 24.50 Private doctor/nurse 8.45 21.13 18 63 18.27 21.72 18.51 Private clinic 12.30 14.39 12.42 19.78 22.52 17.07 Private hospital 0.67 2 45 3 67 5.43 6.77 4.32 Khru Khmer 4.35 3.87 5.06 3.87 3.41 4.07 Other traditional 1 62 1.61 2.03 2.56 0.91 1.78 Other provider 2.20 2.23 3.92 2.34 2.29 2.64 TOTAL 10000 100.00 100.00 100.00 100.00 100.00 Source: Data provided by the Government 23. The rapid rise of the private sector in delivering health services in recent years, while desirable, implies that the MoH needs to attach increasing importance to its regulatory functions. The large number of pharmacies and drug vendors, many of these unregistered, raises issues of public safety and regulation. Since the private sector (other than pharmacies) caters disproportionately to the more affluent groups in Cambodian society, it implies that government health facilities need to increasingly target their services to the poor, who may be unable to afford treatment in private health facilities. Finally, it is important to recognize that the boundaries between public and private health providers in Cambodia are somewhat porous. Very often, the private doctor is employed at a government health facility, and engages in private practice after hours, typically in the same (public) facility. At the same time, as noted earlier, the unofficial charges extracted from patients at public facilities are likely to flow into private pockets. H. Institutional Constraints 24. One of the major problems of the health sector in Cambodia is that health spending decisions are highly centralized. The MoH and provincial and district health managers have limited authority over budget and expenditure decisions. Although the MoH undertakes budget planning and negotiations for the central and provincial levels, the responsibility for accounting and disbursement of funds lies (in the case of provinces and districts) with the provincial authorities. As a result, the provincial health departments need approval from the MEF as well as the provincial governor for any commitment of public funds. Not only does this budgetary system contain too many steps and intermediaries, it also means that district managers and health center staff do not know the amount of national budget they will receive each month. In turn, this makes it difficult for them to plan and prioritize their activities. 25. The vertical budgetary system, whereby resources flow from the MEF to the provincial governors, means that, although the MoH makes its central and provincial budgets, actual health expenditures, especially at the provincial levels, are often beyond its control. The provincial health department has to depend on the amount approved by the provincial governor from the authorized health budget for that province. There is impressionistic evidence that, until October 1997, it was common for a provincial health department to receive authorization from the provincial govrmor for only 70-80 percent of the budgeted amount, since the former were not aware of the actual budgeted health expenditures for -80- the province. Of course, the main reasons for the large discrepancy between actual and budgeted health expenditures are perpetual revenue shortfalls and military expenditure overruns at the central level. However, the important point is that the vertical budgetary system results in consistent underutilization of budgeted expenditures because of the system's cumbersome nature and lack of transparency (see Table 4). For example, in 1996 the MoH was unable to spend 25.6 percent of its budgeted amount for the provinces and 48.1 percent of its budgeted amount at the central level. The large amount of unutilized public health funds is a critical constraint on the implementation of health programs and projects in the country. 26. Recently, the MoH and the MEF designed and released a new cash release process (called the Accelerated District Development (ADD) Program), but this applies only to 22 districts in the country. Under this program, a new budget line (Chapter 13) has been established by the MEF in order to provide sufficient and regular government funds for the ADD program. Chapter 13 pays for the various administrative and running costs normally covered by Chapter 11 (O&M) plus overtime payments for night duty. But it does not cover the cost of basic salaries (Chapter 10), social allowances (Chapter 3 1), or investment in major new equipment or buildings (Chapter 50). The new system allows cash to be provided to the ADD districts, via the MoH and the Provincial Health Departments, before expenditure is incurred, and to be accounted for afterward. Each district must prepare a budget showing the detailed planned expenditure and this budget must be approved by the MoH. In addition, monthly reports have to be provided by each district health department that compare actual expenditure to the planned (budget) expenditure. 27. By reducing the number of administrative layers in the budget disbursement system, the ADD program is expected to speed up disbursements and reduce the discrepancy between budgeted and actual health expenditures at the district level. In addition, the program reduces the scope for "leakage" of budgeted funds at successive administrative levels. There is already some evidence that the program is yielding significant results at the ground level. Household survey data indicate that the level of utilization of health services at health centers and referral hospitals is significantly higher in ADD than in non-ADD districts. Eventually the ADD program is planned to be extended to all districts. This will be essential as the current budget system is seriously hampering the effective utilization of health funds in the country. To date, however, the ADD program, limited as it is to 22 districts, remains only a pilot program. 28. Sector-Wide Approach. Preparation toward adopting a Sector-Wide Approach (SWAP) in the health sector is being initiated. Under this approach, donors and NGOs would give up some of their rights to select individual projects for financing and instead would develop, in close collaboration with the MoH, a sector-wide policy and strategy that would guide the selection of all investments and health projects in the country. The available resources of the MoH, donors, and NGOs would be pooled to finance the projects deemed consistent with the sector strategy. The advantage of this approach is that donor and government interventions in the sector are rationalized; duplication and waste is reduced; and complementarity between projects is maximized. While donors and NGOs give up their rights to finance individual projects of their choice, they gain the right to participate in developing sector-wide policy and strategy. 29. In close collaboration with donors and NGOs, the MoH has already formed a strategy for the development of the health sector in the near future. This is embodied in the Health Coverage Plan and in the National Health Policy and Strategies 1996-2000, both of which could serve as the basis of all donor/NGO assistance to the health sector. However, the main constraint to moving forward on a SWAP in the health sector is government capacity. While the MoH has been building up implementation and planning capacity during the last few years, local capacity for carrying a SWAP apenda fnrw.ari renm,ns limited, not only at the provincial level but also at the centfal level. As all the dono, azsiaii are to -81- flow through the government system under a SWAP, it could overwhelm capacity (both central and local). Thus, a number of system-wide issues, such as managerial decentralization, reforms in accounting and budgeting systems, and changes in staffing incentives, need to be addressed for a SWAP to be adopted effectively. It should be also noted, however, that moving toward a SWAP itself would help strengthen the weak institutional capacity of the MoH. 30. It is critical that the institutional linkage between the MoH and the MEF be strengthened. In close collaboration with the MEF, the MoH would be able to develop reform issues such as decentralization, SWAP, and evaluation of the ADD system more effectively. I. Evidence of the Impact of Public Expenditure/Donor Aid 31. Cambodia has been receiving large amounts of donor assistance in the health sector during the last few years. An important question is the extent to which this donor assistance and public spending on health in general have been effective in improving health outcomes. As is well known, evaluating the impact of public expenditures or policy interventions is not always straightforward in the absence of a "before-and-after" experiment and control for other variables that may have affected health outcomes. 32. There is evidence that donor assistance and government health expenditures, especially on the Expanded Program of Immunization (EPI), have been very effective. Table 8 shows the percent of children immunized by 12 months of age and the percentage of pregnant women immunized against tetanus from 1992 to 1997. There has been an impressive increase in immunization rates since 1992; in the case of polio, DPT and measles, the immunization coverage has more than doubled (from about 32-33 percent in 1992 to 68-72 percent in 1997). In the case of tetanus, immunization coverage increased sixfold--from a mere 6 percent of pregnant women immunized against tetanus in 1992 to as many as 30 percent being immunized in 1997. Even in the case of BCG vaccines, coverage increased from 50 percent in 1992 to 80 percent in 1997. These are all impressive rates of increase in immunization coverage. Table 8: Immunization Coverage (percent of children immunized by 12 months of age and percent of pregnant women immunized against tetanus), 1992-97 1997 Antigen 1992 1993 1994 1995 1996 (expected) BCG (TB) 50 57 78 95 90 82 OPV3 (Polio) 32 36 54 80 75 72 DPT3 (Diphtheria, Pertussis. Tetanus) 32 35 53 79 74 72 Measles 33 36 53 75 72 68 TT2+ (tetanus for pregnant women) 6 22 28 36 36 30 Source: Ministry of Health, 1997. 33. Table 9 shows the number of reported cases of EPI diseases for the same years. One would expect the increase in immunization coverage to be reflected in a decline in the number of reported cases of the preventable diseases. This certainly appears to be the case; the number of reported cases of childhood tuberculosis, polio, diphtheria, measles, pertussis, and neonatal tetanus all fell dramatically between 1992 and 1997. 34. In addition to EPI, malaria control has been a successful vertical program in Cambodia. Data on recorded malaria cases and deaths at hospitals show a sharp drop in case fatality rate, as the numbers of hospital laboratory and health center staff clinically trained in malaria detection and treatment increased from 1992 to 1997 (Table 10). For example, in 1992 only 5 percent of hospital laboratory staff were trained in malaria detection, and only three provinces (out of a total of 22) had received hospital -82- clinicians training. Furthermore, there were no health centers with CNM-trained staff. By 1996, 53 percent of all hospital laboratory staff were trained in malaria detection and 13 provinces were receiving hospital clinicians training, and in 1997 there were 109 health centers nationwide with CNM-trained staff. Correspondingly, the case fatality rate (i.e., the number of recorded deaths from malaria as a percentage of reported malaria cases) observed in hospitals fell from 1.7 percent in 1992 to 0.7 percent in 1996. Table 9: Reported Cases of EPI Diseases, Cambodia, 1992-96 EPI Disease 1992 1993 1994 1995 1996 Poliomyelitis 146 135 296 130 83 Diphtheria 13 22 5 NA NA Measles 2.759 1.226 946 2.038 697 Pertussis 995 611 63 NA NA Neonatal Tetanus NA 88 147 8 2 Source Alinistry of Health. I 997 Table 10: Mortality Reduction in Hospitals through Laboratory and Clinical Training in Malaria Detection and Treatment, 1992-97 No. of hospital lab. No. of provinces No. of health centers No of malaria Recorded malaria staff trained receiving hospital with CNM-trained staff deaths recorded in cases** (case fatality Year clinicians training hospitals* rate) 1992 15/300 (5%) 3/22 0 1.504 91.077 (1 7%) 1993 30/300 (10%) 522 0 1,163 115.701 (1.0%) 1994 50/300 (17%) 7/22 0 1.009 85.012 (1.2%) 1995 143/315 (45%) 10/22 20 624 107.265 (06%) 1996 170/320 (53%) 13/22 35 745 106.614 (0 7%) 1997*** 100/290 (35%) 10/22 109 811 170.387 (0.5%) * An estimated 90 percent of cases and deaths go unrecorded. ** Decreasing proportion of unrecorded cases in 1997 owing to better equipment/training of health centers. Temporary interruption of activities after July events Source: WHO esiumates. 35. In addition, it is possible to relate government health expenditures at the provincial level to morbidity rates in these provinces. The latter can be constructed from the Cambodia Socioeconomic Survey 1997, which permits an estimation of the proportion of the population in each province that reported an illness episode in the four weeks preceding the survey. A regression of reported morbidity (health outcomes) on government health expenditures per capita suggests an inverse relationship (significant at the 10 percent level) between the two variables.5 While no causality can be inferred from the inverse association, the data in Figure 4 are certainly suggestive that higher levels of government health expenditure per capita in a province do indeed contribute to lower levels of reported morbidity. 5 This relationship obtains with and without control for household consumption per capita (which is used as a proxy for household incomes). -83- Figure 4: Relationship between reported morbidity and government health expenditures per capita across Cambodian provinces, 1996 25 23 - 21 - 0 9 - 17 J. vauaio ofte I I I 9 5. S00 1000 IS00 2000 2i111 3000 3SG0 Per capita go-enm,ent health expenditure (rieis) J. Evaluation of the PIP 36. The proposed Public Investment Program (PIP) for 1998-2000 calls for the government to continue implementing the Health Coverage Plan and to continue raising the level of public spending on health. The 1998-2000 PIP plans to raise spending on health to US$239 million, so that it would constitute nearly 20 percent of all government expenditure. Of this, approximately 47 percent would be spent on the revitalization of basic health services (including the construction of health infrastructure), 34 percent on the expansion of key national health programs into the district and commune levels, 10 percent on the development of responses to emerging health priorities, and the remaining 9 percent on the rehabilitation of national hospitals and the extension of provincial blood bank centers. 37. As part of the PIP, the MoH proposes to pilot-test alternative health financing mechanisms (such as user charges, community-based rural insurance, and the contracting-out of services at public facilities) in light of the limits on public spending. To reduce the maternal mortality rate, the PIP also calls for a target of 40 percent of rural births being attended by a trained midwife or nurse (up from a current rate of only 15 percent) by the year 2000. An important element of the PIP is the expansion and strengthening of national programs for tuberculosis, polio, and HIV/AIDS. 38. The projected increase in the PIP over the 1998-2000 period appears overly ambitious, given the weak implementation capacity and the inefficiencies in the existing budgetary system in all but 22 ADD districts in the country. In addition, more than 90 percent of health investment under the PIP is expected to come from donor resources, which represents even more dependency on donors. However, the subsectoral allocation of PIP resources across primary health care, tertiary health care, and vertical disease programs is balanced and in line with the needs of the health sector in Cambodia. K. Summary and Recommendations 39. While privately-financed health expenditures in Cambodia are very large (about US$33.36 per capita per annum), publicly-financed health spending is inadequate. At about US$7.16 per capita per annum, government-financed spending on health is 40 percent lower than the US$12 minimum required by low-income countries for implementation of a basic package of health services, including public health. Thus, there is a need to increase the resources allocated to health so that they reach a level of at least US$12 per capita per year. The increase in funding for health will have to be accomplished with local funds, as donors and NGOs already account for more than three-quarters of all publicly-financed health expenditures and are unlikely to increase allocations to health significantly in the near future. This means that the government will have to develop new and innovative means offinancing health activities in the future, including greater cost recovery, health insurance schemes for formal sector employees, and -84- community prepaid insurance schemes. There exists substantial scope for cost recovery in light of the fact that individuals already pay large amounts for medical treatment at both public and private health facilities. Unfortunately, most of the payments made to public facilities and providers are informal payments that do not contribute to government revenue. One of the highest priorities for the MoH should be to see how the large informal fees that are already being paid by individuals to public health providers can be formalized and internalized by the district and provincial health departments. The initial results of the pilot programs of cost recovery recently introduced at central hospitals in Phnom Penh are encouraging, and should be replicated at other tertiary based health facilities. It is important to ensure that the user fee collection process be transparent and that the collection revenues be kept at the facility level to improve the quality of services. 40. One of the largest impediments to the effective utilization of the government health budget is the vertical budgetary system which concentrates health budgets in the hands of the provincial governors. In this overly centralized system, the MoH, and in particular the provincial and district health managers, are left with very limited authority over budget and expenditure decisions, sometimes not even knowing the amount of health funds actually budgeted for the province or district. The system has too many steps and intermediaries and often makes planning and prioritizing public health activities difficult for local health officials. Another undesirable effect of the vertical budgetary system is that it results in consistent underutilization of budgeted expenditures because of its cumbersome nature and lack of transparency. The MEF and the MoH have designed and implemented, on a pilot basis (in 22 districts in the country), a new cash release process (the ADD program). Early results from this program are encouraging, and suggest that not only does the ADD program expedite disbursements and reduce "leakages" of budgeted funds at the district level, but it also results in higher levels of use of health services at health centers and referral hospitals. It is imperative that the ADD system be expanded to a larger number of districts. 41. Another major problem of the health sector in Cambodia is the low rate of use of health services, especially by the poor. There are many reasons for these low rates: the high cost and lack of affordability of health services and the poor and inadequate physical access (proximity) to health facilities, and also the poor quality of health services offered at public facilities. To some extent, these problems can be alleviated by means of targeting. The MoH has not made enough use of targeting in the past. Targeting is a means of making the most of public expenditures by focusing them on target activities or groups. Evidence from household surveys in Cambodia suggests that targeting public spending on health to poor individuals and underserved provinces would raise the cost-effectiveness and equity of public spending. For example, the provincial distribution of per capita public health spending is inequitable, with richer provinces having higher levels of government health expenditures per capita than poorer provinces. This is not due to a bias in the formula for allocating public health spending across provinces, but instead is the relative ease of implementing health projects in more affluent provinces that have higher capacity and better infrastructure, facilities, and administrative resources. In addition, the inequality in health spending per capita may occur because the scarcity of local resources in poor provinces such as Prey Veng makes reallocation from health spending more likely. Whatever the reasons, the inequitable distribution of health spending across provinces needs to be reversed. Likewise, to the extent that secondary and tertiary based facilities are disproportionately used by the better-off, while primary care facilities, such as health centers, are used mostly by the poor, government health policy needs to shift its focus from secondary and tertiary care to primary care. In this regard, there has in fact been a very desirable shift in public spending away from tertiary hospitals and toward primary health care in the last few years. Such a shift needs to be further encouraged as it improves not only the efficiency of public health spending but also its equity. 42. Even now there is heavy participation by the private sector in health services delivery in Cambodia, with nearly two-thirds of all individuals who report an illness episode visiting nongovernment -85- facilities (i.e., pharmacies, private doctors/nurses, private clinics, and private hospitals) for their outpatient care. The rapid rise of the private sector in delivering health services in recent years, while desirable, also implies that the MoH needs to attach increasing importance to its regulatory functions. The large number of pharmacies and drug vendors, many of these unregistered, raises issues of public safety and regulation. In addition, the rise of the private sector, which typically caters to the more affluent groups in Cambodian society, implies that government health facilities need to increasingly target their services to the poor, who may be unable to afford treatment in private health facilities. 43. Finally, it is important to note that the MoH's capacity for carrying a SWAP agenda forward remains limited, not only at the provincial levels but also at the central level.While preparation toward a SWAP needs to be initiated, a number of system-wide issues, such as managerial decentralization, reforms in accounting and budgeting systems, and changes in staffing incentives need to be addressed for a SWAP to be adopted effectively. It should be also noted, however, that moving toward a SWAP itself would help strengthen the weak institutional capacity of the MoH. -86- Annex 4 PUBLIC EXPENDITURE ANALYSIS OF THE EDUCATION SECTOR A. Sector Overview 1. Cambodia has made considerable progress in expanding schooling opportunities, especially at the primary level, since 1991. While primary and secondary enrollments have increased impressively in absolute terms, several problems remain in the education sector. While the country's gross primary enrollment ratio (94.5 percent) is comparable to that of most countries in the region, its gross secondary enrollment ratio (19 percent), is one of the lowest in the region, lower, for example, than that of Lao PDR (25 percent) or Nepal (37 percent). There is a large discrepancy between gross and net enrollment rates at all levels, reflecting delayed entry into school, high rates of grade repetition, and over-age enrollment. Nearly 20 percent of all students enrolled in urban areas and 26 percent of rural students are repeaters. Dropout rates in grades 5, 8, and II are 19 percent, 38 percent, and 51 percent, respectively. High rates of repetition and dropout indicate enormous wastage in the educational system. 2. There are disparities in enrollment rates between rural and urban areas, especially at the lower and upper secondary levels. The gross lower secondary enrollment rate in urban areas is nearly three times as large as that in rural areas, and the gross upper secondary enrollment rate in the urban areas is 13 times as large.' In addition, there are gender disparities in enrollment: male students enjoy a gross primary enrollment ratio of 102 percent, while the corresponding ratio for female students is only 86 percent. The gender differences widen at higher schooling levels. Males have a gross lower secondary enrollment ratio that is 68 percent higher than that of females, while the male-female enrollment difference widens to 88 percent at the upper secondary level. The economic disparities in enrollment rates are even greater than the regional disparities. For example, the gross primary enrollment ratio for the poorest per capita expenditure quintile is only 85 percent, while that for the richest 20 percent of the population is 113 percent. These disparities widen at the higher levels. B. Government Policies 3. The Government's Socio-economic Development Plan 1996-2000 envisages major changes in the strategies for repositioning the education system to service entry into the market economy. Planning and management for improved access and quality of basic education are seen as essential prerequisites to the training of the skilled human resources needed to modernize Cambodian society. Improved literacy and numeracy are considered the necessary foundation for market-oriented technical and vocational skills training and higher education. A National Training Board has been established to coordinate the activities of government ministries and stakeholders in the development of flexible market-responsive formal and non-formal investments in Technical and Vocational Education and Training (TVET). A ministerial-level cross-sectoral National Higher Education Task Force has recommended the establishment of a National Commission of Higher Education to coordinate government and donor education system planning and, increasingly, donor planning to support institutions managed by four other ministries. The Education Investment Framework 1995-2000, adopted to coordinate sector development at the Education Round Table of December 1994, gives priority (approximately two-thirds) to investments in basic education. Successive Public Investment Programs (PIPs) since then have set targets for implementing these strategies. All enrollment rates as reported here, with the exception of those for different per capita expenditure quintiles, are based on school-facility data complied by the MoEYS (1997). The quintile-specific enrollment rates are based on household survey data from the Cambodia Socioeconomic Survey (CSES) 1997. There is some discrepancy between the overall enrollment rates as reported by the MoEYS and those calculated from the CSES. -87- C. Funding Mechanisms 4. Sources of Funding. Currently, much of the education sector is financed by donor assistance. In 1996 nearly one-half (46 percent) of all educational expenditure (i.e., both public and private) in the country was funded by international donors and NGOs (Figure 1). The Ministry of Education, Youth, and Sports (MoEYS) and the private household sector each contributed 27 percent of total educational spending. Thus, the government's share of educational expenditure is about the same as that of households and only about half as much as the share of NGOs and donors. Figure 1: Financing education expenditures, 1996 2 7 % 27T% 35% OMOEYS DODA ONGOs CHousehold 5. Cost Recovery. Officially, schooling is free all the way from the pre-primary to the university level in Cambodia. Unlike the Ministry of Health, which has introduced cost sharing on a pilot basis at central hospitals, the MoEYS has yet to establish a cost recovery policy to benefil from regularizing or formalizing the currently widespread unofficial fees and contributions. Recently, a study was undertaken under the auspices of the Higher Education National Action Plan to determine the feasibility of imposing user fees at the university level. While the study indicated that there would be substantial private willingness to pay for university education, no policy actions have been taken on this front. 6. However, household survey data from the Cambodia Socioeconomic Survey (CSES) 1997 indicate that there is substantial scope for increased cost recovery in the education sector through charging official school fees. Indeed, the data indicate that households are already paying nontrivial amounts as school fees, or what they consider to be school fees, for their children's schooling at all levels. Table 1, which reports average out-of-pocket expenditure on school fees at public and private schools, shows that, on average, school fees in public schools are CR 3,512 per year at the primary level, CR 8,010 at the lower secondary level, and CR 10,215 at the upper secondary level. Obviously, since public schools officially do not have fees, these amounts represent informal payments to the schools. 7. In addition, Table 1 shows that households are willing to pay substantially more in school fees for better schooling quality. For example, for primary education, households are willing to pay ten times as much in school fees per student at private schools (which presumably provide better quality) than at public schools (CR 33,235 versus CR 3,512). While the ratio of private schools to public schools is 2 Other studies such as ADB's and Bray's found that the household contribution was much higher than the government's as their surveys included only households that have children in school and also included in-kind and labor contributions. In contrast, the calculation here is based on 1997 CSES data which include households that have children in school as well as those that do not. -88- somewhat lower at the secondary level, households appear willing to pay five to nine times as much as they are currently paying for improved school quality at these levels as well. Table 1: Average Out-of-pocket Expenditures on School Fees per Student, 1997 (riel per year) Cambodia Phnom Penh Other Urban Rural Public schools Primary 3.512 5,123 5.635 3,033 Lower secondary 8,010 6.947 7.107 8.680 Upper secondary 10.215 8.417 10.777 12,274 Private schools Primary 33,235 59.950 2,856 2,373 Lower secondary 74,016 79,115 0 Upper secondary 49,356 25.973 131.170 All schools Primary 3,680 7,001 5.569 3.014 Lower secondary 8,638 9,586 7.076 8.634 Upper secondary 10.914 8,871 10.522 13,786 Source Cambodia Socioeconomic Survey (CSES) 1997 D. Expenditure Trends 8. Overall Educational Expenditures. Total public spending on education increased nearly 60 percent from 1994 to 1996 -- from CR 140 billion to CR 225 billion. The increase in outlays on education occurred in both the national (MoEYS) budget and donor/NGO budget, although the increase in donor expenditures was much greater than that in national expenditures (85 percent versus 30 percent) (Figure 2). Since virtually the entire capital budget is provided by donors and NGOs and recurrent expenditures come out of the national budget, Figure 2 suggests that capital or development expenditures in the sector grew nearly three times as rapidly as recurrent expenditures. Figure 2: Aggregate public educational expenditure, 1994-96 (milb,ns of Riels) 230 000 130 000 I 0000 970,000 50 000 110000 - 7o.000 - - - - 1994 1995 1996 - - - MoEYS - - - - - -Donors and NGOs - Total Sources: MoEYS, MEF. and staff estimates. 9. The increase in public spending meant that, in 1996, total spending on education in Cambodia-- by the government, donors, NGOs, and households--was CR 30,079 per capita and CR 133,079 per student (US$11.42 and US$50.50, respectively) (Table 2). Total educational expenditure (by households, government, and donors) constituted about 3.7 percent of GDP in 1996, while total public (i.e., government and donors) spending on education constituted 2.7 percent of GDP and 11 percent of total government expenditure. If only public spending out of the national budget (and not by donors and NGOs) is considered, public spending on education constitutes about 0.9 percent of GDP. -89- Table 2: Overall Educational Spending per Capita and per Student, 1996 Education exp. per capita Education exp. per student Source Riels USS Riels iS$ Government Funded 8,044 3.05 35.589 13.51 Donors and NGOs 14.050 5.33 62.163 23.59 Households 7.985 3.03 35.327 13.41 Total 30,079 11.42 133.079 50.50 Sources MoEYS MEF and staff estimates. 10. Comparative data suggest that Cambodia's spending on education financed from its own source in relation to its GNP (0.9 percent) is very low in comparison with other countries in the region (Table 3). Cambodia spends a smaller proportion of its GNP on education than even Lao PDR, Bangladesh, and Nepal. Myanmar is the only country that has a lower ratio of public spending on education to GNP than Cambodia. Table 3: Comparative data on Government Spending on Education as a Share of GNP. 1995 Public spending as % of Country GNP Bangladesh 2.3 Bhutan 4.3 Cambodia 0.9 China 2.3 India 3.5 Lao PDR 2.4 Myanmar 1.3 Nepal 2.9 Pakistan 2.7 Sri Lanka 3.1 Source: UNESCO. 1998. 11. Distribution across Subsectors. Table 4 shows the subsectoral distribution of public spending on education from all sources. About 40 percent of the recurrent spending on education went to primary education, 15 percent to secondary education, 13 percent to tertiary, and 18 percent to vocational education in 1996. (It should be noted that language education, French in particular, provided by bilateral donors was the predominant component of the tertiary education.) The sectoral allocations fluctuate a great deal over time. For example, the allocation to primary education dropped from 34 percent in 1994 to 25 percent in 1995 and increased to 40 percent in 1996. Correspondingly, the allocation to tertiary education went down from 26 percent in 1994 to merely 7 percent in 1995 and then increased to 13 percent in 1996. 12. On closer examination, it appears that the large fluctuations in subsectoral allocations are largely due to changes in donor/NGO priorities from year to year. Table 5, which shows the subsectoral allocation of the national MoEYS (recurrent) budget, indicates much greater stability in subsectoral allocation over time. Further, a comparison of Tables 4 and 5 suggests that the large allocation to tertiary education comes about because of donor funding of technical and managerial education. The share of tertiary education in the national MoEYS recurrent budget has remained only in the 2-5 percent range between 1994 and 1996. -90- Table 4: Public Spending on Education by Subsector, 1994-96 (From all Sources-Budget, ODA, and NGOs) 1994 1995 1996 Subsector Mn. riels % of total Mn. riels % of total Mn. riels % of total M. of Education 140,056 100.00 227.801 100.00 225.101 100.00 Sector Policies 14.991 10.70 25,385 11 14 20.324 903 Primary 47,769 34 11 56.257 24 70 90.827 40.35 Secondary 19,155 13.68 30.212 13 26 34.343 15.26 Tertiary 36.287 25.91 15,669 688 28.421 12.63 Techn & Manag. Edu. & Training 18.325 13.08 94.903 41.66 39.565 17 58 Other 3.528 2.52 5.376 2.36 11.621 5 16 Source: MoEYS. 1997 Table 5: Public Recurrent Spending on Education by Subsector, 1994-96 (From the national MoEYS budget) 1994 1995 1996 Function Mn. riels % of total Mn. riels % of total Mn riels % of total M. of Education 62.230 100.00 76.021 100.00 81.954 10000 Sector Policies 6.991 11 23 4.561 6 00 10.599 1293 Primary 35.090 56.39 40.180 52.85 41.757 50.95 Secondary 18,039 2899 25.704 33 81 26,147 31.90 Tertiary 1.471 2.36 4.078 5.36 2.764 3.37 Other 639 1.03 1.498 1.97 687 0.84 Source: MoEYS. 1997 13. Financing of Recurrent Inputs. Table 6, which reports the MoEYS (national) budget by line item in 1995 and 1996, indicates that wages --mostly teacher salaries-- account for nearly three-quarters of the MoEYS budget, with operations and maintenance (O&M) accounting for the remaining quarter of the budget. The proportion of the national education budget spent on wages is high in comparison with other countries (see Table 7). The dominance of teacher salaries in the recurrent education budget means, of course, that complementary inputs, such as textbooks, are underfunded. Inadequate provision of textbooks and other learning materials in turn reduces the effectiveness of teachers.' Table 6: MoEYS Budgetary Expenditures, by Line Item, 1994-96 1994 1995 1996 Exp. Exp. Exp. Line Item Mn. riels % of total Mn. riels % of total Mn. riels % of total Total MoEYS Budgetary Expenditure 62.230 100.00 76.021 100.00 81,953 100.00 Wages 53.213 85 51 59.133 77.78 58,753 71.69 O&M 7.377 11.85 13.302 17.50 20,462 24.97 Social Subsidies 1,032 1.66 1.398 1.84 1.271 1.55 International Subsidies 14 0.02 6 0.01 66 0.08 Construction & Equipment 594 0.95 2.182 2.87 1,401 1.71 Source: MoEYS. 1997. For example, the CSES 1997 revealed that fewer than 19 percent of primary school students and 15 percent of lower secondary students in the country were enrolled in schools that had adequate books for all students. -91- Table 7: Teacher Emoluments as Percentage of National Recurrent Education Budget, 1995 Country Wages as % of recurrent total Cambodia 71.7 Lao PDR 65.0 Malaysia 62.1 Myanmar 56.8 Thailand 61.6 Source: Figures for countries other than Cambodia are from UNESCO (1998). 14. There are some encouraging signs in that the share of O&M in the national budget has more than doubled between 1994 and 1996. Correspondingly, the share of wages (mostly teacher salaries) has fallen from 86 percent of the national recurrent budget in 1994 to 78 percent in 1995 and 72 percent in 1996. A continuation of this shift will be highly desirable, as it will assure a better supply of non-teacher inputs in schools. Despite the large drop since 1994, the wage share in the national recurrent education budget remains high not only in comparison with other sectors but also in comparison with other countries in the region (Table 7). 15. Center-Province Distribution. Only 18 percent of the national education budget is distributed at the central level; the remaining 82 percent goes to the provinces. Central spending consists primarily of budgets for tertiary and research institutions based in Phnom Penh and for the operating expenses of MoEYS headquarters. The heavy concentration of public education expenditure in the provinces is an anomaly in Cambodia, since the provinces account for only 14 percent of total government expenditure. This suggests that public spending in the education sector is considerably more decentralized than in other sectors. 16. Provincial Distribution. Two observations can be made about provincial distribution. First, there is substantial variation in government education expenditures per capita and per student across provinces. Low population provinces such as Mondulkiri, Ratanakiri, and Steung Treng have significantly higher per student government education expenditures (CR 198,196, CR 57,010, and CR 54,574, respectively) than large provinces such Phnom Penh (CR 28,634). In part, this reflects the difficult (mountainous) terrain as well as the scattered populations in these provinces, which results in a low number of pupils per school and thus drives up the cost per student. 17. Second, there appears to be a positive relationship oetween government education expenditures per capita/per student and provincial GDP per capita (as proxied by household consumption expenditure per capita). This is observed in Figures 3 and 4, which plot the two variables against each other. What this suggests is that the formula for allocating government education expenditures across provinces unintentionally favors the richer provinces. One reason for this may be that it is easier for the more affluent provinces to attract and retain teachers and, since teachers constitute the single largest item of expenditure on education, this tends to drive up the per student and per capita education expenditures.4 This trend is contrary to the principle of targeting social sector expenditures to the poorest provinces and creates inequity in the provincial distribution of government education expenditures. 4 Another reason for the observed positive relationship could be that wealthier provinces demand more education or contribute more to education, and government expenditures on education are complementary to private household spending on education. -92- Figure 3: Public education expenditures per student and household Figure 4: Public education expenditures per capita and household consumption expenditure per capita, Cambodian provinces, 1996 consumption expenditure per capita, Cambodian provinces. 1996 60.000 8.000 55,000 7 S00 50000 ~~- 450000 20,000 00 o35.000 S 30.0W0 -sw~ .0 25.000 - .000) iL '0.000 .2 13,000 17,000 21,000 25000 29.000 13 000 17000 21,000 2S,000 29.000 Household consumption expenditure per capita (riels) Household consumption expenditure per capita (riels) E. Efficiency and Equity of Public Spending 18. Given the large social returns to schooling in low income countries, more public spending on education is generally desirable. Indeed, the success of the East Asian tigers is often attributed to their high levels of public spending on education. However, an equally important question is whether Cambodia is getting its money's worth from its public spending on education. One way of answering this question is by looking at the outcomes "produced" by public spending on education for a sample of Asian countries for which data are available, which shows that countries that spend a larger share of their GNP on education tend to have significantly higher gross enrollment rates. Cambodia is slightly above the regression line in the graph, which suggests that it enjoys the overall enrollment rate consistent with its public spending on education. Figure 5: The relationship between combined first. second and third-level gross enrollment ratio and public expenditure on education in selected Asian countries, 1995 80 - 75 70 -65 Cambodia- 7 60 41 1 2 1 4 6 7 8 9 Piblii cfpi ire on eduaiin is U Nof G'SP 19. Figure 5 does not necessarily suggest that there are no inefficiencies in public education expenditures in Cambodia or that there is no scope for raising the effectiveness of public spending on education. It merely indicates that Cambodia has been as successful as other Asian countries in obtaining the maximum enrollments out of its limited public expenditures. Indeed, as the discussion below points out, there is substantial scope for improving the effectiveness of public education spending even further, making educational opportunities more equitable, and for enhancing the quality of schools. 20. Allocation between Teacher Salaries and Other Recurrent Inputs. The current allocation of resources within the education budget is too heavily concentrated in wages, with 72 percent of the national MoEYS budget devoted to teacher salaries. Within the primary and secondary sector budgets, teacher salaries are likely to account for an even greater proportion of recurrent expenditure. As a result, -93- hardly any public resources are left for other school inputs such as learning materials and textbooks, and communities and households are held responsible for providing these complementary inputs. In practice, communities and households, especially those at the bottom of the income distribution, cannot afford to provide these inputs. Thus, when they could be teaching, teachers spend their time copying material verbatim from their own textbook to the blackboard, and students must recopy this material in their exercise books. 21. Distribution across Schooling Levels. A large body of evidence from other developing countries shows that the social returns to schooling are greatest at the primary level and smallest at the tertiary level. Even for Cambodia, the private returns to schooling have been estimated to be significantly higher at the primary level (33 percent for males and 40 percent for females) than at the secondary level (5.7 percent for males and 19.9 percent for females) (UNDP, 1997). Thus, the efficiency of public spending on education can be increased by shifting expenditures from secondary and tertiary education to primary education. 22. In addition, there is a strong equity consideration in having a larger share of the government budget devoted to primary education. The representation of the poor is much greater in the primary student population than in the secondary or tertiary student population. This is particularly true of Cambodia, where 20 percent of primary students, but only 2 percent of upper secondary students are drawn from the poorest 20 percent of the population. In contrast, 61 percent of the upper secondary students in the country come from the richest 20 percent of the population (Table 8). The representation of the poor in tertiary education is zero, while the richest 20 percent of the population accounts for 57 percent of tertiary education. This means that public expenditures on upper secondary and tertiary education will benefit disproportionately the more affluent sections of society, while expenditures on primary education will relatively benefit the poor. Table 8: Distribution of Enrolled Students, by Quintile and by Schooling Level, 1997 (percent of total population in each quintile) Per capita expenditure quintile Poorest II Ill IV Richest Total Primary 20.32 21.22 21.59 18.87 1801 100.00 Lower secondary 4.85 13.34 18.51 24.85 38.44 100.00 Upper secondary 1.96 4.23 11.98 20.83 61.00 100.00 Post-secondary 0.00 3.74 4.88 34.35 57.02 100.00 Sources: AloEYS, AEF and staff estimates. 23. As in many developing countries, in Cambodia tertiary and technical education (including language, technical, and management training) takes up a disproportionately large share of public spending on education. While tertiary and vocational students account for merely 0.5 percent of all students in Cambodia, public spending (financed by government, donors, and NGOs) on tertiary and technical education took up 30 percent of total public spending on education in 1996. In contrast, primary education, which accounts for 85 percent of all students in the country, received only 40 percent of the total allocation. As noted earlier, at this stage of development such an allocation is not only inequitable, it is also inefficient, given the higher private (and presumably social) returns to primary schooling in Cambodia. It could be argued that R&D from higher education would have big externalities and that there would be a capital market constraint (market failures) since higher education is expensive and human capital is hard to borrow against. But Cambodia is still far away from any basic and applied research and almost all of the people obtaining higher education come from the richest income quintile who would not be credit-constrained. -94- 24. Unit Public Costs of Schooling. Public spending per pupil varies a great deal across levels -- from CR 40,248 (US$15) at the primary level to CR 1,949,181 (US$740) at the tertiary level (Table 9). As noted earlier, the large per pupil expenditure at the tertiary level is mainly due to the large capital expenditures on tertiary education by international donors. The variation in per pupilrecurrent costs is much smaller--from CR 21,263 (US$8) at the primary level to CR 242,053 (US$92) at the tertiary level. Table 9: Public Spending on Education per Pupil, by Subsector, 1996 Total Recurrent spending spending No. of Total spending per pupil Recurrent spending per pupil Level Mn. riel Mn. riel students in riels in USS as % of GDP in riels in US$ as % of GDP Primary 79.039 41.757 1.963.799 40.248 15.27 4.97 21.263 8.07 2.63 Secondary 32,374 26.147 327.566 98.832 37.51 12.21 79.822 3029 9.86 Tertiary 22.258 2.764 11.419 1.949.181 739.73 240.71 242.053 91.86 29.89 Note: Tertiary does not include management training and technical education. Sources: MEF AfoEYS. and World Bank estimates 25. To compare these unit public costs of schooling with those in other countries, the recurrent spending per pupil is been shown in Table 10 as a percent of GNP per capita for nine developing countries in Asia for which data are available. Cambodia's per pupil recurrent spending on primary education in relation to its GNP per capita is the lowest -- even lower than that of Lao PDR and Myanmar. Cambodia's per pupil recurrent spending on secondary education is also the lowest in the group; however, it is comparable to Myanmar's spending and only slightly lower than the ratio for secondary expenditure in Thailand. While Cambodia spends more on tertiary education than Myanmar and Thailand, it still spends much less than other countries in the region. What these data suggest is that Cambodia's recurrent public spending per pupil at all levels is low in comparison with its GNP per capita. Table 10: Recurrent Public Spending per Pupil as a percent of GNP Per Capita, by Level, 1995-96 Country Primary Secondary Tertiary Bangladesh 6 23 30 Cambodia 3 10 30 China 6 14 81 India II 13 78 Lao PDR 5 25 55 Malaysia 9 22 77 Myanmar 4 10 21 Nepal 8 12 156 Thailand I I 11 25 Sources UNESCO (1998), except figures for Cambodia. which are based on AIfEF AoE YS. and Iorld Bank estimates. F. Evidence of Expenditure Leakage/Ineffective Targeting 26. An important element of cost recovery is to protect the poor and vulnerable from the burden of user fees. Exempting the poor from school fees in public schools can help improve a:cess to school for poor students from underprivileged backgrounds. Conversely, leakage of benefits intended for the poor to the better-off is a serious waste of resources and is contrary to the principle of targeting fee exemptions to the poorest. -95- 27. Data from the CSES 1997 on the proportion of students from different economic backgrounds who reported paying nothing for school fees suggest that school fee exemptions are given out sparingly in Cambodia. At the primary level, only about 7 percent of students pay nothing in school fees, in large part because there are no official school fees at this level (Table 11). As noted earlier, most of the fees collected from primary students in public schools are informal fees and special supplements. While a slightly larger percentage of primary students from the poorest quintile are exempted totally from paying school fees than from the richest quintile (9.8 percent versus 3.5 percent), this level of exemption is too small to make much of a difference. At the lower and upper secondary levels, there is no systematic pattern of fee exemptions with respect to economic status. Indeed, the data suggest that a slightly larger percentage of students from the richest relative to those from the poorest pay zero school fees. The conclusion from these data is that no mechanism currently exists in Cambodia for systematically exempting poor students from school fees, and that there is substantial leakage of benefits to the nonpoor that are in fact intended for the poor. Table 11: Percentage of Students Paying Nothing for School Fees in Public Schools, 1997 (in percent) Per capita exp. quintile I 11 Ill IV V Cambodia Phnom Penh Other Urban Rural Primary 9.83 6.58 7.42 5.72 3.52 6.72 2.77 5.68 7.33 Lower secondary 3.83 3.67 3.00 4.32 5.58 444 4.92 4.93 4.12 Upper secondary 9.08 9.68 2.49 6.24 743 6.69 7.07 11.82 4.23 Source: Cambodia Socioeconomic Survey (CSES) 1997. G. Scope for Private Sector Participation 28. Until 1991, the state was the exclusive provider of education at all levels. Since 1991, there has been some growth in the private sector in education. Nevertheless, the private sector's role in the provision of education is still relatively small. The CSES data indicate that, overall, less than 1 percent of students at all levels are enrolled in non-government schools. 29. Table 12 shows a generally strong, positive relationship between private schooling and per capita expenditure. The top two quintiles (representing the most affluent 40 percent of the population) are the only groups having any significant demand for private schooling. What this means is that, although the proportion of students enrolled in private schools is small at this time, it can be expected to grow significantly with economic growth and rising incomes in Cambodia. Much of the demand for private schooling in the early stages of economic growth will come from the richest quintiles. Table 12: Private Schooling (percent of students in private schools) Per capita expenditure quintile Poorest II Ill IV Richest Cambodia Phnom Penh Other Urban Rural Primary 0.66 0.16 0.15 0.37 2.30 0.68 3.67 0.72 0.30 Lower secondary 2.21 0.00 0.00 0.50 2.13 1.06 3.94 0.44 0.00 Upper secondary 0.00 0.00 0.00 224 2.75 2.14 3.44 0.00 1.27 Post secondary . 0.00 0.00 42.12 7.65 19.51 4.57 23.91 40.51 All levels 0.71 0.15 0.13 0.87 2.38 0.88 3.72 0.73 0.40 Source: Cambodia Socioeconomic Survey (CSES) 1997. 30. This is not altogether undesirable. Given the budgetary constraints that face the education sector, it is unlikely that the public sector can provide education at all levels to all children. As there exists -96- considerable scope for growth in the private provision of education, the government should look for ways to form partnerships with the private sector in order to maximize the amount of high quality education to the people. To this end, the government needs to focus spending more of its own scarce resources on the provision of basic and lower secondary education. At the same time, it needs to put in place policies which encourage private sector involvement in secondary and tertiary education. Such policies need to include a regulatory and quality control framework, public information systems, and policies which ensure access for the poor. 31 The regulatory role is an important one, as the government will need to set curriculum standards, see that they are upheld in public and private schools via school inspections, and establish quality control via annual national primary and secondary school leaving examinations. In order to satisfy the enormous growth in demand for secondary schooling that will accompany economic growth and rising incomes, the government needs to enlist the active support of the private sector in expanding secondary school coverage in the country. As in other developing countries, there exist explicit and hidden restrictions on the private sector's participation in the education sector in Cambodia; rationalization or removal of these barriers will be important if a partnership with the private sector is to be fostered. H. Institutional Constraints 32. A major problem in the education sector, as in most of the other sectors in Cambodia, is that spending decisions are highly centralized. The MoEYS, as well as the provincial and district education officers, have limited authority over budget and expenditure decisions. Although the MoEYS plans and negotiates for its budgetary allocations at the central and provincial levels, the responsibility for accounting and disbursement of funds lies (in the case of provinces and districts) with the provincial authorities. Thus, the provincial education departments need approval from the Ministry of Economy and Finance (MEF) as well as the provincial governor for any commitment of public funds. This is true of higher education institutions based in Phnom Penh as well. For example, a university needs to obtain approvals from the MoEYS and MEF to purchase even a photocopying machine. Not only does this budgetary system have too many steps and intermediaries, it results in district officers and school-based staff not knowing how much national budget they will receive each month, and thus having difficulties in planning and prioritizing their activities. 33. Another unique feature of the budgeting process is that a large number of capital expenditures are not reflected either in the national budget or in the PIP. In the period 1994 to 1997, for example, there were large capital outlays on construction and renovation of schools from several sources (international and national, political and private), bringing the total amount of expenditure on school building construction to over US$70 million over the period 1994-97. Since virtually none of this considerable building activity has been government funded, the national budget figures do not reflect these expenditures. While these school construction and renovation efforts are commendable, enough thought may not be paid to their long-term recurrent cost implications and obligations, which necessarily fall on the MoEYS. The MoEYS needs to exercise formal powers of approval over capital expenditures in education in the context of careful school mapping and planning exercises. 34. Finally, comprehensive documentation of the investment situation is further fragmented by the funding of certain universities and faculties through ministries other than Education (the Royal University of Agriculture through the Ministry of Agriculture, Forests and Fisheries; the Royal University of Fine Arts through the Ministry of Culture; and the Faculty of Medicine through the Ministry of Health). Since universities are simply departments of their ministries rather than autonomous institutions with their own budgets, the identification of university costs calls for a disaggregation of the budgets of several ministries, institution by institution and line by line. In some cases, this is impossible owing to the -97- ministry practice of retaining all day-to-day decisions on budget operations at the center. As a result, the university does not know what budget it has to spend, and goes cap in hand to its parent ministry's finance department for each individual expenditure proposal, even of the most trivial kind. 35. Teacher Salaries. A high priority for additional funding has been teacher salaries. Teacher salaries are so low that they cannot provide a minimum standard of living for a single person, let alone an entire family. For example, in 1997 the average monthly salary of MoEYS staff (teachers and administrators) was CR 55,032 (US$20.10), when the poverty line for a family of five was estimated to be CR 181,500 in the rural areas and CR 272,850 in Phnom Penh. The MoEYS is, therefore, under enormous pressure from teachers to raise their wages. In March 1994 all employees of the MoEYS, whether teachers or not, were granted the Prime Pedagogue (Teachers' Supplement) of CR 20,000 (US$8) per month, which works out to an approximately 50 percent raise. In February 1998 the Second Prime Minister granted a further raise to all teachers, of CR 11,000 per month. An unfortunate feature of both pay raises is that they have been politically determined on a single rate for all teachers. It would have been more prudent, and certainly more efficient, if the pay raises had been given out as incentive payments for better training and performance, longer teaching hours, and service in the rural areas. I. Evidence of the Impact of Public Expenditure/Donor Aid 36. The education sector in Cambodia has been receiving large amounts of donor assistance during the last few years. As noted earlier, international donors and NGOs spend nearly twice as much on education in Cambodia as does the government (although almost all of the donor/NGO funding is on capital expenditure). An important question is the extent to which this donor assistance -- and public spending on education in general -- has been effective in improving educational outcomes. 37. Figure 6 shows the trends in total student enrollment and the public Figure 6: Indenumben of tal enrollment and total public expenditures on education from the expenditureoneducation, 194-97([994=100) MoEYS budget for the period 1994-97. It 3s is clear from the figure that student enrollments have grown as public 20 expenditure on education has increased, although the rate of increase in MoEYS expenditure has been greater than that in 1994 1995 996 997 enrollments. Of course, the growth in enrollments could have occurred because of other factors, such as income growth. 38. Table 13 suggests that provinces in Cambodia that have higher levels of public spending on education per capita tend to have significantly higher gross enrollment ratios. At the sample mean, a I percent increase in public spending on education per capita is associated with enrollment ratio increases of 0.3 percent at the primary level, 2.2 percent at the lower secondary level, and 4.7 percent at the upper secondary level. The elasticities for secondary enrollment are extremely large, and indicate that, controlling for household consumption per capita, public expenditure on schooling can influence secondary enrollment rates in a considerable way. -98- Table 13: Provincial Regressions of Gross Enrollment Ratios on Public Expenditure on Education Per Capita and on Househol Consumption Expenditure Per Capita, 1996-97 Gross enrollment rates (%) Primary Lower secondary Upper secondary Independent Variable Parameter T-ratio Parameter T-ratio Parameter T-ratio Public educ exp. per capita ('000 riels) 4 672 1 67 8 573 4.61 4 364 4.11 Household cons exp. per capita ('000 nels) 0 002 0,02 0 168 2 70 0 115 3.23 Intercept 62 091 3.37 -32.121 -2.62 -24.280 -3 48 F-ratio 1.44 12 41 11 67 R-squared 0152 0.608 0.593 Source Staffestimates. 39. Table 14 reports the results of a similar regression, but with grade repetition rates as the dependent variables. A 1 percent increase in public spending on education per student is observed to reduce grade repetition rates by 0.06 percent in lower secondary schools and by 0.70 percent in upper secondary schools. Thus, higher levels of public educational spending per student reduce wastage in education by lowering repetition rates at the secondary school level (perhaps by improving the quality of instruction). Table 14: Provincial Regressions of Grade Repetition Rates on Public Expenditure on Education Per Student and on Household Consumption Expenditure Per Capita, 1996-97 Grade repetition rate (%) in: Lower secondary Upper secondary Independent Variable Parameter T-ratio Parameter T-ratio Public educ exp. per student ('000 riels) -0.041 -262 -0 108 -3 14 Household cons. exp. per capita ('000 riels) -0 056 -3 11 -0.070 -2 49 Intercept 13 022 5 86 14778 8 80 F-ratio 5.10 3.98 R-sq uared 0389 0.332 Source Staff estimates. 40. In addition, various donor-driven initiatives have made significant contributions to subsectors in education. Among the inputs financed by donor investments, the most striking and most easily quantifiable is the growth in the output of textbooks since the collapse, at the beginning of the decade, of Eastern bloc support to school publications. UNICEF responded to the urgent need created by commissioning the CODE study (1991), and then investing in the provision of paper, equipment, fuel. and training to establish a modem printshop capable of large-scale quality print runs. 41. Equally important is the qualitative dimension, realized through the UNICEF-led development of professionally designed competency based curricula and the associated rewriting of textbooks and teacher manuals. With the development of the Textbook Master Plan (led and majority financed by the Asian Development Bank), the MoEYS now has a professional publishing house managing a plan to supply schools with textbooks adequate in both quantity and quality (for example, the number of textbooks and teachers' manuals printed for grades I to 9 increased from 432,475 in 1986-87 to 3,794,000 in 1997-98. 42. Another large-scale donor-led input -- the European Union's Program of Assistance to Primary Education (PASEC) -- has resulted in major system-wide investments in the improvement of primary education. These inputs include the establishment ofa program of distance education for 45,000 primary teachers, supported by the training of 650 trainers of teachers for posts in teachers colleges and -99- demonstration schools and as field trainers and supervisors. Complementary activities of the project include the financing of the Social Sciences component of the Textbook Master Plan, the initiation of community school links and of the teachers' association, and the establishment of a system of small grants to schools to foster community led school initiatives. 43. In 1995 UNESCO, with the support of UNICEF, initiated a professionally designed Education Management Information Systems (EMIS) program which has generated two years' worth (1996-97 and 1997-98) of comprehensive school system statistics (providing much of the data needed to evaluate system performance in this review). UNESCO complemented the EMIS activity with a capacity building program which has provided professionally designed and comprehensive on-the-job training to 43 MoEYS administrators. The training program started with English training (for access to professional training in the region as part of the capacity-building program), training in administrative skills (covering planning, budgeting, human resource development, organization, and management), and leadership, motivation, and communication skills development. The high quality training materials covering these skill areas are now at the disposal of the 28 national staff trained as trainers to extend these skills at the provincial, departmental, and institutional levels. 44. In cooperation with Redd Barna and a number of other NGOs, UNICEF has promoted, since 1991, a system of cluster-based school support which has now been adopted as Ministry policy and is increasingly being used by various donors working at the primary school level. Evidence collected by UNICEF suggests that the introduction of a sub-district level of school support may be an effective means of ameliorating one of the system's most pervasive problems, that of high repetition and dropout rates. The analysis of statistics from three sample clusters6 of five schools -- one urban, one rural, and one minority area -- indicates significant improvements in average cluster enrollment, repetition, and dropout rates over a four-year period (Table 15). Table 15: Improvements in Participation Rates in Three Cluster Samples, 1993-94 to 1996-97 Net enrollment rates Repetition rates Drop-out rates Type of cluster 1993-94 1996-97 1993-94 1996-97 1993-94 1996-97 Urban 63.92 85.84 28.16 20.50 2.14 0.29 Rural 72.62 90.87 38.19 19.82 7.85 3.80 Minority cluster 58.80 83.11 40.10 30.79 6.57 5.19 Source: UNICEF. 1998. 45. Higher education has also benefited considerably from substantial investments by a number of donors. Chief among these have been French Cooperation, which has given consistent priority over a six- year period to the development of quality programs at the Institute of Technology, the Faculties of Medicine and of Law and Economic Sciences, and the Royal University of Agriculture at Chamear Daung. 46. At the level of non-formal education, UNESCO's Community/Temple Learning Center (CTLC) Program is a five-month program being pilot-tested in seven districts in Battambang Province, five in Siem Reap Province, and one in Kandal Province. Records indicate a substantial level of participation -- for example, of 5,948 adults in Siem Reap and adjoining districts in seven sessions between October 1994 and March 1998. Of these, 3,924 were deemed, on the basis of an interview at the end of the course, to While all primary schools in the country are now nominally grouped into clusters, the benefits of clusters are only readily apparent in those which have benefited from organized and regular support and supervision, generally sponsored by an international organization such as UNICEF or an interested NGO. -100- have achieved literacy along with other knowledge in such applications as health and agriculture. It is notable that, as with other literacy programs, the majority of the program participants are women, perhaps reflecting the needs created as a result of the relatively high dropout rate of girls from schools. J. Evaluation of the PIP 47. The proposed Public Investment Program (PIP) for 1998-2000 calls for the government to raise public expenditure on education from 1.4 percent to 2.8 percent of GDP and from 12 percent to 15 percent of the national budget by the year 2000. The PIP sets a target of having 65 percent of all 12-year olds complete Grade 6 and 50 percent of all 16-year olds complete Grade 10. The PIP calls for enhancing the cost-effectiveness of delivery by reducing non-instructional staff levels and increasing the pupil/teacher ratio through redeploying teachers in administrative positions; extending hours of instruction from 600 to 700; establishing basic school services based on the cluster model; and improving the monitoring of standards through more regular examinations. There is a commitment to exploring greater cost recovery, mainly via the granting of tax incentives to individuals to provide financial support to educational institutions and via encouragement of payment for vocational and higher education by future employers of graduates. The PIP also calls for implementation of a Text Book Master Plan and a Teacher Training Master Plan to improve teacher capacity and educational effectiveness. Finally, the PIP calls for encouraging the entry of the private sector into the provision of schooling by a more transparent accreditation program. 48. The PIP targets and goals appear overly ambitious, given the limited resources, weak implementation capacity, and inefficiencies of the existing budgetary system in the MoEYS. Increasing the share of government educational expenditures to 2.8 percent of GDP will mean spending at least an additional CR 82 billion each year (and probably much more if GDP expands over time) in recurrent education costs out of the national budget. It is not clear how these resources will be sourced, given that there is little, if any, support for recurrent expenditures by donors. In addition, there is the real question of sustainability in the sector as the bulk of the (capital) investment to be made under the PIP is expected to come from donor resources. K. Summary and Recommendations 49. While Cambodia has made impressive progress in expanding access to basic education in the last two decades, the education sector is still beset by numerous problems (less than universal primary enrollment, especially among the poor; extremely low rates of enrollment at the secondary and tertiary levels; and a very poor quality of education). The poor quality of schooling has led to very high rates of dropout and grade repetition. The low enrollment rates of the poor arise primarily from the high private costs of public schooling and the lack of physical access of the poor to good quality schools. There is a severe shortage of secondary and post-secondary institutions in the country, and this acts as a constraint on the expansion of post-primary enrollment rates. 50. The evidence in this Annex indicates that Cambodia underspends significantly on education relative to other countries in the region. In addition, education spending per pupilas a share of GNP per capita is significantly lower in Cambodia than in virtually any other country in the region. 51. It is thus clear that it will be imperative for Cambodia to increase the amount of resources going into education. However, the increase in resources will have to come out of the national budget for two reasons. First, donors and NGOs already account for nearly one-half of all expenditures on education in the country. It is unlikely that they will substantially increase their contributions in the near future. Second, almost all of the donor assistance is on capital expenditure, such as the construction and -101- rehabilitation of new school facilities and equipment. However, the new school facilities that are created or rehabilitated have major recurrent cost implications for the government. If a minimum commitment of recurrent expenditure cannot be made from the national budget, the new school facilities are likely to be underutilized and ineffective. Thus, the MoEYS needs to explore new and innovative mechanisms for financing recurrent educational expenditures. One such mechanism would be greater cost recovery, especially in higher education. 52. A recent Higher Education Taskforce recommended that it would be possible to introduce cost recovery in a phased manner in institutions of higher education in Cambodia. In addition, the experience of the Ministry of Health in introducing user fees at central and specialty hospitals based in Phnom Penh suggests that limited cost recovery is an option that needs to be explored by the MoEYS in the near future. Introducing user fees is also important for equity. Public expenditure on upper secondary and tertiary education is currently largely benefiting the most affluent groups. At the same time, although primary schools are free in principle, the actual cost of primary schooling (in terms of informal fees and expenditure on textbooks and uniforms) is high and unaffordable for the poor. By introducing formal user fees in upper and post-secondary institutions, the MoEYS may be able to make basic (primary and lower secondary) education truly free for the poor over the longer term. 53. Another means of expanding access to basic education for the poor is by allowing the private sector to grow in the provision of education. The current share of the private sector in school enrollments (less than I percent) is much smaller in Cambodia than in other countries in the region. The private sector can play a very important role in reducing the cost burden on the government by catering to the schooling demand of the better-off segment of the population, especially in tertiary and technical education, and leaving government schools to target their services to the poor. The growth of the private sector in the provision of schooling could be constrained by various hidden restrictions, including complicated accreditation requirements. The MoEYS needs to examine whether such restrictions are inhibiting private sector participation, and if they are, they should be rationalized or removed. At the very least, the MoEYS will need to put in place a more transparent school accreditation program, as envisaged in the PIP. 54. In addition to increasing the amount of resources going into education out of the national budget significantly, some reallocation of resources will be needed to increase the effectiveness and equity of public spending on education. Three types of reallocation are needed. First, the recurrent education budget in Cambodia is too heavily concentrated in wages. As a result, there are few public resources left for school inputs such as learning materials and textbooks. To raise the effectiveness of public spending on education, the allocation to non-teacher recurrent inputs will need to be raised substantially. A second type of reallocation needed is a shifting of resources from tertiary and technical education (including language, technical, and management training) to basic and lower secondary education. Not only is the current allocation inequitable at this stage of development, it is also inefficient given the higher private (and presumably social) returns to primary schooling in Cambodia. At the same time, it needs to put in place policies which encourage private sector involvement in tertiary and technical education. A third type of reallocation would be a shift of public spending from well-served to underserved provinces. The current provincial distribution of public spending on education is inequitable, with richer provinces having higher levels of government educational expenditures per capita and per student than poorer provinces. Equity considerations demand that government expenditures be targeted toward the poorest provinces with the lowest educational achievements. 55. Another important and overdue policy reform relates to teacher salaries. Despite recent raises, teacher salaries remain so low that they cannot provide a minimum standard of living for a single person, let alone an entire family. Low teacher salaries have a direct adverse impact on educational effectiveness -102- and school quality by reducing teacher motivation and morale and, in turn, increasing rates of teacher absenteeism and attrition. It is very common for teachers in Cambodia to supplement their incomes by other pursuits, a practice that increases teacher absenteeism. Teacher absenteeism reduces student learning and achievement, while teacher attrition increases the cost of teacher training. Increasing teacher salaries when teacher emoluments already take up 72 percent of the recurrent education budget is difficult. However, there are two ways to accomplish this. First, a substantial increase in the recurrent education budget, as recommended earlier, will allow greater maneuverability to raise teacher salaries without necessarily increasing the share of teacher emoluments in the total recurrent budget. Second, some savings could be realized from increasing the average pupil-teacher ratio from its current level of 45 at the primary level and 18 at the secondary level, and these savings could be used to finance teacher salary increases. There is a large literature on schooling effectiveness which suggests that smaller pupil- teacher ratios are not necessarily associated with better student learning or greater student achievement. Certainly, at the secondary level, there appears to be some scope for increasing the pupil-teacher rat . in Cambodia. 56. Finally, as in the case of other social sectors, the vertical budgeting process employed in Cambodia needs to be streamlined for greater effectiveness of public educational spending. The overly centralized budgetary system prevents the MoEYS from planning and prioritizing effectively, and often results in underutilization of budgeted resources. The MoH and the MEF have jointly developed and implemented, on a pilot basis, a new cash release process in the health sector, the Accelerated District Development (ADD) system. The new system allows cash to be provided to the ADD districts, via the MoH and the Provincial Health Departments, before expenditure is incurred and accounted for afterward. By reducing the number of administrative layers in the budget disbursement system, the ADD program speeds up disbursements and reduces the discrepancy between budgeted and actual government expenditures at the district level. It may be useful for the MoEYS to negotiate a similar cash release process with the MEF, and pilot-test it in a few districts. - 103 - Statistical Appendix List of Tables Table 1: Key Economic Indicators Table 2: Gross Domestic Product by Industrial Origin at 1989 Constant Prices Table 3: GDP by Industrial Origin at Current Prices Table 4: Growth Rates of GDP by Industrial Origin at 1989 Constant Prices Table 5: Shares of GDP by Industrial Origin, at 1989 Constant Prices Table 6: Consumer Price Index and Exchange Rates Table 7: Monetary Survey, 1991-1998 Table 8: Balance of Payments, 1991-1997 Table 9: External Debt, 1991-1996 Table 10: Summary of Budget Operations, 1991-1998 Table I1: Cambodia: Structure of Revenue, 1994-1998 Table 12: Budgetary Expenditure by Economic Classification, 1994-1998 Table 13: Consolidated Public Expenditure by Sector in 1994 Table 14: Consolidated Public Expenditure by Sector in 1995 Table 15: Consolidated Public Expenditure by Sector in 1996 Table 16: Consolidated Public Expenditure by Sector in 1997 Table 17: Consolidated Public Expenditure by Sector and Chapter in 1994 Table 18: Consolidated Public Expenditure by Sector and Chapter in 1995 Table 19: Consolidated Public Expenditure by Sector and Chapter in 1996 Table 20: Consolidated Public Expenditure by Sector and Chapter in 1997 Table 21: Consolidated Public Expenditure by Region in 1994 Table 22: Consolidated Public Expenditure by Region in 1995 Table 23: Consolidated Public Expenditure by Region in 1996 Table 24: Consolidated Public Expenditure by Region in 1997 Table 25: Number of Civil Employees and Their Salary Level in Cambodia, 1997 Table 26: Pay and Allowance of Military Personnel in Cambodia, 1997 -104 - Table 1: Key Economic Indicators 1998 1991 1992 1993 1994 1995 1996 1997 (Et. (Est.) (Percentage Change) Real GDP 7.6 7.0 4.1 4.0 7.6 7.0 1.0 0.0 CPI (final quarter basis) 150.4 112.5 41.0 17.9 3.5 9.0 9.1 12.0 Domestic Liquidity 28.6 214.1 31.7 37.7 44.2 40.5 16.4 6.8 Net Credits to Government 22.9 141.8 5.6 -7.9 1.1 -3.1 -7.9 12.2 Velocity of Money 21.8 16.3 19.6 15.4 13.7 10.5 9.4 10.1 (Million of US$) Export of Goods (USS. excluding re-export) 81 101 168 234 269 298 404 469 Import of Goods (USS, retained import) 113 160 361 509 673 749 707 775 Gross official reserves .. 30 71 100 182 234 264 386 (Months of import of goods and services) .. 1.0 1.8 1.4 1.5 2.1 2.5 3.5 (Percentage of GDP unless otherwise specified) Budget Revenue 4.4 6.2 5.4 9.6 8.9 9.1 9.7 8.1 Tax 2.3 4.4 4.3 5.9 6.2 6.5 6.6 5.9 Non-tax (including capital revenue) 2.1 1.8 1.0 3.7 2.7 2.6 3.0 2.1 Budget Expenditure 7.8 9.8 11.2 16.5 16.7 16.3 13.9 11.7 Current expenditure 7.4 9.5 6.9 11.0 9.6 9.9 9.0 8.3 Capital expenditure 0.4 0.3 4.3 5.5 7.1 6.4 5.0 3.4 Current budget balance -3.0 -3.3 -1.5 -1.4 -0.6 -0.6 0.6 -0.5 Overall budget balance -3.4 -3.6 -5.9 -6.9 -7.8 -6.5 -4.3 -3.7 Domestic Investment 9.4 9.8 14.3 18.5 21.2 20.4 16.1 Government Investment 0.4 0.3 4.3 5.5 7.1 6.4 5.0 Non-government investment 9.0 9.5 10.0 13.0 14.1 14.0 11.1 Financing oflnvestment 9.4 9.8 14.3 18.5 21.2 20.4 16.1 National savings 7.9 7.3 5.6 4.8 5.4 4.9 4.7 Government savings -1.2 -4.3 -1.4 -1.4 -0.6 -0.8 0.6 Non-Government savings 9.1 11.6 6.3 6.2 5.7 5.7 4.2 Foreign savings 1.5 2.5 9.4 13.7 16.1 15.5 11.4 External Current Account Deficit (US$ millions) -28 -50 -190 -329 -474 -485 -346 -329 (in percent ofGDP) -1.5 -2.5 -9.4 -13.7 -16.1 -15.5 -11.4 -11.6 Memorandum items: Nominal GDP (billions ofRiels) 1,336 2,508 5,414 6,131 7,200 8,250 9,100 10,750 Official exchange rate (Riels/USS) 703 1,253 2,470 2,543 2,462 2,624 2,989 3,700 Source: Cambodian authorities. - 105 - Table 2: Gross Domestic Product By Industrial Origin at 1989 Constant Prices (in billions of riels) 1991 1992 1993 1994 1995 1996 1997 Agriculture 135.9 138.5 137.1 137.1 145.9 149.4 151.2 Crop & Rubber 79.3 79.0 77.2 70.9 82.2 83.4 83.1 Rice 47.6 46.7 48.2 38.6 47.9 48.0 46.8 Other Crop & Rubber 31.7 32.3 29.0 32.4 34.3 35.4 36.3 Livestock 34.5 36.9 38.0 39.5 41.1 44.2 46.4 Fishery 13.5 12.7 12.0 11.8 12.2 12.1 11.9 Forestry 8.6 9.9 9.9 14.9 10.4 9.7 9.8 Industry 39.5 45.7 51.7 55.6 61.1 72.2 70.1 Mining and Quarrying 3.0 3.2 3.4 3.7 4.0 4.8 4.5 Manufacturing 18.4 19.0 20.5 22.1 24.3 27.4 29.5 Electricity and Water 0.5 0.6 0.7 0.7 0.8 0.9 1.0 Construction 17.6 22.9 27.1 29.1 32.0 39.0 34.6 Services 86.8 96.5 103.3 111.0 119.7 128.1 131.4 Transport and Communication 7.0 8.1 8.9 9.7 10.7 11.8 12.4 Wholesale and Retail Trade 34.4 39.9 42.3 44.8 48.8 53.3 54.8 Hotels and Restaurants 1.0 1.3 1.4 1.7 1.8 2.2 2.1 Govt. Admin., Education & Health 11.0 11.0 11.5 12.5 12.9 12.7 12.8 Home Ownership 16.3 17.4 18.8 20.3 21.8 23.1 23.6 Other Services 17.1 18.8 20.4 22.1 23.7 25.1 25.5 Gross Domestic Product 262.2 280.7 292.1 303.7 326.8 349.7 353.2 Memorandum Items Nominal GDP 1,336 2,509 5,414 6,131 7,200 8,251 9,100 Nominal GDP in million US$ 1,900 1,981 2,013 2,409 2,938 3,144 3,044 Source: Cambodian authorities and World Bank/IMF estimates. - 106 - Table 3: GDP By Industrial Origm at Current Prices (in billions of riels) 1991 1992 1993 1994 1995 1996 997 Agriculture 666 1,198 2,742 3,140 3,753 4,185 4,655 Crop & Rubber 300 721 1,724 1,853 2,374 2,628 2,831 Rice 139 425 1,017 961 1,354 1,478 1,556 Other Crop & Rubber 161 296 707 891 1,020 1,150 1,274 Livestock 225 303 678 803 922 1,065 1,208 Fishing 64 92 140 158 197 220 237 Forestry 78 83 200 326 260 272 326 Industry 162 330 743 853 999 ',262 1,325 Mining and Quarrying 5 9 16 18 22 28 29 Manufacturing 71 127 286 325 367 440 523 Electricity and Water 7 16 39 44 50 63 77 Construction 79 178 403 467 560 730 699 Services 508 981 1,929 2,138 2,448 2,804 3,139 Transport and Communication 43 86 173 195 248 292 343 Wholesale and Retail Trade 196 395 788 878 1,007 1,154 1,285 Hotels and Restaurants 6 13 24 31 35 46 50 Govt. Admin., Education & Health 63 108 216 232 269 266 296 Home Ownership 93 172 312 336 384 432 477 Other Services 108 206 416 466 505 614 687 Gross Domestic Product 1,336 2,509 5,414 6,131 7,200 8,251 9,100 Memorandum Items Average Official Exchange Rate (riels/USS) 703 1.267 2,689 2,545 2,451 2.624 2.989 Nominal GDP (US$ miion) 1.900 1,981 2,013 2,409 2,938 3144 3_! Source: Cambodian authorities and World Bank/IMF estimates. - 107 - Table 4: Growth Rates of GDP By Industrial Origin at 1989 Constant Prices (in percent) 1991 1992 1993 1994 1995 1996 1997 Agriculture 6.7 1.9 -1.0 0.0 6.4 2.4 1.2 Crop & Rubber 7.2 -0.4 -2.3 -8.2 15.9 1.5 -0.3 Rice -4.0 -1.9 3.2 -19.9 24.1 0.2 -2.5 Other Crop & Rubber 29.9 1.9 -10.2 11.7 5.9 3.2 2.6 Livestock 1.2 7.0 3.0 3.9 4.1 7.5 5.0 Fishery 8.9 -5.9 -5.5 -1.7 3.4 -0.8 -2.0 Forestry 24.6 15.1 0.0 50.5 -30.2 -6.7 1.1 Industry 8.8 15.7 13.1 7.5 9.9 18.2 -2.9 Mining and Quarrying 7.1 6.7 6.3 8.8 8.1 20.0 -5.5 Manufacturing 7.0 3.3 7.9 7.8 10.0 12.8 7.7 Electricity and Water 0.0 20.0 16.7 0.0 14.3 12.5 14.9 Construction 11.4 30.1 18.3 7.4 10.0 21.9 -11.4 Services 8.5 11.2 7.0 7.5 7.8 7.0 2.6 Transport and Communication 9.4 15.7 9.9 9.0 10.3 10.3 5.3 Wholesale and Retail Trade 12.8 16.0 6.0 5.8 9.1 9.2 2.9 Hotels and Restaurants 42.9 30.0 7.7 21.4 5.9 22.2 -2.5 Govt. Admin., Education & Health 0.0 0.0 4.5 8.3 3.6 -1.6 1.1 Home Ownership 2.5 6.7 8.0 8.0 7.4 6.0 2.3 Other Services 10.3 9.9 8.5 8.3 7.2 5.9 1.8 Gross Domestic Product 7.6 7.1 4.1 4.0 7.6 7.0 1.0 Source: Cambodian authorities and World Bank/IMF estimates. - 108 - Table 5: Shares of GDP By Industrial Origin, at 1989 Constant Prices (in percent) 1991 1992 1993 1994 1995 1996 1997 Agriculture 51.8 49.3 46.9 45.1 44.6 42.7 42.8 Crop & Rubber 30.2 28.1 26.4 23.3 25.2 23.8 23.5 Rice 18.2 16.6 16.5 12.7 14.7 13.7 13.3 Other Crop & Rubber 12.1 11.5 9.9 10.7 10.5 10.1 10.3 Livestock 13.2 13.1 13.0 13.0 12.6 12.6 13.1 Fishery 5.1 4.5 4.1 3.9 3.7 3.5 3.4 Forestry 3.3 3.5 3.4 4.9 3.2 2.8 2.8 Industry 15.1 16.3 17.7 18.3 18.7 20.6 19.8 Mining and Quarrying 1.1 1.1 1.2 1.2 1.2 1.4 1.3 Manufacturing 7.0 6.8 7.0 7.3 7.4 7.8 8.4 Electricity and Water 0.2 0.2 0.2 0.2 0.2 0.3 0.3 Construction 6.7 8.2 9.3 9.6 9.8 11.2 9.8 Services 33.1 34.4 35.4 36.5 36.6 36.6 37.2 Transport and Communication 2.7 2.9 3.0 3.2 3.3 3.4 3.5 Wholesale and Retail Trade 13.1 14.2 14.5 14.7 14.9 15.2 15.5 Hotels and Restaurants 0.4 0.5 0.5 0.6 0.6 0.6 0.6 Govt. Admin., Education & Health 4.2 3.9 3.9 4.1 3.9 3.6 3.6 Home Ownership 6.2 6.2 6.4 6.7 6.7 6.6 6.7 Other Services 6.5 6.7 7.0 7.3 7.3 7.2 7.2 Gross Domestic Product 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Cambodian authorities and World Bank/IMF estimates. - 109 - Table 6: Consumer Price Index and Exchange Rates Consumer Price Inflation Parallel Official Official Rate Index(CPI) Rate I/ Market Exchange /Parallel (March 1988=100) (M) Exchange Rate Rate 2/ Rate(%) 1990 March 293.2 128.5 380 345 9.2 June 378.7 143.2 480 380 20.8 September 558.9 150.5 675 510 24.4 December 604.5 152.3 613 600 2.1 1991 March 925.0 215.5 691 600 13.2 June 1137.4 200.3 905 800 11.6 September 1767.8 216.3 1125 1000 11.1 December 1135.6 87.9 550 520 5.5 1992 March 1360.6 47.1 935 800 14.4 June 1971.1 73.3 1275 1000 21.6 September 3224.3 82.4 2142 2000 6.6 December 3143.0 176.8 2310 2000 13.4 1993 March 6003.8 341.3 4200 3800 9.5 June 5651.2 186.7 2950 2700 8.5 September 4832.4 49.9 2700 2617 3.1 December 4117.0 31.0 2470 2390 3.2 1994 March 4299.0 -28.4 2527 2480 1.9 June 4588.0 -18.8 2542 2515 1.1 September 5064.0 4.8 2626 2606 0.8 December 5192.0 26.1 2601 2575 1.0 1995 March 4794.0 11.5 2392 2400 -0.3 June 5030.0 9.6 2371 2341 1.3 September 5463.0 7.9 2526 2496 1.2 December 5197.0 0.1 2560 2526 13 19963/ March 105.7 5.4 2591 2558 1.3 June 107.7 6.6 2667 2619 1 8 September 113.3 5.9 2697 2687 0.4 December 115.9 10.1 2731 2713 0.7 1997 March 111.7 5.7 2745 2735 0.4 June 115.6 7.3 2773 2761 0.4 September 1266 11.7 3343 3192 4.5 December 126.6 9.2 3523 3452 2.0 1998 March 1300 16.4 3618 3585 0.9 June 137.0 18.5 4051 4000 1.3 September 143.0 13.0 3882 3830 1.3 1/ Inflation rate is the percentage change in the CPI from the same month of the previous year. 2/ End-period buying rate. 3/ From January 1996, National Institute of Statistics'CPI was used; base: July-September 1994=100. Source: National Bank of Cambodia and National Institute of Statistics. -110 - Table 7: Monetary Survey, 1991-1998 (in billions of riels) 1991 1992 1993 1994 1995 1996 1997 Oct-98 Net Foreign Assets -1.7 24.2 101.9 390.8 550.0 881.0 1,177.0 1,665.6 Foreign Assets 16.1 203.2 298.5 631.5 897.0 1,230.0 1,602.0 2,121.6 Foreign Liabilities 17.8 179.0 196.6 240.7 347.0 350.0 425.0 456.0 Net Domestic Assets 80.7 223.9 224.9 59.2 99.0 31.0 -115.0 -531.5 Domestic Credit 87.6 267.3 355.3 386.3 446.0 567.0 697.0 886.9 Net Claims on Government 55.6 156.8 187.5 143.2 148.0 127.0 54.0 184.1 Claims on Government 55.6 170.0 218.5 215.0 217.0 213.0 211.0 289.2 Deposits of Government 0.0 13.2 31.0 71.8 69.0 86.0 157.0 105.1 State Enterprises 22.1 11.7 6.2 6.0 5.0 5.0 6.0 6.2 Private Sector 9.9 98.8 161.6 237.1 293.0 435.0 637.0 696.7 Other Items (Net) -6.9 -43.4 -130.4 -327.1 -347.0 -536.0 -811.0 -1,418.5 Total Liquidity 79.0 248.1 326.8 450.0 649.0 912.0 1,062.0 1,134.1 Narrow Money 77.6 171.8 203.8 199.0 279.0 329.0 385.0 478.5 Currency outside Banks 64.8 158.7 189.7 176.0 251.0 300.0 356.0 445.2 Demand Deposits 12.8 13.1 14.1 23.0 28.0 29.0 29.0 33.3 Quasi-Money 1.4 76.3 123.1 250.0 371.0 583.0 678.0 655.6 Time and Savings Deposits 0.7 11.0 8.5 18.0 5.0 8.0 13.0 15.9 Foreign Currency Deposits 0.7 65.3 114.6 233.0 366.0 575.0 665.0 639.7 Source: National Bank of Cambodia. - 111 - Table 8: Balance of Payments, 1991-1997 (in millions USS) 1991 1992 1993 1994 1995 1996 1997 Trade Balance -32.5 -86.2 -203.0 -275.4 -404.6 451.4 -303.3 Total Exports 212.5 264.5 219.0 461.7 808.6 658.8 733.8 Domestic Exports 67.3 51.3 102.0 233.7 268.6 297.6 403.6 Re- Exports 145.2 213.2 117.0 228.0 540.0 361.2 330.2 Total Imports 245.0 350.7 422.0 737.1 1213.2 1110.2 1037.1 O.w. Retained Imports 1/ 99.8 137.5 305.0 499.0 673.3 749.0 706.9 Services Balance 1.3 27.6 10.0 -74.6 -89.0 -56.6 -62.8 Receipts 14.0 49.7 66.0 59.7 121.1 171.7 186.2 Payments 2/ 12.7 22.1 56.0 134.2 210.1 228.3 249.0 Private transfers 3.5 9.0 4.0 20.0 20.0 23.0 20.0 Current Account Balance -27.7 -49.6 -189.0 -330.0 -473.6 485.0 -346.1 (excluding official transfers) Official Transfers 2.5 5.0 149.0 235.0 345.5 299.3 170.7 Current Account Balance -25.2 -44.6 -40.0 -95.0 -128.1 -185.7 -175.4 (including official transfers) Capital Account 15.2 38.0 75.0 136.0 167.6 238.1 204.7 Official medium-and long-term loans(net) 5.2 -2.1 5.0 54.0 71.4 59.0 37.3 Disbursements 5.2 0.0 5.0 61.0 75.3 75.2 37.7 Amortization 2/ 0.0 2.1 0.0 7.0 3.9 16.2 0.4 Foreign Direct Investment 20.0 40.1 0.0 80.0 150.7 239.9 135.0 Short-term flows and errors & omissions -10.0 0.0 70.0 2.0 -54.5 -60.8 32.3 Overall Balance -10.0 -6.6 35.0 41.0 39.5 52.4 29.3 Financing 10.0 6.6 -35.0 -41.0 -39.5 -52.4 -29.3 Change in gross official reserves 8.0 3.6 -41.0 -72.0 -82.0 -52.4 -29.6 Debt rescheduling 0.0 0.0 0.0 0.0 239.0 0.0 1.0 Arrears reduction 0.0 11.0 0.0 10.0 -239.0 0.0 0.3 IMF 2.0 -8.0 6.0 21.0 42.5 0.0 0.0 Purchase/disbursement 2.0 00 6.0 21.0 42.5 0.0 0.0 Repurchase/repayment 0.0 -8.0 0.0 0.0 0.0 0.0 0.0 Memorandum Items: Current Account Balance/GDP (excluding official transfers) -1.5 -2.5 -9.4 -13.7 -16.1 -15.4 -11.4 Current Account Balance/GDP (including official transfers) -1.3 -2.3 -2.0 -3.9 -4.4 -5.9 -5.8 Gross Official Reserves 3/ In Millions Of US Dollars .. 30 71 100 182 234 264 In Months Of Imports of Goods and Services .. 1.0 1.8 1.4 1.5 2.1 2.5 I/ Total imports (including project aid imports and estimate of unrecorded imports) less re-exports. 2/ Accrued interest on external debt and scheduled amortization on the existing stock of debt vis-A-vis official creditors are excluded. 3/ For 1992-1993, gross official reserves are the gross foreign assets of the National Bank of Cambodia and the Foreign Trade Bank. From 1994 onward, gross official reserves are of National Bank only. Source IMF estimates based on data provided by Cambodian authorities. -112 - Table 9: External Debt, 1991-1996 (in millions of US$) 1991 1992 1993 1994 1995 1996 Total Debt Stocks 1/ 1861.8 1839.8 1829.3 1914.9 2035.3 2100.5 Long-term debt 1688.6 1679.8 1685.4 1745.2 1946.3 2012.5 Public and publicly guaranteed 1688.6 1679.8 1685.4 1745.2 1946.3 2012.5 Private nonguaranteed 0.0 0.0 0.0 0.0 0.0 0.0 Memo: of which interest arrears 114.4 113.1 103.6 110.1 5.2 5.2 Official creditors 113.2 111.9 102.4 108.9 4.0 4.0 Private creditors 1.2 1.2 1.2 1.2 1.2 1.2 of which principal arrears 222.5 328.1 447.1 576.9 587.8 707.9 Official creditors 222.1 327.7 446.7 576.5 587.4 707.5 Private creditors 0.4 0.4 0.4 0.4 0.4 0.4 Use of IMF credit 26.8 14.9 8.6 29.6 71.7 69.4 Short-term debt 146.4 145.1 135.3 140.1 17.3 18.6 Principal Payments 0.0 11.2 15.1 0.0 4.1 4.2 Interest Payments 16.1 1.8 19.2 1.9 2.3 5.6 Debt Indicators Debt/Exports (%) 834.6 594.5 653.4 378.8 225.2 226.2 Debt/GNP (%) 99.7 93.6 92.5 82.1 71.3 72.3 Debt Service/Exports (%) 7.1 4.1 12.1 0.4 2.7 3.7 1/ The official exchange rate of USS= 0.554 rubles is used in converting ruble debt into US dollars. This valuation method does not constitute an endorsement of its appropriateness. Source: World Debt Tables. - 113 - Table 10: Summary of Budget Operations, 1991-1998 1/ 1991 1992 1993 1994 1995 1996 1997 1998 Estimated (In Billions of Riels) Revenue 58.8 156.1 290.1 590.3 643.0 754.7 881.0 867.6 Tax 31.1 109.7 234.1 364.6 445.5 534.4 597.4 638.2 O/w Customs Duties 22.0 79.3 159.5 280.9 320.8 344.1 347.4 354.6 Non-tax 27.8 46.3 56.0 224.5 189.8 181.1 271.3 200.0 O/w SOE Transfer 25.8 42.3 53.6 42.7 26.7 45.0 35.9 31.6 O/w Forestry Exploitation .. 1.4 3.9 86.0 52.8 27.5 37.4 16.4 Capital Revenue .. .. .. 1.2 7.7 39.2 12.3 29.5 Expenditures 104.1 245.6 608.4 1009.1 1200.7 1342.8 1267.9 1262.3 Current Expenditure 98.9 238.6 373.2 673.8 689.6 812.9 815.9 895.7 Defense & Security 46.8 118.6 219.4 398.2 398.2 399.9 419.3 460.6 O/w Salaries 22.5 56.9 105.3 192.1 214.8 220.6 252.3 315.0 Civilian 52.2 119.9 153.8 275.6 291.4 413.0 396.6 435.1 O/w Salaries 25.1 57.6 73.8 101.4 110.9 130.2 133.6 160.7 Capital Expenditure 5.2 7.1 235.2 335.3 511.1 529.9 451.9 366.6 O/w Domestically Financed 5.2 7.1 5.0 78.5 56.9 61.6 111.3 76.6 Current Balance -40.1 -82.5 -83.1 -83.4 -46.5 -63.8 52.8 -57.6 Overall Balance -45.3 -89.6 -318.3 -418.7 -557.7 -593.7 -386.8 -394.7 Financing 45.3 89.6 318.3 418.7 557.7 593.7 386.8 394.7 Foreign Financing 6.1 1.5 239.1 432.1 , 559.3 575.6 445.9 303.5 Project Aid 6.1 1.5 230.2 267.1 450.7 467.1 349.9 300.0 Budget Support 0.0 0.0 8.9 121.7 81.8 149.9 96.0 3.5 Domestic Financing 39.2 88.1 79.2 -13.4 -1.6 18.1 -59.0 91.2 O/w Bank Financing 14.5 112.1 30.7 -26.6 5.5 -17.0 -75.3 127.0 (In Percent of GDP) Revenue 4.4 6.2 5.4 9.6 8.9 9.1 9.7 8.1 Tax 2.3 4.4 4.3 5.9 6.2 6.5 6.6 5.9 O/w Customs Duties 1.6 3.2 2.9 4.6 4.5 4.2 3.8 3.3 Non-tax 2.1 1.8 1.0 3.7 2.6 2.2 3.0 1.9 Expenditures 7.8 9.8 11.2 16.5 16.7 16.3 13.9 11.7 Current Expenditure 7.4 9.5 6.9 11.0 9.6 9.9 9.0 8.3 Defense & Security 3.5 4.7 4.1 6.5 5.5 4.8 4.6 4.3 Civilian 3.9 4.8 2.8 4.5 4.0 5.0 4.4 4.0 Capital Expenditure 0.4 0.3 4.3 5.5 7.1 6.4 5.0 3.4 Current Deficit -3.0 -3.3 -1.5 -1.4 -0.6 -0.8 0.6 -0.5 Overall Deficit -3.4 -3.6 -5.9 -6.8 -7.7 -7.2 -4.3 -3.7 Financing 3.4 3.6 5.9 6.8 7.7 7.2 4.3 3.7 Foreign Financing 0.5 0.1 4.4 7.0 7.8 7.0 4.9 2.8 Domestic Financing 2.9 3.5 1.5 -0.2 0.0 0.2 -0.6 0.8 O/w Bank Financing 1.1 4.5 0.6 -0.4 0.1 -0.2 -0.8 1.2 1/ Cash basis Source: Cambodian authorities. -114 - Table 11: Cambodia: Structure of Revenue, 1994-1998 (In billions of riels) 1994 1995 1996 1997 1998 (Est.) Total revenue 590.3 643.0 754,7 881.0 867.6 Tax receipts 364.6 445.5 534.4 597.4 638.2 Direct taxes 8.6 20.9 26.5 43.6 49.5 Wage tax 0.0 1.1 2.7 5.7 8.1 Profit tax 6.3 17.8 21.8 35.0 37.0 Property tax 0.9 1.2 2.1 2.6 3.9 Others 1.5 0.8 0.0 0.3 0.5 Indirect taxes 75.0 103.8 163.8 206.5 234.1 Turnover tax 8.1 17.1 27.4 46.5 65.9 Consumption tax 46.9 60.0 70.4 75.2 83.4 Excise taxes 2.9 9.0 56.6 74.1 69.9 Others 17.0 17.7 9.4 10.7 14.9 Taxes on international trade 280.9 320.8 344.1 347.4 354.6 Taxes and duties on imports, of which 257.6 300.8 334.8 335.8 350.7 Taxes on exports 18.2 17.3 8.1 9.9 2.7 Others 5.1 2.7 1.2 1.7 1.2 Nontax receipts 224.5 189.8 181.1 271.3 200.0 Receipts on public property 148.6 120.5 96.7 133.5 84.6 Fisheries 4.8 9.1 7.1 7.4 8.8 Forests 86.0 52.8 27.5 37.4 17.9 Receipts from public enterprises 42.7 26.7 45.0 35.9 37.1 Royalties and concessions 4.2 17.0 5.1 33.1 4.4 User fees 0.1 6.5 6.9 6.6 7.6 Building leases 8.6 6.2 3.0 12.6 7.3 Other 1.0 2.0 2.1 0.5 1.5 Other nontax receipts 75.9 69.3 84.4 137.8 115.5 Capital Revenue 1.2 7.7 39.2 12.3 29.4 Source: Data provided by the Cambodian authorities. - 115 - Table 12: Budgetary Expenditure by Economic Classification, 1994-1998 1/ 1994 1995 1996 1997 1998 (Est.) (In billions of riels) Salaries 293.4 325.7 350.8 385.8 475.7 Operating costs 324.3 284.4 344.9 325.2 286.6 Social transfers 41.3 44.8 63.9 68.6 61.5 Economic transfers 4.7 14.5 16.1 6.6 14.3 Capital 335.3 511.1 529.9 451.9 366.6 Other 2/ 10.1 20.2 37.2 29.6 57.6 Total 1009.1 1200.7 1342.8 1267.9 1262.3 (in percent of total expenditure) Salaries 29.1 27.1 26.1 30.4 37.7 Operating costs 32.1 23.7 25.7 25.7 22.7 Social transfers 4.1 3.7 4.8 5.4 4.9 Economic transfers 0.5 1.2 1.2 0.5 1.1 Capital 33.2 42.6 39.5 35.6 29.0 Other 2/ 1.0 1.7 2.8 2.3 4.6 Total 100.0 100.0 100.0 100.0 100.0 (In percent of GDP) Salaries 4.8 4.5 4.3 4.2 4.4 Operating costs 5.3 4.0 4.2 3.6 2.7 Social transfers 0.7 0.6 0.8 0.8 0.6 Economic transfers 0.1 0.2 0.2 0.1 0.1 Capital 5.5 7.1 6.4 5.0 3.4 Other 2/ 0.2 0.3 0.5 0.3 0.5 Total 16.5 16.7 16.3 13.9 11.7 1/ Cash basis. 2/ Includes interest payments and subsidies to provinces. Source: Data provided by the Cambodian authorities. -116 - Table 13: Consolidated Public Expenditure by Sector in 1994 (Million Riels) All Sources Government 1/ ODA NGO 21 Sector/Subsection Amount % Amount % Amount % Amount % Total All Sectors 1,475,308 100.0 740,940 100.0 690,10 100,0 44.261 100.0 General Administration 184,959 12.5 123,477 16.7 58,794 8.5 2,688 6.1 Royal - - 11,023 1 5 - - - National Assembly 6,110 0 8 - Council of Ministers 47,452 64 - M of Foreign Affairs - - 22,212 3 0 - - - M of Interior 18,909 26 M of Eco& Finance - 13,560 18 - M of Justice 1,682 02 - M of Planning - 1,961 03 - Defense & Security 416,280 28.2 391,458 52.8 24,466 3.5 356 0.8 M of Defense 330,212 224 307,478 41 5 22.734 3 3 Public Security 86,068 5 8 83,980 11 3 1,732 03 356 0 8 Social Sector 396,406 26.9 136,679 18.4 236,708 34.3 23,019 52.0 M. of Education 140,056 95 62,230 84 69,111 100 8,7151 9 7 Sector Policies 14,991 I 0 6,991 09 7,654 1 1 346 0 8 Primarv 47,769 3 2 35,090 4 7 9,251 1 3 3428 7 7 Secondary 19,155 1 3 18,039 2 4 1.116 0 2 Tertiary 36,287 2.5 1,471 0 2 31,198 4 5 3,619 8 2 Techn & Manag. Edu. & Training 18.325 1 2 - - 18,325 2 7 - Other 3,528 0 2 639 0 1 1.566 0 2 1.322 3 0 M. of Health 87,792 6 0 30,079 4 I 45.393 6 6 12.321 27 8 Sector Policies & Planning 39.310 2 7 22,585 3 0 15,113 22 1,612 3 6 Primarv Health Care 14,101 1 0 814 0 1 6.805 1 0 6 482 146 Disease Control 7030 0 5 716 0 I 3,520 0 5 2.795 6 1 Family Planning 1,383 0 1 - - 1.231 0 2 153 0 3 Hospital 21.952 1 5 1,949 0.3 18,724 2 7 1,279 2 9 Construction, Equipment & Counterpart 3.984 0 5 - M of Culture 17.530 1 2 2.773 0 4 14,757 2 1 M of Religious Affairs 818 0 1 818 0 1 - M of Information 10.393 0 7 8,590 1 2 1.747 0 3 56 0 13 M of Social Action 134.575 9 I 30.864 4 2 101,911 14 8 1.800 4 07 M of Environment 4,690 0 3 774 0 1 3,789 0 5 127 0 29 M of Women Affairs 551 0 0 551 0 1 - - - Economic Sector 477,662 32.4 89,326 12.1 370,139 53.6 18,198 41.11 M of Agri , Forestry & Fishery 70.551 4 8 13.135 1 8 50.634 7 3 6.782 1532 PolicY & Planning 2,467 0 2 - - 2.263 0 3 203 0461 Research and Development 9,206 0 6 - - 9.206 1 3 0 0_00 Support Services 40.828 2 8 - - 34.542 5 0 6.286 14201 Food Crops 572 0 0 526 0 1 46 0 101 Industrial Crops 1,475 0 1 - - 1.228 0 2 247 0 56 Livestock 1,877 0 1 - - 1.877 0 3 Forestry 193 0 0 - - 193 00 - - Fishery 799 0 1 - - 799 0 1 M of Rural Development 72.862 49 1,129 0 2 61.220 8 9 10 513 23 75 Integrated Rural Development 25,092 I 7 24.466 3 5 626 1 411 i/lage & Communit Development 46,603 3 2 - - 36.716 5 3 9,887 22 34 Inter Country Area 38 1)0 - - 38 0 0 - - Infrastructure 241.228 164 45,905 6 2 195,323 28 3 Public Work 176,767 12 0 36,485 49 140.282 203 Water Supply 48U121 3 3 - - 48,121 7 0 Civil A viation 16.340 1 1 9,420 1 3 6,920 1 0 M oF Industry, Mines and Energy 69.212 4 7 7,524 I 0 60,785 8 8 903 2 04 M Post & Telecommunication 19.848 1 3 18,426 2 5 1.422 0 2 - M of Commerce 2,077 0 1 2,077 0 3 M of Tourism 1.885 0 1 1,130 0 2 755 0 1 it includmfg external 5udget support 2/ NGO's own budget Assistance provided by bilateral and multilateral donors through NGO is included in ODA Source: World Bank staff estimates in collaboration with MfEF and CDC. - 117 - Table 14: Consolidated Public Expenditure by Sector in 1995 (Million Riels) All Sources I1 Government 21 ODA NGO 3/ Sector/Subsector Amount % Amount % Amount % Amount % Total All Sectors 1,821,686 100.0 793,652 100.0 940,285 100.0 87,749 100 General Administration 215,442 11.8 134,915 17.0 80,527 8.6 - - Royal - - 14,098 L8 - - - - National Assembly - 9,800 1.2 - - - - Council of Ministers - - 31,130 3.9 - - - - M. of Foreign Affairs - - 19,899 25 - - - - M. of Interior - 26,076 3 3 - - - M. of Eco.& Finance - - 29,425 37 - M. of Justice - - 2,802 04 - M. of Planning - - 1,685 0.2 - Defense & Security 435,656 23.9 430,71 54.3 4,941 0.5 - M. of Defense 326,779 179 326,779 41 2 - - - - Public Security - 108,877 6 0 103,936 13.1 4.941 05 - Social Sector 406,479 22.3 154,619 19.5 251,860 26.8 - - M. of Education 227,801 12.5 76,021 9.6 151,780 16I - - Sector policies 25,385 1 4 4,561 0,6 20,824 2 2 - - Primary 56,257 3 1 40,180 5.1 16,077 1.7 - Secondary 30,212 1 7 25,704 3.2 4,508 05 - - Tertiary 15,669 0.9 4,078 0.5 11,591 12 - Technical & Manag. Edu & Training 94,903 5.2 - - 94,903 10.1 - - Other 5,376 03 1,498 0.2 3,878 04 - - M. of Health 81,865 4.5 26,322 3.3 55,543 59 - - Sector Policies & Planning 39,413 2.2 22,524 2.8 16,889 18 - - Primary Health Care 8,142 0.4 714 01 7,428 08 - - Disease Control 3,196 02 628 01 2,568 0.3 - - Family Planning 911 0.1 - - 911 0.1 - - Hospital 9,838 05 1,708 0.2 8,130 09 - - Training 2 2,899 0.2 28 0.0 2,871 03 - - Construction, Equipment &Counterpart 10,141 06 197 0.0 9,944 1.1 - - M. of Culture 17,831 1.0 4,167 0.5 13,664 1.5 - - MiofReligiousAffairs 1,118 0.1 872 01 246 0.0 - - M. of Information 9,452 0.5 7,337 0.9 2,115 - 0.2 - - M. of Social Action 62,510 3.4 36,900 4 6 25,610 2.7 - - M. of Environment 04, .6 03 2,208 03 2,548 - M. of Women Affairs 1,147 0.1 792 0.1 355 0.0 - - Economic Sector 675,116 37.1 73,403 9.2 601,713 64.0 - - M. of Agriculture, Forestry & Fishery 93,277 5.1 15,313 1.9 77,964 8.3 - - Policy & Planning 3,191 0.2 - 3,191 0.3 - Research and DevelopmWeo 8,972 0.5 - - 8,972 1.0 - - Support Services 65,085 36 - 65,085 6.9 - - Food Crops 652 0.0 - - 652 0.1 - Forestry 12 0.0 12 0.0 Fishery 52 0.0 - - 52 0.0 - - M. of Rural Development 176,576 9.7 2,254 0.3 174,322 18.5 - Integrated Rural Developmem 35.332 1.9 - - 35,332 3.8 - - Village & Community Development 133,268 7.3 - - 133,268 14.2 - - Inter Country Area 5,099 0.3 - - 5,099 0.5 - - Regional Planning & Developmen 623 0.0 - - 623 0.1 - - Infrastructure 244,358 13.4 0 0.0 244,358 26.0 - Public Work 173,869 9.5 0 0.0 173,869 18.5 - - Water Supply 62,121 3.4 0 0.0 62,121 6.6 - - CivilAviation 8,368 0.5 0 0.0 8,368 0.9 - - M. of Industry, Mines & Energy 95,299 5.2 0 0.0 95,299 10.1 - - M. Post & Telecommunication 8,954 0.5 0 0.0 8,954 1.0 - - M. of Commerce 815 0.0 0 0.0 815 0.1 - - M. of Tourism 1,442 0.1 0 00 - - - Non-Classified -4 0 - - 1,243 0.13 - - 1/ Sectors do not add up to total because NGOs contribution cannot be allocated to sectors. 2/ Including external budget support. 3/ NGO's own budget, Assistance provided by bilateral and multilateral donors through NGO is included in ODA. Source: World Bank staff estimates in collaboration with MEF and CDC. - 118 - Table 15: Consolidated Public Expenditure by Sector in 1996 (Million Riels) All Sources Government 1/ ODA NGO 2/ Sector/Subsector Amount % Amount % Amount % Amount % Total All Sectors 2.047,246 100.0 851,350 100.0 1,084,166 100.0 111,73 100.1 General Administration 278,811 13.6 158,849 18.7 117,243 10.8 2,718 2.4 Royal - - 19,215 2 3 - - - National Assembly I2,037 I 4 - - Council of Ministers 36,242 4 3 - M. of Foreign Affairs 16,293 1 9 - M. of Interior 30,413 36 M of Eco & Finance - - 37,925 45 - M of Justice 5,290 03 2,571 0 3 2,718 '4t M. of Planning 3,656 04 - Defense & Security 407.933 19.9 406,585 47.8 1,349 0.1 - M. of Defense 291,779 14 3 291,779 343 - Public Security 116,155 5 7 114,806 13 5 1.349 0I - Social Sector 548,865 26.8 185,919 21.8 288,858 26.6 74,0891 66.3 M of Education 225,101 11 0 81.954 96 108,760 10 0 34,3881 308 Sector Policies 17,988 09 10.599 1 2 7,389 0 7 Primary 79,039 3 9 41.757 49 37,282 34 - Secondary 32,374 1 6 26.147 3 1 6,227 0 6 Tertiary 22.258 1 1 2.764 0 3 19,494 1 8 - Technical, ManagerialEdu. & Training 30,061 1 5 30.061 28 - Other 8.995 0 4 687 0 1 8,308 08 - M ofHealth 192,074 94 43,178 5 I 124,506 11 5 24.390 21 8 Sector Policies & Planning 30,621 1 5 18,889 2 2 11,732 1 1 - Primary Health Care 44,742 22 1,170 0 1 43,572 4 0 Disease Control 4,625 0 2 1,030 0 1 3,595 0 3 - Family Planning 2.690 0 1 2,690 0 2 - Hospital 5,103 0 2 2.802 0 3 2,301 0 2 Drug Supply 59,198 2 9 18,644 2 2 40,554 37 - Training 3.242 0 2 45 0 0 3,196 0 3 Construction. Equipment. Counterpar 17,459 0 9 592 0 1 16 867 1 6 M of Culture 4.016 0 2 3.717 0 4 299 00 - M of Religious Affairs 914 0 0 914 0 i M of Information 9,954 0 5 9.261 1 1 693 0 1 M of Social Action 102.221 5 0 43.868 5 2 45.458 4 2 12,8941 1 1 M of Environment 11,922 0 6 2.124 0 2 8.533 08 1265 M of Women Affairs 2.663 ( 1 903 0 1 00 1,152 1 Economic Sector 811,637 39.6 99,998 11.7 676,716 62.4 34,923 31.3 M of Agriculture, Forestry & Fishery 196,656 9 6 18.536 2 2 162.998 i5 0 151221 13 5 Pohcy & Planning 7893 0 4 - - 7893 0 7 Research and Developmen 11,039 0 5 - - 11.039 10 - Support Services 136.855 6 7 136 855 12 6 Food Crops 359 00 - - 359 00 - Industrial Cropj 1.008 0 0 - - 1.008 0 1 - Livestock 1.813 0 1 1.813 0 2 Forestri 1,089 01 1.089 01 - - Fisher 2.942 0 1 2.942 0 3 M of Rural Development 253.253 12 4 4,928 0 6 228,524 21 I I ;.801 17 71 Integrated Rural Developmen 27.828 1 4 21828 2 6 [illage & Community Development 191.11 1 9 3 - 19 i 176 - Inter Country Area 4,466 02 -4.466 0 4 - Infrastructure 2 15.558 10 5 27,590 3 2 187,968 173 - Public Work 176.710 86 23,421 2 8 153.289 14 1 - Water Supply 28.565 1 4 28.565 2 6 - Civil Aviation 10.283 0 5 4,169 0 5 6,114 06 - M of Industry, Mines & Energy 48.281 24 12.553 1 5 35.728 33 - M Post & Telecommunication 86,325 4 2 30.557 3 6 55,768 5 1 - M of Commerce 9.535 0 5 3.804 04 5,731 05 - A of ourism 00230 0 2- 1/ Including external budget support 2/ NGO's own budget. Assistance provided by bilateral and multilateral donors through NGO is included in ODA Source: World Bank staff estimates in collaboration with MEF and CDC - 119 - Table 16: Consolidated Public Expenditure by Sector in 1997 (Million Riels) All Sources Budget ODA NGO(') Sector/Subsector Amount % Amount % Amount % Amount % Total All Sectors 2,040,493 100 926,304 100 965,106 100 149,082 100 General Administration 261361 12.81 148,006 15.98 111,053 11.51 2,302 1.54 Royal 16,640 0.82 16,640 180 . National Assembly 10,818 053 10,818 117 . Council of Ministers 36,831 1 81 36,831 3,98 . M. of Foreign affair 20,563 101 20,563 2.22 . . ._. M. of Interior 27,392 1.34 27,392 296 . M of Eco.& Finance 31,214 1.53 31,214 3.37 . M. of Justice 4,808 0.24 2,507 0.27 . . 2,302 1.54 M. of Planning 1.638 0.08 1,638 0 18 . Defense & Security 440,477 21.59 440,477 47.55 . M of Defense 304,744 14.93 304,744 3290 . Public Security 135,733 665 135,733 1465 . Social Sector 666,100 32.64 204,237 22.05 360,835 37.39 101,028 67.77 M. of Education () 299.831 1469 86,261 931 166.475 17.25 47,095 31.59 Sector policies 9,140 0.45 9,140 095 . . Primary 57,440 2.81 57.440 5 95 . . Secondary 837 004 837 009 ._. Tertiary Education 15.292 075 15.292 L58 Techn, Manag Edu & training 52,161 256 . 52,161 5.40 . Other 31.606 1 55 . 31,606 3 27 M. of Health121 181,916 8.92 54,689 5.90 95,729 9.92 31,498 21.13 M. of Culture 4,046 020 4,046 044 M.ofReligionAffair 1.316 0.06 1,316 0.14 M. of Information 4,821 0.24 4.821 0.52 M. of Social action 147,667 7.24 50.525 545 81,163 841 15,979 10.72 M.of Environment 22,950 1 12 1,551 0 17 . 1 81 3,931 2.64 M. of Women affair 3.553 0 17 1.028 0 11 . 2,526 1 69 Other Economic Sector 672,554 32.96 133,584 14.42 493,218 51.11 45,753 30.69 NI. of Agri.Forestry & Fishery 92.669 454 24.105 260 53,838 5.58 14,727 9.88 Policy Planning 6,615 0.32 6,615 069 .. Research and Development 1,533 008 1,533 0 16 . . Support Services 42.746 209 1 42,746 4 43 Food Crops 577 003 577 006 ._. Industrial Crops 389 002 . 389 004 . . Livestock 209 001 209 002 . . Forestry 1.769 009 . 1,769 0 18 . . Fishery 0 000 . NI. of Rural Development 237,719 II 65 3.686 040 203,007 21 03 31,026 20.81 Integrated Rural Development 23.763 1 16 23,763 246 Village & Community Development 178.216 8 73 178,216 1847 Inter Country Area River Basin 1,028 005 1,028 0 11 Infrastructure 150.162 736 38.119 4 12 112,043 11 61 Public Work[3] 132.685 650 24.134 261 108,552 11 25 _ Water Supply 741 0.04 000 741 0 08 Civil Aviation 16.735 082 13,985 1.51 2,750 028 .N. of Industry, Mines & Energy 58.667 2 88 6.850 0.74 51,817 5 37 N1. Post & Telecommunication 103.533 5 07 53,431 5.77 50,102 5 19 N1. of Commerce 28.053 1 37 5,641 061 22,412 2 32 _ NI. of Tourism 1.751 0.09 1,751 0 19 Other I/ Including external budget support. 2/ NGO's own budget. Assistance provided by bilateral and multilateral donors through NGO is included in ODA Source World Bank staffestimates in collaboralion with AIEF and CDC. Table 17: Consolidated Public Expenditure by Sector and Chapter in 1994 (Million Riels) Government 1/ ODA NGO 2/ 10 11-20 30 31 32 50 51 51 Sector/Subsection RoaaI,2 1102 6,6 240 -,3 - 7II - Nain lA s m la,1 ,1 1 ,5 11 -42 a- - U a M.o ltrir1,0 1890 4,5.269 1 -a69 - a. - . C a U w IL Dees-euiy 1,8 9,5 18882 19,34 -3,0 - 446 5 Mao ees 3,1 0,7 5,1 1529 - 2,92 a- - 2273 0 r a ~ Z. C V, t C r uotal All Sectors 1,475,308 740,940 288,447 327,341 4,46 40,599 1,396 78,692 690,107 44,261 Scneral Administration 184,959 123,477 19,119 61,864 6,436 1,381 34,677 58,794 2,688 Royal 11,023 11,023 6,466 2,410 1,436 - 711 - National Assembly 6,110 6,110 617 4,950 118 - 425 Council of Ministers 47,452 47,452 373 1435299 3,364 16. 29,170 M. o Foreign Affirs 22,212 22,212 2.776 17,143 42 888 1,363 M. ofInterior 18,909 18,909 4,954 12,693 -613 -649 M. ofeco & Finance 13,560 13,560 2579 7,837 806 477 1,861 M. of Justice 1,682 1,682 522 862 17 281 4 M. of Polann ing 1,961 1,961 754 1,043 34 130 Defense & Security 416,280 391,458 188,820 198,934 3,704 24,466 356 M. ofDltnse 330,212 307,478 152,314 152,239 - 2,925 - 1922,734 - Pubic Security 86,068 83,980 36,506 46,695 - 751,732 356 Social Sector 396,406 136,679 66,610 35,338 29,583 Is 5,133 236,708 23,019 M. of Education 140,056 62,230 53.213 7,377 ,032 14 594 69,111 8,715 Sector Policies 7,654 - - -----7,654 Primary 9,251 ----- 9,251- Secondary 1,116 ------1,116 - 7erftary 31,198 ---- --31,198 7hI & Alanag Ed/u raining 18.325 ---- -18,325- Oither 1,566 .- --1,566 M. o'llealth 87,792 30.079 8.378 21,466 152 83 45,393 12,321 Sector Policies& Planning 15,113 - --15,113- Primary Healtha Care 6,805 - --- 6,805- Disease Control 3.520 ---- -3,520- Family, Planning 1,231 - -- 1,231 Hospital 18,724 - - -.18,724- M. of Culture 1-7,530 2,773 1,459 1,040 -85 -189 14,757 _ _____ M. of'Religious Affairs 818 818 315 460 15IS 28 - M. offIniorination 10,393 8,590 1,869 2,948 -102 I3,670 1,747 56 M. ol'Social Action 1349575 30.8641 =,0=64 - 1,426 - 28,07L4 30 1191I8 Table 17: Consolidated Public Expenditure by Sector and Chapter in 1994 (Million Riels) Government I/ ODA NGO 2 10 11-20 30 31 32 50 51 51 Sector/Subsection Pc & R.ofEsearhondmeepe nt 4,9,27 830 6 IO39,2067 Support Services 40.828 . 34,542 6,286 Food Crops 572 . 526 46 Industria/ Crops I,475 ....--- 1,228 247 Livestock 1,877 ....--- 1,877 Forestry I193 ....--- 193 - Fishery " 799 - M. of Rural Development 72,862 1,129 310 7331 9 77 61,220 10,513 Integrated Rural Development 25,0921 - 24,466 626 Village & Community Development 46,603 . 36,716 9,887 Inter Coumiry Area 38 . 38 - Infrastructure 241,228 45,905 3,828 5,636 1,407 103 -34,931 195,323 - Public Work 176,767 36,485 3,441 2,578 1,407 102 -28,957 140,282 - WaterSupply 48,121 - 48,121 - Civil Aviation 16.340 9,420 387 3,058 - 1 15,974 6,920 M. of Industry. Mines & Energy 69,212 7,524 1,169 1,041 2,531 47 - 2,736 60,785 903 M. of Post and Telecom. 19,848 18,426 1,255 16,956 - 15 - 200 1,422 M. of Commerce 2,077 2077 952 958 27 32 - 108 -36 M. of Tourism I,885 1,130 293 686 - 14 - 137 7551 - II Including external budget support. NGO's own budget. Assistance provided by bilateral and multilateral donors through NGO is included in ODA. Surc World Bank staf7estimates in collaboration with MEF and CDC, Table 18: Consolidated Public Expenditure by Sector and Chapter in 1995 (Million Riels) Government 21 ODA NGO 31 10 11 to 13 20 30 31 32 50 51 SI Domestically Financed Sector/Subsection t -C - Z 3- zC - 0. z w Total All Sectors 1,821,687 793,652 736,789 56,863 346,198 323,972 1,877 15,204 47,198 2,340 50,482 6,381 940,286 87,749 General Administration 215,442 134,915 109,860 25,055 29,330 65,392 1,877 5,373 6,436 1,452 20,177 4,878 80,527 - Royal Palace 14,098 14,098 12,947 1,152 5,889 2,118 4,940 1,152 -- National Assembly 9,800 9,800 9,430 370 6,410 2,780 140 100 370 Prime Ministers' Office 31,130 31,130 25,745 5,385 2,772 17,719 4,154 688 412 5,385 - - M. of Foreign Affairs 19,899 19,899 19,840 59 6,044 12,755 100 941 59 - - M. of Interior 26,076 26,076 25,355 721 4,327 20,659 368 721 - - M. of Eco & Finance 29,425 29,425 12,503 16,922 2,552 6,721 1,877 1,218 135 12,043 4,878 - M. of Justice 2,802 2,802 2,464 338 545 1,877 42 3381 M. of Planning 1,685 1,685 1,576 109 792 762 22 109 Defense and Security 435,656 430,715 430,715 229,602 196,601 4,513 4,941 - M. ofDefense 326,779 326,779 326,779 191,211 131,759 3,810 - Public Security 108,877 103,936 103,936 38,391 64,842 703 4,94I -. Social Sector 406.482 154,619 148,647 5,972 72,977 39,078 35,847 745 5,581 391 251,863 - M. of Education 227,802 76,021 73,839 2,182 59,133 13.302 1,398 6 1,791 391 151,7811 Sector Policies 25,385 4,561 20,824 - Priaty 56,257 40,180. 16,077 - SecOndur) 30,212 25,704 4,508 - Tertiary 15,669 4,078 11,591 'ech & Mgt. 1du. & lraing 94,903 94,903 - 0ther 5,376 1,498 3,878 M. ofIlcalth 81,866 26.322 26,125 197 8.646 16,705 134 640 197 55,544 - Sector Pelicie, & Plani 39,413 22,524 36,889 Priury Ileali (are 8,142 734 7,428 - /Divave (inrol 3,196 628 2,568 - Fan laung 911 0 911 - lipfai/ 9,838 3,708 8,130 - irinmng 2,899 28 2,871 (onorution,I Equipmeni & (ounterpart 10,141 197 9,944 - M. ofCulture 17,831 4,167 4,096 71 l,995 2,019 83 71 13,664 - M. of Religious A Tairs 1,118 872 842 30 381 445 17 30 246 M. of Information 9,452 7,337 4,506 2,831 1,253 3,195 56 I 2,831 2.115 - M. of Social Action 62,510 36,900 36.801 99 1,105 1,550 34,049 98 99 25,610 - M. offinvironment 4,756 2,208 1,739 469 325 1,384 30 469 2.5481 M ofWomen AITairs 1,147 792 698 94 140 478 80 94 355 EconomicSector 675,115 73,403 47,567 25,836 14,289 22,902 9,831 402 142 24,724 1,112 601,712 Table 18 Consolidated Public Expenditure by Sector and Chapter in 1995 (Million Riels) Government 2/ ODA NGO 3/ 10 li tol3 20 30 31 32 50 51 51 Domestically Financed Sector/Subsection - V t r C M. ofAgriculture, Forestry & Fishery 93,277 15,313 13,118 2,195 5,827 5,826 1,301 164 2,195 77,964 - M. of Rural Development 176,576 2,254 2,224 30 697 1,523 4 30 174.322 - (iYdAsiston 10,552 2,184 2,184 0 401 1,638 2 142 8,368 - ublic Work 203,380 29.511 8,011 21,500 3,600 1,568 2,755 87 21,349 151 173.869 - (abinet ofMintster, Policy & Planitng 1,276 483 793 rutsprL Alwsagement 562 562 , (otlruction Alsnagement 345 345 - Roadv & Bridges Ainagenten 3,709 3,709 - raining 157 157 - Rail WaY 10,312 1,895 8,418 - Other 860 860 - (onviruction. I*quipiew & (Counterpart 186,159 21,500 164,659 - Water Supply 62,121 62,121 - M. of Post and Telecommunication 19.997 11,043 10,668 375 1,362 9,288 19 375 8,954 - M. of Industry, Mines & Energy 101,080 5,781 4.692 1,089 1,218 684 2.746 44 128 961 95,299 - M. ofCommerce 6,691 5,876 5,574 302 875 1,603 3,030 65 302 815 - M. ofTourism 1,442 1,442 1,097 346 309 771 17 346 - pon-Liassied 1,243 1 1,2431 I/ Sectors do not add up to total because NGO's cortribution cannot be allocated to sectors. 2/ Including external budget support. 3/ NGO's own budget. Assistance provided by bilateral and multilateral donors through NGO is included in ODA. Source: World iank staffesiimales in collaboration with AlEf and ( IX Table 19: Consolidated Public Expenditure by Sector and Chapter in 1996 (Million Riels) Government I/ ODA NGO 2 10 11ks13 20 30 31 32 50 51 Domestically Financed Sector/Subsection - --t E~ E 0 00tJu Total All Sectors 2,047,245 851,349 789,771 61,578 343,979 351,643 12,844 15,035 60,204 6,067 51,890 9,689 1,084,166 111,730 General Administration 278,811 158,849 134,261 24,588 32,211 68,301 12,844 4,841 10,417 5,647 21,566 3,022 117,243 2,718 Royal Palace 19,215 19,215 15,829 3,387 7,194 2,343 6.292 3,387 - National Assembly 12,037 12,037 11,804 233 8,789 2,865 ISO 233 Prime Ministers' 011ice 36,242 36,242 27,499 8,743 2,525 20,826 2.036 1,641 471 8,743 - M. of Foreign Affairs 16,293 16,293 16,160 133 4,142 10,449 118 1,452 133 M. of Interior 30,413 30,413 27,836 2,577 5,036 21,528 950 321 2,077 500 M. of Eco. & Finance 37,925 37,925 30,213 7,712 2.599 7,343 12,844 1,855 1,848 3,724 6,512 1,200 - M. of Justice 5,290 2,571 2,489 82 908 1,558 23 82 1 2,718 M. of Planning 3,656 3,656 1,935 1,721 869 1,045 20 400 1,322 - Defense and Security 407,933 406,585 406,585 222,326 177,575 6,684 1,349 M. of Defense 291,779 291,779 291,779 180,805 104.973 6,001 - - - - Public Security 116,155 114,806 114,806 41,521 72,602 683 - - 1,349 - Social Sector 548,864 185,918 179,668 6,250 73,707 62,859 42,771 331 5,608 642 288,858 74,089 M. of Education 225,101 81,953 80,552 1,401 58,753 20,462 1,271 66 961 441 108,760 34,388 Sector Policv 17,988 10,599 7,389 - Primary 79,039 41,757 37,282 - Secondary 32,374 26,147 6,227 - Tertiary 22,258 2,764 19,494 - Tech, Aig. Amd. l Training 30,061 30,061 - Other 8,995 687 8,308 - M. of Ilealth 192,074 43,177 42,605 573 9,140 33,151 205 109 387 186 124,506 24.390 Sector Palicies & Ilamung 30,621 18,889 I 1.732 Primary Heaih (are 44,742 1,170 43,572 - biseave (onrol 4,625 1,030 3,595 Famiy Planning 2,690 2,690 . Hospitals 5,103 2,802 2,301 - DrugSupplv 59,198 18,644 40,554 - Training 3,241 45 3.196 (onviruction, Iqupnent & (Counerpari 17,459 592 M. ofCulture 4,016 3,717 3,702 16 2,118 1,504 79 16 - M. of Religious Affairs 914 914 893 22 493 387 13 22 - M. of Information 9,954 9,261 5,391 3,870 1,338 3,977 76 3,870 693 M. of Social Action 103,129 43,868 43,819 49 1,131 1.509 41,022 156 49 46.366 12,894 M.ofEnvironment 11,922 2.124 1.858 266 469 1,363 26 266 8.533 1,265 M.ofWomen Affairs 2,663 903 849 53 265 507 78 53 - 1,152 Table 19:.Consolidated Public Expenditure by Sector and Chapter in 1996 (Million Riels) Government I/ ODA NGO2/ 10 11 to 13 20 30 31 32 50 51 Domestically Financed Sector/Subsection V t a, E. 0 U ~0 0. ~~t,. Economic Sector 811,637 99,998 69,257 30,741 15,734 42,908 10,194 332 89 24,716 6,025 676,717 34,923 M. of Agriculture, Forestry & Fishery 196,656 18,536 16,099 2,437 5,939 9,087 933 140 2,426 11 162,998 15,122 M. of Rural Development 253,253 4,928 2,242 2,686 923 1,316 3 2,686 228,524 19,801 Civil Aviation 10,283 4,169 2,380 1,789 1,187 1,118 75 1,779 10 6,114 - Public Work 176,710 23,421 7,003 16,418 3,532 1,453 1,951 68 14,650 1,767 153,289 - Cahinelt ofMiniser, Polky & Planning 1,913 538 1,375 - rawar Managentei 578 578 (nnvrucion Alagennt 438 438 Roads & 11ridges Management 3,375 3,375 -raining 124 124 - - Rail Way 11,615 1,610 10,005 - . Other 341 341 - - (anstruction, E'quipment & (ounlerpart 158,326 16,418 141.909 - Water Supply and Sanitation 28,565 28,565 - M. of Post & Telecommunication 86J25 30,557 27,385 3.173 1.476 25,899 10 2,000 1,172 55,768 - M. of Industry, Mines & Energy 48,281 12,553 8,625 3,928 1,333 913 6,310 55 14 917 3.011 35,728 - M. ofCommerce 11,564 5,833 5,523 311 1,343 3.123 1,000 56 257 53 5,731 I/Including external budget support. 2/ NGO's own budget. Assistance provided by bilateral and multilateral donors through NGO is included in ODA. Source: World fank staffestimales in) collahoration witi Aff and CIX Table 20- Consolidated Public Expenditure by Sector and Chapter in 1997 (Million Riels) Budget Expenditure I/ ODA NGU Z 10 11to13 20 30 31 32 40 50 51 St Domestically Financed Sectors/Subsecturs no .o C 0) 0 It X t~ X U1 U1 GrandTotalASectors 2,040,493 926,304 816,0101 110,295 385,8101 342,1361 9,5481 6,6321 68,647 1,150 2,087 92,9841 17,3111 965,1061 322,86 149,082 GeneralAdministration 1 261,3611 148,0061 145,6631 2,344 35,051 84,069 9,548 1,937 11,821 1,150 2,087 2231 113 111,0531 37,154 2,302 Roal Palace16,640 16,640 16,346 294 5,119 3,969 7,259 2941 2 NationalAssembly 10,818 10.818 10,818 7,803 2,865 150 Prime Ministers' Office 36,831 36,831 36,831 . 3,392 28,385 1,247 3,281 527 Relations with Parliament 159 159 159 . 70 86 2 Civil Service 245 245 245 . 80 165 0 Foreign Affairs 20,563 20.563 20,563 . 8,560 11,303 77 623 Interior GeneralAdministratio 27,392 27.392 27,226 166 5.695 21,227 304 166 Economy & Finance 31,214 31,214 29,331 1,883 2,657 13,645 9,548 690 704 2,087 1,770 113 Justice 4,808 2,507 2,507 0 770 1,712 25 2,302 Planning 1,638 1,638 1,638 0 906 712 20 Defense and Security 440,477 440.477 419.336 21,141 252,258 159,883 0 0 7,195 0 20,145 996 0 0 National Defense 304,744 304,744 304,744 0 202,567 95.556 6,620 Security 135,733 135,733 114,592 21,141 49,691 64,327 575 20,145 996 Social Sector 666,100 204,237 189,700 14,S37 81,433 58,975 0 16 49,276 0 11,225 3,312 360,835 120,721 101,028 Education 299,831 86,261 83,418 2,843 64,766 16,198 16 2,438 1,212 1,631 166,475 55,696 47,095 Sector policy 9,140 * * 9,140 3,058 Primary 57,440 . * 57,440 19,217 Secondary 837 . . 837 280 Higher 15,292 . * 15,292 5,116 Technical and Training 52,161 - * 52,161 17,451 Other 31,606 - . 31,606 10,574 Health 181,916 54.689 45,299 9,390 10,531 34,555 214 7,792 1,598 95,729 32,027 31,498 Sector policy planning 0 Primary health care 0 Disease control * * * 0 Family planning . . . 0 Hospitals . . * 0 Drug Supply . * . 0 Training , 0 Construction, Equipment & Counterpart Fund *0 Social Action 147,667 50.525 48,755 1,770 1,139 1,319 46,298 1,770 81,163 27,154 15,979 Environment 22,950 1,551 1,489 62 539 935 14 62 17,468 5,844 3,931 Infornation 4,821 4,821 4,650 171 1,342 3,287 22 88 83 0 Culture and FineArts 4,046 4.046 4,040 6 2.288 1,637 115 6 0 CultsandIReligion 16 1.316 1,105 211 545 530 29 211 0 Table 20* Consolidated Public Expenditure by Sector and Chapter in 1997 (Million Riels) Budget Expenditure I/ ODA NGO 2l 10 li to l3 20 30 31 32 40 50 51 51 Domestically Financed Sectors/Subsectors L 2 z a C 0.. ~~" E.! i6 2 ~, ba 0q -. .2 E~ E R - - - - - K Wonens'aAffairs 3,553 1,028 944 84 283 51S 145 84 0___ 2,526__ Economic Sector 672,554 133,584 61,310 72,273 17,067 39,209 0 4,679 355 0 0 59,383 12,890 493,218 165,0111 45753 Agriculture 92,669 24,105 14,803 9,302 6,562 7,040 ____ 1,059 142 8.644 658 53,838 18,012 14,727 Rural Developmnent 237,719 3.686 2,037 1,649 951 1.069 17 1,320 329 203,007 67.9 18 3 1.026 Air Infrastructure 16,735 13.985 2,781 I1I,204 1,324 1,452 ____5 II,204 ___ 2,750 920 Transport 132,685 24,134 6,944 17,190 3.858 1.295 ____ 1,702 88 12.729 4.461 108,552 36,317 _____ Cabinet of Minister, Policy &0 Planning ______ . ___._ O_________________________________ _____ Transpo Management 0 Construction Management 0 Roads & Bridges Management . 0 E U I Training( School of' Public- Work & Construction) 0 Rail Way . . . O____ Other . . . 0 Construction. Equipment &0 Local Counter Part(Ch.50) 0____ ____ _____ _____ ___ ____ ____ _____ EYater Supply.Sanitation 741 .. . 0 0 0 0 741 248 0 Con nication 103,533 53,431 26,868 26,563 1,613 25.238 1 18 22,597 3,966 50,102 16,762 14,7T Industry,Energy,Miinae 58,667 6,850 3,132 3,718 1,320 859 918 35 316 3,402 51,817 17,336 _____ Industry & Mining . O Energy 51,817 ____ ________ ______________ ________ 51,817 17,336 _____ Trade 28,053 5,641 3,031 2,610 981 1,016 1,000 34 2,536 74 22,412 7,498 Tourism 1,751 1.751 1,714 37 458 1,240 16 37 0 O th e r 0 M i ni ste r, P oli cy & 0 IIIncluding external budget suppor. 2/ NGOs own budget. Assistance provided by bilateral and multilateral donors through NGO is included in ODA. Soure: .World Bank staff eonstiate in collahonnin wide AEFI and (TX Rail Way1 OtherI C ons ruc ion,E q u p m en &I Table 21: Consolidated Public Expenditure by Region in 1994 (Million Riels) All Sources Government II ODA/NGO Code# Province/City Amount % Exp. per Capita Amount % Exp. per Capita Amount % Exp. per Capita Population 2/ 000 Total 1,475,308 100.0 149,686 740,940 100.0 75,177 734.368 100.0 74,510 9,856,000 001 Country wide 917,069 62.2 93,047 599,302 80.9 60,806 317,766 43.3 32,241 9,856,0 002 All Provinces 558,239 37.8 56,639 141,638 19.1 14,371 416,601 56.7 42,269 9,856,00 003 Phnom Penh 292,265 19.8 354,261 15,540 2.1 18,836 276,725 37.7 335,425 825,000 004 Kandal 19,761 1.3 23,112 11,142 1.5 13,031 8,619 1.2 10,080 855,000 005 Kompong Cham 18,778 1.3 12,827 14,599 2.0 9,972 4,180 0.6 2,855 1,464,000 006 Battambang 33,543 2.3 48,967 10,416 1.4 15,206 23,127 3.1 33,761 685,000 007 Prey Veng 13,424 0.9 14,388 10,918 1.5 11,702 2,506 0.3 2,686 933,000 008 Siem Reap 29,645 2.0 43,596 7,569 1.0 11,131 22,076 3.0 32,465 680,000 009 Kompong Thom 7,657 0.5 15,627 7,314 1.0 14,926 343 0.0 701 490,000 - 010 Takeo 18,566 1.3 26,714 9,839 1.3 14,157 8,727 1.2 12,557 695,000 on Oil Svay Rieng 10,285 0.7 23,217 6,497 0.9 14,666 3,788 0,5 8,551 443,000 012 Pursat 11,216 0.8 34,510 5,645 0.8 17,370 5,571 0.8 17,140 325,000 013 Kompong Chhang 9,440 0.6 28,957 5,866 0.8 17,995 3,574 0.5 10,962 326,000 014 Kompong Speu 27,765 1.9 56,663 7,810 1.1 15,938 19,955 2.7 40,725 490,000 015 Kom Pot/Kep 17,732 1.2 36,865 7,450 1.0 15,488 10,282 1.4 21,376 481,000 016 Sihanouk Vill 6,850 0.5 56,610 2,607 0.4 21,542 4,243 0.6 35,068 121,000 017 Koh Kong 2,552 0.2 35,949 2,552 0.3 35,949 0 0.0 0 71,000 018 Preah Vihcar 2,279 0.2 23,252 2,279 0.3 23,252 0 0.0 0 98,000 019 Kratie 4,848 0.3 22,870 3,134 0.4 14,783 1,714 0.2 8,087 212,000 020 Ratanakiri 2,179 0.1 34,046 2,082 0.3 32,536 97 0.0 1,510 64,000 021 Modulkiri 1,149 0.1 52,230 1,125 0.2 51,141 24 0.0 1,089 22,000 022 Bancay Meanchey 22,940 1.6 44,115 5,398 0.7 10,381 17,542 2.4 33,734 520,000 023 Steng Treng 5,365 0.4 95,796 1,856 0.3 33,151 3,508 0.5 62,645 56,000 I/ Including external budget support. 2/ Population data provided by Ministry of Planning. Source: World Bank staffestimates in collaboration with AEF and CDC, Table 22: Consolidated Public Expenditure by Region in 1995 (Million Riels) All Sources Government 1/ ODA 2/ Exp. per Exp. per Exp. per Population 3/ Code# Province/City Amount % Capita Amount % Capita Amount % Capita 000 Total 1,821,686 100.0 170,021 793,652 1.00 77,824 940,285 1.00 92,203 10,198,000 001 Country wide 1,013,863 55.7 99,418 626,725 0.79 61,456 387,138 0.41 37,962 10,198,000 002 Phnom Penh 381,161 20.9 446,324 17,521 0.02 20,516 363,640 0.39 425,808 854,000 003 Kandal 26,267 1.4 29,680 13,557 0.02 15,319 12,710 0.01 14,362 885,000 004 Kompong Chan 27,649 1.5 18,250 16,573 0.02 10,940 11,075 0.01 7,310 1,515,000 005 Battambang 42,061 2.3 59,324 11,051 0.01 15,586 31,010 0.03 43,738 709,000 006 Prey Veng 18,232 1.0 18,873 I I,688 0.01 12,100 6,543 0.01 6,774 966,000 007 Siem Reap 35,818 2.0 50,878 11,895 0.01 16,897 23,923 0.03 33,982 704,000 008 Kompong Thom 11,988 0.7 23,645 7,859 0.01 15,501 4,129 0.00 8,144 507,000 009 Takeo 22,879 1.3 31,820 10,226 0.01 14,222 12,653 0.01 17,598 719,000 010 Svay Rieng 14,637 0.8 31,958 7,199 0.01 15,719 7,438 0.01 16,239 458,000 033 Pursat 12,993 0.7 38,669 6,391 0.01 19,021 6,602 0.01 19,647 336,000 012 Kompong Chhang 14.941 0.8 44,336 6,448 0.01 19,133 8,493 0.01 25,203 337,000 013 Kompong Speu 38,505 2.1 75,947 9,001 0.01 17,754 29,504 0.03 58,193 507,000 014 Kom Pot/Kep 17.254 0.9 34,646 8,993 0.01 18,058 8,261 0.01 16,589 498,000 015 Sihanouk Ville 8,903 0.5 71,222 3.115 0.00 24,923 5,787 0.01 46,300 125,000 016 Koh Kong 3,224 0.2 44.168 3,101 0.00 42,473 124 0.00 1,695 73,000 017 Preah Vihear 3.377 0.2 33,436 3,377 0.00 33,436 0.00 101,000 018 Kratie 4,374 0.2 19,974 4,129 0.01 18,855 245 0.00 1,119 219,000 019 Ratanakiri 3,742 0.2 56,690 2,923 0.00 44,289 818 0.00 12,401 66,000 020 Modulkiri 1,879 0.1 81,692 1,879 0.00 81,692 0.00 0 23,000 021 Banteay Meanchey 23,009 1.3 42,767 7,205 0.01 13,391 15,804 0.02 29,376 538,000 022 Steng Treng 5,696 0.3 98,210 2,795 0.00 48,196 2,901 0.00 50,014 58,000 023 Unclassitied 1,486 0.1 146 0.00 0 1,486 0.00 146 10,198,0 024 Total 1,821,686 170,027 793,652 77,824 940,285 92,203 10.198,0 I/ Including external budget support. 2/ NGO's contribution is not included (CR 87.749 billions) because it could not be broken down by provinces. 3/ Population for 1995 was provided by Ministry of Economy and Finance. Srce: World Bank staff estimates in collaboration with MEF and CDC. Table 23: Consolidated Public Expenditure by Region 1996 (Million Riels) _ All Sources Government I/ ODA/NGO Exp. per Exp. per Exp. per Population 2/ Code# Province/City Amount % Capita Amount % Capita Amount % Capita 000 Total 2,047,245 100 191,331 851,350 100 79,565 1,195,896 100 111,766 10,700,00H 001 Country wide 1,274,254 62.24 119,089 670,474 78.75 62,661 603,780 50.49 56,428 10,700,000 002 All Provinces 772,991 37.76 72,242 180,875 21.25 16,904 592,116 49.51 55,338 10,700,000 003 Phnom Penh 325,665 15.91 404,051 22,188 2.61 27,529 303,477 25.38 376,522 806,000 004 Kandal 35.889 1.75 35.289 16,026 1.88 15,758 19,862 1.66 19,530 1,017,000 005 Kompong Charn 39,638 1.94 24,774 17,551 2.06 10,969 22,087 1.85 13,805 1,600,000 006 Battambang 34,382 1.68 46,400 13,815 1.62 18,644 20,567 1.72 27,756 741,000 007 Prey Veng 35,602 1.74 37,634 11,941 1.40 12,623 23,661 1,98 25,012 946,000 008 Siem Reap 34,130 1.67 57,073 9,649 1.13 16,136 24,481 2.05 40,938 598,000 009 Kompong Thorn 15,971 0.78 27,922 8,257 0.97 14,436 7,714 0.65 13,485 572,000 010 Takeo 38,461 1.88 49,755 12,050 1.42 15,589 26,410 2.21 34,166 773,000 0 01i Svay Rieng 28,625 1.40 60,518 7,746 0.91 16,376 20,879 1.75 44,142 473,000 012 Pursat 15,839 0.77 47,000 6,366 0.75 18,890 9,473 0.79 28,110 337,000 013 Kornpong Chhang 25,771 1.26 70,996 6,812 0.80 18,766 18.959 1.59 52,230 363,000 014 Kornpong Speu 58,237 2.84 96,101 9,171 1.08 15,133 49,066 4.10 80,968 606,000 015 Kom Pot/Kep 25,948 1.27 44,204 9,607 1.13 16,367 16,340 1.37 27,837 587,000 016 Sihanouk Vill 15,346 0.75 119,894 3,004 0.35 23,466 12,343 1.03 96,427 128,000 017 Koh Kong 2,845 0.14 27,896 2,225 0.26 21,816 620 0.05 6,080 102,000 018 Preah Vihear 3,539 0.17 33,071 3,500 0.41 32,707 39 0.00 365 107,000 019 Kratie 6,341 0.31 27,097 4,621 0.54 19,750 1,719 0.14 7,347 234,000 020 Ratanakiri 3,281 0.16 40,509 2,704 0.32 33,382 577 0.05 7,127 81,000 021 Modulkiri 1,921 0.09 71,153 1,882 0.22 69,707 39 0.00 1,446 27,000 022 Bantcay Meanchey 22,109 1.08 42,113 9,286 1.09 17,687 12,823 1.07 24,425 525,000 023 Steng Treng 3,450 0.17 44,802 2,472 0.29 32,107 977 0.08 12,694 77,000 I/ Including external budget support. 2/ Ministry of Planning. Source: World Bank staffestinates in collaboration with MIEFand CDC. Table 24: Consolidated Public Expenditure by Region 1997 (Million Riels) All Sources Budget ODA NGO Province/City All Sources % Exp. per Budget % Exp. per ODA % Exp. per NGO % Exp. per Population(') Capita([)(n Capita Capita (in Capita (in Ricl) (in Riels) Riels) Riels) 000 Total 2,040,493 100 185,516 926,304 100 84,217 965,106 100 87,745 149,079 15 13,554 10,999,000 001 Country wide 777,052 38.08 70,648 742,697 80.18 67,524 34,356 3.56 3,124 10,999,000 002 All Provinces 298,331 14.62 27,123 183,608 19.82 16,693 114,724 11.89 10,430 10,999,000 003 Phnom Penh 42,699 2.09 51,569 19,612 2.12 23,686 23,087 2.39 27,883 828,000 004 Kandal 27,835 1.36 26,611 15,744 1.70 15,052 12,091 1.25 11,559 1,046,000 005 Kompong Cham 24,166 1.18 14,690 19,536 2.11 11,876 4,630 0.48 2,815 1,645,000 006 Battambang 30,903 1.51 40,608 15,844 1.71 20,820 15,059 1.56 19,788 761,000 007 Prey Veng 15,904 0.78 16,345 11,833 1.28 12,161 4,071 0.42 4,184 973,000 008 Siem Reap 16,995 0.83 27,679 10,535 1.14 17,159 . . . 6,459 0.67 10,520 614,000 009 Kompong Thom 15,766 0.77 26,813 8,670 0.94 14,745 . . . 7,096 0.74 12,068 588,000 010 Takeo 15,590 0.76 19,634 12,152 1.31 15,305 . . 3,437 0.36 4,329 794,000 Ol Svay Rieng 10,898 0.53 22,378 7,341 0.79 15,074 3,557 0.37 7,304 487,000 012 Pursat 14,117 0.69 40,801 6,594 0.71 19,058 7,523 0.78 21,744 346,000 013 Kompong Chhang 10,721 0.53 28,744 6,453 0.70 17,301 . . . 4,268 0.44 11,443 373,000 014 Kompong Speu 15,559 0.76 24,974 10,074 1.09 16,171 5,485 0.57 8,804 623,000 015 Kom Pot/Kep 12,110 0.59 20,083 9,752 1.05 16,172 2,358 0.24 3,911 603,000 016 Sihanouk Vill 4,908 0.24 37,178 3,320 0.36 25,154 . . . 1,587 0.16 12,024 132,000 017 Koh Kong 2,749 0.13 26,177 2,288 0.25 21,793 460 0.05 4,384 105,000 018 Preah Vihear 4,558 0.22 41,433 3,604 0.39 32,765 . . . 953 0.10 8,668 110,000 019 Kratie 6,141 0.30 25,481 4,291 0.46 17,804 1,850 0.19 7,677 241,000 020 Ratanakiri 4,176 0.20 50,312 2,520 0.27 30,362 1,656 0.17 19,951 83,000 021 Modulkiri 2,039 0.10 72,823 2,039 0.22 72,823 0 0.00 0 28,000 022 Banteay Meanchey 17,113 0.84 31,690 9,001 0.97 16,668 1 8,112 0.84 15,022 540,000 023 Steng Treng 3,386 0.17 42,862 2,403 0.26 30,414 1 1 983 0.10 12,448 79,000 I/ Including external budget support. 2/ Ministry of Planning. Source: World lank suff estisnates in collaboration with AIEF and C/X - 132 - Table 25: Number of Civil Employees and Their Salary Levels in Cambodia, 1997 Number of Percentage of Employees in Average salary per Employee Permanent Pemoees Different Categories (Mn. Riels) Employees A B C Total A B C Total 1. Administrative Services 26264 25.0 64.6 10.3 100.0 0.9 0.7 0.7 0.8 2. Social Services 99663 2.8 93.6 3.4 100.0 1.0 0.9 1.2 1.0 2.1 Education 75573 0.6 97.7 1.6 100.0 1.0 1.0 1.0 1.0 2.2 Health 16461 4.4 87.7 7.7 100.0 1.0 0.9 1.9 1.0 2.1 Other Social Services 7629 21.6 66.4 11.9 100.0 0.9 1.0 1.0 1.0 3. Economic services 26239 9.5 64.4 26.0 100.0 0.9 0.9 0.9 0.9 3.1 Transport 5863 8.1 38.7 53.1 100.0 0.9 0.9 1.0 0.9 3.2 Communication 2021 5.6 76.7 17.6 100.0 1.0 0.9 1.0 0.9 3.3 Agriculture 11124 7.5 72.8 19.6 100.0 1.0 1.0 0.9 0.9 3.4 Rural Development 1906 15.6 79.2 5.0 100.0 0.9 1.0 1.0 1.0 3.5 Industry, Mining and Energy 2506 11.8 56.1 32.0 100.0 1.0 1.0 1.0 1.0 3.6 Other Economic Services 2819 17.9 72.5 9.5 100.0 0.9 0.8 0.9 0.9 Total 152166 7.2 84.3 8.4 100.0 0.9 0.9 1.0 0.9 Note: Number of employees and average salaries are given in respect of only the civil departments. The details of in security and defense sectors are not included. Source: Ministry of economy and Finance, Royal Government of Cambodia - 133 - Table 26: Pay and Allowance of Military Personnel in Cambodia, 1997 Rank Number of Monthly Monthly Child Food Total Monthly Personnel Wage Bonus Allowance * Allowance Salary A. General 575 3 Star 64500 12900 7500 27375 112275 2 Star 54500 10900 7500 27375 100275 1 Star 49500 9900 7500 27375 94275 Colonel 9321 Colonel 44500 8900 7500 27375 88275 Lt. Colonel 38500 7700 7500 27375 81075 Major 35500 7100 7500 27375 77475 B. Lieutenant 44160 Captain 32400 6480 7500 27375 73755 Lieutenant 28400 5680 7500 27375 68955 2nd Lt. 26400 5280 7500 27375 66555 Serg. Maj. 25400 5080 7500 27375 65355 Soldiers 84450 Sergeant 24400 4880 7500 27375 64155 2nd Serg. 23400 4680 7500 27375 62955 Corporal 22400 4480 7500 27375 61755 2nd Corporal 21400 4280 7500 27375 60555 Private 20400 4080 7500 27375 59355 Note: The table makes an assumption that the average soldier in each rank has three children. Child allowance is R 2500 per child per month. Food allowance is R 900 per day (monthly allowance amounts to 900*365/12 =27375). In addition to these, every soldier gets 22 kgs of rice allowance per month and 2 outfits of clothing per year. Source: Data provided by the Government. MAP SECTION IBRD 28781 102 104° 1 106° LAO PEOPLE S THAILAND EMOCRATIC UBLIC ) PREAN VIHEAR M CHEY SIEM REAP nom Thbeng ST ENG Meanchey Sisophon StunLmpa Siem Reap Tren Ba ang rom, KOMONGTHOM OXKompong KAI BATTAMBANG KpThom MONDULKRI S. Porong Senmonorom Purs Kratie O PURSAT ,Ku-.arx..g ýlur.pc KOMPON r _yjHNANG 12' CHHNANG -1.2' - OMPcONG cH~A KOMPONG SPEU K%CI'G - * PREYVENG--<r- Kong KKNG --g .g Tak <ju'- l VIETNAM Gulf of D Tha ila nd IENG r KAPOT Sihanoukville K .. CAMBODIA N CHINA MYANMAR ROADS PM.RE ----RAILROADS TALN RIVERSTHAIND RIVER Andeon 'VIET -- PROVINCE BOUNDARIES - - - INTERNATIONAL BOUNDARJES Sous MAL.A Sea 0 25 5 100 My Map ILOMMTS SINGAP lamamas= dd. a=VsIs The Wo ~oi G-Ms4. mny ~ @n. nym ~ ofMW w o N 0 N E S A 5/5/1997ny wdmcia C=2=ocx o c. 35 OR 5/5/1997fl1~ii ER3 oö 3T
Groupe de la Banque mondiale · Public Expenditure Review
Cambodia - Public expenditure review enhancing the effectiveness of public expenditures (Vol. 2 of 2) : Main report
Voir le document original
Le texte intégral est hébergé par l’organisation qui le publie. lawenc.com indexe les métadonnées et renvoie vers la source officielle.
Texte intégral
Informations clés
Organisation
Groupe de la Banque mondiale
Type de document
Public Expenditure Review
Pays
Cambodge
Source
Banque mondiale