Report No. 43280-TJ Tajikistan Second Programmatic Public Expenditure Review (In Four Volumes) Volume II: Technical Background Papers June 12, 2008 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank NBT National Bank of Tajikistan NBT National Bank of Tajikistan NDR Nurse-to-Doctor Ratio NPV Net Present Value NSED National Strategy for Education Development OSI Open Society Institute OTL Overhead Transmission Line PARS Public Administration Reform Strategy PBB Policy-Based Budgeting PCF Per Capita Financing PDPG Programmatic Development Policy Grant PEFA Public Expenditure and Financial Accountability PEFA Public Expenditure and Financial Accountability PEIR Public Expenditure and Institutional Review PEMPAL Public Expenditure Management Peer Assisted Learning Network PEP Private Enterprise Partnership PETS Public Expenditure Tracking Surveys PFM Public Financial Management PIP Public Investment Program PPAs Power Purchase Agreements PPER Programmatic Public Expenditure Review PPF Production Possibility Frontier PPG Public and Publicly Guaranteed PPIAF Public-Private Infrastructure Advisory Facility PRGF Poverty Reduction and Growth Facility PRSP Poverty Reduction Strategy Paper QFD Quasi-Fiscal Deficit RED Rayon Education Department REEPS Regional Energy Export Potential Study RRS Rayons under Republican Subordination SAC Structural Adjustment Credit SCO/SDC Swiss Cooperation Office SME Small and Medium Enterprises SUB Rural Hospital SVA Rural Physician Ambulatory Centers SWAPs Sector Wide Approaches TA Technical Assistance TALCO Tajik Aluminum Company TPPs Thermal Power Projects USAID U.S. Agency for International Development WBI World Bank Institute Vice President: Shigeo Katsu Country Director: Annette Dixon Sector Director: Luca Barbone Sector Managers: Kazi Matin Task Team Leader: R. Sudharshan Canagarajah iii CONTENTS PREFACE .........................................................................................................................................IX 1. MACROECONOMIC AND FISCAL DEVELOPMENTS.................................................................... 1 A. INTRODUCTION .............................................................................................................1 B. GROWTH OF OUTPUT AND SPENDING............................................................................2 C. INFLATION AND FOOD PRICES.......................................................................................4 D. COMPONENTS OF DEMAND ...........................................................................................5 E. FISCAL POLICY .............................................................................................................6 F. EXTERNAL DEBT...........................................................................................................9 G. MONETARY POLICY ......................................................................................................9 H. PROSPECTS AND RISKS FACING THE TAJIK ECONOMY ................................................10 I. AGGREGATE DEMAND AND INFLATION.......................................................................10 J. THE BALANCE OF PAYMENTS .....................................................................................11 K. EXTERNAL DEBT DISTRESS.........................................................................................11 L. POLICY REFORMS REQUIRED TO ADDRESS THE MACROECONOMIC RISKS AND CHALLENGES .............................................................................................................11 M. MACROECONOMIC MANAGEMENT..............................................................................12 N. EXTERNAL DEBT STRATEGY.......................................................................................13 O. STRUCTURAL REFORMS ..............................................................................................13 2. EFFICIENCY OF THE EDUCATION AND HEALTH SYSTEMS IN TAJIKISTAN .............................. 15 A. BACKGROUND.............................................................................................................15 KEY EMERGING ISSUES...............................................................................................16 B. RECENT REFORMS IN SOCIAL SECTORS AND KEY CHALLENGES AHEAD ....................16 KEY BUDGET MANAGEMENT CHALLENGES................................................................19 C. RECENT TRENDS IN PUBLIC SPENDING ON EDUCATION AND HEALTH.........................21 EXISTING INTRA-SECTORAL FINANCING STRUCTURE..................................................21 TRENDS IN EDUCATION AND HEALTH SPENDING ........................................................23 D. PERFORMANCE OF THE EDUCATION AND HEALTH SYSTEM.........................................29 E. DEMAND SIDE.............................................................................................................29 F. SUPPLY SIDE...............................................................................................................33 G. CONCLUSIONS.............................................................................................................35 3. HEALTH SECTOR: PUBLIC EXPENDITURE TRACKING SURVEY RESULTS ............................... 65 A. BACKGROUND.............................................................................................................65 B. ORGANIZATIONAL STRUCTURE OF THE HEALTH SYSTEM ...........................................67 Republican Tier.........................................................................................................68 Oblast Tier ................................................................................................................69 Rayon Tier ................................................................................................................69 Jamoat Tier................................................................................................................69 C. ANALYSIS OF THE HEALTH SECTOR BUDGET MANAGEMENT .....................................70 Resource Allocation in the Health Sector.................................................................70 Management of Budgetary Resources ......................................................................72 Discretion of Resource Managers.............................................................................73 v Management of Human Resources...........................................................................75 Management of Nonhuman Resources Inputs..........................................................77 D. WERE LEAKAGES IDENTIFIED? ...................................................................................79 E. THE WAY FORWARD...................................................................................................80 Specific Recommendations.......................................................................................81 Management of Human Resource Inputs..................................................................82 Management of Other Inputs and Service Outputs...................................................82 F. CONCLUSIONS.............................................................................................................83 4. EDUCATION SECTOR SPENDING: FINDINGS FROM A PUBLIC EXPENDITURE TRACKING SURVEY ................................................................................................................................ 85 A. MOTIVATION FOR AN EDUCATION PUBLIC EXPENDITURE TRACKING SURVEY ...........85 Public Expenditure Trends........................................................................................85 Student Enrollment Trends .......................................................................................88 B. ALLOCATION AND FLOW OF PUBLIC EDUCATION EXPENDITURES...............................91 Per Capita Financing of Education ...........................................................................96 C. MAIN FINDINGS OF THE EDUCATION PETS.................................................................96 Tracking Public Education Expenditures..................................................................98 Teachers....................................................................................................................99 School Infrastructure...............................................................................................103 Extra-budgetary Transfers ......................................................................................105 D. SUMMARY.................................................................................................................107 5. PUBLIC FINANCIAL MANAGEMENT REFORMS AND IMPLEMENTATION OF POLICY-BASED BUDGETING IN TAJIKISTAN................................................................................................. 111 A. INTRODUCTION .........................................................................................................111 Current Status of PFM system................................................................................111 B. PUBLIC FINANCIAL MANAGEMENT STRATEGY IN TAJIKISTAN..................................114 C. POLICY-BASED BUDGETING (PBB)...........................................................................118 Background.............................................................................................................118 What Has Been Achieved So Far?..........................................................................119 Key Outcomes and Lessons Learned......................................................................122 Experience with and Constraints to Strengthening Sector-Based Planning ...........127 Moving Forward .....................................................................................................128 D. SUMMARY.................................................................................................................130 6. THE IMPLICATIONS FOR PUBLIC FINANCES AND FISCAL SUSTAINABILITY OF ELECTRICITY SECTOR INVESTMENTS AND REFORMS................................................................................ 131 A. INTRODUCTION .........................................................................................................131 B. ENERGY INVESTMENTS .............................................................................................133 C. SUPPLY, DOMESTIC DEMAND, AND EXPORTS............................................................134 D. EXTERNAL PUBLIC BORROWING...............................................................................136 E. IMPACT ON GOVERNMENT REVENUE AND EXPENDITURES........................................138 Government Revenues............................................................................................138 Government Expenditures ......................................................................................142 Net Impact on the Budget .......................................................................................142 Fiscal Impact with only CASA 1000 OTL Investment ..........................................143 F. QUASI-FISCAL DEFICITS AND CONTINGENT LIABILITIES...........................................144 vi Quasi-Fiscal Deficits...............................................................................................144 Contingent Liabilities for Unused Power ...............................................................145 G. CONCLUSIONS...........................................................................................................147 REFERENCES................................................................................................................................ 159 List of Annexes Annex 2.1: Analysis of the Efficiency of the Education and Health Systems--Supply Side ...... 38 Annex 2.2: Education and Health Expenditure Budgets for 2006 by Functional Classification.. 57 Annex 2.3: Education and Health Expenditure Budgets for 2006 by Economic Classification... 60 Annex 2.4: Aggregate Three-Year Expenditure Forecast in the Education and Health Sectors .. 62 Annex 4.1: Data Annex............................................................................................................... 109 vii PREFACE The Programmatic Public Expenditure Review (PPER) was launched by the World Bank in May 2006, at a workshop with officials of the Government of Tajikistan (GOT), under the leadership of the Minister of Finance. At that workshop it was also agreed to undertake a Public Expenditure Management and Financial Assessment (PEFA), which is a comprehensive diagnostic assessment of the state of the public financial management (PFM) system based on a comparison with internationally agreed benchmarks. The PPER is supported by some of the development partners of Tajikistan, notably the U.K. Department for International Development (DFID). The work on PFM is carried out in close coordination with the UNICEF and EC. The PPER is closely aligned with ongoing reforms to PFM, such as the introduction of the Medium Term Expenditure Framework (MTEF) and the technical assistance required to support these reforms. The PPER will support the PFM reforms through analytical and diagnostic work. The PPER has been designed to cover five themes that represent the key concerns of/confront fiscal policy and PFM reforms in Tajikistan: Theme 1: Public Expenditure, Fiscal Space, and Growth Theme 2: Policy-Based Budgeting Theme 3: Challenges of Public Expenditure in the Social Sectors Theme 4: Fiduciary Risks Theme 5: Intergovernmental Fiscal Relations The PPER is being undertaken over three fiscal years (2006/07­2008/09) and published in three annual reports, of which this represents the second. The work in PPER 1 covered two of the issues under theme 1--the implications of public expenditures for economic growth and long-term fiscal space for priority expenditures. This work supports the need to analyze how scarce budget resources can best be allocated to meet the GOT's developmental objectives for growth and poverty reduction. PPER 1 also included an outline of the strategy for implementing the MTEF and a summary of the PEFA results, which support the GOT's decision in 2006 to begin implementing the MTEF in 2007. This report, PPER 2, is a sequel to PPER, which was published in July 2007. PPER 2 provides a detailed analysis of key public expenditure issues in Tajikistan and reports on the non-lending policy dialogue and technical assistance programs managed and coordinated by the World Bank. PPER 2 has a special focus on social sectors, especially the health and education sectors. Public Expenditure Tracking Surveys (PETS) carried out for the first time in Tajikistan in 2007 contributed to the findings in this report. The report also updates the macroeconomic and fiscal situation to take account of important ix developments in 2007, and also analyzes the implications of energy sector reforms and investments for fiscal sustainability. This report presents the technical background papers done for the Second Programmatic Public Expenditure Review (PPER 2, 2008) and is part of the four volume report on PPER 2. The report was task managed by Sudharshan Canagarajah. Technical Paper 1 on recent macro economic and fiscal developments was written by Sudharshan Canagarajah. Technical Paper 2 on efficiency of the education and health systems in Tajikistan was written by a team consisting of Shuhrat Mirzoev, Utkir Uamrov, Jariya Hoffman, and Matin Kholmatov. Technical Paper 3 on health PETS was written by Sarosh Sattar. Technical Paper 4 on PETS education was written by Dina Abu Ghaida and Nathan Smith. Technical Paper 5 on Public Finance Management Reforms and Implementation of Policy Based Budgeting was written by Matin Kholmatov with input from Svetlana Proskurovska , Mario Sangines, Shuhrat Mirzoev, Parvina Rakhimova, Armen Yeghiazaryan and Vahram Avanesyan., while Technical Paper 6 on the implications for public finances and fiscal sustainability of electricity sector investments and reforms was written by Martin Brownbridge under the guidance of Raghu Sharma and Sudharshan Canagarajah. We are thankful to Kazi Matin, Luca Barbone, Annette Dixon, Chiara Bronchi and Mehrnaz Teymourian for their overall guidance and support for the preparation of this report. We also would like to acknowledge the active collaboration and overall guidance we received from the Tajik government counterparts ­ namely Mr. Matlubkhon Davlatov (State Advisor on Economic Policy), Mr. Negmatjon Buriev (Senior Advisor, Executive office of the Presidential Administration), Mr. Safarali Najmiddinov (Minister of Finance); Mr. Abdujabbor Rakhmonov, (Minister of Education); Dr. Nusratulloh Faizulloev (Minister of Health). We acknowledge the active collaboration of Minisrty of Fianance officials, especially Shavkat Sohibov (Deputy Minister of Finance) and Nazrullo Abybullayev (Director, Budget Department). Local government officials in the finance, education and health departments of the oblasts and rayons as well as staffs working in the education and health sector generously took their time to meet with, discussed and provided information to the survey team during their visits to the rayons and oblasts in 2005. We also acknowledge the efficient work of the Center of Sociological Research "Zerkalo" who assisted us in fielding the health and education PETS surveys. Last, but note least, we acknowledge the generous financial support received from Department for International Development (DFID), UK, without which the interesting work on health and education PETS would not have been possible x 1. MACROECONOMIC AND FISCAL DEVELOPMENTS A. INTRODUCTION 1.1 Reform of macroeconomic management in Tajikistan was a key factor in stimulating the economic recovery and restoration of macroeconomic stability that began in the second half of the 1990s. A cut in fiscal deficits, the near-elimination of domestic deficit financing, and a sharp reduction in money supply growth enabled inflation to be brought down to single digit levels in 2004. Improved macroeconomic management and the restoration of peace and security following the civil war of the 1990s in turn helped to attract strong inflows of foreign exchange (from donor aid and migrant worker remittances). This supported a recovery of GDP growth, which has averaged around 9 percent per year since the start of the 2000s, although real output is only now reaching the level of the previous peak at the end of the Soviet era. In addition, the provision of debt relief from bilateral and multilateral creditors, together with constraints on external public borrowing in the first half of the 2000s, brought Tajikistan's external public debt down to sustainable levels by the end of 2006. 1.2 Beginning in 2006, macroeconomic performance deteriorated. Although real GDP growth has remained strong, inflation accelerated to almost 20 percent at the end of 2007. Furthermore, trade and current account deficits have widened sharply, unencumbered foreign exchange reserves have fallen to the equivalent of only two weeks of imports, and external public debt has risen sharply. The debt sustainability analysis completed in 2007 concluded that the country is now facing a high risk of debt distress. Tables 1.1 and 1.2 present data on the key macroeconomic and balance of payment (BOP) variables. 1.3 External factors have also contributed to the worsening of Tajikistan's macroeconomic performance. Rising international food and fuel prices have given a supply side boost to domestic inflation. However, the major cause of the deteriorating macroeconomic performance is a weakening of macroeconomic management, in several respects. 1.4 First, macroeconomic policy was procyclical and Tajikistan has experienced a boom in private consumption expenditure in the last few years, driven in part by strong growth of remittances. This has been accompanied by an expansionary monetary policy and fiscal stance, which has exacerbated the growth in aggregate demand to levels that far exceed growth in aggregate supply capacities. Hence inflation has increased and the trade deficit widened. Second, the central bank--the National Bank of Tajikistan (NBT)--lost control of the money supply, with broad money growing by 60 percent in 2006 and 78 percent in 2007. Third, the NBT has undertaken quasi-fiscal activities, providing direct finance (which partly explains the rapid money supply growth) and external loan guarantees to a privately owned financial institution, ostensibly for cotton financing. The external loan guarantees mainly have taken the form of pledging the NBT's own foreign exchange reserves as collateral deposits. Hence, more than two thirds of the NBT's foreign reserves are now encumbered. NBT loan guarantees and new external borrowing by the GOT for infrastructure projects has also led to a sharp rise in 1 outstanding public and publicly guaranteed external debt and an associated worsening of the country's debt indicators. Table 1.1: GDP, Expenditure, Inflation, and Investment 2004­07 2004 2005 2006 2007 Nominal GDP (SM millions) 6,158 7,201 9,272 12,780 Real GDP growth 10.6 6.7 7.0 7.8 Nominal expenditure growth (percent) 31 33 36 50 Inflation (percent) 5.6 7.1 12.5 19.8 Public investment (percent of GDP) 8.1 7.8 6.8 14.1 Private investment (percent of GDP) 5.1 6.0 6.0 7.0 Source: IMF. B. GROWTH OF OUTPUT AND SPENDING 1.5 Real GDP growth of 7.8 percent was recorded in 2007, slightly higher than in the previous two years, but consumer price inflation also has risen for three successive years, to reach 19.8 percent in 2007 (Table 1.1). In addition the deficit in trade in goods and services widened very sharply, by 12 percentage points of GDP, to 46 percent of GDP in 2007 (Table 1.2). 1.6 A key feature of the economy in the last few years has been the buoyancy of domestic demand. This has contributed to the rapid pace of economic recovery, providing a stimulus to bring back into production the factors of production--capital, labor, and land--that had been left underutilized by the severe contraction of the 1990s. However if growth in demand outstrips supply capacity, external and/or internal imbalances will emerge, and this is what has occurred in Tajikistan. The nominal growth rate of total expenditure in the economy--the sum of public and private consumption and investment expenditures--averaged around 30 percent in the mid-2000s and this growth accelerated sharply in 2007 to 50 percent (see Figure 2.1). 1.7 Higher nominal aggregate spending must have, as its counterpart on the supply side of the economy, some combination of higher real output, higher prices, and a larger trade deficit in goods and services. Figure 2.1 shows the annual growth in total expenditure from 2004 to 2007, broken down into its equivalent supply side counterparts. For example, in 2004 total expenditure rose by 31 percent in nominal terms, of which real GDP growth contributed 9.6 percent, the rise in the GDP deflator 17.1 percent, and the widening of the trade deficit 4.3 percent. 2 Figure 1.1: Contributions to Growth in Nominal Expenditure of Real GDP Growth, Growth in the GDP Deflator and Change in the Trade Deficit 60 Change in trade deficit GDP deflator Real GDP growth 50 40 21.8 30 13.7 4.3 Percent 17.8 20 17.1 22.4 17.2 10 9.2 9.6 6.0 5.5 5.8 0 2004 2005 2006 2007 Year Source: derived from IMF data. Note: In each year, the growth of real GDP, the growth of nominal GDP attributable to the change in the GDP deflator, and the trade deficit are expressed as percentages of the previous year's total expenditure. 1.8 The salient point to note from figure 1.1 is that the tradeoff between real GDP growth on the one hand, and inflation and the trade deficit on the other, has worsened as the growth of nominal spending has accelerated. In 2004, almost a third of the growth in nominal spending could be met by higher real GDP, but with the acceleration of growth in spending, the contribution of real GDP growth to the increase in total spending has fallen to about a sixth. Most of the growth in nominal expenditure in 2007 was manifested in higher inflation (the GDP deflator increased by 22 percent and consumer price inflation reached 20 percent by the end of 2007) and a widening of the trade deficit in goods and services (from 34 percent of GDP in 2006 to 46 percent of GDP in 2007). This is not surprising. Capacity constraints limit the rate at which real GDP can grow in the short term; hence if nominal spending growth accelerates beyond the maximum potential rate of supply side growth, as it did in 2007, the higher spending must find its outlet either as a larger trade deficit or as higher domestic inflation. The widening of the trade deficit mainly was driven by a huge increase in imports, which more than doubled in dollar values between 2005 and 2007. 3 Table 1.2: External Sector Data, 2004­07 2004 2005 2006 2007 Balance on goods and services ­12 ­27 ­34 ­46 (percent of GDP) Current account balance (percent ­4.0 ­2.7 ­3.0 ­9.5 of GDP) Net private remittances 141 526 903 1,416 (US$ millions) Unencumbered foreign exchange 189 91 111 107 reserves (US$ millions) External PPG Debt 824 897 896 1,216 (US$ millions) Source: IMF. PPG = public and publicly guaranteed. C. INFLATION AND FOOD PRICES 1.9 Inflation accelerated in 2006 and 2007 as a result of both the buoyant demand discussed above and supply side shocks. Consumer price inflation rose from 7 percent in 2005 to 12.9 percent in 2006 and then 19.8 percent in 2007.1 Food prices rose fastest, with food price inflation reaching 15.8 percent in 2006 and 27.5 percent in 2007. Within the food basket, the price of bread, a staple component of the diet in Tajikistan (comprising about one third of the total food basket) rose by 5.9 percent in 2006 but by a staggering 49.6 percent in 2007, reflecting the rapid global rise in wheat prices. Such a large increase in the price of bread will have had a large impact on the living standards of the poor, because it comprises a large share of their consumption basket. Moreover, the increase in inflation was not confined to food and fuel prices; inflation was strong among domestic services as would be expected if domestic demand was driving up prices. The prices of services rose by 8.3 percent in 2006 and 17 percent in 2007. Only among nonfood goods items was inflation relatively subdued, albeit accelerating, with prices rising by 5.3 percent and 7.7 percent in 2006 and 2007 respectively, which probably reflects the impact of the relative stability of the exchange rate on the prices of imported manufactured goods. 1.10 Figure 1.2 below shows the price indices for bread, all food items, services, and all consumer items over the period January 2005 to February 2008. Note the very steep climb in bread prices in the second half of 2008. 1These inflation rates are measured as the 12 month, December to December, rise in prices. 4 Figure 1.2: Price Indices for Bread, All Food Items, Services, and the CPI, January 2005 to February 2008 (December 2004 = 100) 200 Bread 180 Food Services CPI 160 140 120 100 80 05 06 07 7 Jan-05 r-05 -05 l-05 -06 -06 l-0 -07 ay v-05 ar- ay ar- ay-07 v-07 Ma Ju M Jul-06 Ju M Sep- No Jan-06 M Sep Nov-06Jan-07M M Sep No Jan-08 Source: Derived from GOT inflation data. D. COMPONENTS OF DEMAND 1.11 A pertinent question to ask is which of the components of domestic expenditure made the greatest contributions to its growth. Unfortunately, reliable and comprehensive GDP by expenditure estimates are not available, so it is not possible to give a definitive answer to that question. However, we have some estimates of government and private (in fact nongovernment) investment expenditures from the IMF. By deducting private investment and total GOT spending, minus interest payments and recurrent spending on social protection (a proxy for transfer payments to the private sector), we can derive an admittedly crude estimate of private consumption. Estimated on this basis, private consumption, which comprises about 80 percent of total expenditure, appears to have grown very rapidly in nominal terms, and at an accelerating pace, over the last three years, by about 33 percent in 2005, 39 percent in 2006, and 44 percent in 2007. 1.12 The rapid growth of private consumption expenditure partly was driven by a large increase in workers' remittances, mainly from Tajiks working in the Russian Federation. Workers' remittances are estimated to have almost tripled in nominal U.S. dollar values from US$526 million in 2005 to US$1,416 million in 2007 (Table 1.2), although these estimates are inevitably, given the difficulties in measuring this variable, not very reliable. By 2007 workers' remittances amounted to the equivalent of 38 percent 5 of GDP, and thus clearly contribute a substantial, and growing, share of household incomes. 1.13 It is likely that private consumption spending also was stimulated by the rapid growth in the domestic money supply. Estimated private sector savings fell in 2007, by almost five percentage points of GDP; hence the increase in workers' remittances cannot explain all of the rise in private consumption. Households must also have increased their net borrowing. Private sector credit growth (from the banking system) was very rapid in 2006, when it increased by 44 percent, and especially in 2007 when it rose by 58 percent (see Table 1.3). It appears likely that part of this credit growth financed higher private consumption. 1.14 Faced with very buoyant private sector consumption demand, a prudent macroeconomic policy would have entailed dampening aggregate demand growth through a tightening of fiscal and/or monetary policy, to maintain macroeconomic stability. Instead, both fiscal and monetary were expansionary and hence acted in a procyclical manner to intensify the spending pressures in the economy. Fiscal and monetary policies are discussed below. E. FISCAL POLICY 1.15 The fiscal performance in the last two years was characterized by strong growth in both domestic revenues and state budget expenditures, and by a large increase in externally funded capital expenditures in 2007 (tables 2.3 and 2.4). The latter led to a sharp increase in the overall fiscal deficit in 2007. 1.16 Domestic revenues almost doubled in nominal terms between 2005 and 2007. The growth in indirect taxes, especially VAT, made the largest contribution to the rise in domestic revenue. As a share of GDP, domestic revenues increased by just over one percentage point, from 16.4 percent in 2005 to 19.2 percent in 2007, but this largely reflects the growth in aggregate expenditure especially private consumption expenditure as a share of GDP. The nominal growth in indirect taxes (about 100 percent) was no faster than that of private consumption during 2005­07. Therefore it is doubtful whether the improvement in revenue performance will prove sustainable once the current boom in expenditures subsides. 1.17 State budget expenditures also rose quickly in nominal terms, by 23 percent in 2006 and 27.3 percent in 2007, with the growth in government salaries (42 percent in 2006 and 27 percent in 2007) making a major contribution to the expansion of the state budget. In each of the last three years, actual state budget expenditures exceeded budgeted amounts; by 2 percent in 2005, 7 percent in 2006, and 14 percent in 2007. Nevertheless, state budget expenditures rose slightly less in nominal terms than domestic revenues and grants. Therefore there was an increase in the fiscal surplus, excluding the externally financed public investment program (PIP), from 0.5 percent of GDP in 2005 to 1.8 percent of GDP in 2007, and hence also a reduction in net domestic financing of the budget (Table 1.3). 6 Table 1.3: Fiscal Aggregates, 2004­07 (SM millions) 2004 2005 2006 2007 Domestic revenue 1,063 1,390 1,754 2,621 Grants 41 50 436 142 Total expenditure 1,250 1,656 2,032 3,524 Current expenditure 747 1,087 1,393 1,710 Capital expenditure 500 562 634 1,806 PIP 189 245 291 1,100 Overall deficit ­146 ­209 158 ­761 Deficit excluding PIP and 43 36 70 232 MDRI debt relief Net external financing 249 187 145 884 Domestic financing ­103 22 ­303 ­124 State budget expenditure 1,061 1,411 1,741 2,424 Sources: IMF. Notes: In 2006, grants include SM 330 million from the Multilateral Debt Relief Initiative (MDRI). This is excluded from the deficit excluding the PIP. 1.18 PIP expenditures, which had been constrained to less than 4 percent of GDP in the first half of the 2000s, mainly because of concerns over the sustainability of external debt, increased very sharply in 2007 to 8.6 percent of GDP. The increase was attributable to the commencement of large infrastructure projects (a road and overhead power transmission line) financed by loans from China, which were contracted in 2006. Table 1.4: Fiscal Aggregates as Percent of GDP, 2004­07 2004 2005 2006 2007 Domestic revenue 17.3 19.3 18.9 20.5 Grants 0.7 0.7 4.7 1.1 Total expenditure 20.3 23.0 21.9 27.6 Current expenditure 12.1 15.1 15.0 13.4 Capital expenditure 8.1 7.8 6.8 14.1 PIP 3.1 3.4 3.1 8.6 Overall deficit ­2.4 ­2.9 1.7 ­6.0 Deficit excluding PIP 0.7 0.5 0.8 1.8 Net external financing 4.0 2.6 1.6 6.9 Domestic financing ­1.7 0.3 ­3.3 ­1.0 Source: IMF. 1.19 Table 1.5 provides a sectoral breakdown of government expenditures, including the externally funded projects, over the years 2005­07.2 Total expenditures more than doubled in nominal terms in this period. The main beneficiaries were in the economic block, where total expenditures tripled. Within the economic block, the fuel and energy sector and the transport sector together received most of the increase, reflecting the capital expenditures funded from Chinese loans that were undertaken in these sectors in 2The total expenditures shown in Table 1.5, which are sourced from MOF data, are higher than those in Table 1.3, which are sourced from IMF data. In particular, the MOF reports higher externally funded project expenditures than the IMF. 7 2007. The social block also received a substantial increase in resources, with education spending rising by 91 percent and health spending rising by 111 percent in nominal terms between 2005 and 2007. There were much more modest increases for the public administration and governance, defense, and law enforcement sectors, of 21 percent, 64 percent, and 47 percent respectively between 2005 and 2007. The sectoral shifts in budget allocations have, therefore, been generally positive, with the composition of the budget shifted towards infrastructure that can support productive activity and social services that contribute to human capital development. 1.20 Because of the large rise in externally funded project expenditures in 2007, the overall fiscal deficit widened sharply, from 2.9 percent of GDP in 2005 to 6 percent of GDP in 2007.3 This was the largest fiscal deficit incurred since the mid-1990s. Table 1.5: Sectoral Allocation of Expenditures, 2005­08 (SM millions) 2005 2006 2007 Total expenditure 1,713 2,127 3,563 State budget 1,313 1,643 2,215 Special funds 98 135 80 External projects 311 349 1,268 By sector Governance and public administration 243 293 293 Defense 96 105 157 Law enforcement agencies 115 138 169 Social bloc 676 907 1,219 Education 290 373 553 Health 90 125 190 Social insurance and protection 251 342 382 Culture and sport 45 67 94 Economic block 518 605 1,521 Housing and communal services 121 125 196 Fuel and energy 103 176 572 Agriculture, fishing, and hunting 84 103 184 Industry and construction 24 45 22 Transport 176 152 539 Other economic services 10 5 8 Other expenditures 64 78 203 Source: Ministry of Finance. 3The overall fiscal surplus of 1.7 percent of GDP recorded in 2006 is somewhat misleading because it includes as revenue and grants the US$99 million (3.5 percent of GDP) of debt relief provided under the Multilateral Debt Relief Initiative. If this were excluded from revenue and grants, the overall fiscal balance would be a deficit of 1.8 percent of GDP. 8 F. EXTERNAL DEBT 1.21 The financing of the increased PIP expenditures in 2007 had a profound impact on the country's external debt dynamics. Debt relief provided by bilateral creditors (the Russian Federation and Pakistan) and the IMF under the Multilateral Debt Relief Initiative reduced the net present value of public and publicly guaranteed external debt to 23 percent of GDP and 99 percent of exports at the end of 2006. These ratios were well within the thresholds of 40 percent of GDP and 150 percent of exports under which debt is deemed sustainable for a low-income country that is a medium performer in terms of its Country Policy and Institutional Assessment (CPIA) rankings under the IDA/IMF low-income countries debt sustainability framework. The external loans contracted from China, with a nominal value of more than US$600 million, will sharply raise Tajikistan's debt indicators over the medium term (the loans will be disbursed over the 2007­09 period). Consequently, the Joint Fund-World Bank Debt Sustainability Analysis (DSA) completed in early 2007 concluded that "as a result of the rapid debt accumulation ... debt distress risk has increased significantly and could be now classified as high" (IMF and World Bank, 2007, p. 1). 1.22 Subsequent to the DSA, the NBT disclosed that it had guaranteed US$328 million of private sector external commercial loans, which had been extended mainly by foreign banks to Kredit Invest, a financial institution involved in financing the cotton trade. Consequently, although precise data on the debt indicators are not yet available for end-2007, it is almost certain that the net present value of external public and publicly guaranteed (PPG) debt will exceed both the sustainable threshold of 40 percent of GDP and 150 percent of exports in 2008. G. MONETARY POLICY 1.23 The NBT lost control over the money supply in 2006 and 2007. Reserve money expanded by 48 percent in 2006 and 40 percent in 2007, while broad money, which includes foreign exchange deposits held by Tajik residents in commercial banks, rose by 60 percent in 2006 and 78 percent in the following year. Money supply growth was accompanied by rapid growth in credit to the private sector, by 44 percent and 58 percent in 2006 and 2007 respectively (see Table 1.6). It is virtually inevitable that such rapid growth of monetary aggregates will create destabilizing internal and external imbalances. 1.24 Given that the GOT's domestic borrowing requirement was negative in 2006 and 2007, the rapid growth in reserve money requires some explanation. The main contributor to this growth appears to be domestic credit of approximately SM 700 million (about 7 percent of GDP) which the NBT extended to Kredit Invest (the same financial institution to which it provided external loan guarantees). Furthermore, the NBT's efforts to mop up the created liquidity, using sales of securities (NBT bills), were not very aggressive. The interest rate on NBT bills barely rose above 7 percent in 2006 and 2007, even though inflation was rising steadily, which meant that the real yield on NBT bills became progressively more negative. 9 1.25 The NBT deposited US$245 million of its own foreign exchange reserves as collateral for the external commercial loans of Kredit Invest (which is unable to service its loans). These reserves are, therefore, encumbered. Consequently the NBT's reserves, which were not very high even before the loan guarantees were extended, were reduced to only about US$100 million, or the equivalent of about two weeks of imports. Table 1.6: Monetary Data, 2004­07 2004 2005 2006 2007 Broad money (including FX 14.3 29.9 60.2 78.3 deposits) growth (percent) Broad money (excluding FX 146.0 55.2 38.3 deposits) growth (percent) Reserve money growth (percent) 10.0 28.8 48.5 40.2 Private sector credit growth 57.9 29.7 43.8 57.8 (percent) NBT bill interest rate (percent) 6.2 7.0 7.7 7.3 Source: IMF. H. PROSPECTS AND RISKS FACING THE TAJIK ECONOMY 1.26 The medium-term prospects for the economy have become less optimistic since 2006. The economy faces several serious medium-term risks as a result of recent policy mistakes. The most serious risks are (i) an acceleration of inflation, (ii) a BOP crisis, and (iii) external debt distress. Real GDP growth is likely to fall in the short term because of the supply side shocks that hit the economy this winter (in the first quarter of 2008): widespread electricity shortages disrupted several sectors of the economy and agriculture was adversely affected by the severe winter weather. In the medium term, economic growth could be curtailed by continued high inflation, which is likely to undermine the confidence of private sector investors, and which could discourage inflows of remittances and encourage capital flight. Furthermore, real growth certainly will be hit hard if there is a BOP crisis. I. AGGREGATE DEMAND AND INFLATION 1.27 Inflationary pressures are likely to remain strong over the short to medium term, for several reasons. First, even if the NBT reigns in money supply growth in 2008, the huge increase in the money supply that took place in 2007 will provide an impetus to spending and hence inflation in 2008, given the lags in the transmission mechanism of monetary policy. Second, the budgeted fiscal stance is more expansionary in 2008 than in 2007. The state budget expenditures are budged to increase by 41 percent (which includes a 53 percent increase in the wage bill) in 2008, alongside a 30 percent increase in PIP expenditures. Consequently, the overall fiscal deficit will widen and the fiscal balance excluding the PIP will move from a surplus in 2007 to a deficit in 2008. The budgeted increase in domestic borrowing in 2008 will impede the NBT's efforts to reduce the growth of reserve money. Third, rising international prices for grains and fuel, together with planned increases in electricity tariffs, will continue to exert pressure on inflation. 10 Inflationary pressures will be greatly exacerbated if there is a sharp nominal depreciation of the exchange rate. J. THE BALANCE OF PAYMENTS 1.28 The Tajik economy has been rendered vulnerable to a BOP crisis, for two reasons. First, the large increase in the domestic currency money supply over the last three years, driven in part by the domestic credit extended to Kredit Invest by the NBT, has created an overhang of domestic currency assets. Private agents could seek to change these assets into foreign currency assets if they believe that the domestic currency will fall in value relative to foreign currencies: hence there would be a huge increase in demand for foreign exchange and a fall in the nominal exchange rate. Second, because of the NBT's pledging of its foreign exchange reserves to secure external borrowing for Kredit Invest, it has almost no unencumbered foreign reserves left to defend the exchange rate in the event that it faces depreciationary pressures. A sharp depreciation of the exchange rate is likely to have serious adverse economic consequences, through, the following: (i) inflationary pressure arising from higher prices of imported goods and services, (ii) increased costs of imported inputs that will undermine the viability of businesses in the non-traded sectors, (iii) potentially serious risks to the soundness of commercial banks that have very large foreign exchange denominated liabilities, and (iv) an undermining of confidence of private sector investors in the stability of the economy. K. EXTERNAL DEBT DISTRESS 1.29 The contracting of new concessional loans from China and the guarantees extended by the NBT to private sector debtors (which are almost certain to be called because the borrower has defaulted on the servicing of these loans) will push Tajikistan's debt indicators back above the thresholds for the external debt stock indicators at which external debt is regarded as sustainable. Whether or not external debt sustainability can be restored will depend on the strategy that the GOT adopts towards new borrowing. If it restricts the volume of new borrowing to around 4 percent of GDP per year and only contracts loans on concessional terms, then the debt stock indicators can be brought back to sustainable levels over the medium term. If, on the other hand, the GOT contracts much higher levels of external credits, or contracts loans on commercial terms, then the debt stock indicators will remain above the sustainable thresholds and the debt servicing indicators will also rise, leaving the country vulnerable to a debt crisis in the long term. L. POLICY REFORMS REQUIRED TO ADDRESS THE MACROECONOMIC RISKS AND CHALLENGES 1.30 The priorities for macroeconomic policy should be to restore macroeconomic stability, avert the risk of a balance of payments crisis, and ensure long-term external public debt sustainability. The key policy objective of promoting sustainable economic growth should be pursued through supply side reforms that aim to attract more private investment and improve the efficiency of resource use in the economy (especially in the agricultural sector), which is crucial for poverty reduction. 11 M. MACROECONOMIC MANAGEMENT 1.31 Macroeconomic management in the short to medium term should be focus on the objectives of dampening the growth of aggregate demand, in order to reduce inflation and the current account deficit, and minimizing the risk of a BOP crisis. These objectives are interlinked, in that reducing aggregate demand and inflation will help to avert a BOP crisis while a BOP crisis would seriously damage efforts to control inflation. 1.32 Reducing aggregate demand growth requires a tightening of monetary policy, with much lower rates of money supply growth than were recorded in 2006 and 2007, together with a tightening of fiscal policy. In addition, the NBT must begin rebuilding its foreign exchange reserves, which have been depleted to dangerously low levels by its guaranteeing of the Kredit Invest loans. Until this is done, the NBT will have limited policy options available for stabilizing the exchange rate in the event that a run on the domestic currency takes place. 1.33 The monetary aggregate that the NBT can control is reserve money.4 If it is to control reserve money and at the same time increase its foreign exchange reserves, government borrowing from the NBT must be reduced: in effect, government must save in order to create the resources for the NBT to purchase foreign assets. Hence a monetary policy strategy that has the twin objectives of controlling reserve money growth and rebuilding foreign exchange reserves requires a supporting fiscal policy that involves the GOT saving (that is, negative borrowing) with the NBT. The budget for 2008 envisages the government borrowing from the NBT about 1.5 percent of GDP. Therefore fiscal policy should be tightened to ensure that government can save with the NBT. How the fiscal adjustment should be implemented is beyond the scope of this Technical Background Paper but possible components include the following: (i) cutting the growth of state budget expenditures, especially wages and salaries; (ii) raising personal income tax rates (the top marginal rate is only 13 percent, so there is scope for mobilizing more income tax from higher marginal rates); and (iii) enforcing tax payments by state-owned enterprises such as the Tajik Aluminum Company (TALCO). The objective of rebuilding the foreign exchange reserves requires a medium-term timeframe to implement; hence fiscal policy will have to be tighter over the medium term, and not just in 2008, to support this objective. 1.34 The NBT's foreign exchange reserves should be rebuilt in a gradual manner, with the NBT allowing an orderly depreciation of the exchange rate. An orderly depreciation will reduce the risk that the exchange rate will be subject to a sharp and economically disruptive fall at some point in the future. While the risk of a BOP crisis cannot be eliminated, it can be reduced if (i) macroeconomic policies dampen aggregate 4 The NBT has imperfect control over somoni broad money, because it does not control the money multiplier, although it can exert some influence over it through the cash reserve ratio imposed on commercial banks. The NBT has no effective control over the foreign currency component (foreign exchange deposits) of the money supply. Foreign currency deposits comprised 47 percent of broad money at end-2007. In the mean time, a recent Resolution of the Tajik Government has beep adopted "On approval of the inflation control in the Republic of Tajikistan throughout 2008-09." 12 demand and thus reduce the external sector deficits; (ii) if money growth rates are sharply cut, which reduces the amount of domestic liquidity that is available for purchasing foreign currency; (iii) if the nominal exchange rate depreciates in an orderly manner, which makes purchase of foreign currency more expensive; and (iv) if the foreign exchange reserves are rebuilt. It would be very helpful for macroeconomic management if the country's international donors were to provide budget support to the GOT, without insisting that this budget support is used to expand state budget expenditures, because this budget support could then be used to supplement the NBT's foreign exchange reserves. N. EXTERNAL DEBT STRATEGY 1.35 The GOT should implement a strategy for new external borrowing to ensure the long-term sustainability of external PPG debt5. There are three key elements of such a strategy. First, the GOT should formulate, and adhere to, quantitative annual ceilings on new external borrowing, calibrated to bring the external debt indicators back down to the thresholds at which external debt is regarded as sustainable for an economy like that of Tajikistan. These ceilings should cover all PPG debt, including that contracted by parastatals. Second, new borrowing should only be contracted on concessional terms. Third, there should be legal constraints on the authority of the NBT and other parastatals to contract or guarantee external debt. O. STRUCTURAL REFORMS 1.36 The limits to which GDP growth can be generated simply by boosting aggregate demand are probably now close to being reached. Since the late 1990s, when the economy began its recovery, high rates of GDP growth have been the result of a move back towards the production possibility frontier (PPF), from a point well within the PPF. As the economy approaches its PPF, further growth will require an outward movement of the PPF--that is, an expansion of aggregate supply. Raising private investment rates, which in spite of the economy recovery have remained at low levels of around 6­7 percent of GDP (see Table 1.1), together with improvements in the efficiency of resource allocation, are the keys to a sustainable expansion of aggregate supply. 1.37 To boost private investment, reforms to improve the business climate for private investors are needed, including reforms to business licensing and regulation to create a more transparent and less complex regulatory framework that gives less scope for public officials to impose arbitrary decisions on private investors. This issue was discussed in chapter 2 of PPER 1 (World Bank, 2007). Sectoral reforms are needed to improve the efficiency of resource use, especially in the agricultural sector, where farm productivity (especially in the cotton sector) is impeded by restrictions on land use (which force farmers to grow specific crops irrespective of their economic viability), and monopolistic agricultural marketing arrangements. These issues will be explored in forthcoming World Bank analytical work. 5 External Debt Management Strategy for 2008-10 is currently pending approval in the Government. The strategy identifies key debt parameters and will gradually ensure a constant flow of resources under sustainable thresholds. 13 2. EFFICIENCY OF THE EDUCATION AND HEALTH SYSTEMS IN TAJIKISTAN A. BACKGROUND 2.1 This Technical Background Paper examines efficiency of education and health spending in Tajikistan. The efficiency of spending on education and health can be examined from two perspectives: balance of resources and input-output efficiency. First, analysis of the composition (including economic and functional classification) and geographical distribution of spending can help inform policy makers and stakeholders on the overall balance and progress of service delivery in education and health. Such analysis could also show ways to improve efficiency through reallocation of public spending towards priority areas and achieve desired coverage and impact. Second, an analysis of spending against performance benchmarks at the national or sub-national levels can help identify potential technical and allocative issues related to improving efficiency. The efficiency analysis agenda in Tajikistan is further discussed in Public Expenditure Tracking Surveys (PETS) for both education and health sectors, presented as Technical Background Papers 4 and 5 of this PPER. 2.2 Education and health sectors play vital roles in the achievement of the Millennium Development Goals (MDGs) in Tajikistan. A recent World Bank report,6 however, assessed the prospects of Tajikistan achieving MDGs related to education and health as unlikely. Tajikistan first needs to improve enrollment and completion rates to achieve universal primary education indicators, and also improve infant and maternal mortality in order to tackle health MDGs. These are very challenging tasks for a country like Tajikistan. 2.3 To achieve MDGs related to education and health targets, Tajikistan needs balanced and prudent management of its limited resources. There are four MDGs directly related to education and health.7 According to the UNDP Need Assessment, the cost of meeting the education MDG under the baseline scenario is estimated at about US$1.76 billion (or US$160 million annually or US$84 per student per year) and the cost of meeting the health MDGs is estimated at about US$3.6 billion or US$46 per person during 2005­15.8 The total cost of achieving the MDGs (halves poverty and hunger and achieve sustainable human and social development) is approximately US$13 billion over the next 11 years in the baseline scenario.9 The financing gap would stand at US$2.12 billion, or 16 percent of total needs in the high-growth scenario.10 Potential sources of 6 The World Bank (2005), "Millennium Development Goals: Progress and Prospects in Europe and Central Asia Region," Washington, D.C. 7 They include achieving universal primary education (MDG 2), lowering child (MDG 4) and maternal mortality rates (MDG 5), and combating the spread of infectious diseases (MDG 6). 8 UNDP (May 2005), "Investing in Sustainable Development: Millennium Development Goals Needs Assessment in Tajikistan," Dushanbe, Tajikistan. 9 The scenario assumes GDP growth of 5 percent, tax revenues as a share of GDP at 19 percent, and a ratio of government expenditures to MDG investments equal to 31 percent of the total budget. The total financing gap stands at US$ 4.7 billion, or 36 percent of total needs. 10 The high-growth scenario assumes GDP growth of 7 percent, tax revenues at 24 percent of GDP, and budget allocation to MDG investment equaling to 50 percent. 15 financing include domestic (public and private payments) and international donors. However, there remains a financing gap of US$552 million in the education sector and US$1.76 billion (or US$160 million annually) in the health sector. Key Emerging Issues 2.4 Current economic and social circumstances in Tajikistan (with two thirds of people living below poverty line) are such that Tajiks cannot afford a private education or health care system. Therefore, it is important to maintain the public education and health system inherited from the former Soviet Union and halt the current decline of quality and coverage. There is a clear need to streamline and sharpen policies in both sectors and link them to priority actions that are backed up by resources and at the same time help the neediest and marginalized people. This is especially relevant given the rising demand for these services and continuing declines in equity, coverage, and quantity of services currently being provided. 2.5 These demands cannot be met without leadership and commitment by the government to streamline the supply side of public service delivery. Delivery mechanisms need to be reviewed for efficient resource allocation, composition, coverage, and so forth, and actionable measures put forth to achieve better outcomes. Increasing public spending on both education and health, although commendable and evident in Tajikistan, does not guarantee improved outcomes. Given limited resources, the main focus needs to be more efficient use of existing resources. 2.6 Finally, both service delivery structures and accountability need to streamlined and strengthened. Improved organizational structure should include linkages between national and local levels and within sectors to balance priorities, remove inefficiencies, and improve technical efficiency of spending. 2.7 This Technical Background Paper is organized into four sections. Section 2 provides an overview of the recent reform effort in Tajikistan and outlines key institutional and budget management challenges based on sector PETS. Section 3 examines the trends and the composition of public expenditures in the health and education sectors. Section 4 discusses performance-related issues of public service provision, including quality and access in urban and rural areas. The section also focuses on budget accountability as an essential component for enhancing access to public services, especially at the local level. Section 5 summaries key challenges and provides final remarks on efficiency of public spending in health and education, followed by technical Annexes with background data and detailed analysis. B. RECENT REFORMS IN SOCIAL SECTORS AND KEY CHALLENGES AHEAD 2.8 Tajikistan's education and health indicators were impressive following independence, but the government was unable to maintain the comprehensive social service delivery systems inherited from the Soviet era. Before independence, the education system provided comprehensive teacher education and in-service professional development and heavily subsidized pedagogical publishing and learning materials. The 16 health system was highly specialized and centralized and placed heavy emphasis on curative and in-patient care with an oversupply of hospitals and specialized doctors to provide services to citizens. However, the loss of transfers from the Soviet Union after independence combined with external shocks and civil war led to inadequate funding of public services. As a result, the physical infrastructure deteriorated, schools and health facilities did not have sufficient inputs to operate effectively, skilled education and health personnel left the sectors, and widespread informal payment practices limited access of the poor to services. 2.9 The decline in the quality of education and health services prompted the government to initiate reforms in both sectors. For example, the government amended the constitution through referendum in 2004 to formalizing service fees and introduce a Basic Benefit Package (BBB) free of charge for vulnerable population groups and selected health services. The second PRSP for 2006­08 incorporated comprehensive reform strategies that aimed at reversing the declines, and the government has begun implementing the strategies with donors' assistance. 2.10 The government has adopted a "National Strategy for Education Development." The strategy aims at ensuring access to basic education for everyone, raising secondary education enrollment, and improving the quality and relevance of education at all levels. To achieve these objectives, the strategy focuses on the following reforms: (i) reform the curriculum of basic education; (ii) increase the provision of textbooks; (iii) develop new textbooks aligned with the new curriculum; (iv) retrain teachers; (v) pilot a new scheme of school financing based on number of students; and (vi) reorganize the Ministry of Education by transferring the responsibilities for vocational and technical education back to the Ministry of Education and by establishing new units responsible for planning, policy analysis, and monitoring and evaluation. Implementation of the strategy has been supported during 2006­07 by a multidonor grant "Education for All--Fast Tract Initiative" (EFA-FTI) of US$18 million. Other donors have also supported the reforms. 2.11 Education reform has progressed gradually. A new national curriculum for basic education for all grades of basic education has been approved and is being implemented. Modern textbooks have been developed to align the new curriculum and a textbook rental scheme has been introduced to ensure adequate supply for students. Teachers' work loads (stavka) have been reduced through a combination of increasing teaching hours per week (from 14­16 to 16­18) and reducing the teaching plan (syllabus) by 300 hours. Since 2005, the government has piloted the per capita financing for general education and now expanded to 17 rayons (districts) to increase the autonomy and financing of the school system. 2.12 Health sector reform is also currently underway. Spending for the health sector increase from 0.9 percent in 2002 to 1.3 percent and 1.4 percent of GDP in 2005­06. Per capita public spending on health care was US$3.8 in 2005 (up from US$1.7 in 2002). In April 2007, the government adopted a formula for per capita financing of public health care and introduced the system in seven rayons on a pilot basis. A guaranteed Basic Benefit Package (the BPP) of health services was introduced in 2005, but was suspended 17 as health workers and patients raised concerns about the variation and the high rate of copayments and numerous exemption categories. The Ministry of Health revised the BBP and reintroduced it in three pilot rayons in June 2007. 2.13 Between 1994 and 1996, the Ministry of Health began rationalizing the health system, leading to shorter hospital stays, fewer beds, and consequently the reduction in the number of medical personnel. This was accompanied by opening nursing hospitals and administration, one-day-stay hospitals, daytime hospitals, and in-home healthcare. Funds saved from bed reduction (US$5.21 million) were absorbed by state budget, which unfortunately reduced interest and incentives for future rationalization and cost cutting. Furthermore, the hospital rationalization adversely affected the quality of health services delivery because primary health care was not reformed in parallel. Limited access and high costs led to a deterioration of key health indicators. 2.14 The government's focus is now on strengthening primary care to reverse the decline in health status. International experience has shown that primary health care is cost effective in improving the health status of the population and achieving health outcomes. However, underfunded primary care has led patients, especially in urban areas (80 percent), to seek care at hospitals. Since 2003 the government has increased the number of primary health care medical personnel by increasing salaries, developing institutional capacity and retraining existing medical personnel as family doctors and nurses, improving management and performance, improving conditions in primary health care facilities through repair and maintenance, and adopting a Health Financing Strategy. 2.15 The government's Health Financing Strategy11 aims to improve equity, efficiency and cost-effectiveness of the health system through health financing reforms. The strategy includes an overall increase in public funding to the health sector and in particular primary health care, an introduction of per capita financing for primary health care, and the introduction of the BBP. This will help the Ministry of Health incrementally to move away from input- to output-based financing, and provide a better match between needs and resource allocation. The BBP provides free services for vulnerable population groups and provides a legal framework for developing the copayment policy for selected health services in hospitals. The majority of covered services are in primary health care. The implementation of BBP in hospitals will support efforts to formalize informal payments by allowing hospitals to charge for services not covered by the state under the BBP (paid services). A medium-term objective of the BBP is to move hospitals away from budget finance to a combination of budgetary funds and fee revenues from patients. Budgetary funds thus freed would be reallocated to other parts of the health system such as primary health care and public health. 11 Key directions of reforms proposed in the strategy include: Establishment of an institutional structure of a single- payer for health care; Pooling of sources of public funds for free health care; Development and implementation of new provider payment mechanisms; Regulation of informal payments in the health system and introduction of formal co- payments; Increasing health personnel salaries; Reorganization of the system of health services delivery; Improving the quality of health care; Increasing public financing for health care; Improving donor aid coordination in the health sector. 18 2.16 However, reforms in health and education sectors have been slow. Despite ongoing reforms in education and health, traditional and ineffective methods of financing and procurement are still practiced in most secondary schools and hospitals, also relying heavily on local discretion. Furthermore, assumed leakages in the health care and education systems and lack of coordination between donor agencies providing external financing are slowing down any reforms in the sectors. Thus weaknesses in the budget management system and excessive discretionary authority at the local government level poses challenges in the implementation of joint reform strategies in the health and education sectors. Key Budget Management Challenges 2.17 Common challenges in health and education sectors hinder the benefits of ongoing reforms. Fragmentation of the budget into republican and local tiers and complex financing mechanisms at the local level (including respective end-user institutions, jamoats, rayons, and oblasts) create multiple sources of public funding and challenges in monitoring efficiency. In addition, weak accountability and record-keeping practices in local governments and the procurement and disbursement inefficiencies were observed by PETS. Other challenges pertaining to health and education include frequent delays in salary payments and practically no expenditure planning at facilities or by jamoat and rayon level officials. 2.18 The national health and education budgets do not reflect the budget outside the respective ministries and hence reflect only a small portion of total public expenditure. A consolidated budget that includes all financing sources of the entire health and education sectors (including public investment) is not readily available. One would have to consolidate all health budgets (the republican health budget, the local health budget, and capital investment for health in the Public Investment Program [PIP]). The republican heath budget includes funding for the Ministry of Health, republican hospitals, the national diagnostic center, state medical university, and public health services. It also includes other health care facilities operated by other ministries and state agencies (Ministry of Agriculture, Ministry of Irrigation and Water Supply, Ministry of Environment, the Republican Center for Healthy Lifestyles [established in 1999], and a National Inter-Agency Coordinating Committee).12 These facilities are financed through the budgets of their respective ministries and agencies. Local budgets finance health services delivered by facilities in oblasts, cities, rayons, and jamoats. Finally, capital investment of the health sector is financed by the PIP (foreign-financed projects) and Centralized State Investment Program (CSIP). 2.19 Procurement of goods and services in health and education is unsystematic and usually in kind. Goods and services generally are procured by the oblast, rayon, or Central Rayon Hospital (CRH) and further distributed to health facilities via the CRH or jamoat. As health facilities do not have bank accounts, they receive in kind inputs in the form of drugs, dressing materials, food, etc. from the CRH and the jamoats. For instance, medicines are procured centrally by either rayon or CRH and distributed to health 12 The World Bank, "Tajikistan: Health Sector Policy Note," 1994. 19 facilities directly or through jamoats, while food is procured by CRH from local branches of agricultural state-owned enterprises (SOEs). Rarely, a school gets some cash in order to pay for goods and services. Frequently, school principals do not even know the amount of the respective transfer. 2.20 There is practically no expenditure planning among education and health care providers, and spending plans are not revised properly during the year. In-year budget revisions or amendments are common in Tajikistan. The Public Expenditure and Financial Accountability (PEFA) assessment in 2007 found that budget revisions were mostly upward due to better revenue collections than expected due to conservative forecasts of GDP growth. In addition, the annual Law on State Budget provides incentives13 to underestimate revenues.14 This enabled an increase in spending at various levels. Wide deviations in reported revenues, ranging from 73 to zero percent in 2005, reduce the reliability of the budget and hinder the ability of local governments to plan ahead. Health facilities find it difficult to plan health spending due to budget amendments during the year, which results in suboptimal use of available resources. 2.21 Lack of knowledge on formal budget procedures and unclear rules and processes allow discretion in allocation of public resources, leading to nonoptimal outcomes. In budget management, the rayon chairman usually has the last say in resource allocation, reallocation, and the use of unspent funds. Health facilities and educational institutions, which receive funding through local budgetary channel, may remain solely at the discretion of the rayon and oblast level administrations. After the budget is approved, funds can be reallocated but there are restrictions. According to the recently amended Law on Public Finance, funds can be reallocated across line items15 with an exception of protected expenditure. According to PETS survey findings, the majority of respondents at the rayon and the school/CRH levels reported that funds can be allocated across line items, while only 17 out of 74 respondents at the jamoat level reported that fund can be reallocated across line items. 2.22 In 2005­06, there were frequent delays in payment of wages and salaries in both sectors. About 12, 19, and 29 percent of health care workers respectively at the rayon, CRH, and jamoat levels reported delays in cash payment for wages and salaries and 26 percent of health facilities reported such delays in 2005. The frequency of delay was higher at CRH or facility levels than at rayon or jamoat levels. More than one half of the CRH and health facilities experienced delays three times or more in 2005. However, 52 percent of respondents at the health facility reported that the longest delay in payment of wage and salary lasted for two months or more. The delays mostly were due to shortfalls of local government revenues. In the education sector, 47 percent of schools reported that there were salary delays in 2006, and 28 percent reported that salaries were 13 Article 17 of the Law on Budget System and Process (2002) states that revenues received in excess of budget projections as well as any excess of local revenues over local expenditures resulting from "saving" in planned expenditures are retained by the local councils and will be spent at the discretion of local executives in cities and rayons. Additional revenue has no impact on the amount of subventions or subsidies. 14 "Tajikistan: Public Expenditure and Financial Accountability Assessment--Performance Report," SECO (Swiss Economic Cooperation Office) and World Bank, June 2007, p. 23. 15 Apart from these protected items, vehement limits are wide (up to 20 per cent of a budget organization's allocation; and this limit can be exceeded if so agreed between the MOF and the budget organization. 20 delayed more than once. In Dushanbe and Rayon of Republican Subordination (RRP), less than 9 percent of principals reported payment delays in 2006. The situation is completely reversed in Soghd and Khatlon, where respectively 72 percent and 56 percent of principals reported delays. Delays are also more frequent among rural schools than urban ones. C. RECENT TRENDS IN PUBLIC SPENDING ON EDUCATION AND HEALTH 2.23 The Tajik government is committed to increasing spending over the medium term. Since the implementation of the Poverty Reduction Strategy Paper (PRSP) in 2003, public spending on education and health has steadily increased to 4 percent and 1.4 percent of GDP respectively in 2006, which accounts for one fourth of the total pubic expenditure (equivalent to 5.4 percent of GDP in 2006). However, health and education spending remains low compared to international standards. The government has also sought to improve spending efficiency through ongoing reforms outlined in the government's National Strategy for Education Development (NSED, 2005) and Health Financing Strategy (2005). Existing Intra-sectoral Financing Structure 2.24 Provision of education and health services is the responsibility of both central and local governments.16 The central government provides tertiary services with the financing from the republican budget, while local governments provide secondary and primary education and health care services financed by the local government budget (from own tax revenues and transfers from the republican budget). Financing from the private sector, especially in the form of informal payment or fee-for-services, is difficult to estimate as no current data are available. Finally, donor financing (grant and credit) as well as in-kind assistance (technical and humanitarian) is sizable but the government statistical system only provides information on donors' funds that are allocated through the PIP. 2.25 The bulk of public expenditure on education and health was carried out by local government. Local governments are responsible for provisioning of primary and secondary education and primary health care as well as public health services. Education spending by local governments remained steady between 2002­07, accounting for around 81 percent of the national education budget and 46 percent of aggregate local budgets. By contrast, health spending by local governments declined to 71 percent of the total health budget in 2004 (from 79 percent in 2002) before climbing to about 77 percent of the total in 2007. Despite the trends, local governments still play key roles in the financing and provision of education and health care services in Tajikistan. 16 Local governments in this Technical Background Paper refer to oblasts, rayons, and towns (excluding jamoats unless otherwise specified). 21 Table 2.1: Public Spending on Education and Health (percent of total) 2002 2003 2004 2005 2006 2007 Education Total 100 100 100 100 100 100 Republican 19 18 19 17 19 19 Local 81 82 81 83 81 81 Health Total 100 100 100 100 100 100 Republican 21 25 29 24 24 23 Local 79 75 71 76 76 77 Source: Ministry of Finance, Republic of Tajikistan. 2.26 The organizational structure of the health system is complex at the local government level, creating difficulties in monitoring the efficiency of resource allocations. The Ministry of Finance provides resources to the oblast finance departments for health care expenditures. These resources are given to the Oblast Health Department responsible for providing funds to region-wide facilities such as oblast-level hospitals and large polyclinics. The 61 rayons do not depend on budget transfers from oblasts because a major part of their budgetary revenue is from local taxes and shared taxes, which are under their control. Therefore the rayon transfers the majority of health funds to the CRH. The 356 jamoats are responsible for disbursing the funds transferred from the rayons based on line items to health facilities operating within the jamoat territory, mostly in the rural area. As jamoats are more closely connected to the community base, they are responsible for implementing rayon policy and do not have their own source of financing. Mostly, the rayon health budget is executed though the CRHs and the jamoat, while about 60 different rayon chairmen negotiate for health resources. 2.27 Rayons are responsible for allocating among agencies and defending their budgets at the Ministry of Finance and thus have more flexibility in managing their budgets. Rayon finance departments (RFDs) and rayon education departments (REDs) are responsible for executing a complex budgetary process. Allocation among schools is the rayons' responsibility. It is worth noting that schools participating in the per capita financing reform program have very accurate records of their budgets. However, resources for schools not participating in the per capita financing reform program are channeled sometimes through rayons and sometimes through jamoats. In addition, the rayon governments have more flexibility in managing their budget because of the ability to collect and retain local taxes and the share-tax without sending all revenue back to the Central Treasury of the Ministry of Finance. The ability to retain excess revenue over the forecast provided greater scope to local governments to spend additional revenues on other priorities that were left out at the beginning of the year. They are empowered to make final resource allocation to sectors at the beginning of the year, reallocate additional revenue to sectors during the year, and reallocate resources across sectors and line-item expenditures. 22 2.28 Private out-of-pocket payments for education and health service have increased to make up for shortfalls in public spending. They pay for compensation of service providers, necessary inputs to provide services, and even for communal services to keep schools and health facilities operating. Out-of-pocket payments for education are roughly equivalent to public spending on education (2.1 percent of GDP in 1999) and account for about 2.4 percent of household spending. Private out-of-pocket payments accounted for 71 percent of total health spending in 2003, compared to 16 percent by the government budget and 13 percent by donors. The practice has become widespread and acceptable as a norm and consequently the poor can no longer afford education and health services. 2.29 Tajikistan's health workers consider themselves grossly underpaid and employ coping strategies--some of which are illegal. A salary considered "fair" is 7.7 times higher than the actual average salary. In order to compensate for the inadequate salary, approximately 18 percent of health workers either seek employment outside the facility or informally charge patients. On average employees work 20 hours per week at second jobs, which may explain the high absentee rates seen in Tajikistan. Doctors and administrative staff are more likely to hold a second job. Besides working outside the facility, 46 percent of the health workers admit receiving informal payments (cash or in kind) from patients to supplement their income. Informal payments are most prevalent in Dushanbe, where, in the month preceding the survey. 72 percent of workers received them, 60 percent of doctors, and 50 percent of nurses. The average health worker is able to extract as much as 28 somoni per month from patients, with a peak of 124 somoni for doctors in Dushanbe. Trends in Education and Health Spending 2.30 Public spending on education and health has been on the rise but mostly driven by nominal increases in wages, salaries, and public investment. An increase in revenues generated by strong economic growth and strengthened tax administration enabled the government to increase education and health spending respectively to 18.8 and 6.3 percent of total budget expenditures in 2007 (from 13.5 and 4.8 percent in 2002) or 3.4 and 1.1 percent of GDP in 2007 (from 2.6 and 0.9 percent of GDP in 2002). Such increases are consistent with the government's priorities stated in the PRSP to improve the living standards of the population. The share of education in total public expenditure has continued to be the largest, while health spending ranked the fifth (after social protection, public administration, law enforcement, and defense). Per capita spending on education and health reached respectively US$17.7 and US$6.7 in 2007, compared to US$4.9 and US$1.7 in 2001. Additionally, the government, in consultation with the IMF and World Bank, [[what arrangement with the IMF stipulates raising wages?]] has used additional tax revenue to raise wages and salaries, especially for teachers and health workers. As a result, the wage bill increased by close to 260 percent in the education sector and by 340 percent17 in the health sector during 2002­06. These wage increases 17 The education sector wage bill increased by 47 percent in 2002, 25 percent in 2003, 30 percent in 2004, 63 percent in 2005, 50 percent in 2006, and about 24.8 percent in 2007. The health sector wage bill increased by 44 percent in 2002, 22 percent in 2003, 32 percent in 2004, and 96 percent in 2005 alone. 23 were accompanied by implementation of reforms in the sectors.18 Similarly, public investment has increased significantly in response to the need for rehabilitation and maintenance of education and health infrastructure that was destroyed during the civil war and had been neglected in early years after the war ended in 1997. In 2006, capital expenditure in the education and health sectors increased by several times compared to 2001. Table 2.2: Public Spending on Education and Health (excluding PIP financing) 2002 2003 2004 2005 2006 2007 Percent of GDP Education 2.6 2.4 2.7 3.5 3.4 3.4 Health 0.9 0.90 0.9 1.1 1.1 1.1 Percent of total government expenditure Education 13.5 12.3 13.1 15.8 19.3 22.1 Health 4.8 4.8 4.7 5.1 6.4 6.3 Spending per capita (US$) Education 4.9 5.5 8.2 11.8 14.3 17.7 Health 1.7 2.1 2.9 3.8 4.8 6.7 In 2000 constant prices (somoni) Education 78.4 101 134.2 193.9 n.a. Health 27.6 35.2 42.9 57 67.4 n.a. Source: Ministry of Finance, Republic of Tajikistan. 2.31 The 2004 Public Expenditure and Institutional Review (PEIR) highlighted key developments of public spending in education and health during the 1999­02 period. The PEIR found that intrasectoral allocations in the education sector conformed to international standards while allocation of budgetary resources in the health sector was in favor of hospital services. Wage bills in both sectors accounted for the largest share of local government budgets for education and health services. Importantly, wages were protected while other important expenditure categories such as medicines, utilities, repairs, and maintenance were not. Repair and maintenance as well as capital investment were low and inadequate; as a result, physical infrastructure sharply deteriorated. Per capita public spending in both sectors was inequitable across geographical areas. The rich quintile benefited from public spending on higher education and hospital care. The PEIR concluded that there was scope for improving allocative and technical efficiency in both sectors. 2.32 The consolidated health budget at the oblast level financed mostly hospital services, with primary care given much less attention than in previous years. In 2005, spending on hospital services was four times that of public spending on polyclinic services. It accounted for 71 percent of the total oblast health budget, while polyclinic services accounted for only 16 percent. Spending on other health care and health affair services comprised the residual. It is not clear why primary health care is also given short 18 In the education sector, since the government approved its reform plan in June 2004 the class size increases and changes in working practices have been implemented. Reform in the health sector includes implementation of the Health Financing Reform Strategy adopted in 2004. 24 shrift at the oblast level. There could be many reasons, including that the bias reflects the unwillingness of oblast officials to allocate funds to a system that is so underfunded that it would require significant investment to upgrade; that the Ministry of Finance would not allocate as many resources for lower-level health care facilities as compared to hospitals; or that hospitals have greater influence of health care officials as compared to primary care professionals. Alternatively, it could be that the oblasts lack the capacity to design an effective health care system and allocate resources based upon past formulas, especially given the changing environment. Economic Composition of Expenditure 2.33 Wages and salaries towered over other expenditures in the education and health budgets during 2002­06 but the trends in both sectors were opposite. In 2005, wages and salaries as a share of the local health spending increased to 50 percent from 40 percent in 2004. Wages as a share of education spending declined to 58 percent in 2006 (compared to 74 percent in 2002), due largely to a reduction in the wage bill at the local government level. Wages are a share of total health expenditure remained at 37 percent in 2006 (similar to 2002), due to an increased wage bill at the local government level, which largely was negated by high inflation rates in 2005­06. The local wage increase likely reflects government attempts to halt the departure of skilled personnel (doctors and paramedical personnel per 10,000 of population declined during this period.) According to the State Statistical Committee, despite faster-growing wages in the health sector compared to the education sector (3.3 times versus 2.9 times of the 2002 levels), the average wage in the health sector remained lower than in the education sector (41 and 121 somoni per month in 2006, respectively). Table 2.3: Public Expenditure on Education and Health by Economic Classification, 2002­06 (percent) Education Health FY 2002 FY 2006 FY 2002 FY 2006 Rep. Local Total Rep. Local Total Rep. Local Total Rep. Local Total Current expenditure 89 98 96 82 95 91 66 97 90 67 89 84 Labor compensation and contrib. 34 83 74 31 5 71 20 42 37 23 53 46 Goods and services 51 15 22 39 81 18 46 54 53 44 36 38 Subsidies and current transfers 5 4 5 12 14 2 0 0 0 0 0 0 Capital expenditure 11 2 4 18 5 9 34 3 10 33 11 16 Total 100 100 100 100 100 100 100 100 100 100 100 100 Source: Ministry of Finance, Republic of Tajikistan. 2.34 Spending on goods and services remained limited and declined in 2006 compared to 2002, accounting for around 18 percent of total education expenditure and 38 percent of health. Spending on goods and services (material inputs, utilities, training, and so forth) provides complementary inputs necessary to enhance the quality of education and health service delivery. However, an increase in the wage bill has crowded out other expenditure, and as a result, the government has been unable to allocate sufficient resources for material inputs. For example, expenditure on goods and services, 25 especially on medicine and food, fell to 38 percent of total health budget in 2006, compared to 53 percent in 2002. 2.35 However, the government has allocated more resources for repair and maintenance in both sectors during the past few years. Spending on goods and services in both sectors includes communal services (gas, electricity, water) and repair and maintenance. Tajikistan's Budget Law protects utilities expenditure such that its allocation can not be spent for other types of expenditure. Health expenditure on communal services and repair and maintenance fell slightly to 5 and 10 percent respectively in 2006, from 8 and 12 percent of total health expenditure at the local level in 2004. Education spending on communal services and on repair and maintenance increased during 2002­06, while other expenditure for goods and services declined. In the education sector, spending on learning materials (such as textbooks and visual aids) and food has declined and in health spending on food and drugs declined. Insufficiency of such inputs could lower the efficiency of education and health services. 2.36 Capital expenditure also increased in both sectors during 2002­06, mostly due to renovations and capital rehabilitation of schools and health care facilities. Several years of poor maintenance and inadequate investment coupled with the civil war deteriorated the education and health infrastructure. Recent improvements in the government's fiscal position have provided additional resources for increased capital expenditure. Capital investment increased in 2006 to 8 and 16 percent in education and health budgets respectively (reflecting, in particularly, increased health investments for Khatlon19 and Soghd oblasts). The World Bank Facility Survey conducted in 2005 indicated that more than 60 percent of the schools included in the survey reported undertaking some small-scale renovations between 1999 and 2004 and about 7 percent of the schools report undertaking major/capital rehabilitations. Local governments have benefited more from the capital increases. Table 2.4: Selected Expenditures on Goods and Services (percent of sector budget) 2002 2003 2004 2005 2006 Total G&S on education 100.00 100.00 100.00 100.00 100.00 Stationary, textbooks and other visual aids 11.50 11.20 8.40 5.40 3.50 Food 23.10 20.40 16.40 13.80 20.10 Communal services 12.60 17.70 17.10 19.20 23.10 Repair and maintenance 22.80 29.80 38.00 37.90 29.40 Total G&S on health 100.00 100.00 100.00 100.00 100.00 Medicine 18.60 15.30 14.90 14.70 Food 31.30 25.50 24.80 22.80 Fuel 2.40 2.20 2.40 2.40 Communal services 11.80 17.00 15.70 15.40 Repair and maintenance 19.40 25.20 27.50 30.30 Source: Ministry of Finance, Republic of Tajikistan. 19 Capital expenditure in Gorno-Badakhshan Autonomous Oblast (GBAO) was financed by Aga Khan Health Service, which rehabilitated Khorog Oblast General Hospital and three rayon hospitals (Murghab, Ishkashim, and Vanj). 26 Functional Composition of Expenditure 2.37 Spending on priority programs that help the poor in both sectors remained unchanged during 2002­06. In 2006, the share of spending on primary and secondary education in the total education budget increased slightly to 4 and 77 percent respectively as overall spending by local government increased. However, the share of polyclinics in total health budget remained unchanged, while its decrease in percent of GDP is visible. These results imply that intrasectoral resource reallocation in both sectors has not yet materialized. 2.38 Intrasectoral allocation patterns in the education sector did not change substantially throughout 2002­06. The government allocated over two thirds of the education budget to primary and secondary education (81 percent of total education spending in 2006, compared to 73 percent in 2002). However, spending on higher education both by the republican and local government budgets increased slightly during 2002­06 and comprised 43 and 2 percent respectively in 2006. Both increases were offset by a reduction in spending on preschool (at local government level) and other education affairs and services. Table 2.5: Education Spending by Source and Level of Education, 2002­06 (percent of total) 2002 2003 2004 2005 2006 Rep. Local Total Rep. Local Total Rep. Local Total Rep. Local Total Rep. Local Total Pre-school 0 6 5 0 5 4 0 5 4 0 5 4 0 5 4 education Secondary education 36 90 80 44 90 82 33 90 79 31 91 81 28 89 77 affairs and services Primary and secondary 17 85 73 23 86 75 13 85 72 11 87 74 10 85 70 education Secondary education-- 18 0 3 19 0 3 17 0 3 16 0 3 15 0 3 vocational and technical Secondary education not indicated in 2 5 4 2 4 4 2 4 4 3 4 4 4 4 4 other categories Higher education 36 1 8 38 1 8 37 2 8 41 2 9 43 2 10 affairs and services Education services not 2 0 1 3 0 1 6 0 1 4 0 1 3 0 1 definable by level Other education 26 3 7 15 3 5 25 3 7 23 3 6 26 3 8 affairs and services Total: 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Source: Ministry of Finance, Republic of Tajikistan. 27 2.39 Secondary professional education received the largest budget share per student in 2002­07, following by preschool. Both shares were double per student spending on general education. A high rate of return to specialized/vocational education compared to general education20 led to increased enrollment for secondary professional and higher professional education of 3.4 and 6 percent, respectively, during 2002­07.21 Table 2.6: Public Spending on Education by Level of Education (both central and local budgets) (somoni per student) 2001 2002 2003 2004 2005 2006 2007 Preschool 52 68 78 108 157 161 218 General education 30 40 52 72 113 141 205 Secondary professional education 79 125 151 182 240 348 348 Higher professional education 59 81 93 127 185 142 212 Source: Ministry of Finance, Republic of Tajikistan. 2.40 Intrasectoral allocation patterns in the health sector have not changed significantly. The majority of public health care funds--whether at the republican, oblast or rayon level--are devoted to hospital services (63 percent in 2006) with the residual amount financing primarily health care facilities. This occurs despite the high cost of running hospitals in comparison to, for instance, polyclinics and other health care facilities. In 2003, spending on hospital affairs and services in the republican budget declined and then recovered to previous levels. Reduced spending on hospital services in the republican budget was balanced by an increase in public health affairs and services in the republican budget. At the local government level, spending on hospital services remained unchanged in 2004­05 but increased slightly by about 5 percent of total sector expenditure in 2006. Spending on polyclinics, mostly at local government level, has also not changed as it declined in 2003 and subsequently increased in 2004­06. The increase was balanced by a reduction in other health affairs and services. Table 2.7: Public Spending on Health by Functional Classification (percent of total) 2002 2003 2004 2005 2006 2007* Rep. loc Tot Rep. loc Tot Rep. loc Tot Rep. loc Tot Rep. loc Tot Rep. loc Tot 5.1 Hospital affairs and services 44 71 65 40 71 63 43 71 63 47 71 65 54 66 63 78 65 67 5.2 Polyclinics affairs and services 1 16 13 1 15 11 1 16 11 1 16 13 1 17 13 2 24 20 5.3 Public affairs and services 3 6 5 5 6 6 4 6 5 4 7 5 5 7 7 7 7 7 5.4 Other health affairs and services 52 8 17 54 8 20 52 8 21 49 5 16 40 10 17 12 4 5 · Data for 2007 are only for operational expenditures. Source: Ministry of Finance, Republic of Tajikistan. Allocation of resource across regions 2.41 Allocation of public resource across regions is highly inequitable and remained unchanged during 2002­06, which reflect the continued use of norm-based budgeting. Public spending per capita still favors Gorno-Badakhshan Autonomous Oblast (GBAO), 20 According to the Welfare Dynamics and the State of Service Delivery in Education and Health Sector, The World Bank, 2007, a worker with specialized secondary and university education earns a premium of 19.7 percent in the private sector and 15 percent in the public sector. 21 State Statistical Committee. 28 the remote and mountainous region. For example, GBAO received 23 somoni per capita which was 2.6 times the average per capita health spending for all oblasts. According to recent PETS health surveys, resource allocation per capita was not correlated at the rayon level with income per capita. Hence, poorer rayons did not receive more money. In the general education sector, per student expenditure increased significantly in all regions during 2002­05. However, no significant changes were recorded in patterns of financing in the health and education sectors across different regions in Tajikistan throughout 2002­06. The average per student spending in GBAO region remains the highest, followed by Regions under Republican Subordination (RRS), Khatlon and Sughd oblast regions, and Dushanbe. Furthermore, budget planning and formulation is still based on the previous year's budget adjusted by inflation in both the education and health sectors. As the previous year budget was formulated based on norms such as the number of workloads per teacher and number of subjects taught (in education), and the number of doctors per population and number of hospital beds (in health), the norm-based practice continues to affect the current year budget in both sectors. 2.42 The annual Law on State Budget stipulates that unspent funds must be returned to the Treasury at the end of the year. The survey found that more than half of respondents at the rayon, CRH, and jamoat level (respectively 64, 64, and 53 percent) returned funds to the Treasury. With respect to payment of wage and salary based on approved workloads (stavkas), only 35 percent of rayon, 26 percent of CRH, and 25 percent of jamoats reported that funds from floating stavkas had to be returned. Unreturned funds were required in most cases to be used in the same facility. The majority of facilities citing such a requirement were found at the jamoat level (for 44 percent of rayon, 44 percent of CRH, and 79 percent of jamoats. D. PERFORMANCE OF THE EDUCATION AND HEALTH SYSTEM 2.43 This section assesses the performance of the education and health systems from both the demand and supply side. On the demand side, performance criteria include quality, access, and cost of services as perceived by end users. On the supply side, performance criteria derive from analysis of the production frontier to identify ways of improving technical efficiency by service providers. E. DEMAND SIDE 2.44 The health sector appears to be a low priority for all levels of government as indicated by the budget. Even when in previous years additional funds became available due to higher annual revenue (for example, at the rayon level), the health sector gained little. Therefore, performance-based budgeting that links budget resources with sector policy priorities should be created gradually in the health sector. A similar medium-term expenditure framework currently is being piloted in the education sector. 2.45 Budget management challenges in the education sector lead to leakages in the system and salary delays. Almost half of secondary schools reported salary delays in 2006, which is a particularly serious issue in rural areas. The weak link between jamoats and schools ensues from absence of proper record keeping and reporting, including 29 unrecorded allocation of goods and services among schools, non-matching procurement records across jamoats and schools, and others. However, the rate of inconsistencies in reporting on allocations and deliveries varies with the budget category. 2.46 Despite increasing financing in the education sector, subsector budgets are sometimes left underexecuted. The PETS study verified that about 72 percent of secondary schools spent less than the revised budget. The balance is carried forward as "free reserves," reported by almost 37 percent of schools at the beginning of 2006. Free reserves increase the budget flexibility of rayons, which is nevertheless offset by the trend of rising budgets, giving rayons scope to assume more present than future spending 2.47 Some key players in the health system have little or no discretionary authority. The PETS health findings confirm that the jamoat officials responsible for polyclinics and rural facilities primarily act as disbursing agents for the rayons. Heads of health facilities do not control their operating resources, whether personnel or medical inputs (medicine or medical supplies). Most do not even know their budget allocation, and receive their resources mainly in kind. As a result, they are unable direct budgets to identified priorities at the local level. Hence, it is difficult to infer any output-based activity at the local government level, and it is unclear how desired outcomes at the local and republican tiers are achieved. 2.48 Misallocation of limited public resources for health care exacerbates key health indicators. Limited resources are mostly spent on wage and salaries of health personnel. Little funding remains for other inputs such as medicine and medical equipment that would improve the impact of service delivery on the general health of the population, particularly morbidity and mortality indicators that are regarded as health sector reform outcomes. 2.49 Some education Table 2.8: Education--Intermediate Indicators output indicators recently have improved. Tajikistan's 2002 2003 2004 2005 2006 2007 Number of school rapidly growing population establishments 3,729 3,745 3,775 3,807 3,830 3,810 (about 2 percent) since the Number of end of the civil war has students increased the demand for (thousands) 1,637 1,660 1,674 1,682 1,711 1,713 education services. The Number of teachers 101 101.5 99.1 98.5 99.9 99.4 education budget has been Student/Teacher increased to increase the ratio 16.2 16.4 16.9 17.1 17.3 17 number of institutions and Source: State Statistical Committee, Tajikistan. classrooms and subsequently the number of teachers, with the exception of preschool institutions, especially in rural areas. Average student-teacher ratios in general education (from 1:13 in 1990 to 1:17 in 2007) have increased in spite of new hires and expanded construction activity. Teaching loads have increased from 14 to 16 hours per week for teachers at higher levels and from 16 to 18 hours at primary level. 30 2.50 Intermediate health outputs have declined, reflecting government efforts to adjust inputs to the availability of resources. The number of doctors, paramedical personnel, and hospital beds declined in 2003, and subsequently increased in 2004­05 due to an increase in hospital institutions and medical facilities. The number of doctors per 10,000 of inhabitants declined during 2003­04 before increased slightly in 2005. The number of hospital beds per 10,000 of inhabitants declined from 62.1 in 2002 to 57.1 in 2006. In 2007 this trend continued because of ongoing health sector reforms and closures of medical houses, along with a shift from hospital service to more-effective primary health care. Table 2.9: Health--Intermediate Indicators 2002 2003 2004 2005 2006 2007 Number of doctors (thousands)* 13.1 12.7 13 13.3 13.3 13.4 Per 10,000 of population 20.1 19.1 19.1 19.4 18.8 Number of paramedical personnel (thousands) 30.3 28.6 28.7 28.9 29.4 29.6 Per 10,000 of population 46.5 43.1 42.4 42.2 41.7 Number of hospital institutions 449 452 453 454 449 450 Number of hospital beds (thousands) 40.4 40.2 40.6 40.6 40.3 38.8 Hospital beds per 10,000 of population 62.1 60.6 59.9 58.6 57.1 Number of medical houses 1709 1714 1710 1712 1712 1692 Source: State Statistical Committee, Tajikistan. Note: Number of paramedical personnel includes dentists. 2.51 Reforms and strong Table 2.10: Education - Gross Enrollment Rates economic growth recently have 1990 1995 2000 2005 begun to reverse the decline in Gross enrollment rates (%) education indicators. Some Primary education 91.0 92.6 98.2 101.2 indicators have been stabilized Basic secondary 102.1 80.6 73.5 81.8 during the past few years; Full secondary 22.8 20.4 14.0 17.3 however, they remain below levels Source: EDSTAT, Education Profile, the World Bank; Ministry of before independence. Declining Education. enrollment rates in basic education seem to have been reversed lately as the number of school-age children increases. Overall enrollment rates have increased at all levels of education. However, enrollment among girls has dropped. 2.52 Some health indicators have improved compared to 2002. According to the recent poverty update, infant mortality decreased by 6.4 percent between 2002 and 2004 and maternal mortality decreased by 24.1 percent during the same period. However, there has been an increase of infectious and communicable diseases. Tuberculosis rates have been stabilized through the implementation of DOT22 and the incidence of measles has declined due to the national immunization campaign supported by UNICEF. 22 DOT is a national program to improve patient registration and enhance access to preventive treatment. 31 Quality of Services 2.53 Uneven funding for rural and urban health care makes providing quality services to patients difficult. Most resources are allocated to hospitals, then to urban polyclinics, with little left over for rural facilities. Low remuneration for rural health employees may explain the difficulty of getting doctors to work in these areas. Rural primary health care facilities also receive less funding and fewer in-kind allocations (such as drugs and medicines), which has a negative impact on service quality. 2.54 The quality of education has been affected by inadequate infrastructure and lack of qualified teachers, especially in rural areas. Broken windows, lack of electricity and heating, and lack of water connections make for a poor learning environment, and parents are concerned about their children's safety. Further school closures, especially in the winter due to lack of heating, affects learning outcomes. Based on a facility survey conducted by the World Bank in 2005, 60 percent of urban schools rely on electricity as the main source of heating, but less than 20 percent of rural schools rely on electricity due to shortages.23 About 70 percent of rural schools report no electricity in winter for 7 hours per day or more as opposed to 45 percent of urban schools. However, lack of electricity is more often cited as a reason for urban school closures (65 percent versus 30 percent). (Rural schools heat with wood and do not have to close.) Low salaries of teachers and frequent delays in salary payments make it difficult to attract and retain qualified teachers. About 9 percent of teachers revealed that they never been paid on time (14.5 percent in rural schools, 4.5 percent in urban schools), and many have to hold second job to supplement their income, thus lowering their attendance and performance. 2.55 Health care quality has been reduced by shortages or absenteeism of doctors, poor availability of drugs, lack of medical equipment, lack of water and electricity, and other conditions necessary for targeted public service delivery. According to the 2006 poverty update, 15 percent of doctors in the top quintile were absent, compared to 2 percent of doctor in bottom quintile. Only 50 percent of health facilities had drugs in stock. About 50 percent of facilities reported 7 hours or more shortage of electricity in the winter season as opposed to 35 percent of outpatient facilities and less than 10 percent of inpatient facilities. These factors lessen the quality of services and discourage patients from seeking care. Access to Services and Accountability 2.56 There is a disconnect between policy and budgetary institutions, which undermines accountability and provides opportunities for the diversion of resources. Despite being responsible for policy, sector ministries have limited involvement in budget formulation either at republican or local levels. Because central government resources are channeled to either RFDs or CRHs, it is difficult to hold the Ministry of Education or the Ministry of Health accountable for outcomes. In addition, local government discretion in managing budgets weakens transparency and increases the possibility for diversion of funds. 23 In contrast, about 40 percent of rural schools identify wood as their main source of heating. 32 2.57 There is no evidence that physical factors such as location and operating hours hinder access to education and health services, and extensive networks of facilities were inherited from the Soviet period. (Access to service can be affected by three factors: physical, economic, and cultural.) However, access to specialists may be limited as a district hospital may be too far. 2.58 However, economic factors such as fees can impeded access to services, particularly among the poor. According to the Constitution amended in 2003, the Tajik population is entitled to basic education and health care services free of charge. However, several years of inadequate funding of schools and health facilities has led to a practice of informal payment for drugs, material inputs, and to service providers. The Tajikistan Living Standards Survey (TLSS) indicated that inequity of access to health care between the top and the bottom quintile has widened.24 The utilization rate for health service is much lower among the poor quintile compared to the richest quintile. 2.59 Deterioration of the physical infrastructure in education poses a serious challenge for the sector. Deferred maintenance and damage to infrastructure during the civil war has led to a shortage of schools and classrooms. Only 14 percent of school principals reported that a capital renovation of existing school infrastructure was undertaken. According to a recent PETS education survey, only 18 percent of schools have added classrooms and/or premises. In, addition a growing school-age population has led to more intensive use of school premises throughout the country. 2.60 Many health facilities do not have basic infrastructure necessary to provide services. For example, an estimated 53 percent of facilities reported having access to water, but only 42 percent of rural primary care facilities had access to water, either piped or delivered. On average, over 90 percent of facilities had access to electricity by region. However, only 23 percent of all health facilities reported having access to a radio and/or telephone, and most of these were in urban areas. Finally, only 19 percent of health facilities reported having at least one vehicle available for transporting patients in emergencies or for collecting important inputs such as drugs and vaccines.25 2.61 Because nonwage expenditures in education and health facilities are paid mostly in kind by the CRH and/or RFDs, budget allocations remain effectively at the discretionary of local government. Many health facilities report receiving very limited funding. Whether this is because rayons keeps most resources for themselves or that goods are mostly diverted for private gain of local government officials needs further investigation. F. SUPPLY SIDE 2.62 Tajikistan ranks among the lowest of all International Development Association (IDA) countries in per capita spending on education and health, and the lowest of all among CIS countries. Given its narrow economic base, the country cannot afford 24 A more recent TLSS study and findings will be available by the end of 2008. 25 As a result, an estimated 33 percent of facilities reported not receiving drugs (budget or actual pharmaceuticals) from government resources (excluding GBAO). 33 ambitious and expensive social development programs. Therefore, spending wisely and efficiently becomes critical (see Annex 1 for technical details). 2.63 Analysis of efficiency on the supply side involves benchmarking expenditures (supply side), both capital and recurrent, against (i) other low-income IDA countries (national level) and (ii) among the agents themselves (health and education service providers), comparing them to the average-performing agents at the sub-national level (see Farrell, 1957). 2.64 Efficiency measures in both sectors are based on two-step analysis: input-access (for example, expenditures--immunization coverage in health and expenditures-- gross enrollment rates in education) and access-output (for example, immunization coverage--infant mortality in health and gross enrollment rates--literacy rate in education). The purpose of this two-step approach is to track the impact of initial resources (inputs) on results (outputs) via intermediate (access) step. This captures the effect of other intermediate factors since input-output may not strongly correlate and is also dependent on other factors. Efficiency could be either technical or allocative. Analysis of technical efficiency (of spending as an input) can lead to better input or output efficiency. Analysis of allocative efficiency can reveal a suboptimal input combination given input prices and corresponding productivities to reach given output level (e.g. allocation of funds between hospital care to primary health care and immunization to increase output efficiency). Due to data limitations only a technical efficiency analysis was conducted for Tajikistan. The key messages for policy makers could be grouped as follows. Health Sector: National and Sub-national Level 2.65 Despite significant progress in 2000­05, public spending remains sluggish and inefficient. Efficiency could be achieved through further policy framework improvement, better fiscal policy, more equitable public resource use, and broader public administration reform. Tajikistan spends 30 percent more than the most efficient IDA countries to achieve same immunization rates (see S. Herrera). [[need date and full citation in footnote]] This mean that with better efficiency through the measures above, and with the existing expenditures on health (immunization), Tajikistan could improve mortality rates by 4 percent. The key to efficiency improvements is a strong link between development program objectives and public resource allocations. 2.66 Results at the sub-national level are mixed. In particular, input efficiency (or public spending across rayons) is uneven. Indicators of efficiency also changed unevenly but input efficiency (spending) remains relatively equal for all sub-national service providers while output efficiency (mortality rate) has a wider range. In addition, rayons seem to heavily overspend their available resources to achieve marginal changes in output efficiency performance. Additional spending has not generated improvements in access, showing that systemic shortcomings need to be addressed in order to improve health outcomes. 34 Education Sector: National and Sub-national Level 2.67 Tajikistan still enjoys high secondary enrollment rates given the total spending in education. However, these historically high rates are declining. As with health care, countries that lead in spending do not lead in input and output efficiency. Given its limited resource base, Tajikistan needs to find ways to improve literacy rates through improving efficiency. At the sub-national level, education spending in Tajikistan is the largest in the total social expenditure portfolio. However, output-access indicators are lacking and spending efficiency analysis is limited to a very few input-access indicators. As opposed to health care, increased spending in education seems to improve enrollment and teacher-per-student ratio, but national spending is 70 percent more than that of most efficient rayons to get the same enrollment rates. This means that even some existing rayons (intra-rayon comparison) are currently much more efficient at reaching enrollment rates, given available funding, than the national average. The student-teacher ratio is 38 percent. This may be because most spending goes to secondary education and almost all spending is financed through the local budgets. In short, public spending in education does not increased enrollment enough, but it does affect the student-teacher ratio more significantly. G. CONCLUSIONS 2.68 Tajikistan can make substantial progress toward achieving MDGs by addressing the allocative and operational efficiency concerns raised in this Technical Background Paper. Allocative efficiency gains can be achieved by aligning resources with requirements in subsectors. This Technical Background Paper shows that the past allocation history does not reflect the government reform program discussed in the NDS and PRSP sector strategies. This is especially clear from the allocation patterns in recurrent and capital expenditure over the last few years. Basically Tajikistan can not improve outcomes by just changing the allocation, it also needs reforms in the systems of education and health which transform the resources into outputs. The demands on technical efficiency requirement are both time consuming and require system reforms that require additional technical capacity. 2.69 Key reforms are ongoing in both education and health sectors as part of the government's comprehensive sector development strategies, but implementation has been slow. Reforms have included the piloting and introduction of per capita financing, which focused on primary education and health service delivery, and the introduction of basic resource planning and management tools that link policy and budgeting. Although resources for implementation are scarce, the government and donors have developed good coordination mechanisms (such as FTI donor coordination). It is important to maintain the pace of these reforms, and to increase focus on institutional aspects and on linking policy making and budgets, which will improve expenditure efficiency and outcomes. 2.70 Both sectors show strong and commendable increase in spending as a percentage of GDP and total budgets, but these are mostly static increases driven by recurrent costs (salaries) and some public investment. Sector budget increases have 35 resulted from overall economic growth and other ad hoc spending increases, rather that well-planned and well-costed interventions. This shows continued reliance on norm- based budgeting. Expenditures continue to be heavily tilted toward and implemented through local budgets, giving them great discretion without accountability for outcomes that are in fact deteriorating. There is also a great degree of variation of public expenditures, both interregional (geographically) and intraregional. This is virtually unchanged since 2002. Expenditures overall in both sectors continue to be dominated by the wage bill with only marginal spending increases on capital expenditures, maintenance, and goods and services. 2.71 Tajikistan inherited a complex health infrastructure that may be beyond the country's means to maintain. Accordingly, there are three serious challenges to the country's institutional rehabilitation. First, the health and education sectors continue to be severely underfunded, leaving few resources public financing of reforms in both sectors. Second, fewer resources reach frontline providers, which leads to nonperformance of key health care facilities and educational institutions that are aligned with both national and local priorities, and with regional and international development indicators outlined in other strategic documents, including the MDGs. Third, the local government officials' discretion for allocation of scarce resources is greater than needed for addressing fiduciary and accountability concerns. On the other hand, some key interlocutors clearly do not have a say in budget allocation and formulation. 2.72 In the education sector, Tajikistan needs to reform institutions and resource allocation mechanisms. Pilots of per capita spending help address resource allocation inefficiency but are rather slow in implementation. In addition, education services could be delivered more efficiently if the operational aspects of education service delivery can be addressed. 2.73 Health and education resource allocations vary enormously by oblast and rayon even when controlling for type and size of an institution (or a facility). This shows the high degree of discretionary power at the oblast and rayon level. It is also indicative of the inability of sectoral ministries to apply norms, and thus the ability of the central government to control the education and health care systems. 2.74 Control over funds is very poor and it is likely that there are leakages. There are many avenues through which inefficiencies in procurement and distribution of public resources in the system can occur. Government monitoring is weak, which is exacerbated by the fact that frontline providers recipients of funding or in-kind inputs do not have approved budgets. Since most facilities receive inputs (such as medicines and food) in kind rather than in cash, monitoring the use of these resources is difficult. The Technical Background Papers 4 and 5 based on the PETS in the health and education sectors discuss these issues in more detail. 2.75 Challenges in the health and education sectors also result from weakened institutional structures and a significant lack of public resources. Even limited reforms in health and education have had positive effects on institutional performance and efficiency gains in public resource allocations. However, if a large part of the system is 36 barely functioning, options may include downsizing certain facilities and institutions or retargeting large donor-funded baskets of resources (for example, as part of future SWAP activity). However, if institutions are not functioning effectively and accountably, then it is not clear to what extent the additional resources will be used to provide more or better services. 2.76 Analysis of the supply side in Tajikistan's health and education expenditures shows that outcomes (mortality and enrollment rates) can be greatly improved through efficiency gains. Even with current levels of funding, Tajikistan can attain better outcomes by improving input or output efficiency--that is, by using less inputs to obtain similar outputs (on the national level), or by producing more outputs with given inputs. While commendable increases in health and education spending are taking place, Tajikistan has limited capacity for creating additional fiscal space. Therefore, efforts to improve efficiency will go a long way toward improving outcomes in the short to medium term. 37 Annex 2.1: Analysis of the Efficiency of the Education and Health Systems--Supply Side26 This section will employ benchmarking techniques that involve the estimation of a frontier to evaluate the efficiency of service providers in the education and health sectors.27 Performance will be evaluated at national and subnational levels. The performance of the education and health systems will be analyzed using an efficiency frontier based on Free Disposable Hull (FDH) and Data Envelopment Analysis (DEA). A service provider is considered efficient when it operates on the estimated frontier and inefficient if it operates outside (below or above) the frontier. Such analysis can show policy makers how far the system is from the benchmarked countries, given available resources for the sector. Spending Efficiency in Health and Education: Methodological Remarks Farrell (1957)28 identified two types of efficiency of productive agents. The first type is technical efficiency, in which agents are deemed inefficient if they use more inputs than technically required to obtain a given level of output. The second type is allocative efficiency, in which agents are considered inefficient if they use a suboptimal input combination given the input prices and their marginal productivities. Due to limited data, this analysis covers only technical efficiency. Tajikistan has some of the lowest per capita spending on health and education among IDA countries. It spent US$PPP 57 on average during 1999­2005. Among IDA countries, it only outspent Sierra Leone, Burundi, Madagascar, the Republic of Congo, and Niger--the poorest Sub-Saharan African countries. Its spending was the lowest among other CIS countries. Tajikistan is one of the poorest IDA countries. Per capita GDP of US$PPP 1,173 (2005), equivalent to 51 percent of the 1990 level, was slightly higher than in poor Sub- Saharan African countries that form the lowest decile in IDA, and ranks with Congo, Nigeria, Kenya, and Benin.29 Given a narrow economic base, the government cannot afford ambitious and expensive social development programs. Thus, spending wisely and efficiently is critical. 26 The analysis and write up in this annex was carried out by Utkir Umarov (ECSPE). 27 This annex is based on the techniques employed by Herrera and Pang (2005) for an inter country analysis. 28 M.J. Farrell (1957), "The Measurement of Productive Efficiency," Journal of the Royal Statistical Society, Series A (General), Part III. 29 2000­05 average per capita GDP (US$ PPP 985) was equivalent to almost 70 percent of IDA median, but 28 percent of the richest (top decile) IDA country average. 38 Figure 2A.1.1: Per Capita Spending in Health and Education in IDA, 1999­2005 Source: own calculations. IDA countries' spending on health is proportional to the level of per capita income. During 2000­05, IDA countries spent the equivalent of 5 percent of income on a per capita basis on health. The spending range was 2­10 percent, and overall it was highly correlated with per capita GDP (correlation coefficient was 0.61). At 3 percent of per capita GDP, Tajikistan spent less on health than the average IDA country in 2000­05. Similar to health, per capita spending on education is proportional to per capita income in IDA countries in absolute terms. On average, IDA countries spend 3.6 percent of GDP-worth funds (2000­05 average) on education; the range varies from 1.8 to 9 percent of GDP. As in the health sector, Tajikistan underspends on education, averaging 2.45 percent of GDP in 2000­05, which is 1.1 percentage points lower than the IDA average. Efficiency Measurement: Statistical Issues and Methodological Remarks Two techniques of efficiency analysis are used. First, the Free Disposable Hull (FDH) ranks the efficiency of service delivery agents (rayons) by comparing each individual performance with a production possibility frontier that only the agents with the highest level of output/outcome for a given level of input can meet. Conversely, one can determine the lowest level of input necessary to attain a given level of output/outcome. This allows identification of inefficient service delivery agents both in terms of input efficiency and in terms of output/outcome efficiency. Second, the Data Envelopment Analysis (DEA) is used to evaluate the efficiency of a number of producers. A typical statistical approach is characterized as a central tendency approach and it evaluates service delivery agents relative to an average performer. In contrast, DEA is an extreme point method and compares each service delivery agent with only the "best" performers. The efficiency will be measured at the national and sub-national level. National efficiency will be benchmarked against IDA countries; sub-national efficiency will be benchmarked against 65 districts (rayons). The sub-national system is responsible for 39 service provisioning in Tajikistan. The analysis is based on 2000­05 data for the national level and 2002­05 for the sub-national level. Efficiency is measured with a two-layer (or two-step) analysis: input-access and access- output. In reality, input and output indicators may not be strongly correlated, as outputs/outcomes depend on a number of other factors. For example, reducing infant mortality rate is not only dependant on health financing but also on the quality of services, as well as other factors not related to health (change in levels of incomes, change in nutrition patterns, and so forth). The low quality of services at an intermediate stage may have a greater impact on mortality. This briefly describes the rationale for two- layer efficiency analysis: the objective is to track the impact of initial resources (input) on results (output) through intermediate (access) steps. The analysis at national and sub-national levels involves various statistics sources. World Development Indicators (WDI) and other databases of the international institutions are used for the analysis at national level; national health and education statistical data is used for the analysis at the sub-national level. In an ideal situation, one would use the same set of indicators for both national and sub-national efficiency analysis. However, current education and health statistics are not comprehensive enough, so the analysis will employ standard input and output indicators at national level, while for sub-national analysis we use basic indicators (available for the time being) that reflect milestones in both sectors. Data comparability across rayons/countries has been taken into consideration. This mostly applies to price disparity among rayons/countries. Prices vary among rayons, given difficult landscape and physical constraints in internal trade. Also, teachers in mountainous rayons are eligible for pay adjustments (supplements). Therefore, original budget (and copayments in health care) data are adjusted for the above factors. All amounts are given in terms of Dushanbe prices (remuneration supplements) at the sub- national level analysis. Nationally, we use spending amounts in terms of U.S. dollar purchasing power parity. The Free Disposable Hull (FDH) framework identifies inefficient service delivery agents both in terms of input efficiency and in terms of output/outcome efficiency (see Figure 2A.1.2). Data Envelopment Analysis (DEA) is commonly used to evaluate the efficiency of a number of producers. A typical statistical approach evaluates service delivery agents relative to an average performer. In contrast, DEA is an extreme point method and compares each service delivery agent with only the "best" performers. 40 Figure 2A.1.2: Free Disposable Hull (FDH) and Data Envelopment Analysis )tup (Out Y S e rv ic e D e liv e ry A g e n ts ­ D e c is io n M a k in g U n its (D M U ) X ( I n p u t ) D E A F ro n tie r ­ C o n s ta n t R e tu rn o n S c a le D E A F ro n tie r ­ V a ria b le R e tu rn o n S c a le F D H F ro n tie r DEA may employ variable and constant returns to scale. Since this analysis uses a constant input value of 65 rayons/towns and data for all IDA countries, it avoids the complications caused by allowing different returns to scale. Returns to scale refers to increasing or decreasing efficiency based on size. For example, a rayon/country can achieve certain economies of scale by increasing access to primary health care with less per-unit resources it needs now: this is an example of increasing returns to scale (IRS). On the other hand, increasing per-unit inputs may lead to a lower output-to-input rate: this illustrates decreasing returns to scale (DRS). Combining the two extreme ranges would necessitate variable returns to scale (VRS). Constant Returns to Scale (CRS) means that rayons are able to linearly scale the inputs and outputs without increasing or decreasing efficiency. This is a significant assumption. The assumption of CRS may be valid over limited ranges but its use must be justified. As an aside, CRS tends to lower the efficiency scores while VRS tends to raise efficiency scores. The recent analysis is based on variable returns to scale. Results of the analysis are given in the form of input and output efficiency scores (from 0 to 1, where 1 is the highest rating). For example, an input efficiency score of 0.5 indicates that a specific service delivery agent (rayon) uses inputs in excess of the most efficient agent to achieve the same output level. An output efficiency score of 0.5 indicates that inefficient agent attains 50 percent of the output obtained by the most efficient agent with the same input intake. 41 Spending Efficiency in Health Efficiency Analysis at National Level Efficiency assessment of health spending in IDA countries is based on analysis of two sets of data: Data Set 1: · Input: health expenditures (public) · Access: immunization (measles and cough-diphtheria-tetanus ) coverage · Output: infant mortality Data Set 2: · Input: health expenditures, policy performance, adult literacy rate · Access: immunization (measles and cough-diphtheria-tetanus ) coverage · Output: infant mortality and life expectancy Detailed results of simulations with using the above data is given in Annex 3. Improving the policy Figure 2A.1.3: Changes in Policies (X) and Public framework is a significant Spending on Health (Y) in IDA countries factor in efficient provision of public service delivery in countries that progressed in reforms. The main drivers of improving public spending efficiency are fiscal policy, equity of public resource use, policies of building human capacity, social protection and labor, quality of budget and financial management, efficiency of revenue Source: own calculations. mobilization, efficiency of public administration and transparency, and accountability and corruption in the public sector. These items have a significant impact on the quality of public service delivery. The World Bank's Country Policy and Institutional Performance Assessment (CPIA) can be used to quantify relative changes in policies across countries. During 2000­05, Georgia, Armenia, and Angola led most reformist group among IDA, while Tajikistan was also in the top decile of reformists (but lagged far behind the top three). Analysis of 62 IDA countries (for which data was available) shows that the following: (i) in 24 countries improving policy brought increased public spending (for example, Armenia); (ii) in 17 countries where progress on reforms was made, public spending on health either changed negligibly (Tajikistan) or declined (Georgia); and (iii) in other 21 countries policies deteriorated (mostly Sub- 42 Saharan Africa), but despite that in 12 countries there still was an increase in health spending (Figure 2A.1.3). Tajikistan has made significant progress over last five years, but the "static" picture shows it still has one of the poorest policy performing indicators. Tajikistan needs to increase greatly transparency and the quality of public finance management in a broad context of the public administration reform strategy. In the context of input-output analysis it means changing the budget process and ensuring linkages between development program objectives and public resource allocation. IDA countries that spend more on health than peer countries show lower efficiency. The analysis shows that, except Sri Lanka, the top 10 input and output efficient countries were not listed as the top 10 spenders on health. Despite the fact that most countries had fairly high input-access efficiency rates (closer to the efficiency frontier), their performance was uneven (see the first chart in Figure 2A.1.4). One main reason for input inefficiency might be the earmarking of a fixed amount of funding for immunization: even if spending on health increases, this amount either does not change or changes marginally. Figure 2A.1.4: Input-Access Frontier (DEA) Data Set 1: Input (HTL) ­ Access (IMN) 120 100 LKA MNG MDA HNDARM TZA GUY NIC TJKERI WSM 80 RWA ZMBBGD GHA BEN MWI GEO CPV BDI LSO YEMCOMBFA KENMOZ NPL UGA T MRT U GIN KHM PT MLIETH TGO 60 MDG GNB CMR CIV SEN AGO OU SDN HTI NER LAO COG TCD 40 CAF NGA 20 - - 50 100 150 200 250 300 350 INPUT Data Set 2: Input (HTL+PLC+LRA) ­ Access (IMN) 120 100 LKA MNG MDA HNDARM GUY TZA NIC WSM TJK ERI ZMB 80 BGDRWA GHA GEO BEN MWI CPV BDI LSO COM NPL YEM KEN MOZ UGA BFA T U GIN MRT KHM PT TGO MLI ETH 60 GNB CMR CIVMDG SEN AGO OU SDN HTI LAONER COG TCD 40 CAF NGA 20 - - 5 10 15 20 25 30 INPUT Source: own calculations. *see abbreviations below. 43 Tajikistan's input-access (funds-to-immunization) performance was fairly strong among IDA countries. However, Tajikistan's spent 30 percent more funds to achieve the existing level of output compared to other IDA countries, or with existing funds it could increase the level of output by 4 percentage points (DEA input and output efficiency). Particularly, higher immunization rates in Tajikistan are linked with extensive door support in this area. Policy performance has been critical to ensuring better access to health services in IDA countries. If policy performance indicators are factored in, the correlation between input and access becomes very strong, 0.7. A regression analysis on Data Set 2 produces the following equation linking inputs (In) with Access (Ac): Ac = 3.439108691*In + 7.759720366 Correlation rate = 0.6879 Since, In = ln(HEX) + PLC2 + ln(LRA), we have Ac = 3.439108691 * ln(HEX) + 3.439108691 * PLC2 + 3.439108691 * ln(LRA) + 7.759720366 where HEX is total spending on health, PLC is average CPIA ratings for fiscal policy, equity is public resource use, policies of building human capacity, social protection and labor, quality of budget and financial management, efficiency of revenue mobilization, efficiency of public administration and transparency, accountability, and corruption in public sector, and LRA is adult literacy rate. The equation above demonstrates that the better is policy performance and the higher is the adult literacy rate, the higher is the impact of spending on improving access to health services. 44 Output-access is less correlated Figure 2A.1.5: Access-Output Frontier (DEA) (than input-access) in IDA countries. One of the main reason Access (IMN)-Output (INM) is the relatively contrasting picture 80 70 LKA within IDA countries: life 60 expectancy is high and infant 50 T U mortality is low in transition PT 40 WSM MDA OU CPV ARM 30 NIC HND economies and other countries GEO MNG 20 GUY COM ERI SDN NPL BGD GHA such as Sri Lanka (Figure 2A.1.5), LAO TJK COG HTI MDG SEN TGO MRT YEMUGA KEN LSO CMR TZA 10 BEN NGA BFA ZMB CAF ETH KHMGIN TCD CIV MOZBDI MWI GNB MLI RWA NER AGO while the situation is much - - 20 40 60 80 100 120 INPUT different in Sub-Saharan Africa. Other reasons for that the contrast Access (IMN) ­ Output (INM+LEB) include the following: (i) output 8 0 LKA or outcome indicators are 7 0 6 0 dependent on multidimensional 5 0 factors that are related both to T U WSM M D A PT 4 0 C PV AR M N IC H N D OU access to medical services, such as 3 0 GEO M N G GU Y C OM ER I 2 0 SD N N PL BGD level of income, livings standards, LAO GH A T JK C OG H TI MSEN D G TGO M R T YEM KENLSO C M R U GA TZA BEN N GA KH MGIN BFA ZM B C AF TC D C IV M OZBD I M WI GN B M ETH LI R WA 1 0 N ER AGO environment, and so forth; (ii) - - 2 0 4 0 6 0 8 0 1 0 0 1 2 0 transition economies inherited IN P U T good social infrastructure facilities and high quality of Source: own calculations. *See abbreviations below education and medical services while in most African countries infrastructure is still being built; and (iii) life expectancy and infant mortality and other indicators are improved in longer time horizons, while inputs' affect on access might be observed in a relatively short period of time. Nevertheless, the overall trend shows significant correlation between immunization and infant mortality and life expectancy in IDA countries. Correlation between aggregate access and output indicators (Data Set 2) is significant at 0.5792. A regression analysis on Data Set 2 for outputs and access produces the following equation linking Access (Ac) with Outputs (Q): Q = 0.4139148715*Ac - 8.974655163 Since, Ac = (MSL+DPT)/2, we have Q = 0.2070 * (MSL+DPT) - 8.974655163 or where MSL and DPT are Measles and Cough-Diphtheria-Tetanus immunization coverage percentage, INM is infant mortality rate, and LEB is life expectancy at birth. The equation above claims immunization reduces (inversely correlated) infant mortality over a longer period of time and increases life expectancy at birth. 45 Efficiency Analysis at Sub-national Level Improvements in medical statistics are observed in Tajikistan in recent years. Tajikistan has been improving medical statistics that would help to access efficiency of public spending in this sector. For instance, the Center for Medical Statistics and Information under the Ministry of Health has been publishing health statistics yearbooks with a breakdown at rayon level. These statistics could be an input for spending efficiency and monitoring developments in the sector. 1. For a sub-national-level health spending efficiency analysis, we have chosen the following set four input-access-output indicators: Data Set 1: · Input: total public spending for health: current + capital · Access: immunization coverage · Output: infant mortality Data Set 2: · Input: total public spending for health: current + capital · Access: number of physicians, midwives and nurses per population · Output: infant and maternal mortality Data Set 3: · Input: total spending for health: public + private · Access: births attended by skilled health staff · Output: maternal mortality Data Set 4: · Input: number of beds and physicians · Access: initial medical application and bed occupation · Output: life expectancy at birth The reasons for choosing these data are (i) that they reflect main developments in the sector in recent years; and (ii) data constraints. 46 The spread of public spending Table 2A.1.1: Recipients of Budget Allocations, in Tajikistan across rayons is Top and Bottom Deciles not even. The difference between highest and lowest per TOP 10 Recipients BOTTOM 10 Recipients Rayon Spending* Rayon Spending capita spending is almost nine- SRS 11.62 RDK 1.31 fold, with a standard deviation DNG 7.50 BGR 1.31 (1.72) equivalent to two thirds of KHJ 6.03 MST 1.44 the national average (see Table SRB 5.74 HSR 1.75 2A 1.1 for details). During RSN 5.61 VRZ 1.80 2002­05 the top decile almost TVL 5.59 QAB 1.80 CKL 5.36 VHD 1.81 did not change--eight out of ten MRG 5.02 BHT 1.87 rayons remained in the top DRV 4.96 KLH 1.87 decile. The lowest decile VNC 4.84 GNC 1.92 significantly changed--only Source: own calculations. Mastchoh, Gissar, and Gonchi * 2002-2005 average per capita spending. remained in lowest decile of recipients. Changes in indicators of efficiency for all data sets during 2002­05 were uneven. Figure 2A.1.6 illustrates the uneven shifts in efficiency frontiers (single input-two output model) in 2005 versus 2002. "Input-Access" frontiers shifted towards improvement, except Data Set 4, while "Access-Output" frontier shift indicates improvement, except Data Set 3. More specifically, according to analysis of the data sets, spending on health facilitated the increase in number of births attended by skilled medical staff; also rationalization of assets (beds) in some rayons led to more accessible primary health care services. On another hand, fast-growing spending on health did not bring much improvement in immunization coverage: the spread was uneven across rayons. Similarly, more spending on health did not lead to increasing numbers of medical staff--doctors, midwives, and nurses--per population. The overall impact of access on outcomes was also limited. A marginal increase in immunization led to declining infant mortality. However, the increasing share of births attended by medical staff indicates limited correlation with maternal mortality. Similarly, improving efficiency and access to primary health care was not directly correlated with increasing life expectancy. As discussed earlier, the last two are dependent on a number of other factors including remittances flows. 47 Figure 2A.1.6: Shifts in Efficiency Frontiers--Single Input and Two Output Models 35 AST 30 TKB 25 ut npI 20 o-t- put 15 ut RDK O 10 5 0 - 50 100 150 200 250 300 Access-to-Input 700 600 500 nputI- 400 t-tou 300 tpu O 200 KRG 100 0 - 200 400 600 800 1,000 1,200 1,400 1,600 Access-to-Input 180 160 140 120 nputI- 100 t-tou 80 tpu O 60 40 KRG 20 MST BGR 0 - 5 10 15 20 25 30 Access-to-Input 1 VRZ 0.9 NUR 0.8 0.7 nputI- 0.6 0.5 t-tou tpu 0.4 O 0.3 0.2 0.1 0 - 1 1 2 2 Access-to-Input Source: own calculations. *see abbreviations above. 48 The first step of the analysis is to identify efficiency benchmarks--that Figure 2A.1.7: Input-Access Frontier (DEA) is, the most efficient service delivery Data Set 1: Input (HTL) ­ Output (IMN) agents. Efficiency benchmarks 9 MRG 9 represent a group of most-efficient MST AST VHS QUM KTB PNDPNJ DYUISK KHJ TRS KYR BGR HSR BHTIST KLH HRS JRS KVL SPT YVN AJM ZFRTBS JLK SBD 9 RDK QAB GNCSRT ISF AYN TJK service delivery agents--rayons, in VHDFZBKBD FRK RGN VSEKRG MMS HMD VNC DNG SRS SRN TMR SRB T 9 KUL NRK U CKL BLJ MMN this case. Benchmark rayons have the PT RSN DRV NHS OU 9 VRZ JRG NUR RSQ SGN highest input and output efficiency TKB 9 RST scores and other less efficient ones are 9 TVL compared against the performance of 9 - 2 4 6 INPUT 8 10 12 14 those rayons. The analysis of 2002­05 Data Set 2: Input (HTL) ­ Output (PYS+MDW+NRS) averages indicates that just a few 18 rayons in the highest per capita KHJ 16 KRG KTB CKL DYU 14 JRS KYR allocation decile (the top 10 of budget KBD AST KUL SPT TRS BGR ISF RSN SRS TJK PND IST TBS 12 SRN ZFR SBDMRG VNC DNG MST HSRGNCHMD AYN FZB FRK PNJ DRV TVLSRB VHD SGN SRT NRK TMR MMNISK KLH VSE JLK NHS funds recipients) are on efficiency QABJRG TKB T RST YVN VHS HRSKVL 10 BHT U RDK RSQBLJ RGN PT AJM QUM VRZ MMS OU 8 NUR frontiers: Murgab, Tavildara, 6 Chkalovsk, and Sarband. Efficiency 4 2 frontiers mostly include rayons that - - 2 4 6 INPUT 8 10 12 14 fall in lowest decile in terms of Data Set 3: Input (HTL) ­ Output (BRT) provision with public resources: 5.0 BGR MST JRSSPT KBD TBS KYR CKL KHJ SRS ZFR DYU KTB HSRSRN GNCKRGTRS ASTISF 4.5 PND IST AYN NRK Rudaki, Bobojon Gafurov, Mastchoh, MRG RDK HMD TJK RSN VHS 4.0 KLH HRS JLK SGN SBD KUL QABFZB BHT SRT TKBQUM TMR DNG YVN FRK PNJAJM RSQ VNCDRV and Bohtar. The frontier indicates that VHD 3.5 VRZJRG NHS BLJ RST VSE MMS ISK MMN SRB 3.0 KVL T TVL those these rayons are much more U PT NUR 2.5 OU RGN 2.0 input efficient. 1.5 1.0 Input efficiency of access to health 0.5 - - 10 20 30 40 50 60 services is relatively compressed, INPUT Data Set 4: Input (BED+PYS) ­ Output (IMA+BOC) while output efficiency has a wider 140 range. Input efficiency indicates 120 TKB TVL relatively high degree of dispersion in 100 RST AYN T 80 JRG VNC output indicators among rayons. U SBD PT FRK DRV SRN CKL HMD OU 60 PND ZFR KBD KHJ HSR Figure 2A.1.7 illustrates that most RGN KRG VRZ NUR IST KYR VHSYVN RSNJRS GNC FZB TJK 40 VSE PNJ SPT KVL BGR ISF HRSKLH SRS SGN VHD AST TRS KTB RSQ RDK ISK SRB JLK MST KUL DYU NHSMMS TMR rayons fall under Input Zone 1, which SRT TBS AJM QAB BHT MRG QUMNRK 20 BLJ MMN DNG is closer to the most input-efficient - - 10 20 30 40 50 60 70 80 90 100 INPUT agent while far enough from relevant Source: own calculations. output-efficient agents. Moreover, *See abbreviations above. except the top 5 recipients (per capita)--Shahristan, Dangara, Khujand, Sarband and Rushan--all other rayons are concentrated around per capita allocation of US$3 per person a year (with standard deviation of 1.05); this also explains compressedness of input efficiency. At the same time, this critical mass of rayons is located far below than output-efficient rayons. There is an output-compressed and input-decompressed relationship between immunization, number of medical staff per population, share of births attended by skilled medical staff (access), and public spending (Data Sets 1, 2, and 3). In this specific case input efficiency ranges between 0.10 and 1.00 while output efficiency 0.50 to 1.00. This indicates a very input-intensive way of achieving marginal change in output 49 efficiency, or asymmetric performance of most rayons to achieve more efficiency. That is, rayons heavily overspend to achieve little output efficiency performance. Data Set 4 reflects uneven use of Figure 2A.1.8: Access-Output Frontier health facilities assets across rayons. (DEA) Figure 2A.1.8 clearly illustrates that most of output-efficient rayons are Data Set 1: Access (IMN) ­ Output (INM) 0.30 concentrated in Input Zone 1. It HRS signals about excessive inputs in 0.25 SGN health care facilities that have not 0.20 VHD T NUR been generating efficiency to improve U FRK PT RSTISK 0.15 OU VSE QAB access to health services. BLJ CKL VRZ RDK ZFR HSR 0.10 BHT MMN JLK AJM TBS SRB SRT KYR DNG VHS DYU TKB SRN HMD FZB TJK RSN KUL KBD VNC SBD TRS RSQ TMR BGR IST KRG PND DRV NRK 0.05 SRS KLH MST JRGMRG RGN SPT YVN ISF PNJ KVL QUM MMS AYN GNC NHS Access-output efficiency analysis and JRS KTB AST KHJ TVL - its "integration" with the findings of - 1 2 3 4 INPUT 5 6 7 8 9 10 Data Set 2: Access (PYS+MDW+NRS) ­ Output input-output efficiency shows very (INM+MTM) uneven performance on the access 450 side of service delivery. We 400 KRG 350 mentioned earlier about asymmetric 300 input and output efficiency T U 250 PT performances under "Input-Access" OU 200 KHJ 150 JRG TBS analysis. It mostly indicated huge TKB CKL TVL 100 MRG KYR KTB AST RSN RGN SPT NUR TMR KUL RST dispersion of input efficiency ZFR 50 SRT SRBHMD SBD BGR VRZ MMS RSQ DYU QAB MSTAYN AJMRDKVHSYVNKLH TJK SRS TRS QUM VHD FZBHSR DNG PND JRS VSENRK SRN IST JLK FRKGNC - BHTHRS ISF BLJKVL NHSISK SGN PNJ MMN DRV VNC KBD indicators among rayons under very - 2 4 6 8 INPUT 10 12 14 16 18 concentrated output efficiency; in Data Set 3: Access (BRT) ­Outputs (MTM) 25.0 BHT other words this would mean that most 20.0 rayons have been overconsuming ISF inputs to produce the same level of 15.0 T U PT output which other rayon(s) produce OU 10.0 FRK JLK with using fewer amounts of GNC NRK 5.0 VSE SRNIST FZBQUM resources. Since the picture of the VHD DNG HSR RDK TRSPNDJRS SRS AJM YVN KLHVHS QAB TJK VRZ AYN MST DYU SRB MMS RSQ SRTSBD HMD NUR RST TMR ZFRBGR RGN KUL KTBAST SPT TVL TKB RSNMRG CKL KYR JRG TBS KHJ - KRG KVL MMN ISK BLJ NHS VNCDRV PNJ HRS SGN KBD access-output side is a bit mixed, we - 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 INPUT will discuss each specific data-set Data Set 4: Access (IMA+BOC) ­Outputs (LEB) 140 trend in separate paragraphs below. 120 TKB More details of simulation results are TVL 100 RST given in Annex 2.3. AYN T VNC 80 JRG U SBD PT FRK DRV SRN CKL HMD OU 60 PND ZFR KBD KHJ HSR RGN KRG 2.77 Infant mortality does not VRZ IST NUR KYR RSN JRS VHSYVN GNC FZB TJK 40 VSE PNJ SPT KVLHRS BGR ISF KLH SRS SGN VHD AST TRS KTB RSQ RDK ISK JLK SRB MST KUL DYU NHSMMSSRT TMR TBS AJMQAB much depend on immunization (Data MRG BHT NRK 20 QUM BLJ MMN DNG Set 1); the correlation coefficient is - - 10 20 30 40 50 60 70 80 90 100 INPUT minimal, 0.06. It is mostly due to the Source: own calculations. fact that infant mortality does not too *see abbreviations below. much depend on immunization, which has pretty much "spread" across the country. Infant mortality depends on a number of factors, including spread of various illnesses, nutrition, level of income of households, accessibility of primary health care services, and so forth. This is the main reason for 50 having a "nonstandard" efficiency frontier (Figure 2A.1.8.) where the curve itself is located too far from output axis and input efficiency is highly concentrated. The impact of the number of medical staff per population on infant and maternal mortality (Data Set 2) has been very limited. One of the main reasons is the high concentration of doctors and other medical staff in big towns that serve neighboring rayons ("cross-subsidization"). Dushanbe is one of the most extreme examples: per capita physician, midwives and nurses is the highest in Dushanbe--50 percent higher than country average--but it serves neighboring rayons (mostly in hospital services). Note also that the spread of maternal and infant mortality varies from rayon to rayon. For instance, Sogd Oblast has higher infant and maternal mortality rates than the rest of the country. The number of births attended by skilled medical staff has been uneven among rayons and has had little correlation with maternal mortality. Maternal mortality per 100,000 births averaged 50 during 2002­05 when more than 60 percent of births were attended by skilled medical staff. During 2002­05, the share of births attended by skilled medical staff significantly increased from 62 to 77 percent, despite the fact that some mountainous rayons and the eastern part of RRP had very few births attended by skilled medical staff (for example, in Rogun it is 9 percent). Theoretically, births attended by skilled medical staff and declining maternal mortality rate are correlated. Since improving coverage of births attended by skilled medical staff has been uneven among rayons, the phenomenon of efficiency asymmetry is likely to be observed in near future. As primary health care accessibility rises and bed occupation rates fall, they will have an impact on life expectancy. Life expectancy at birth is a multidimensional function and improving access to medical services may affect on it in the long term. Life expectancy at birth depends on death rates in specific age groups and population age structure and improving medical services will affect it through declining death rates in various age groups. Therefore one may observe very fluctuating "behavior" of access to primary health care and bed occupation against "stable" life expectancy at birth across rayons. Overall, input-access efficiency performance of rayons for selected data sets is very asymmetric. Input and output efficiency for each chain depends on a number of other factors apart from the ones described above. We have also to take into consideration that the following: (i) A limited absolute amount of funds is earmarked for immunization. An increase in health financing does not mean an automatic increase in immunization financing on a pro rata basis. Moreover, immunization is widely supported by donor agencies, funding for which is not reflected in budget. (ii) An increase in financing and staffing has very limited positive elasticity. Because most funds are being spent for wages of existing staff, fresh funds do not necessarily go for recruitment of new staff. 51 The analysis of access-output is multidimensional and cannot be explained with the behavior of one or two indicators. Infant and maternal mortality as well as life expectancy depend on a number of factors--not only access to health services. However, input efficiency of all data sets is statistically significant and reflects higher concentration of access indicators around national average. Nevertheless, output efficiency is relatively low for Data Sets 1-3 because infant and maternal mortality rates are still high and vary significantly from one rayon to another (Table 2A.1.2). Table 2A.1.2: Health--Summary of Input and Output Efficiency for Data Sets Input (HTL) -Access (IMN) Access (IMN) - Output (INM) Input efficiency Output efficiency Input efficiency Output efficiency MAX 1.0000 1.0000 1.0000 1.0000 MIN 0.1123 0.9721 0.9730 0.1034 Average 0.5054 0.9942 0.9813 0.2543 St. Deviation 0.2065 0.0059 0.0064 0.2112 Access (PYS+MDW+NRS) - Output Input (HTL) -Access (PYS+MDW+NRS) (INM+MTM) Input efficiency Output efficiency Input efficiency Output efficiency MAX 1.0000 1.0000 1.0000 1.0000 MIN 0.1171 0.4941 0.5558 0.0002 Average 0.5062 0.7923 0.6241 0.1531 St. Deviation 0.2080 0.1104 0.0964 0.2230 Input (HTL) -Access (BRT) Access (BRT) - Output (MTM) Input efficiency Output efficiency Input efficiency Output efficiency MAX 1.0000 1.0000 1.0000 1.0000 MIN 0.1119 0.4777 0.4777 0.0213 Average 0.5025 0.9188 0.5413 0.1093 St. Deviation 0.2130 0.1587 0.1306 0.1792 Input (BED+PYS) -Access (IMA+BOC) Access (IMA+BOC) - Output (LEB) Input efficiency Output efficiency Input efficiency Output efficiency MAX 1.0000 1.0000 1.0000 1.0000 MIN 0.0785 0.1618 0.1614 0.9502 Average 0.2859 0.3736 0.4319 0.9638 St. Deviation 0.2092 0.1973 0.2060 0.0153 Source: own calculations. * see abbreviations above. Spending Efficiency in Education Efficiency Analysis at National Level Due to limited data availability and a need to ensure compatibility of national and subnational data, the analysis is extended to two sets of education input-output data. The national-level analysis compares efficiency of public spending in education among IDA countries based on the following indicators: 52 Data Set 1: · Input: total education expenditures (ETL) · Access: gross secondary enrollment (GSE) · Output: youth literacy rate (LRY) Data Set 2: · Total spending for education (ETL), adult literacy rate (LRA), and policy performance (PLC) · Gross secondary enrollment (GSE), gross primary enrollment (GPE), and teacher-to-student ratio (TST) · Youth literacy (LRY) ECA IDA countries including Figure 2A.1.9: Input-Access Frontier Tajikistan are one of the best (DEA) performers in input-access Data Set 1: Input (ETL) ­ Access (GSE) efficiency. Tajikistan is one of the benchmark countries in the IDA 120 along with Georgia and the Kyrgyz TON 100 GUY Republic. This means that those KGZ ARM MDA 80 GEO WSM TJK MNG T countries, along with four other IDA U CPV PT 60 MDG NIC OU BGD countries, have the highest input YEM KEN LAO 40 GMB NPL CMR VUT COG TGO LSO efficiency on secondary enrollment. MWI ERI COM ZMB BENETH GIN KHM CIV 20 SLE MRT UGA SEN AGO In the case of FSU countries it is RWA BDI NER - mostly due to the fact that the literacy - 50 100 150 200 250 300 350 400 450 INPUT rate is traditionally high and there is Data Set 2: Input (ETL+LRA+PLC) ­ Access relatively adequate infrastructure in (GSE+GPE) place (inherited from Soviet times) 6 which put them in a more MWI TONGUY NICARM MNG TJK LAOKHMBGDCMRGEO MDG LSO WSM CPV NPL UGA MDA KGZ COG SLE TGOBEN GMB RWA KEN VUT 5 ZMB COM ETHYEM MRT BDI GINERI SEN CIV AGO advantageous position compared with NER 4 the rest of IDA. T U PT 3 OU 2 Input-access efficiency of public 1 spending versus gross secondary enrollment indicates symmetric but - - 2 4 6 8 10 12 14 16 INPUT uneven input and output efficiency Source: own calculations. among IDA countries. This is mostly *See abbreviations above. due to FSU IDA countries where enrollment rates have been higher than in other IDA countries. This factor largely decompresses input and output efficiency; both indices vary from 0.10 to 1.00. Countries' access performance significantly improves once policy measures are factored in. Data Set 2 takes into consideration the impact of policies--along with resources and the adult literacy factor--on access to education services. The aggregate input-access correlation rate is statistically significant. A regression analysis on Data Set 2 for inputs and access produces the following equation linking Inputs (In) with Access (Ac): 53 Ac = -0.1174161469*In + 5.756934956 Since, In = ln (ETL*LRA) + PLC and Ac = ln(TST+GSE+GPE), we have ln(TST+GSE+GPE) = -0.1174161469 * (ln (ETL*LRA) + PLC) + 5.756934956 or TST+GSE+GPE = e-0.1174161469 * (ln (ETL*LRA) + PLC) + 5.756934956 where TST, GSE and GPE are teacher-per-student ratio, gross enrollment and gross primary rates, ETL is total education expenditures, LRA is adult literacy rate and PLC is average CPIA ratings for fiscal policy, equity os public resource use, policies of building human capacity, social protection and labor, quality of budget and financial management, efficiency of revenue mobilization, efficiency of public administration and transparency, accountability and corruption in public sector. Countries that spend around the IDA Figure 2A.1.10: Access-Output Frontier average on education have stronger (DEA) input-output efficiency performance. The analysis indicates that in countries Data Set 1: Access (GSE)-Output (LRY) where public spending on education is 120 comparable to the IDA average and 100 MDG TJK WSM GEO MDA KGZ ARM GUY GIN COG MNG LSO CMR ZMB ERI KEN CPV there is stronger policy performance RWA LAO 80 KHM UGA MWI NIC CIV TGO BDI AGO T COM VUTGMB TON have the higher input-output efficiency ETH U YEM BEN PT MRT NPL 60 BGD OU SEN indicators than other IDA countries. SLE 40 NER None of countries that form the decile 20 of top spenders belong to best- - - 20 40 60 80 100 120 performing (benchmark) countries in INPUT terms of input-output efficiency. Data Set 2: Access (GSE+GPE) ­ Output (LRY) 5.0 GIN COGCMR GEO ARM MDA TJKMDG WSM GUY KGZ 4.5 LSO MNG ERIZMB KEN CPV RWAUGA LAOKHM CIV TGO NIC MWI AGO BDI Efficiency Analysis at Subnational COM VUT GMB MRTBEN NPL BGD YEM ETH TON 4.0 SEN SLE NER 3.5 Level 3.0 T U 2.5PT OU The education budget has the highest 2.0 1.5 share in the social spending portfolio, 1.0 but very poor access and output 0.5 - statistics. General secondary education - 1 2 INPUT 3 4 5 6 Source: own calculations. accounts for 80 percent of overall *see abbreviations above. spending on education; 92 percent of secondary education budget is financed from local government budgets. However, education statistics are very scarce. If we apply the same approach as in health sector--classify indicators as input, access, and output/outcome indicators--education has no statistics of outputs. Key output/outcome indicators such as average years in school, primary and secondary completion rate, learning scores, and literacy rates do not exist. Moreover, most of indicators published are given at oblast level, with no breakdown at rayon (KBO) level. 54 The spread of public spending Table 2A.1.3: Recipients of Budget Allocations, in education is less dispersed Top and Bottom Deciles than in health; the difference TOP 10 Recipients BOTTOM 10 Recipients between highest and lowest per Rayon Spending* Rayon Spending capita spending is 7.7 times,30 MRG 48.27 AJM 12.21 with a standard deviation (7.59) TVL 41.78 QAB 12.83 equivalent to more than one RSN 35.67 QUM 13.37 third of the national average31 RSQ 34.04 SRT 13.59 SGN 33.12 IST 13.86 (see Table 2A.1.3 for details). DNG 32.61 JLK 14.21 During 2002­05 the top decile BLJ 30.77 BHT 14.29 almost did not change. The DRV 30.62 HMD 14.31 education sector "by default" ISK 30.46 VHS 14.39 VNC 29.75 KLH 15.02 continued a flat increase of *2002­05 average per capita spending. education budgets across all rayons and the top 10 recipients; consequently, the decile remained the same. Note that most of the top decile is represented by mountainous rayons. Given the data constraints described Figure 2A.1.11: Single Input­Two Output above, the scope of spending efficiency Efficiency Frontier Shift in 2005 vs. 2002 analysis in education sector is limited to very few input-access indicators. Since there is no data available for access-output analysis, our analysis at rayon levels will cover only the following indicators in primary education: (i) public expenditures; (ii) student-teacher ratio; and (iii) gross enrollment in primary and general education. The efficiency frontier of spending efficiency in terms of improving Source: own calculations. *See abbreviations below. enrollment and the teacher-to-student ratio deteriorated during 2002­05. Figure 2A.1.11 illustrates shifts in efficiency frontiers (single input­two output model) in 2005 versus 2002. Both sides of "input-access" couples--public spending in terms of gross enrollment and teacher to student ratios-- shift towards deterioration. The shift reflects declining enrollment rate by 1 percentage point (from 88 to 87 percent), and number of teachers per 1,000 student by 2 percentage points (from 60.8 to 58.8) in 2005 versus 2002. 30 Per capita spending in secondary education is based on number of students; amounts spent are adjusted against price disparity and allowances to teachers in mountainous areas. 31 2002­05 average. 55 Input efficiency and output efficiency for "input-access" Figure 2A.1.12: DEA Frontiers analysis indicates a relatively higher 1. Input (ETL) - Access (GSE) Frontier (than in the health sector) degree of 180 160 concentration of efficiency indices in SGN RSN 140 ISK RSQ the sector. In other words it means a VNC MRG 120 TVL KRG AST AYN higher degree of correlation between T U 100 SRS DRV PT MMS BGR 80 TBS GNC financing, one the one hand, and OU KBD SPTPNDJRSKYR KHJ JRG ISF ZFR NHS SRN SRB MST BLJ 60 PNJTRS TKB CKL IST TJK RST VRZ HRSHSR FRK SRT QAB JLKKLH HMD KTB BHT KUL NURMMN TMR enrollment and teacher-to-student YVN RDK VHD FZB RGN VSE AJM VHS KVL 40 DNG QUM SBD DYU ratio, on the other hand. One of the 20 NRK - main reasons for the correlation is that - 10 20 30 40 50 60 INPUT education is mostly funded by the 2. Single Input (ETL) ­ Two Output (GSE and TST) budget. Informal payments take place 7.0 AST mostly in big towns, while the 6.0 BGR 5.0 countryside it has negligible share, if GNC SGN KBD ISK IST RSNVNC NHS 4.0 QAB RSQ SPT SRT any. KRGMMS ISF SRS SRBPNJJLK X2/ AYN PNDSRN HMD JRS TRS KYR HRS BHT AJM ZFR KLH Y HSRFRK DRV MST RDK YVN 3.0 TBS TKB VHS QUM TVL KHJ RST VSE MRG JRG VHD KUL VRZ FZBKVL SBD DYU MMN NUR CKL NRK BLJ KTB 2.0 However, input (public spending) RGN TMR DNG 1.0 efficiency of general education - enrollment and the teacher-to- - 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 Y / X1 student ratio represent a mixed Source: own calculations. picture. For instance, overall the * see abbreviations above. sector (nationwide) spends 70 percent more than the most efficient rayons; for the teacher-to-student ratio it is 38 percent. In other words, public spending in education does not contribute enough to improving enrollment rates, while the spending impact on the teacher-to-student ratio is more significant Table 2A.1.4: Education--Summary of Input and Output Efficiency Input (ETL) -Access (SER) Input (ETL) -Access (TST) Input efficiency Output efficiency Input efficiency Output efficiency MAX 1.0000 1.0000 1.0000 1.0000 MIN 0.1057 0.6778 0.2755 0.2733 Average 0.2824 0.8770 0.6237 0.5463 St. Deviation 0.2043 0.0647 0.1590 0.1723 Source: Own calculations 56 % 2% ange ch 98.2% 98.6% 99.3% 97. 99.3 96.8% 94.0% 97.2% 78.8% 99.7% 101.7% 101.7% 98.5% 98.4% 96.5% 92.6% % al 9.32 2.66 0.57 2.09 24.4 2.63 9.81 13.70 13.70 12.83 32.02 20.74 11.28 Actu 317.72 244.93 222.77 BUDGET 7 9.38 2.76 0.61 2.15 3.34 9.84 STATE Adj. 13.47 13.4 13.20 32.54 20.89 11.65 26.34 323.69 248.56 225.98 5 .72 .52 .84 01 2.52 9.33 2.61 0.60 2. 3.45 11.96 11.96 12 31.36 19 11 35.99 14.5 Appr. 33 250.58 228.53 ange .5% .0% .4% ch 98.1% 98.5% 86.6% 98.8% 93.9% 91 84.2% 75.7% 98.9% 101.7% 101.7% 98 94.9% 124.2% 88 86.2% % al Classification 5.67 0.25 5.42 0.68 0.37 0.31 8.68 2.19 2.32 13.70 13.70 0.005 10.56 Actu 256.19 227.44 216.87 BUDGETS L 7 98 35 5. 0.21 5.77 0.77 0.41 0.36 2.89 2. Adj. 13.4 13.47 0.006 10.69 10.07 LOCA 261.09 230.79 220.09 Functional 0 2 53 19 26 4.89 0. 4.69 0.67 0.41 0. 8.59 3.01 2.77 by 11.96 11.96 10.2 Appr. 254. 228.40 218.18 % A 2006 % N/A N/A .6% N/ 99.3% 90.6% 99.0% 99.7 99.6% 99.8% change 98.3% 98.4% 100.2% 99.2% 99.8% 96 for 100.0% 2.2: 0 0 8 al 98 19 somoni) 5.90 9.32 2.27 5.86 1. 0. 1.78 0.44 7.48 BUDGET 61.53 17.4 26.35 20.49 15.71 57 Actu Annex Budgets (mln 0 0 38 5.88 9. 2.50 5.87 1.99 0.19 1.79 0.44 7.49 Adj. 62.59 17.77 26.56 20.68 16.27 REPUBLICAN 0 0 98 .34 9.33 2.51 7.14 1.94 0.18 1.75 0.44 77/ 22.18 10 26.47 19.32 27.39 11.78 Appr. Expenditure Health es no and gorie atic ation cat edu educ education education of other onal) ea in ssie ar de vocational courses institutions eht education Education of utions utions cat rofp( nt in professional ation CLASSIFICATION al ea indi instit instit gher velsel ar educ institutions not hi secondary eht nalo nalo no by of developme development in vocation supervision area vity educati educati aryd skills skills and educati ensuring ensuring FUNCTIONAL educational parablees the acti in y education hoolc education onal secon y ar not olo e-sr catio and no P General V Secondary Institutions Institutions Professional Professional activit Management Subsidi Education Pre-sch 4.1.1. Secondary 4.2.1. 4.2.2. 4.2.3. Higher 4.3.1. 4.3.2. Educati 4.4.1. 4.4.2. Other 4.5.1. 4.5.2. 4 # 4.1 4.2 4.3 4.4 4.5 % ange 65.8 96.9% ch 98.8% 97.6% 90.5% 94.9% 98.8% 84.7% 98.7% 99.6% 99.8% 101.2% 102.2% 100.1% 100.5% 100.6% 108.0% 92.2% 100.2% 100.0% % al 1.69 2.22 1.23 1.12 9.48 0.35 0.52 3.41 0.08 7.07 0.02 0.08 0.04 0.19 10.26 66.86 47.06 15.23 13.84 Actu 105.37 BUDGET .06 2.58 2.27 1.36 1.18 9.60 0.42 0.52 3.15 0.08 7.67 0.02 0.08 0.04 0.19 STATE Adj. 10.59 66 46.03 15.23 13.77 106.61 5 0 49 04 2.53 2.09 1.35 0.73 9.29 0.41 0. 3.04 0.08 0.02 0.08 0. 0.19 15.4 64.26 44.38 15.7 13.32 10.06 Appr. 105.71 % .4% 0.4 ange 43 10 97.6% 75.7% 94.9% 98.8% 71.1 96.5% 98.0% ch 106.1% 101.6% 102.5% 100.4% 100.5% 100.6% 108.0% 90.9% 100.2% 100.0% % 2 0 0 al 0.66 3.52 6.32 2.22 0.40 1.12 9.48 0.15 0.52 3.41 0.03 5.89 0.02 0.04 80.60 53.4 43.37 13.59 Actu BUDGETS L 0 0 0 1.52 3.31 6.29 2.27 0.53 1.18 9.60 0.22 0.52 3.15 0.03 6.49 0.02 0.04 Adj. 80.26 52.57 42.3 13.52 LOCA 0 0 29 .57 84 49 04 1.52 1. 5. 2.09 0.52 0.73 9.29 0.21 0. 3.04 0.03 6.39 0.02 0. 73 50.35 41.15 13.06 Appr. A A % % % .6% N/ N/A N/A .0% N/A N/A N/ N/A 97.7% 92.8% change 94.0% 99 99.3% 99.8% 99.9% 99.2 99 100.0% 99.5 99.6% 100.0% 0 0 0 0 0 0 0 al 1.04 6.75 3.69 8.92 0.83 0.25 0.20 0.05 1.18 0.08 0.18 BUDGET 24.76 13.44 58 Actu 0 0 0 0 0 0 0 72 1.06 7.28 3. 8.94 0.83 0.25 0.20 0.05 1.18 0.08 0.18 Adj. 26.35 13.49 REPUBLICAN 0 0 0 0 0 0 0 1.02 3.22 9.86 0.83 0.25 0.20 0.05 3.66 0.08 0.18 14.16 32.14 13.92 Appr. education other of se in de area gorie eht cat es cat ni indi gorie other not cat CLASSIFICATION archeser no rset in her de ers cen ot cat al in s nters) educati and de ce s cent) indi of s als cat icsn icsn center not area care centers center experiment eht italsp indi (health care center hospit h ycli centers es th usionf (treatment FUNCTIONAL not and ni hos ed s pol healt ria polycli health icesvres ans tiona heal tr aliz pliedp vityitc s spens sinfection A A iesro Care dicale ood al blicu ti-plaguen General Speci P Rehabilitation Hospital General Di Dentistry General M Bl A Di Immuniz 4.5.3. 4.5.4. tegac Health Hospit 5.1.1. 5.1.2. 5.1.3. 5.1.4. 5.1.5. Polyclinics 5.2.1. 5.2.2. 5.2.3. 5.2.4. 5.2.6. Community 5.3.1. 5.3.2. 5.3.3. 5.3.4. 5 # 5.1 5.2 5.3 ange 91.5% 87.3% 98.9% 94.0% 57.7% 93.7% 87.8% 83.8% ch 110.3% 115.8% 92.1% 97.7% 99.8% 99.9% 78.0% % al 3.93 0.35 0.62 0.04 0.37 0.34 0.01 0.38 0.16 0.52 7.66 0.65 4.55 4.72 17.58 Actu BUDGET .10 4.29 0.32 0.71 0.04 0.37 0.36 0.02 0.41 0.18 0.62 7.84 0.65 4.55 6.05 STATE Adj. 19 17 98 4.24 0.32 0.71 0.05 0.34 0.36 2.52 0.39 0. 0.61 6.79 0.62 9.67 18.07 Appr. % % % .3% N/A ange 90.8% 87 98.8 86.4 57.7% 87.4% 87.6% 83.8% 81.5% 99.9 ch 114.0% 115.8% 100.2% 107.9% % 0 al 3.58 0.27 0.62 0.05 0.31 0.14 0.01 0.18 0.13 0.52 7.69 0.72 4.52 2.46 Actu BUDGETS L 0 20 88 3.95 0.24 0.71 0.04 0.32 0.16 0.02 0. 0.14 0.62 7.68 0. 4.52 2.28 Adj. LOCA 0 14 31 3.89 0.23 0.71 0.05 0.29 0.16 0.02 0.20 0. 0.61 3.76 0.50 0.95 2. Appr. A A N/A N/A 0% N/ N/ % 99.8% 99.9% 99.8% 99.9% 88.7% change 100. 86.6% 99.8% 99.8% 59.9% 100.0% 0 0 0 0 al 0.34 0.08 0.05 0.20 0.20 0.03 9.89 6.94 0.65 0.03 2.26 BUDGET 59 Actu 0 0 0 0 34 0. 0.08 0.05 0.20 0.20 0.03 6.96 0.65 0.03 3.77 Adj. 11.42 REPUBLICAN 0 0 0 0.34 0.08 0.05 0.20 2.50 0.18 0.03 6.28 0.62 0.03 7.36 14.30 Appr. cal medi care s.e h education other categories of pplius in s healt purposes other of area cal in ea center eht ar ni medi indicated medical ations cated eht ch to not st centers zation care d e ni oringts care CLASSIFICATION ai centerst ndii th car ion areser aining Tajikistan. ers en centers m sr not aleh and al alth cent gency e populari of eat health care ingd pert he of emiologicald tr cente supervis emer ers of ovi Finance, diseas ne health area pr goods area epi eatmentrt health and of and aln cent seases tiona y the experiment eht FUNCTIONAL and di medici other and y ni ance y riza in y tionsa cal and V/AIDS bul m ti-intestin opi mila pulo mmunito ganiz F Diagnostic P Reproductive C Ministry: Sanitar HI A A Tr activit pliedp vityitc Management A Or A iesro 5.3.5. 5.3.6. 5.3.7. 5.3.8. 5.3.9. 5.3.10. 5.3.11. 5.3.12. 5.3.13. 5.3.14. Other 5.4.1. 5.4.2. 5.4.3. equipment. 5.4.4. tegac Source # 5.4 ange N/A N/A N/A N/A N/A ch 98.1% 97.6% 97.2% 97.1% 97.5% 99.5% 99.9% 90.4% 99.7% 96.9% 96.0% 104.9% 104.2% 96.0% 104.9% 101.6% 113.4% % 0 0 0 0 0 al 0.57 8.83 0.02 8.81 7.25 24.15 42.39 58.69 27.78 10.34 19.98 24.15 16.89 exp. 317.72 293.57 226.05 183.66 Actu 0 0 0 0 0 TOTAL 0.57 9.20 0.02 9.17 6.39 23.03 43.47 59.01 27.81 11.44 19.18 23.03 16.63 Adj. 323.85 300.82 232.62 189.15 2 8 0 0 0 0 0 .23 62 .34 55 34 82 0. 7.35 0.01 7. 7. 34 43.9 60.52 32 13.25 14.3 34.23 26.40 335.63 301.41 233.54 189. Appr. % ange N/A N/A N/A .9% N/A N/A 96.9% 97.3% 89.9% 99.4% 96 86.2% ch 98.1 97.5% 97.0% 109.4% 101.3% 104.8% 107.4% 86.4% Classification 109.4% 102.7% 140.1% % 0 0 0 0 0 al 8.71 0.37 1.45 0.02 1.43 9.99 2.97 12.97 38.77 34.81 14.43 11.30 12.97 exp. 256.19 243.22 206.95 168.18 Actu Economic BUDGETS by 0 0 0 0 0 .83 .53 73 9.68 0.37 1.68 0.02 1.66 9. 2.12 11.86 39 34.36 13.78 10 11.86 Adj. 261.26 249.39 213.34 173.51 LOCAL 2006 for 7 0 0 0 0 0 .99 34 6.03 0. 1.22 0.01 1.20 9.06 2.81 2.3: 11.8 40.33 29 12.24 11.38 11.87 257.65 245.77 214.55 174.22 Appr. somoni) 60 Budgets Annex % (mln ange .3% N/A N/A N/A N/A N/A N/A ch 98 97.9% 99.1 99.0% 99.7% 96.9% 95.2% 93.3% 98.1% 100.1% 100.3% 100.3% 98.1% 100.1% 100.0% 100.2% % 0 0 0 0 0 0 al 3.62 1.64 8.68 0.21 7.37 7.37 6.89 4.28 61.53 50.36 11.18 19.10 15.48 23.88 13.35 11.17 exp. Expenditure BUDGET Actu 0 0 0 0 0 0 3.63 1.76 8.65 0.20 7.51 7.51 6.90 4.27 Health 62.59 51.43 11.17 19.27 15.64 24.64 14.02 11.17 Adj. REPUBLICAN and 0 0 0 0 0 0 3.58 1.86 8.35 0.20 6.14 6.14 5.01 77.98 55.64 22.35 18.98 15.39 30.52 20.09 22.35 17.34 Appr. Education government esc and transfers onal ery,n Expenditures employer services al) (CAPEX) icesvres rvies assets no nati institutions services of and ati machi CLASSIFICATION and current rofit capital and on pairer expenditures payments levels tiona investments payments pment, ributions other fixed goods goods communal and communic other not-for-p popul of (Operation capital equi no cont of of and Expenditures for for interest payments interest to to to s of compensation ECONOMIC compensati oyer nancee iesd ers ers Educati Recurrent (OPEX) Capital Labor contributions Labor Empl Purchases Purchases Payment Maint Payment Interest Domestic External Subsidie Subsi Transf Transfers Transf Acquisition Centralized Purchase 4 1 2 3 4 5 # III IV 1.1 1.2 2.1 2.2 2.3 2.4 3.1 3.2 4.1 4.2 4.3 4.4 5.1 5.2 ange N/A N/A N/A N/A N/A N/A ch 98.7% 98.0% 97.9% 98.0% 97.6% 98.4% 95.6% 99.1% 61.5% 101.9% 104.6% 107.6% 61.5% 101.9% 102.5% 101.2% % 0 0 0 0 0 0 al 9.47 4.95 0.51 0.13 0.13 9.11 8.07 88.18 17.18 48.41 38.94 39.64 23.88 10.31 17.18 exp. 105.37 Actu 0 0 0 0 0 0 TOTAL 9.69 4.99 9.85 0.47 0.22 0.22 8.89 7.97 89.94 16.86 49.43 39.74 40.29 24.97 16.86 Adj. 106.81 2 1 8 0 0 0 0 0 0 2 9 63 58 42 21 13 90.68 15.9 9. 5. 8.15 0. 0.21 0. 4. 106.60 49.24 39.6 41.23 27.0 15.9 11.7 Appr. % ange N/A N/A N/A N/A N/A N/A ch 99.8% 97.8 97.9% 97.4% 99.1% 55.0% 100.2% 103.7% 103.2% 102.6% 106.9% 109.1% 55.0% 103.7% 110.3% 101.4% % 0 0 0 0 0 0 al 8.96 8.34 4.32 7.12 0.44 0.10 0.10 8.96 2.44 6.52 80.60 71.64 42.72 34.38 28.81 16.94 exp. Actu BUDGETS 6 0 0 0 0 0 0 .13 .51 66 8.64 8.56 4.35 6. 0.40 0.18 0.18 8.64 2.22 6.42 80.4 71.82 43.70 35 27.93 16 Adj. LOCAL 0 0 0 0 0 0 05 .73 18 05 5. 8.49 4.75 5.07 0.37 0.18 0. 5. 2.39 2.66 74.47 69.42 43.50 35.01 25 15.53 Appr. 61 ange .1% N/A N/A N/A N/A N/A N/A ch 94.0% 91.3% 99 99.0% 99.7% 87.6% 82.0% 99.0% 99.9% 98.9% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% % 0 0 0 0 0 0 al 8.22 5.68 4.55 1.13 6.94 0.63 3.18 0.07 0.03 0.03 8.22 6.67 1.55 24.77 16.55 10.83 exp. BUDGET Actu 5 0 0 0 0 0 0 8.22 5.73 4.60 1.13 8.46 0.64 3.19 0.07 0.03 0.03 8.22 6.67 1.55 26.3 18.13 12.36 Adj. REPUBLICAN 0 0 0 0 0 0 .87 5.73 4.60 1.13 0.82 3.08 0.05 0.03 0.03 9.40 1.47 32.14 21.27 10 15.50 11.55 10.87 Appr. government tse esc and tse ass transfers onal ery,n ass xed Expenditures employer services fi al) (CAPEX) icesvres rvies assets no nati institutions xed Tajikistan services of and ati machi fi CLASSIFICATION as and current rofit capital as and ed on pairer expenditures payments levels tiona investments ed payments pment, ributions other fixed Finance, goods goods communal and communic other not-for-p popul of (Operation capital equi of consider cont of of for for interest and to to to of consider y Care Expenditures payments interest s y compensation ECONOMIC compensati ers ers ributions oyer nancee iesd inventor Health Recurrent (OPEX) Capital Labor cont Labor Empl Purchases Purchases Payment Maint Payment Interest Domestic External Subsidie Subsi Transf Transfers Transf Acquisition Centralized Purchase inventor Ministry: 5 1 2 3 4 5 # III IV 1.1 1.2 2.1 2.2 2.3 2.4 3.1 3.2 4.1 4.2 4.3 4.4 5.1 5.2 Source Annex 2.4: Aggregate Three-Year Expenditure Forecast in the Education and Health Sectors (mln somoni) 2007 2008 Forecast SECTORS approved actual approved expected 2009 2010 2011 Education 506.4 524.3 691.2 691.2 1,019.4 1,316.4 1,686.1 in percent to GDP 5.3 4.1 4.7 4.3 5.0 5.3 5.5 including: - state budget 428.5 437.2 604.4 604.4 906.6 1,178.6 1,473.2 in percent to GDP 4 3 4.1 4 4.5 4.7 4.8 - recurrent expenditures 237.1 246.1 326.6 326.6 457.2 640.1 896.2 - Centralized State Investment Program (CSIP) 39.3 68.0 69.1 69.1 85.0 100.0 110.0 - Public Investment Program (PIP) 49.3 46.2 29.8 29.8 27.4 9.7 7.9 - extra-budgetary funds (special funds) 28.6 40.9 57.0 57.0 85.4 128.1 205.0 4.2. Primary education 15.2 0.0 18.9 18.90 23.7 29.4 35.8 in percent to GDP 0.16 0.00 0.13 0.12 0.12 0.12 0.12 including: - state budget 15.2 18.9 18.9 23.7 29.4 35.8 - recurrent expenditures 15.2 18.9 18.9 23.7 29.4 35.8 - Centralized State Investment Program (CSIP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 - Public Investment Program (PIP) - extra-budgetary funds (special funds) 0.0 0.0 0.0 0.0 0.0 4.2. Secondary education 389.2 46.2 502.9 502.9 766.2 980.7 1,226.8 in percent to GDP 4.07 0.36 3.40 3.12 3.79 3.91 4.01 including: - state budget 339.9 465.4 465.4 728.7 956.9 1,198.9 - recurrent expenditures 305.9 422.3 422.3 676.7 894.6 1,130.4 - Centralized State Investment Program (CSIP) 34.0 62.7 43.1 43.1 52.0 62.3 68.5 - Public Investment Program (PIP) 49.3 46.2 29.8 29.8 27.4 9.7 7.9 - extra-budgetary funds (special funds) 7.7 7.7 10.1 14.1 20.0 4.3. Higher and secondary vocational education 44.4 0.0 127.0 127.0 176.3 239.7 341.1 in percent to GDP 0.46 0.00 0.86 0.79 0.87 0.96 1.12 including: - state budget 44.4 79.9 79.9 103.9 129.8 162.3 - recurrent expenditures 39.1 53.9 53.9 70.9 92.1 120.8 - Centralized State Investment Program (CSIP) 5.3 5.3 26.0 26.0 33.0 37.7 41.5 - Public Investment Program (PIP) - extra-budgetary funds (special funds) 47.1 47.1 72.4 109.9 178.8 4.4. Other organizations in the education sector 29.0 0.0 42.4 42.4 53.2 66.6 82.4 62 2007 2008 Forecast SECTORS approved actual approved expected 2009 2010 2011 in percent to GDP 0.30 0.00 0.29 0.26 0.26 0.27 0.27 including: - state budget 29.0 40.2 40.2 50.3 62.5 76.2 - recurrent expenditures 29.0 0.0 40.2 40.2 50.3 62.5 76.2 - Centralized State Investment Program (CSIP) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 - Public Investment Program (PIP) - extra-budgetary funds (special funds) 2.2 2.2 2.9 4.1 6.2 Health 178.3 164.9 255.5 255.5 396.5 522.9 675.1 in percent to GDP 1.9 1.3 1.7 1.6 2.0 2.1 2.2 including: - state budget 131.0 145.3 212.7 213 340.3 510.5 664 in percent to GDP 1.4 1.1 1.4 1.3 1.7 2.0 2.2 - recurrent expenditures 51.2 51.2 103.4 103.4 165.4 231.6 324.3 - Centralized State Investment Program (CSIP) 13.0 19.0 12.7 12.7 30.0 36.0 42 - Public Investment Program (PIP) 47.2 18.0 40.3 40.3 53.2 8.8 7.1 - extra-budgetary funds (special funds) 0.027 1.6 2.5 2.5 3.0 3.6 4.4 5.1. Hospitals 131.3 18.0 177.2 177.2 212.6 259.3 337.3 including: - state budget 84.1 134.7 134.7 156.4 246.9 325.8 in percent to GDP - recurrent expenditures 71.6 122 122 126.4 210.9 283.8 - Centralized State Investment Program (CSIP) 12.5 18.5 12.7 12.7 30.0 36.0 42.0 - Public Investment Program (PIP) 47.2 18.0 40.3 40.3 53.2 8.8 7.1 - extra-budgetary funds (special funds) 2.2 2.2 3.0 3.6 4.4 5.2. Polyclinics, dentistry centers, and health centers 21.8 0.0 39.8 39.8 136.1 204.2 265.4 including: - state budget 21.8 39.8 39.8 136.1 204.2 265.4 in percent to GDP - recurrent expenditures 21.8 39.8 39.8 136.1 204.2 265.4 - Centralized State Investment Program (CSIP) 0.0 0.0 0.0 0.0 0.0 0.0 0 - Public Investment Program (PIP) - extra-budgetary funds (special funds) 5.3. Health care protection of the population 8.5 0.0 15.7 15.7 19.7 24.4 29.8 including: - state budget 8.5 15.7 15.7 19.7 24.4 29.8 in percent to GDP - recurrent expenditures 8.0 15.7 15.7 19.7 24.4 29.8 - Centralized State Investment Program (CSIP) 0.5 0.5 0.0 0.0 0.0 0.0 0 - Public Investment Program (PIP) 63 2007 2008 Forecast SECTORS approved actual approved expected 2009 2010 2011 - extra-budgetary funds (special funds) 5.4. Other organizations in the health sector 16.6 0.0 22.8 22.8 28.2 35.0 42.6 including: - state budget 16.6 22.5 22.5 28.2 35.0 42.6 in percent to GDP - recurrent expenditures 16.6 22.5 22.5 28.2 35.0 42.6 - Centralized State Investment Program (CSIP) 0.0 0.0 0.0 0.0 0.0 0.0 0 - Public Investment Program (PIP) - extra-budgetary funds (special funds) 0.3 0.3 0.0 0.0 0 Source: MTEF 2008-2010, Ministry of Finance. 64 3. HEALTH SECTOR: PUBLIC EXPENDITURE TRACKING SURVEY RESULTS 3.1 The health care system in Tajikistan requires substantial reform in order to adequately meet the needs of the population and to ultimately lower morbidity and mortality rates. The public sector is the primary provider of the health care services in Tajikistan. Consequently, an understanding of the allocation of public resources and inputs ­ such as health personnel and medicines ­ is essential for being able to introduce reforms to the health care system. The Government of Tajikistan (GOT) has indicated its willingness to undertake interventions to facilitate improving the effectiveness of the system including revisiting the health financing system. 3.2 In order to support future reform efforts, a Public Expenditure Tracking Survey (PETS) was carried out in 2006. This report shares the results of that survey and draws conclusions based upon the findings. It provides information on where and on what health care monies were spent as well as which institutions within the government--both central and local--had the greatest influence on the allocation of funds. Specifically, the PETS collects budget data at each level that public resources pass through before reaching frontline providers such as hospitals and polyclinics. The PETS is also useful as it provides a baseline to check the performance of the health sector since some important reforms were introduced in the last two years (see Box 3.1). 3.3 The PETS study analyzes budget management, human resources, and other (nonhuman resource) inputs of service delivery. In each of these areas, important insights are gained that may not have been known earlier, or if known, were not substantiated with data. However, it is the detailed identification of how funds are actually managed and how human capital and other inputs are deployed that will potentially add value in the design of system-wide reforms. In addition, a valuable case study of immunization services--a key component of primary health care (PHC) services--is also included. 3.4 Unlike most PETS, the main focus of this one was not to estimate a leakage rate. The majority of PETS focus on the narrow ­ though important ­ issue of leakage of resources as they travel through the system. In the case of the Tajikistan health PETS, this is a not a primary objective for two main reasons. First, since health facilities do not have budgets or record in-kind delivery of goods and services, it is not possible to compute the leakage rate. Second, our objective in undertaking the PETS was to help identify which aspects of the budget system work and which are dysfunctional, in order to come up with recommendations to improve its overall effectiveness to meet the health needs of the population. A. BACKGROUND 3.5 The PETS was carried out against the background of a population that suffers from high rates of ill-health and malnutrition. For example, an estimated 42 percent of children under the age of 5 years were stunted in 2003. Though not captured adequately by official statistics, such common health indicators as infant mortality rates, under-five 65 mortality rates, and maternal mortality rates are high. Furthermore, diseases--such as malaria, tuberculosis, and typhoid--that had been eradicated or brought under control are beginning to affect larger numbers of persons than before independence. An estimated total of symptomatic and asymptomatic malaria cases for Tajikistan is between 300,000­ 400,000. And new diseases such as HIV/AIDS are emerging and advancing rapidly and the burden of noncommunicable diseases is increasing. Box 3.1: Policy Reforms in the Health Sector In 2005, the Ministry of Health (MOH) adopted the Health Financing Strategy, which aims to tackle recognized deficiencies in health care financing. The goals of this strategy are to improve the equity, efficiency, and cost-effectiveness of the health system through health financing reforms. The steps that need to be taken are large and will require government commitment, buy-in from various stakeholders including the medical professional community, and significant capacity to implement the complicated reforms. Key directions of reforms proposed in the strategy include the following: Increasing public financing for health care Improving the quality of health care Formalizing informal payments and introduction of user fees Increasing health personnel salaries Improving donor aid coordination in the health sector Establishment of an institutional structure of a single-payer for health care Pooling of sources of public funds for free health care Development and implementation of new provider payment mechanisms Reorganization of the system of health services delivery The health financing reform is considered an integral component of reforming the PHC system. Since 2005 progress has been made including the partial introduction of per capita financing for PHC and introduction of a guaranteed Basic Benefits Package (BBP) in hospitals; the separation of pooling for hospital and PHC; and an incremental shift from input- to output-based financing, with the objective of output-based resource allocation. The MOH and MOF agreed that the next step in financing reforms will be pooling of funds at the oblast level, as defined in the National Health Financing Strategy 2005­15. In June 2007, the MOH reintroduced the BBP of health care services that guarantees benefits and specifies user fees for services included in the BBP. Under the BBP, vulnerable groups are exempt from paying fees. The implementation of BBP in hospitals will support efforts to formalize informal payments by allowing hospitals to charge user fees for services not fully financed by the government. A medium-term objective of the BBP is to move hospitals away from budget finance to a combination of budgetary funds and fee revenues from patients, and reallocate budgetary funds to other parts of the health system such as PHC and public health. In addition to the above, the GOT has modestly increased financing for the health sector and plans to continue to do so in the context of the PRSP. Health worker salaries have increased, especially those of PHC workers. However, limited policy attention is directed so far to addressing problems of ghost workers, absenteeism, and lack of health worker accountability for results. Without addressing these problems, the impact of increased wages is reduced. The government is making preparations for the design of a comprehensive sector strategy including the health sector financing strategy. This strategy will aim at improving fiscal sustainability of the health system while addressing equity concerns. Still, further efforts are needed to address the misallocation of resources, strengthen PHC, and improve public sector management with the goal to improve service delivery in the health sector. 66 3.6 Though poor health may seem to be the inevitable fate of populations of low- income countries, this need not be the case. All countries have resources, and those with modest resources and a large health deficit must make a considered decision to identify health among its top priorities. Perhaps even more important, the resources that are allocated toward health should be allocated in such a way as to help the most people possible within the allotted resource envelope. The gravest error is when public resources are allocated towards high-cost, low-impact services at the expense of other more effective and lower-cost interventions. This results in short-changing a large section of the population, particularly the most vulnerable population groups such as children and the infirm. 3.7 In Tajikistan, there are some problems that are so prevalent or obvious that public officials, civil society, and donors alike acknowledge their presence first and foremost, public resources devoted to health care expenditures are low in absolute terms though the same as the average for low-income countries.32 An estimated 1.3 percent of GDP was allocated by all levels of government. In addition, the state of health facilities is so poor that underfunding is easily apparent. Second, the 2005 Poverty Assessment identified that the bulk of health expenditures in Tajikistan were paid out-of- pocket by individuals either as formal or informal payments. Private expenditures and external aid on health are estimated at 3.3 percent of GDP for Tajikistan.33 The largest category of private medical spending was on pharmaceuticals. Moreover, even the more affluent members of the population indicated that they did not seek health care in some instances because of the high cost. 3.8 Tajikistan's current health system is operating well below the level necessary for addressing even the most basic health needs of the population. Yet the cost of care to individuals and families is high: even some among the better-off can no longer afford all the health care services they need. Since private spending may have reached its limit yet the health care needs of the population remain inadequately addressed, the GOT will need to step in. There are two options available: (i) the GOT can provide additional funding for health care by redirecting expenditures from other sectors or activities, and/or (ii) it can make significant improvements in the allocation of existing resources to improve outcomes. The PETS results are helpful in identifying which option or which combination of these two options will yield the greatest benefits in terms of improvements in the health care delivery system. B. ORGANIZATIONAL STRUCTURE OF THE HEALTH SYSTEM 3.9 The current structure of the health care system remains similar to the Soviet model, with the state as the main provider of care services. The health services organization mirrors the administrative structure of the country whereby care services are divided into four horizontal levels of administration, and also into separate vertical pillars for national programs. The four tiers of government are republican (central), oblast, 32 World Bank. 2007. Health Development: The World Bank Strategy for Health, Nutrition, and Population Results, p. 153. 33 World Bank. 2007. Health Development: The World Bank Strategy for Health, Nutrition, and Population Results, p. 119. 67 rayon, and jamoat levels. Each of these tiers has its strengths and weaknesses in promoting health care. 3.10 The PETS helped to reveal with greater clarity the roles of each tier of government in the provision of funds for health care as well as their role in influencing the delivery of services health care facilities. Some of the main characteristics of the system are as follows: Republican Tier Figure 3.1: Organizational Structure of Government Health System MINISTER First Deputy Deputy Deputy Minister Minister Minister Advisor to the Minister Department of Department of Department of Sanitary and reforms and Department of pharmaceutics and medical services epidemiological international medical services to medical supplies Department mothers and Department of children, and economy and budget Emergency situations and planning ambulatory medical services Unit in health Department of science and scientific research Legal Department Staffing and Control Unit Internal Audit Unit Reporting and Accounting Unit Human Resources Unit 3.11 There are several ministries involved in the health sector, which can be classified into three categories. First, the MOH is responsible for health care policy as well as for national-level care services. Second, the Ministry of Finance (MOF) allocates the central government's funds for health care directly to the five oblast administrations, 68 with little or no consultation with the MOH. The third category of ministerial involvement in health care is through the hospitals, which fall under other line ministries and state agencies (for example, the Ministry of Agriculture and Ministry of Irrigation and Water Supply). Though most likely these health facilities consume a relatively small share of overall public health resources, these health expenditures are subsumed under the ministries' budgets and are not captured in the overall public health care expenditures. Oblast Tier 3.12 The MOF provides resources to the Oblast Finance Departments for health care expenditures. These resources are given to the Oblast Health Department, which is responsible for providing funds to region-wide facilities such as oblast-level hospitals and large polyclinics. However, a large share of the resources received from the MOF is transferred to the rayons. According to the law on local public administration, oblasts are primarily responsible for supervision of the rayons' performance in provisioning of basic services. Rayon Tier 3.13 The 61 rayons do not depend on budget transfers from oblasts because a major part of their budgetary revenue is from local taxes and shared taxes, which are under their control. The rayon transfers the majority of health funds to the central rayon hospital (CRH), which is headed by a chief doctor whose responsibility includes managing the CRH and allocating the rayon's health funds to all rayon health care facilities. The rayon council is led by the rayon chief and includes representatives from the rayon finance and sectoral departments. The rayon council prepares the requests for republican funding which are then sent to the oblast administration and eventually to the MOF. Jamoat Tier 3.14 The 356 jamoats are responsible for disbursing funds transferred from the rayons based on line items to health facilities operating within the jamoat territory, mostly in the rural area. Health facilities receiving financing from jamoats are mostly rural medical houses (FAPs), rural hospitals (SUBs), and Rural Physician Ambulatory Centers (SVAs). As jamoats are more closely connected to the community base, they are responsible for implementing rayon policy and do not have their own source of financing. Although they collect tax to send to the rayon, their budgetary resources are specified in a separate line item of the respective rayon budgets. The budgetary funds are transferred to jamoats' bank accounts and are managed solely by jamoats. 69 Table 3.1: Public Expenditures on Health by Functional Classification, 2005 (In SM'000) (In percent) Republican Local Total Republican Local Total Hospitals 9,199 44,358 53,557 47 71 65 Polyclinics 137 10,281 10,417 1 16 13 Public Health 712 4,609 5,320 4 7 6 Other 9,720 3,415 13,135 49 5 16 Total 19,768 62,662 82,430 100 100 100 Source: World Bank. 2007 Republic of Tajikistan, Public Expenditure Tracking Survey (PETS): Health." C. ANALYSIS OF THE HEALTH SECTOR BUDGET MANAGEMENT 3.15 One of the key contributions of the PETS is to provide a better understanding of the flow of funds and the obstacles frontline health units face in obtaining funds--and moreover sufficient funds to function at a reasonable level of viability. In this section, we discuss the main findings of the health sector budget management The analysis of the survey sheds light on three aspects of budget management: (i) allocation of resources, (ii) management of budgetary resources, and (iii) discretion of resource managers. Resource Allocation in the Health Sector 3.16 The allocation of public resources should be based upon principals of (i) maximizing benefits to the population, (ii) achieving a balance among inputs, and (iii) striving to achieve equity. All of these choices are made within an envelope of funds and hence striving for cost efficiency is essential. The first principal of maximizing welfare is tested by the type of health facilities and interventions that the GOT chooses to support. The second principal is also called "technical efficiency"--that is, is the combination of inputs such that we get the best result? For example, there is little point in having a surgeon without surgical nurses, a functioning theater, bandages, and medicines. Third, equity is an important attribute in the provision of public services to the population-- should essential public services be given more to one group than another? 3.17 The majority of public health care funds--whether at the republican, oblast, or rayon level--are devoted to hospital services with the residual amount financing PHC facilities. In 2005, 65 percent of total public health expenditures were allocated to hospitals, which has remained stable since 2002 (see Table 3.1). This occurs despite the fact that hospitals are expensive to run--much more so than polyclinics and other PHC facilities. Yet the resources spent on hospitals are inadequate as is readily apparent by the physical condition of the buildings, the poor functioning or absence of diagnostic equipment, and the scarcity of medicines, food, and fuel. Thus, the hospital subsector is consuming a large share of health resources yet yielding only limited benefits to the population. Moreover, in order to bring all hospitals to reasonable standards, the resource requirement would be very high--and unaffordable for Tajikistan. However, by developing a well-functioning PHC system, the burden on hospitals could be reduced. 70 Table 3.2: Public Expenditures on Health By Category, 2005 (In SM '000) In percent of total expenditures) Republican Local Total Republican Local Total Current expenditure 12,021 57,740 69,761 61 92 85 Labor compensation and contributions 4,170 31,549 35,719 21 50 43 Wage and salaries 3,327 25,669 28,996 17 41 35 Employer contributions 843 5,880 6,723 4 9 8 Goods and services 7,823 26,079 33,902 40 42 41 Medicine 1,188 4,590 5,778 6 7 7 Food 1,099 4,806 5,905 6 8 7 Fuel 262 1,209 1,471 1 2 2 Communal services 654 3,977 4,631 3 6 6 Repair and maintenance 2,782 6,794 9,576 14 11 12 Other 1,838 4,703 6,541 9 8 8 Interest payments expenditure - - - - - - Subsidies and other current transfers 28 111 140 0 0 0 Capital expenditure 7,747 4,922 12,670 39 8 15 TOTAL 19,768 62,662 82,430 100 100 100 Source: GOT. 3.18 Wages and utilities are protected but other important expenditure categories, such as medicines, and repairs and maintenance, are not. The health budget was allocated among the wage bill, goods and services, repair and maintenance, and capital expenditures (see Table 3.2). In 2005, 43 percent of public health expenditures were on the wage bill and an estimated 41 percent on other inputs, of which 12 percent was on repairs and maintenance. Among all categories of expenditures, the wage bill is the only "protected" category--that is, health facilities actually received wage and salaries as approved. This was not the case for other categories of expenditures, especially goods and services, which includes medicines, fuel, and food. This category remains the least understood in terms of whether health facilities actually receive necessary (nonhuman capital) inputs or not, and if not, what are the causes for this. 3.19 Resource allocation in the health sector varies across the regions. On a per capita basis, some oblasts receive more health care funds than others as seen in Table 3.3. For example, Gorno-Badakhshan Autonomous Oblast (GBAO) received SM 23 per capita, which was 2.6 times the average per capita health spending for all oblasts. It is not clear whether this reflects the higher cost of providing the same level and types of goods and services to GBAO, which is a mountainous region, or whether it reflects a greater level of goods and services being provided to the residents of this small oblast comparative to other oblasts. Also, within the oblast (as opposed to across oblasts), resource allocation per capita was not correlated at the rayon level with income per capita. Hence, poorer rayons did not receive more money. 71 Table 3.3: Health Expenditures by Region and Functional Classification, 2005 Public Per capita Share of Hospitals Polyclinics health Other Total spending total (In SM '000) (In SM) (In percent) Dushanbe 4,736 2,637 466 1,614 9,452 15 15 Khatlon 13,769 2,958 2,034 615 19,375 8 31 GBAO region 3,982 491 399 228 5,099 23 8 RRS 7,002 2,106 556 72 9,736 6 16 Sogd region 14,869 2,089 1,154 887 18,999 9 30 Total 44,358 10,281 4,609 3,415 62,662 9 100 Source: MOF. RSS = Regions under Republican Subordination. 3.20 The health sector appears to be a low priority for all tiers of government as indicated by the budget. As mentioned earlier, health care facilities are poorly funded. But despite this, the lack of prioritization given to the health sector exists throughout the public sector, whether by the republican government, oblast administration, or the rayon councils. Even in instances where additional funds became available due to higher revenue intake (for example, at the rayon level), the health sector gained few additional resources. There appears to be underexecution of the health budget at least at the jamoat level, which means that all resources allocated to this sector were not used up. This is surprising given the physical state of the health care facilities--both buildings and diagnostic equipment--and the scarcity of inputs such as medicines and dressing material, food, and fuel. However, it is possible that resources are not provided in a timely manner and utilizing the full budget prior to the end of the fiscal year becomes difficult. Management of Budgetary Resources 3.21 The management of the health sector budget has some surprising attributes which are of great importance since they shed light on who controls the funds--and perhaps on who sets health policy as well. Although the assumption is that the MOH should play an important role in the management of health resources, it does not, except for a handful of national-level health institutions. Therefore, who formulates and executes the budget? Who monitors whether it is being used as agreed upon? 3.22 About 60 different rayon chairmen negotiate for health resources. The budget is formulated at the rayon level based upon requests submitted by health facilities. The rayon chairman has the responsibility to negotiate the rayon budget, of which health is but one of the components. He (or she) must also negotiate the health budget with the oblast as well. Though the MOH may offer some guidelines, these are insufficient to influence what specifically gets funded either in terms of which facilities, how much, or for what category of spending. Thus, rayon chairmen who may have little appreciation of the country's overall health policy ultimately have some of the greatest influence on how health resources are allocated. 72 3.23 The rayon health budget is executed through the CRH and the jamoats. The rayon administration generally does not allocate resources to health facilities directly. Rather, it delegates this task primarily to the CRH, which acts effectively as the health department for the rayon. But health facilities receive funds not only from the rayon, but also the oblast and the jamoat. The only category where the amount has been predetermined and resources are allocated, is wages and salaries: payment for these are in cash to the CRH and the jamoat. Consequently, there are almost no cases where public health sector employees report not receiving their compensation. However, nonwage expenditures of health facilities were mostly paid in kind by the CRH and many health facilities report receiving very limited allocations of these items. Whether this is because the CRH keeps the majority of the resources for their own needs or that goods are diverted for private gain needs further investigation. 3.24 Health facilities find it difficult to plan health spending due to ad hoc--and sometimes frequent--budget amendments during the year, which results in suboptimal use of resources available. The rayon budget was generally amended during the year due to better tax collection than originally forecasted. The majority of additional resources were given to the CRHs. The jamoat health budget that financed mostly polyclinics did not get additional resources during the budget revision in 2005. Additional funds unexpectedly inhibit the ability of health facilities' to plan their spending, though at one level, why the lack of foreknowledge is a burden is not clear given the gap in what the facility needs and what it receives is so great. The problem may stem from the need to disburse additional funds by the end of the fiscal year ("use them or lose them") and/or that the additional resources they received were for predesignated categories. In 2005, repair and maintenance was given priority. 3.25 Financial reporting requirements on budget execution are weak in the health sector. Rayons, jamoats, and CRHs prepared end-year financial reports on budget execution and submitted these to the higher levels of government in 2005. Health facilities other than CRHs did not prepare financial reports on budget execution. It would be difficult for individual health facilities to undertake financial reporting if only because they lack an approved budget for nonwage inputs and no mechanism exists for assessing the value of in-kind donations like food from local entrepreneurs or the community. Yet, despite the lack of data for the majority of frontline facilities, the annual budget report is frequently audited by the state financial control and the internal audit unit. Consequently, it is not clear how useful the auditing exercise is from the perspective of understanding whether resources reach their designated recipients. Discretion of Resource Managers 3.26 The rayon governments have more flexibility in managing their budget because of the ability to collect and retain local taxes and the share-tax without sending all revenue back to the central treasury. The ability to retain excess revenue over the forecast provided greater scope to local governments to spend additional revenues on other priorities that were left out at the beginning of the year. They are empowered to make final resource allocation to sectors at the beginning of the year, reallocate additional revenue to sectors during the year, and reallocate resource across sectors and line-item 73 expenditures. This power is exercised by the rayon chairmen, who decide on the allocation of resources among various sectors. In addition, even the jamoat chairmen had some discretionary power, because they distributed wages and salaries and decided on allocation of resource among polyclinic services. 3.27 The chief doctor of each of the CRH has significant discretionary power, yet the system does not require much accountability from this doctor. The majority of the rayon's budget is allocated to the CRH, which is responsible for retaining some of the funds for its operations and allocating the remainder among the other health facilities in the rayon. Consequently, the chief doctor acts both as the manager of the CRH but also as the head of the rayon health system. In particular, the chief doctor is the person with the greatest authority in the allocation of resources in three important areas. He has discretion in personnel management (hiring and firing), allocating floating stavkas (compensated workloads), for bonus and approving overtime for health workers; in essence, he is de facto an employer given the degree of discretion provided to him by the government.34 High-level position such as doctors, administrators, and nurses are almost always filled by a person chosen by the CRH. The chief doctor also is responsible for allocating in- kind resources to rayon health facilities, which can include essential inputs such as medicines and medical supplies. She has the authority to reallocate budgetary funds from other health care services (such as polyclinics and public health affairs services) to hospital care. However, despite these enormous responsibilities, there are few mechanisms in place to ensure accountability of the chief doctor to the oblast administration. For example, there is no mechanism for tracking funds from the CRH to health facilities or for monitoring performance. 3.28 Some key players in the health system have little or no discretionary authority. The jamoat officials responsible for polyclinics and rural facilities primarily act as disbursing agents for the rayons. They have little or no say in how funds should be allocated among health services being provided on their territory. Perhaps of greater concern is that heads of health facilities (with the exception of the CRH) do not control their operating resources, whether personnel or their medical inputs (medicine or medical supplies). Most facilities do not even know their budget allocation. Moreover, since they receive their non-personnel resources primarily in kind, they are unable to allocate their budget among their own identified priorities. Thus, many health facilities, for example have inadequate non-personnel inputs such as utilities and medicines. 3.29 Supervision is exercised by the CRH and SUB (rural hospital) and appears to exert a high financial and opportunity cost for facilities. The CRH and SUB provided the bulk of supervision visits (89 and 78 percent respectively) in 2005. Only 16 percent of facilities that provided supervision reported receiving funds for fuel/transportation to perform these activities, although the average number of trips each quarter by the CRH was 24 (more than once weekly). In addition though, 81 percent of health facilities had to visit the CRH on average 56 times per year (more than once per week). Over 85 percent of all SUBs, SVAs, and medical houses had to visit the CRH on a regular basis, 34 High-level positions such as doctors, administrators, and nurses are almost always filled by a person chosen by the chief doctor. 74 especially those in rural areas, but only 6 percent of facilities reported receiving funds for fuel and transportation. This large amount of supervision appears unwarranted and at best could be an inefficient use of resources and at worst a deliberate misuse of funds. However, if additional information shows that these visits are being undertaken to collect supplies or salaries, then these would translate into supply chain management issues or banking system issues that would need to be addressed.35 Management of Human Resources 3.30 Health personnel--especially doctors, nurses, and technicians--are a key determinant of the effectiveness of a health care system. Without them, health care services cannot be delivered to the population. Proper management of health personnel is critical to ensuring the population's welfare, but it is very difficult. Working conditions (especially outside of Dushanbe) are poor and wages are low. In order to best use human resources, there needs to be complementarities between different types of health workers as well as between health care workers and other inputs (such as medical supplies and medicines). 3.31 Regional variations in the health personnel composition and levels point to the possibility of suboptimal allocation of resources that reflects the preferences of health sector workers rather than serves the needs of the population. These regional variations are also consistent with the findings on budget allocation and are reflected in the characteristics of both health workers and facilities. The average health facility employs 24.4 people, with a peak in Dushanbe of 147.7 employees (both allied personnel and health workers). Urban facilities are on average more than 10 times bigger than rural ones in employment terms, which may reflect both larger populations being served and the greater availability of complementary inputs such as medicines and medical supplies. The nurse-to-doctor ratio (NDR), a popular quality indicator, varies a great deal in Tajikistan. Although there is no standard or optimal NDR, the informal best ratio falls somewhere between 2 and 4. The national NDR is 2.08; it goes from a low 0.95 in Dushanbe, which has more doctors than nurses, to the national high of 2.71 nurses per doctor in GBAO. 3.32 Though there are relatively few wage arrears compared to the past, health sector wages are low. Recent reforms undertaken by the GOT in the wage system have been successful at alleviating delays and almost eradicating arrears for protected salaries, at least in the health sector. Except in Sugd, where a little more than 40 percent of health workers experienced delays, almost all employees received their entire salary on time in the country. However, salaries are still very low in Tajikistan's health sector with an average monthly wage of SM 48. There is a huge premium for working in Dushanbe, where the average total salary is more than twice the national average. Even hospital attendants in Dushanbe make more on average than doctors elsewhere in the country. This factor constitutes a powerful brake for any attempt to send doctors in remote areas 35 The high supervision costs are unusual and require further study and additional interviews with staff in order to understand the purpose of these visits, what was ultimately accomplished from them, how the cost to personnel was recovered, and how this impacted the ability of the facilities to deliver services. 75 and may partly explain the very high proportion of doctors per facility in Dushanbe and its very low NDR. 3.33 Tajikistan's health workers consider themselves grossly underpaid and employ coping strategies, some of which are illegal. The average salary considered as "fair" is 7.7 times higher than the actual average salary. In order to compensate for the inadequate salary, many health workers either seek employment outside the facility or informally charge patients. Approximately 18 percent of the health workers admit that they work outside the facility to supplement their low income. On average health sector employees worked 20 hours per week in a second job, which may explain the high absentee rates seen in Tajikistan. Doctors and administrative staff are more likely to hold a second job. Moonlighters mostly work in the agricultural sector (55 percent) or privately provide health care (28 percent). 3.34 Besides working outside the facility, 46 percent of the health workers admit to receiving informal payments (gifts in cash or in kind) from patients to supplement their income. Informal payments are more prevalent in Dushanbe, received by 72 percent of workers, 60 percent of doctors, and 50 percent of nurses. The average health worker is able to extract as much as SM 28 per month from patients, with a peak of SM 124 for doctors in Dushanbe. 3.35 The chief doctor of the CRH is primarily responsible for allocating stavkas (additional work loads and salaries) to health personnel. At the CRH level, 93 percent of the chief doctors recognize that they wield most of the power in that process. Heads of large facilities like oblast hospitals or polyclinics seem to have some leverage in that decision, whereas the heads of the medical houses are completely out of the loop with only 2.6 percent who think they are the ones with the most say. Jamoats also seem not to be involved much in the decision. Thus, the distribution of stavkas gives the chief doctor significant power over human resource decisions as well as over the budget allocations for wages and salaries in the health sector. However, stavka allocation may have perverse effects, such as reducing the incentive to hire new health workers to fill vacancies. 3.36 Although official wages are low, there is a sizeable amount of unallocated funds in the rayon's wage budget that can be used to allocate stavkas as bonuses for the staff. The PETS findings show that the average facility has 13 extra stavkas, of which only 8 were redistributed to the employees, leaving 5 stavkas unaccounted for. In Dushanbe there are on average 123 extra stavkas for the facility with 148 health workers. Doctors and nurses receive the highest numbers of extra stavkas. Men are more likely to be granted extra stavkas. Experience in the health sector and longevity in the facility also boost the chances of receiving extra stavkas. 3.37 The PETS finds that approximately 30 percent of health workers were absent from the facility at the time of the survey. Controlling for facility and staff characteristics, health workers in rural areas are 31 percent more likely to be absent than their urban counterpart. Khatlon's employees are 7 percent less likely to be absent than workers in RRS, Sugd, and GBAO. Absenteeism rates are lower for medical houses, SVAs, and SUBs when compared to small polyclinics and other facilities. Higher salaries 76 and more stavkas reduce substantially the likelihood of absence, indicating that perhaps high absenteeism may be a result of health personnel pursuing alternative employment. 3.38 Nevertheless, there are a high number of unallocated stavkas. There are two possible reasons for this. On the one hand, it could stem from the lack of incentive from the CRH chief doctor to replace health workers who leave the facility for any reason. The CRH controls the wage budget and reductions in health workforce mean more "free" resources under their control. An alternative reason could be that that CRH cannot find qualified people to hire because the university's cohort are not large enough to meet the demand or that doctors do not want to go to remote areas. Management of Nonhuman Resources Inputs 3.39 In order for health care facilities to function adequately, they must receive an array of inputs including communal services (such as water, heating, electricity, and phone), fuel for vehicles, drugs, medical supplies, and material supplies. Moreover, facilities that provide inpatient care and have beds will need food. Health facilities serving larger populations obviously have greater input requirements, but each facility requires a minimum level of inputs in order to function. As discussed below, in Tajikistan, the PETS reveals that many health facilities do not receive adequate inputs, though it is not clear to what degree this is a result of insufficient funds for health care or mismanagement of funds (including corruption). 3.40 Many health facilities do not have the basic infrastructure necessary to provide health services. Tajikistan's crumbling infrastructure is adversely impacting the health sector to the point of making many facilities almost nonfunctional. Though rural health care facilities have less access to infrastructure services than urban facilities, without exception all categories of facilities are lacking in some important service. Access rates across the board are discouraging: · An estimated 53 percent of facilities have access to water. All urban and rural CRHs, hospitals, and polyclinics had access to water, compared with only 42 percent of rural primary care facilities (piped or delivered). · Regional average rates of winter heating are low, with the exception of GBAO's rate of almost 100 percent of surveyed facilities (due possibly to its severe winters). In the other regions, only 30­50 percent of facilities were heated in winter. Only 6 percent of facilities that receive heat are connected to central heating while 64 percent use coal or wood. · On average over 90 percent of facilities have access to electricity by region. However, rural facilities and outpatient care facilities received it for fewer hours than urban and inpatient care facilities. Rural PHC facilities received electricity for only about 3­4 hours per day. · Only 23 percent of health facilities reported having access to a radio and/or telephone. Most of these facilities were in urban areas. However, in rural areas 77 where it would be most needed, only 9 percent of facilities had access to a means of communication. · Only 19 percent of health facilities had at least one vehicle (such as a car or ambulance) for transport of patients in emergencies to and from the facility, for outreach and supervision, or for the collection of important inputs such as drugs and vaccines. 3.41 The lack of basic utilities calls into question not just the quality of care, but the viability of the rural PHC network. 3.42 Inputs of drugs, food, fuel, and other materials varied greatly by region and type of facility, despite the norms established by the MOH. The high level of variation suggests an undue degree of discretion by the persons allocating resources to various facilities. Similar to the patterns of access to communal services and vehicles, urban facilities and hospitals were more likely to receive greater quantities of inputs. · An estimated 33 percent of facilities reported not receiving drugs (budget or actual pharmaceuticals) from GOT resources.36 The average expenditure on drugs per facility is SM 8,044, ranging from SM 108 for medical houses compared to almost SM 42,000 for CRHs. · Only 57 percent of facilities with beds received funds or in-kind deliveries of food from the GOT. There was great variation even among the CRHs and the food allocation per bed also varied greatly. Average monetary value of food per facility is SM 25,623. · Only 11 percent of health facilities received funds or in-kind allocation for fuel. The average allocation was SM 5,342 per facility with high standard deviation around the mean. · Other materials such as office supplies were given to 55 percent of the health facilities. However, the average allocation was SM 9,945, which appears to be high. · An estimated 75 percent of urban facilities and 50 percent of rural facilities were renovated in the past year. 3.43 Variation in funding or in-kind resources was very high among facilities with 16 percent of health facilities receiving no assistance other than wages. Urban facilities received most of the funding with a strong bias in favor of CRHs followed by other hospitals and then polyclinics. In rural areas, SUBs received the majority of the funds with very few resources going to SVAs and medical houses to cover their needs for drugs, fuel, utilities, and other materials. Lack of assistance from the GOT indicates the 36This figure excludes GBAO health facilities that reported receiving no drugs at all. There may have been some misunderstanding that resulted in survey discrepancies. 78 low level of public support for rural PHC facilities receive and also the possibility of high leakages in the flow of funds. 3.44 Unique to Tajikistan, health care staff spend out-of-pocket to purchase needed medicines and supplies for their patients and facilities. The PETS survey shows significant out-of-pocket expenditures by health staff to cover costs. Possibly, these expenditures were financed by additional stavkas and bonuses given specifically to cover additional needed expenditures, from salaries, or through collecting informal payments. Nevertheless, having to use own resources is burdensome and unfair to the public employees. Moreover, the use of stavkas to fund nonwage costs lessens transparency in the use of public funds. 3.45 Donations from international and local organizations partially compensated for insufficient government support. An estimated 55 percent of facilities received in-kind support from external sources. The main source of external support was international organizations (90 percent) and in a few instances national NGOs and the local community. As with government support, CRHs and medical houses had the highest and lowest external support, respectively. However, since external donors did not coordinate their support to the health sector, there is a concern of misallocation of resources; external support might be concentrated in some areas while bypassing more needy areas. D. WERE LEAKAGES IDENTIFIED? 3.46 One of the PETS objectives is to identify whether funds reached frontline service providers, e.g., central rayon/city hospitals and primary health care facilities and how much. However, tracking of health expenditure faces challenges due to a lack of an approved budget for a health facility and poor financial records. Central Rayon or City Central Hospitals were the only type of health facility that had the approved budget, while the other types of health facilities providing primary health care (FAPs, SUBs, SVAs, medical houses and Polyclinics) did not have approved budgets. The survey, therefore, asked for the amount of fund sent and the amount of fund received recorded by various levels (rayon, CRH, jamoat, and health facility). 3.47 There were potential leakages of the resources flowing from the rayon to the CRH. The survey data show discrepancies of funds that flowed from the rayon to the CRH budget and from the rayon to the jamoat budget. Nearly 50 percent of CRH reported discrepancies between the amount executed as reported by themselves and the amount of fund sent to these CRHs as reported by the rayons. Wage and salary constituted the main source of these discrepancies. The discrepancies in the jamoat budget as reported by jamoats themselves and by the rayons were difficult to interpret due to limited budget information. 3.48 Wage expenditure estimated at about 4.8 stavkas per health facility was not accounted for, thus implying potential leakages. Records of salary payments including approved and actual work loads allocated to individual medical staffs were readily available at all level of health facilities. The PETS is in fact the first survey that collected data on the number of stavka held by a representative sample of health workers. On 79 average, a Tajik facility on an average has 52 percent more approved stavkas than actual stavkas based on employees. Therefore over 1/3 of the wage bill in the health sector is unallocated and the CRH director had discretionary power in allocating these stavkas as bonuses for the staff. The PETS findings show that the average facility has 12.7 extra stavkas of which only 7.9 were redistributed to the employees and the remaining 4.8 stavkas are unaccounted for. If they were not returned to the Treasury at the end of the budget year, this represents leakages of the wage expenditure. 3.49 The survey was unable to determine leakages of non-wage inputs flowed from a higher administrative level (CRH and Jamoat) to PHC service providers. Tracking the total amount of fund for non-wage inputs (drugs, food, and other expenditures) received by facilities is more complicated. Health facilities other than CRHs (medical houses/FAPs, SUBs, SVAs, polyclinics) did not know their approved budgets. They received both cash and in-kind inputs from CRHs, jamoats or rayons and have poor financial records. These together with weak capacity of the survey team led to triangulation of the budget data reported by various administrative levels. 3.50 In conclusion, the PETS is not an effective tool for tracking leakages in Tajikistan. This is due to the decentralization of the budget financing and decision making process and fragmentation of financing. The financing of health facilities at the local level is independent from the center as the local budget is approved separately by local councils. Local authorities including the rayon chairmen and Chief doctors of CRHs have more power to decide on resource allocation. Further, only CRH is consider a budget entity while a PHC service provider is not a legal entity and therefore has no approved budgets. There created variations in the flows of fund as PHC service providers in rural areas could be financed by jamoats rather than CRHs. These factors make it more complicate to design the survey to track these expenditures. E. THE WAY FORWARD 3.51 Tajikistan's health sector faces many challenges in addition to the limitations imposed by low income. The country inherited a complex health infrastructure that may be well beyond the country's means to maintain. In addition, the needs of the population are high and the provision of social services is costly given the geography of the country. An estimated 75 percent of the population lives in rural areas and much of the country is mountainous. 3.52 The problems of regional variability of health expenditures, poor control of resources, and underfunding of rural PHC are a manifestation of certain fundamentals of the health care financing system. They represent the rules established, the abuse of these rules by those with discretionary power, and finally the choices made by policy makers to allocate resources to urban hospitals at the expense of PHC. Specifically, the PETS has helped to highlight three important issues that deserve consideration by Tajikistan's policy makers. 3.53 First, if a large part of the system is functioning poorly should certain facilities be downsized and closed? As mentioned, the SVAs (rural physician ambulatory 80 facilities) and medical houses get few resources besides health care personnel. Moreover, relatively few doctors are willing to work in such conditions, which leads to a poor mix of inputs. Though the rural population desperately needs health care services (no matter how inadequate), it is probably the poor that visit these facilities whereas wealthier patients bypass the PHC to reach the functioning hospitals. Does this imply that some facilities could perhaps be closed with rural hospitals providing better outreach services? 3.54 Second, greater public funding as well as more efficient use of funds can provide much-needed resources for the improving the health care system. Public resources devoted to the health sector are low and more are needed to reach even minimally acceptable standards of care to the population. In addition, the existing public resources are so poorly spent that without substantial changes, a large portion of any additional resources channeled into the system will be wasted. Consequently, an increase in public resources to the health sector must be accompanied by large reforms, which may face hurdles in implementation given the competing interest of various stakeholders in the sector. 3.55 Third, though it is apparent that frontline facilities do not receive sufficient funds, it is not obvious how large their budgets should be to provide reasonable services. Even a rough estimate would require better data on numbers of patients and why they visit a health care facility. This data could come from better record keeping at the facility level or through exit surveys. Specific Recommendations Management of the Health Budget · Increase allocation to PHC facilities to ensure that other basic inputs are appropriately financed so that health personnel are able to provide quality health care to citizens. · Separate the PHC and hospital service budgets in pilot rayons to ensure that their allocated resources will not be captured for hospital services. The GOT has already begun such pilots. This effort needs to be supported by training of health staff on financial management and accounting. · Improve dissemination of budget information to include breakdown both by line economic and functional classification and by individual health facilities. Such information will help in planning service delivery activities. Given that the GOT is planning to gradually implement per capita financing, health facilities should have a legal status and autonomy to manage their own budgets in the long run. · Encourage health facilities at all levels to regularly keep records of budgetary and nonbudgetary funds received and consumed. Health facilities should keep records of the amount of cash received for payments of wage and salary and the quantity and value of goods and services received from the budgetary 81 sources (CRH and jamoats). Similarly, they should keep records of nonbudgetary resources (cash and the value and quantity of goods) that they received from donors, businesses, and the local community. · Train the health facility staff that are responsible for budgeting and financing on the budget laws (rules and procedures) and basic financial management and accounting. If jamoats continue to be involved in distributing health resources, this training should be applied to financial officers at the jamoat level as well. Such training can be supported by the World Bank's Community and Basic Health Project. · Develop transparent guidelines and processes for budget preparation and execution, especially the reallocation of expenditures during budget execution, and publish them widely at all levels of local governments and in all types of health facilities. Management of Human Resource Inputs · Though the country has the right minimum NDR of 2:1, there should be reallocation of doctors from Dushanbe to other oblasts and from the polyclinics to other facilities to correct strong rural-urban imbalances. · Given the amount of resources involved, the allocation of stavkas should be reviewed to ensure transparency and accountability. The heads of health facilities and jamoats could be more involved in the decision making on distribution of extra stavkas rather than leaving all the decisions to the chief doctors. Management of Other Inputs and Service Outputs · The GOT should focus on improving the basic infrastructure of rural PHC facilities and their access to communal services, especially water, electricity, and heating. · A management information system (MIS) for the health sector could help provide information to health managers to improve implementation of per capita funding and the BBP. · Donor coordination in the health sector needs strengthening to eliminate duplication and thus ensuring that scarce resources are utilized efficiently. A Sector Wide Approach (SWAp) could be considered as an instrument for strengthening coordination among donors. However, the SWAp will require that the government develop a comprehensive health sector strategy that articulates clear goals, objectives, and policy measures to be implement in the medium term, accompanied by realistic cost estimates. Currently, Tajikistan does not have a comprehensive health sector strategy. 82 · Improve the efficiency of resource utilization in several areas. For example, reallocating funds from food to other health inputs could improve intermediate outcomes. Given current underutilization, consider reducing the number of inpatient beds at CRH level. Utilization of drugs provided by the state for inpatient services is not very efficient and there is scope for efficiency gains. F. CONCLUSIONS 3.56 The gap between the health needs and the health care system in Tajikistan remains large. Without substantial changes and improvements in the provision of public services, this health deficit will grow and the citizens of Tajikistan will be worse off. However, the GOT has shown a commitment to revisiting its current policies and has taken steps to reform the health care system. The health PETS can provide some of the input necessary to help guide policy in the right direction. 83 4. EDUCATION SECTOR SPENDING: FINDINGS FROM A PUBLIC EXPENDITURE TRACKING SURVEY A. MOTIVATION FOR AN EDUCATION PUBLIC EXPENDITURE TRACKING SURVEY 4.1 The present discussion of public expenditures for the Tajik education sector is primarily based on the findings from an education public expenditure tracking survey (PETS) conducted in 2007. While the 2005 Public Expenditure and Institutional Review (PEIR) identified several important aspects of public expenditure in the education sector, such as the crowding out of nonwage current expenditures by the wage bill, its findings remained at a relatively high level of aggregation. At the same time, there are important trends in the education sector that pose challenges and warrant more in-depth analysis. These trends include rising public education expenditures combined with projected substantial increases in student body size. In addition, public education expenditures are not equally distributed across oblasts, rayons, and Regions under Republican Subordination (RRS). Finally, the level of capital investment in the education sector is relatively low--this in the context of already deteriorating infrastructure that will be experiencing further pressures given the increases in student body size. There are, therefore, clear upward pressures on public education expenditures, and a closer inspection of potential efficiency gains is warranted. The PETS is a useful instrument in this regard, whether through assessment of potential leakage of public funds or the need for better deployment of staffing resources, for example. In addition, the PETS questionnaire was designed to shed some light on the important issues of equity of expenditures and the state of education infrastructure in the country. Public Expenditure Trends 4.2 A trend mentioned above is that the share of public education expenditure in GDP increased from 2.3 percent in 2000 to 3.9 percent in 2007 (see Figure 4.1). Although this is a substantial increase, it remains below the OECD average of 5 percent of GDP in 2004.37 Education expenditures as a share of total public expenditures also increased over 2000­07 from 15 to almost 19 percent (see Figure 4.2). This figure, in contrast to the share of education expenditures in GDP, is higher than the average OECD share of 13 percent in 2004.38 The combination of relatively high share for education in public expenditures and relatively low share in overall GDP reflects a strong GOT commitment to the education sector in the larger context of relatively low levels of overall public expenditure. 37 OECD 2008, Table B2.4, p. 208. 38 OECD 2008, Table C4.1, p. 230. 85 Figure 4.1: Public Education Expenditure as a Share of GDP, 2000­07 4.5 4 3.9 3.5 3.5 3.4 3 2.8 P D 2.6 Gfot 2.5 2.4 2.4 2.3 cen 2 er P 1.5 1 0.5 0 2000 2001 2002 2003 2004 2005 2006 2007 Year Source: Ministry of Finance (MOF). 2007. Medium Term Expenditure Framework (MTEF) for 2008­10. Figure 4.2: Public Education Expenditures as a Share of Total Public Expenditures, 2000­07 25.0 20.0 19.3 19.3 erutid 18.8 en 15.5 exp 15.0 15.2 15.0 cilbup 13.1 12.3 altotfot 10.0 cenre P 5.0 0.0 2000 2001 2002 2003 2004 2005 2006 2007 Year Source: MOF. Equity of public education expenditures 4.3 Disparities in the distribution of financial resources are observed across rayons. Thus, Table 4.1 shows that across rayons nationwide, students in the lowest-funded rayon 86 are allocated less than half the amount received by students in the highest-funded rayon. Among regions, the greatest disparities are observed in RRS, and the lowest disparities in Dushanbe. In addition to a low level of disparities, Dushanbe exhibits the lowest regional mean per student allocation, so that students from the better-off rayons in Dushanbe receive a lower level of funding than students from the lowest-funded rayons in Sogd. Some differences in the allocation of funds across rayons may be justifiable if they are compensating for cost disadvantages resulting from, for example, school location, or if certain schools need more funding per student because they serve disadvantaged or poor students. However, funding allocations are not made taking such factors explicitly into account. Table 4.1: Per Student Current Expenditure Disparities at the Rayon Level Khatlon Sogd RRS Dushanbe All rayons Mean 100.85 101.81 106.22 79.89 100.66 High 141.99 129.31 180.23 84.93 180.23 Low 80.92 86.40 77.68 72.63 72.63 No. of rayons in the sample 25 15 11 4 55 Source: Avanesyan 2005. Education sector capital expenditures 4.4 In the context of increasing total public education expenditures, Figure 4.3 shows that, beginning in 2004, capital expenditures have constituted an increasing share of total education expenditures. Starting at the very low share of roughly 3 percent of total education expenditures over 2000­04, capital expenditures in 2007 amounted to a high of 15 percent. Clearly, GOT has not only shown a commitment to the education sector as a whole over the last few years but also to its infrastructure and investment needs. International donors also are active in school rehabilitation and construction. As the discussion below will highlight, however, the needs are great and a lot remains to be accomplished in the area of capital investment in the education sector. Figure 4.3: Capital Expenditure as a Share of Total Public Education Expenditure, 2000­07 16.0 15.1 14.0 eruitd 12.0 11.8 en exp noi 10.0 9.3 catude 8.0 licbupl 7.6 ta 6.0 tofot 4.0 cen 3.6 er 3.1 3.1 P 2.7 2.0 0.0 2000 2001 2002 2003 2004 2005 2006 2007 Year Source: MOF. 87 Student Enrollment Trends 4.5 Like other countries of the former Soviet Union, Tajikistan has a system of four levels of education (World Bank 2002): · Preschool. · general schools, including primary grades 1­4, secondary grades 5­9, and upper-secondary grades 10­11; the term "general secondary" is also used to cover grades 5­11. · professional training schools (PTUs) and specialized secondary schools that offer vocational and technical training. · higher education institutions. 4.6 "Upper secondary" education, following the compulsory first 9 grades (or basic education) requires some explanation. It can be the last two years of general academic schooling (grades 10­11); it can be in vocational and technical schools controlled by the Ministry of Labor (PTUs), taking students who have completed grade 9 (for three years) or grade 11 (for two years); and it can be in specialized secondary schools under both the Ministry of Education and other ministries and agencies, taking students after grade 9 (for four or more years) or after grade 11 (for two years). 4.7 All age groups corresponding to the above-mentioned levels of education are projected to increases in size for the foreseeable future. Figure 4.4 shows the projected change in size, relative to 2005, for the age groupings of years 3­6, 7­15, 16­18, and 19­ 24 (corresponding roughly to the education levels of preschool, basic education, upper secondary education, and higher education, respectively). For the next five years or so, the greatest increase is projected for the basic education age group, but then the older age groups (upper secondary and higher) show the largest changes. In addition, the increases are significant: by 2015, it is projected that the basic, upper secondary, and higher education age groups will be on the order of one third bigger than they were in 2005. Figure 4.4: Trends in Population Size Relative to 2005 225 200 ) 100=( 175 2005 year 150 ot veitaler 125 oi at R 100 75 2005 2010 2015 2020 2025 2030 2035 Year Age 3-6 Age 7-15 Age 16-18 Age 19-24 Source: World Bank Health, Nutrition, and Population Statistics. http://www.worldbank.org/hnp/. 88 4.8 Focusing on general education, enrollments have trended upward for grades 1­ 4 and grades 5­9 following a dip in the early 1990s (see Figure 4.5). For grades 10­11, on the other hand, the dip in enrollments in the early 1990s was more sustained, lasting into the late 1990s. Although enrollments have increased again since 2000, they are still today only roughly at what they were in 1991. In addition, despite the increases in enrollment, as Figure 4.6 shows, gross enrollment rates (GERs) for basic and higher education are roughly at the level they were pre-transition. Of course, in the case of basic education, this means near universal enrollment. But in the case of higher education, the 2005 GER of 15 percent compares with a GER of 45 and 37 percent in Kazakhstan and the Kyrgyz Republic, respectively (UNICEF 2007). Furthermore, upper secondary education GERs have experienced a sharp decline since the transition, although they may be showing an upward trend since 2002. Here again, Tajikistan's 2005 upper secondary education GER is at 31 percent, compared to a GER of 41 and 43 percent in Kazakhstan and Kyrgyzstan, respectively (UNICEF 2007). Thus, despite the increases in enrollment in general education, further improvement in enrollment rates is required, implying further increases in student body size given the population projections. Figure 4.5: Trends in Enrollment in General Education, 1991­2006 1,800.0 1,600.0 1,400.0 stn 1,200.0 udets 1,000.0 Grades 1-4 of Grades 5-9 dsna Grades 10-11 800.0 Grades 1-11 Thous 600.0 400.0 200.0 0.0 91/92 2/93 93/94 4/95 95/96 6/97 97/98 8/99 99/00 0/01 01/02 2/03 03/04 4/05 05/06 19 199 19 199 19 199 19 199 19 200 20 200 20 200 20 Academic year Source: National Statistics Committee of the Republic of Tajikistan 2006. 89 Figure 4.6: Trends in General Education GERs, 1989­2005 (percent) 120.0 100.0 80.0 Basic education 60.0 Upper secondary education Higher education 40.0 20.0 0.0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source: UNICEF 2007. Equity in student enrollments 4.9 Indeed, increases in female enrollment in particular are required, since gender parity in enrollment has declined for all levels of general education (see Figure 4.7) and especially so for upper secondary education (no doubt contributing to the lagging enrollments at that level). One of the main findings of the 2008 poverty assessment (World Bank 2008a) is that the poor and females are less likely to enroll in school and starting to drop out of school earlier. More specifically, educational attainment is lower for the poor and women: only 5 percent of poor women complete grade 11 compared with 26 percent of poor men, and 11 percent of the richest women complete grade 11 compared with 35 percent of men from the wealthiest quintile. In addition, only 1.6 percent of girls of secondary school age from the poorest quintile are enrolled in grades 10­11, compared to 23 percent of boys. At the same time, 19 percent of girls from the richest quintile are enrolled in grades 10­11 compared with 66 percent of boys. In other words, addressing equity concerns would lead to increases in enrollment and enrollment rates. 4.10 In sum, therefore, the trends in expenditure and enrollment described above result in several questions around possible efficiency gains, equity considerations, and capital investment needs that warrant further exploration. The implementation of the education PETS seeks to shed some light on these questions. First, however, some more detail on the structure and flow of public education expenditures is necessary in order to better motivate the areas of focus of the PETS. 90 Figure 4.7: Trends in Gender Parity in Enrollment, 1991­2006 1.10 1.00 s 0.90 boy ot Grades 1-4 slr Grades 5-9 0.80 gi Grades 10-11 of Grades 1-11 oita R0.70 0.60 0.50 91/92 2/93 /94 /98 /01 /05 93 94/95 5/96 96/97 97 8/99 99/00 00 01/02 02/03 3/04 04 05/06 19 199 19 19 199 19 19 199 19 20 20 20 200 20 20 Academic year Source: National Statistics Committee of the Republic of Tajikistan 2006. B. ALLOCATION AND FLOW OF PUBLIC EDUCATION EXPENDITURES 4.11 In 2006, fully 72 percent of total public education expenditures went to general education (grades 1­11; see Table 4.2). Thus, it seems justified for the PETS to focus on general schools, since they absorb the majority of public education funds (see box 5.1 for details on the PETS sample). The great majority of funds for general education came from local budgets (97.3 percent) as opposed to the republican budget (2.7 percent). As a corollary, spending on general education constitutes the lion's share of spending by local government (84.5 percent) and only 11.1 percent of spending at the republican level. 4.12 Wages and salaries account for 69 percent of general education expenditures (Table 4.3) and--together with social and retirement insurance--add up to 85 percent of expenditures and are 99 percent funded out of local budgets. Here again, the fact that staff remuneration constitutes such a large share of general education expenditures meant that the PETS contained a particular focus on human resource deployment and wage issues. Stationery, textbooks, visual aids, and library stocks--all important inputs to the education process--represent a very small share of general education expenditures. By contrast, the maintenance and repair of buildings captures 3.6 percent of total expenditures in the subsector and is almost entirely funded out of local budgets, and similarly for new construction at 3.3 percent of total expenditures. 4.13 In terms of resource management in general education, although the Ministry of Education is in charge of education policy throughout the country, the lines of 91 financial accountability do not run to the Ministry of Education, but from schools and jamoats up through rayons and oblasts (where applicable) to the MOF. Not all schools are subject to jamoats. Whereas rural schools are supplied with goods, services, and nonteaching staff salaries by jamoats, city and town schools are not under jamoat jurisdiction, but are supplied directly by the rayon. The execution of the general education budget is under the responsibility of local governments, with rayon governments in particular enjoying some discretionary power in allocating resources within the education sector. Rayons must execute funds under the "protected" budget lines, including salaries and utilities, within the economic classifications stipulated, but they have discretion to determine the allocation of these among jamoats and schools. There is also a "lump sum" budget allocation for all the nonprotected lines, which rayons can allocate according to their own discretion both among economic classifications and among jamoats and town schools. 4.14 Rayons are not, for the most part, financed by the central government, even though the planning and execution of their budgets is part of the national budgeting process. Instead, revenues are collected at the local level, but some rayons are also "subsidized" with funds from the center. When rayons collect more revenues than anticipated, or when there are unspent funds left over from the previous year, they have more discretion as to how to spend these than other funds. 92 4.5 3.1 4.0 0.9 5.5 78.7 71.7 10.4 100 Total budget 0.0 3.1 3.7 within 31.8 11.1 17.6 49.2 15.3 100 budget share Republican 5.4 0.0 4.2 2.2 0.3 3.5 Percent 88.6 84.5 100 Local budget 0.0 7.0 2.7 somoni) 13.4 82.8 74.0 48.1 17.4 100.0 (in across budget Republican 2006 share budgets 0.0 93.0 97.3 86.6 17.2 26.0 51.9 82.6 100.0 Percent Local Budget, budget of 517 545 318 445 436 Type Total 9,373,424 2,675,087 and 13,673, 12,113, 31,598, 16,876, 239,864,287 218,377, 304,687,772 0 Level 592 477 268 93 8112669 136, budget 5,890,133 9,373,424 1,627,035 1,978,530 16,890, 26,154, 53, Republican Education 0 7 by 517 283 budget 696,55 5,443,968 8,763,767 13,673, 10,486, 222,973,695 212,487,412 251,551,504 Local Expenditures Public other 4.2: institutions in (vocational) Table indicated levels educational not by institutions professional activities separated education, education education educational secondary not EDUCATION and education MOF.: General Professional-technical Secondary Preschool Secondary -- -- -- categories Higher education Education, Other TOTAL Source 0.7 0.3 1.2 0.7 0.2 1.4 0.1 0.4 3.6 0.0 3.3 0.7 0.0 0.0 2.3 69.1 15.9 100.0 Total budget 6.3 3.4 1.6 0.2 0.5 0.1 2.0 0.3 0.0 4.3 0.9 0.1 2.8 27.0 12.6 15.3 22.6 somoni). within 100.0 (in budget share Republican 2006 0.4 0.2 0.8 0.7 0.2 1.5 0.1 0.4 3.1 0.0 3.4 0.6 0.0 0.0 2.3 Percent 70.2 16.2 100.0 Local budget Budget, of 1.1 1.1 6.0 2.7 0.9 5.4 0.0 3.3 2.7 47.1 34.8 34.0 13.1 17.0 32.8 16.0 69.6 100.0 Type across budget Republican and share budgets 0.0 98.9 98.9 52.9 65.2 66.0 94.0 97.3 99.1 94.6 86.9 83.0 67.2 84.0 30.4 96.7 97.3 100.0 Percent Local budget 20 Classification 957 412 3,000 7,545 56,559 74,750 Total 571,411 599,890 334,934 125,629 915,459 256,336 94 1,578,889 2,651,1 1, 3,163, 7,842,568 7, 1,591,597 5,059,290 34,759, 150,792,745 218,37 Economic 0 440 by 8,966 6,764 3,000 96,367 29,361 18,572 52,050 368,888 743,083 198,740 901,515 119,947 330,707 255,295 167,438 budget 1,589, 1, 5,890,133 Republican 5 0 069 37,987 22,700 Expenditures Local 391, 835,806 372,671 325,968 118,86 795,512 336,302 budget 1,749,605 1,503,523 3,134,051 6,511,861 7,256,336 1, 4,891,852 34, 149,203,305 212,487,412 y Education machiner aids General inventory visual 4.3: and insurance and heavy-duty removal forms and repair Table and sewage and structures assets salaries retirement services textbooks, services machinery, and vehicles means and compensation and and and stocks supply capital inventory construction MOF.: Labor Wages Social Goods Stationery, Soft Food Communal Electricity Gas Heating Garbage Water Maintenance Buildings Transport Fixed New Equipment, Transport Library Other TOTAL Source Box 4.1: The Tajik Education PETS The PETS is a survey instrument initially developed in Uganda in 1996, later widely used in Africa and to a lesser extent in other regions of the world. The premise of the PETS methodology is that money officially committed to education, for example, in formal national budgets does not necessarily reach its intended beneficiaries in schools. The PETS seeks to "track" public funds using a sample survey methodology at multiple tiers of the administrative hierarchy down to the individual school. Budgetary data are collected as much as possible from documents, typically more reliable than ordinary respondents' answers to questions. Using data from multiple levels, PETS analysts can "triangulate" by comparing amounts that ought, in principle, to be the same--for example, a transfer from a rayon to a jamoat should be recorded in rayon and jamoat documents as having the same value. The Tajik education PETS was implemented by a local survey firm contracted by the World Bank, and data collection from rayons, jamoats, and schools took place in April­May 2007. The sample of schools was selected to be nationally representative: starting with the complete list of schools in the country (a total of 3,776 schools), two rayons were excluded due to difficulty of access, resulting in a list of 3,735 schools, from which 200 schools were randomly selected. In order to address policy questions of particular interest, two subsets of schools were oversampled: (i) 49 schools were selected from the per capita financing (PCF) pilot rayons, and (ii) 57 schools were selected from the sample of the 2004 Education Facilities Survey (EFS), resulting in a panel (since the selection was random, in the end, 15 sample schools were both PCF and EFS schools). The remaining 109 schools were neither PCF nor EFS schools. In each school, the school principal and four teachers were interviewed. When there were fewer than four teachers in the school, all of them were interviewed. The PETS collected information also at the rayon (both education and finance departments) and jamoat level since these are important links in the resource flow chain for the education system. The table below presents the complete PETS sample. PETS Sample Non-PCF PCF Total Rayons 54 5 59 Jamoats 110 -- 110 Schools 151 49 200 Teachers 598 194 792 Source: Education PETS Report. 4.15 This system of local discretion is characterized by strong formal accountability down to the rayon or jamoat level, with jamoats keeping accurate records of incoming budgets. Yet there is little accountability for the actual deliveries of goods and services to schools. Also, since budget cuts may take place at each stage in the approval process, the system has a bias in favor of the budget lines that are "protected" from such cuts, especially salaries, as opposed to goods and services. Allocations of goods and services to specific schools are not only made at the lowest levels, rayons and jamoats, but are not systematically recorded. This leaves central policy makers without the means to pursue or even observe equity in educational funding between schools. Also, schools usually know little or nothing about their budgets, and have little say in what goods and services they receive. 95 Per Capita Financing of Education 4.16 The per capita financing (PCF) approach--introduced in 5 districts in 2005, expanded to currently a total of 17 rayons, with nationwide expansion planned--was intended to give more leverage to school administrators in the use of resources allocated to them, thereby leading to efficiency, equity, and transparency gains. The main features of the reform are as follows: · School autonomy. Instead of allocating funds to rayons to be distributed among schools at the rayon's discretion, funds are transferred directly to treasury accounts, out of which the schools themselves can spend them. Additionally, school budgets are no longer pre-allocated by economic classification. Instead, schools receive the funds as a lump sum, and can allocate them among salaries, utilities, maintenance and repairs, stationery, and other expenses as they see fit. · Formula funding. The lump sum amount received by a school is determined by a formula as a function of the number of students in the school. The formula includes a constant factor, the same for all schools in a given rayon, and a variable coefficient, which is multiplied by the number of students. The constant factor helps small schools to cope with the lack of economies of scale. 4.17 According to this system, the annual financing of a school is calculated from the Rayon or the City Education Department (RED/CED) budget for recurrent expenditure according to a formula that takes into account factors such as the total amount of funds available for education in the rayon/town budget and the agreed planned enrollment statistics of a school. The allocation of resources for recurrent expenditures for PCF schools is different from non-PCF schools: once the formula is used to calculate the resources available to each school, schools are informed about the amount that is available to them and manage it according to their needs. There is an attempt at assuring equity, since the formula supplies schools with funds proportionally to the number of students. There are also incentives to manage money efficiently, because (i) decision making is located in schools, which have the best information, and (ii) schools face budget constraints and can reallocate across expenditure categories. In addition, the availability of budgets in the schools opens the door to greater civil society participation in substantive school decisions, and also serves as an extra safeguard against leakage and corruption, since parents can find out the budget and make sure it is being used appropriately. C. MAIN FINDINGS OF THE EDUCATION PETS 4.18 Other research in Tajikistan has looked into the question of transparency and accountability in public service delivery. For example, based on data from the 2006 Life in Transition Survey, over 60 percent of surveyed individuals think there is more corruption in the country compared to 1989 (World Bank 2008a). In addition, the education and health sectors are particularly affected by the prevalence of unofficial 96 payments, and the poor and the wealthy alike bear the burden of unofficial payments in order to receive public services (Figure 4.8). Figure 4.8: Unofficial Payments for Public Services in Tajikistan, by Poverty Status 40 30 tn rce 20 Pe 10 0 poorest 2 3 4 wealthiest Public health Public education Unemployment benefits Other social security benefits Source: World Bank 2008b. 4.19 Figure 4.9 shows the high rates of dissatisfaction with the quality and efficiency of public education services in the country. In other words, despite having to pay extra to receive services, the Tajik population still considers the services received to be substandard. Figure 4.9: Dissatisfaction with Quality and Efficiency of Public Education and Health Services in Tajikistan, by Poverty Status 50 40 t 30 rcen Pe 20 10 0 poorest 2 3 4 wealthiest Public health Public education Source: World Bank 2008b. 97 Tracking Public Education Expenditures 4.20 The PETS provides highly disaggregated data--collected at the rayon, jamoat, and even school level, where possible--which can be used to check the accuracy of the education financing system and test whether funds allocated to education are reaching schools. Thus, a "triangulation" exercise was conducted, whereby the budgets from different levels of the hierarchy were compared. Based on the definitions of budget lines and the rules of the budget process, it was possible, at least from the MOF level down to the incoming budgets at the jamoat level, to anticipate where budget lines from the MOF and a rayon, or a rayon and a jamoat, ought to agree. It was also possible to anticipate many cases where one budget line ought to equal the sum of a group of several other budget lines, the latter being an allocation of the former. A confirmation of these expectations would indicate, at the least, that the budgetary accounting is accurate and consistent, and be highly suggestive of an effective internal control process for public monies in the education sector. 4.21 Budget processes in the Tajik education sector appear to be highly accurate down to the jamoat level: in almost all cases, the budget data that the PETS collected from the MOF, from rayons, and to a lesser extent from the incoming budgets of jamoats, proved to be accurate and consistent. From the jamoat level down, however, the picture is less clear. The well-ordered system of recordkeeping and control of funds that exists, for example, in the rayons does not exist at the stage of the budget process where goods and services are being distributed, in most cases by jamoats, to particular schools. Thus, jamoat accounting processes do not adequately carry financial accountability all the way down to the schools. The main problem with the jamoats is not that they fail to execute their responsibilities, but lies rather in how their responsibilities are defined. Jamoats are not expected to keep track of how funds are allocated among schools. This is harmful both to "top-down" and "bottom-up" transparency; that is, (i) it becomes exorbitantly difficult, or impossible, to track public resources all the way to their final destination in the schools, and (ii) schools do not know, at least not formally, what they will get from the jamoat. Although some records are kept in jamoats and schools, jamoat accountants and school zavkhozy (resource managers) were not able to reconstruct, on the basis of these records, accounts of what was allocated to and received by schools that agreed with each other. There was an important exception to this rule: in the PCF rayons, the jamoat no longer plays a role, and schools know their own budgets and can report how they are executed. Thus, in PCF rayons the public expenditure tracking exercise was carried all the way down to the school level, and a high degree of consistency was found among budgets at all levels. 4.22 As a survey instrument, the PETS is designed, among other goals, to diagnose the presence or absence of a particular kind of corruption or leakage--that is, failure by part or all of the funds allocated to the education sector to reach frontline providers (accounting for administrative costs). In other countries, PETS have shown heavy leakage, especially of nonwage funding, merely by comparing records at the school level with records at other levels of the hierarchy. In Tajikistan, recordkeeping at the school 98 level is too haphazard for a quantitative estimate of leakage, while recordkeeping related to rayon-jamoat transactions does not suggest that leakage is occurring. The inadequacy of records that makes it impossible to calculate the amount and value of resources transferred from jamoats to schools, as well as the lack of agreement between what jamoats report allocating and what schools report receiving, suggest that the current transmission system for nonwage inputs is unreliable. 4.23 At the same time, the supply of nonwage inputs to schools is certainly very limited. Most schools reported no deliveries of stationery and inventory. Only a small minority of schools reported receiving basic maintenance materials like paint (19 percent of schools) and lime (17 percent). Even heating fuel, the most commonly delivered item, was not received in all schools, and those that received it typically reported that it was insufficient for their needs. However, the lack of nonwage inputs probably owes more to the low level of budget allocated to these inputs than to delivery problems. Teachers 4.24 Given the large share of staff Table 4.4: Student-Teacher Ratio in compensation in the education sector Grades 1­11 by Geographic Location, budget, a description of the staffing 1991­2007 intensity39 in general education is useful in order to assess the potential Year Urban Rural Total for efficiency gains. Table 4.4 reports on 1991/92 14.6 12.8 13.2 trends in student-teacher ratio based 1992/93 14.4 12.2 12.7 predominantly on administrative data, 1993/94 14.7 11.5 12.3 1994/95 15.4 12.7 13.3 but with the most recent figures as 1995/96 15.6 13.6 14.1 provided by the PETS. The table shows 1996/97 15.4 14.0 14.3 that student-teacher ratios have increased 1997/98 18.6 14.0 15.0 steadily since 1991, and rural areas have 1998/99 18.2 14.3 15.1 lower ratios than urban areas. At an 1999/00 18.1 14.3 15.1 average student-teacher ratio of 16 in 2000/01 16.9 14.8 15.3 general schools, Tajikistan is comparable 2001/02 19.1 14.6 15.6 to the OECD 2005 average of 17 pupils 2002/03 19.5 15.0 16.0 per teacher in primary education and 13 2003/04 19.6 15.1 16.2 students per teacher in secondary 2004/05 20.0 15.7 16.7 education,40 although the OECD figures 2005/06 21.0 15.6 16.9 are expressed in full-time equivalents. 2007 (PETS) 21.7 15.6 16.4 Source: National Statistics Committee of the Republic of Tajikistan , 2006 ; World Bank , 2008a. 4.25 Using PETS data, it is possible Note: The ratios reported are based on numbers of teachers to look more closely at regional rather than full-time equivalents, and include schools for variation not only in student-teacher disabled children in the case of administrative data. ratios but also school size, teaching 39 "Staffing intensity" is defined here as the quantity of human resource "inputs" needed per pupil-hour of a school's "output," so that increasing efficiency implies lowering staffing intensity, that is, keeping class sizes large by some combination of relatively high pupil-teacher ratios and/or relatively low teacher workloads, without adversely impacting quality of education. 40OECD 2008, Table D2.2, p. 382. 99 hours, and class size. As Table 4.5 shows, there is great regional variation in particular in school size and student-teacher ratio, with the latter showing a high of 26 in Dushanbe compared to a low of 11 in GBAO. The Tajik average class size of 18 is lower than average class size in the OECD, which was, in 2005, 22 for primary education and 24 for lower secondary.41 Table 4.5: Regional Variation in Hours, Class Size, and Pupil-Teacher Ratios Oblast Average school size Average Student- Average class (number of teaching hours teacher ratio size students) Dushanbe 1,658 26.5 26.0 27.6 RRS 510 23.8 17.6 19.5 Khatlon 452 25.0 18.7 18.4 Sogd 423 20.4 12.4 17.3 GBAO 251 18.4 10.8 15.6 Total 476 23.0 16.4 18.4 Source: World Bank 2008a. 4.26 As Table 4.5 further shows, teachers teach, on average, more than the official stavka or standard workload, which is set at 18 teaching hours per week for the primary level and 16 for the secondary level. While a majority of teachers (64 percent) are satisfied with their actual teaching hours, more teachers want to teach more than want to teach less (see Table 4.6). Teachers who wanted more hours tended to have shorter- than-average workweeks, and desired rather longer-than-average ones. Those who wanted fewer hours tended to have longer-than-average workweeks, and desired rather short ones. Furthermore, there is a general perception among principals (75 percent) and teachers (65 percent) that teachers want to teach more hours in order to increase their salaries. Table 4.6: Teachers' Attitudes on the Workweek Would you like to teach more than, less Share of Average hours Average than, or the same number of hours as you teachers sample taught by these teaching hours teach now? (percent) teachers desired by these teachers More 27 20.3 27.8 Less 9 30.8 21.2 As many as now 64 27.2 - Source: World Bank 2008a. 41OECD 2008, Table D2.1, p. 381. 100 4.27 Under the traditional Table 4.7: Who Makes the Decisions on Class budgeting system, with the Splitting? locus of discretion in rayons and jamoats, schools do not Who makes the decisions on Number of rayons Percent* internalize the budget splitting classes? reporting constraints, and have no Principal 16 27 reason to recognize the cost RED head 44 75 implications of their School inspector (under RED) 10 17 preferences. However, if RED economist 25 42 All RED schools were allowed to get 52 88 Teachers' union 3 5 what they want--more Other 1 2 teachers and, in many cases, Source: World Bank 2008a. longer workweeks--class size would fall, and education * The percentages do not add up to 100 because some RED heads gave multiple answers. spending would have to increase and/or be diverted from other expense categories to salaries. To prevent this, rayons generally have the final say on when classes can be split, as may be seen in Table 4.7. Class splitting is the mechanism by which schools, were they allowed to do so, could create more hours of teaching work for their teachers. Controls on class splitting are the means by which officials at the rayon level can keep staffing intensity under control. Thus, 88 percent of RED heads report that they, or another RED official, such as an inspector, make the decisions (in some cases jointly) on class splitting, while only 27 percent say that principals (even jointly) make the decisions. 4.28 Of course, the official class splitting rule stipulates that, for grades 1­4, classes be split once class size reaches 25, while for grades 5­11, the threshold class size is 30. However, the PETS finds significant variation in the class splitting rules reported by rayons, and less than half of rayon heads reported the official class splitting rules described above. Since the PETS collected data on the size of every class in every sample school, it is possible to apply any proposed class size rule to the data, "predict" the number of classes which ought to be observed in any grade in any school, and then compare this to the number of classes which are actually observed. In this manner, it is possible to determine whether the class splitting rules are accurately applied. 4.29 The results show that class sizes are consistent with the class splitting rule in 45 percent of Tajik general schools. Of these, almost half are very small schools where there are not enough pupils to do any class splitting under the class splitting rule. In 26 percent of schools, there are only "excess" classes. That is, some grades have "too many" classes, which should not have been created if the class splitting rule were strictly followed, but there are no classes that should have been split and were not. In 18 percent of schools, there are only "missing" classes, that is, some classes that ought to exist if the class splitting rule were followed do not, meaning that some classes are "too large"; but none are too small. Finally, in 10 percent of schools, there are both "excess" and "missing" classes, with the number of classes in some grades exceeding, in other grades falling short of, the norms. 101 4.30 Given that deviations from the class splitting rule are noisy, with some deviations in both directions, the question remains as to the overall effect. As shown in Table 4.8, the effect depends on the region of the country. In the RRS, Khatlon, and GBAO, more classes were actually observed than were predicted by the rule, so that "excessive" (compared to the rule) class splitting results in "too many" classes, and presumably too high a budget for teachers' salaries. In Dushanbe and Sogd, the opposite holds: fewer classes are actually observed than are predicted by the rule, so that a strict application of the rule would require class splitting and would increase the number of classes and the wage bill. Nationwide, however, the total number of classes in Tajik general schools is close to the number that would exist if the class splitting rule were consistently applied, so that the budgetary impact of the current state of affairs is probably small. Table 4.8: Deviations from Class Splitting Rules Result in a Few "Too Many" Classes Oblast Total classes Total classes "Missing" classes; "Excess" classes; actually observed predicted by class classes predicted results of splitting rule by rule but not excessive class observed splitting Dushanbe 5,405 5,976 698 127 RRS 18,177 17,905 570 842 Sogd 22,269 22,459 795 606 Khatlon 30,087 29,684 881 1,284 GBAO 2,744 2,331 3 415 Total 78,683 78,356 2,947 3,274 Source: World Bank 2008a. Consolidation of classes in per capita finance rayons. 4.31 One expected efficiency gain from implementation of per capita finance of education was the consolidation of classes, which would in turn reduce the total number of teaching hours, and therefore the amount of salaries paid to teachers. Under the traditional financing system, it would not be advantageous for schools to do this, because any money saved would revert to the local budget. Using the rich PETS data on class sizes in the sample schools, which include both participants and nonparticipants in the PCF reform, the class consolidation hypothesis was tested. In all regressions, the PCF coefficient was positive and significant at the 1 percent level, constituting statistical evidence that schools are responding to the PCF reform by consolidating classes. The size of the coefficient varies depending on the controls included, with the lowest value just under 1, meaning that the per capita finance schools are increasing class sizes, on average, by one student. 4.32 Despite the evidence on class consolidation reported above, however, it is important to note that the composition of PCF and non-PCF school budgets is similar. No budget category differs by more than a few percentage points between PCF and non- PCF schools. Either PCF schools have not deemed it advantageous to make dramatic changes in budget composition compared to the old system, or else they have not had time to do so. In particular, the largest categories, salaries and social insurance, have not 102 been reduced as a share of the budget. In other words, although PCF schools consolidate classes, this has not led to a reduction in the wage bill as a share of the total budget. Indeed, it appears to have increased it. Instead, the main source of savings in PCF schools appears to be utilities, especially heating and garbage removal. In addition, PCF schools spend a larger share of the budget on purchases of several types of goods and services. They spend almost twice as much on stationery, half again as much on inventory, and more than twice as much on travel expenses (possibly because, with greater responsibilities for procurement and money management, school administrators have more need to travel). Summary of PETS findings on teachers 4.33 Tajik schools generally have reasonable student-teacher ratios and teacher workloads. On average, class sizes are probably smaller than optimal, but there is notable variation across schools and regions. Decisions on class sizes, which are the key determinants of the per student budgetary cost of teachers' salaries, and indeed of general education as a whole, do not seem to strictly follow the standard rule that classes in grades 1­4 should have at most 25 pupils, and grades 5­11 at most 30. Rather, class splitting decisions are subject to local discretion, and sometimes result in "too few," sometimes "too many" classes per grade, relative to the rule. Teaching hours vary widely, and while most teachers are satisfied with the hours they teach, about one quarter to one third would like to teach more, which may create pressures for class splitting that are constrained more by a lack of classrooms than by budgetary considerations (see next section). School Infrastructure 4.34 Despite the large increases in enrollment over the last years, average school size has only gone up from 412 in 1991 to 442 in 2006, reflecting the fact that the number of schools has increased significantly in the same period from a total of 3,179 to 3,761 schools in the country to keep up with demographic trends. Here, as Figure 4.10 shows, Dushanbe stands out in having significantly increased average school size. Figure 4.10: School Size by Oblast and RRS, 1991­2006 1400 1200 1000 stned GBAO 800 tusfo Khatlon Sogd reb Dushanbe 600 RRS mu Total N 400 200 0 92 93 94 95 96 97 98 99 00 01 2 /0 03 04 05 06 91/ 92/ 93/ 95/ 96/ 97/ 98/ 99/ 00/ 01 02/ 03/ 04/ 05/ 19 19 19 1994/ 19 19 19 19 19 20 20 20 20 20 20 Academic year Source: National Statistics Committee of the Republic of Tajikistan 2006. 103 4.35 Nonetheless, Tajikistan's rising student population will require ongoing investment if its needs are to be met, since school buildings are already used intensively. Thus, estimates from the PETS show that 96 percent of schools operate in two or three shifts. In addition, nearly half of school principals (49 percent) reported that their school experienced overfull classes (with variation across oblasts). One of the options available to principals is, of course, to split classes in order to reduce their size. Here, however, principals reported that lack of classrooms and teachers prevented them from doing so: 85 percent of principals who reported overfull classes said that lack of classrooms was an obstacle to class splitting and 54 percent lack of teachers. 4.36 Capital renovations are rare: during the last five years, only 14 percent of school principals reported that a capital renovation was undertaken. Nearly three out of four principals (72 percent) reported that there had not been any capital renovation in the school since it had been built. Capital renovations are more likely to be undertaken in urban than in rural areas: 40 percent of principals from urban areas, compared to 24 percent from rural areas, reported that a capital renovation was undertaken. 4.37 The PETS sample was designed to oversample schools that were visited in 2004 as part of an EFS, and some questions in the EFS questionnaire were repeated in the PETS questionnaire. As a result, a small panel can be constructed of schools that were included in the sample of both surveys, and the evolution of principals' answers to certain questions can be observed. One important result emerges: the concern regarding poor facilities increased substantially relative to other problems schools face between 2004 and 2007. 4.38 Thus, in both surveys, Table 4.9: Proportion of Principals that Reported principals were asked about five Problems as Being Very Important specific issues that impacted the (percent) functioning of their school: lack Problem 2004 (EFS) 2007 (PETS) of teachers, professional skills of Shortage of teachers 56.0 54.4 teachers, poor infrastructure, Quality/skills of teachers 38.4 49.3 inadequate supplies, and lack of Poor infrastructure 25.5 47.9 heating in classrooms. The Inadequate supplies 50.6 66.2 response criteria were subjective, No heating of classrooms 63.3 49.7 with principals asked to report Source: World Bank 2008a. how important--very important, quite important, somewhat important or not important--each issue was for their school. Poor infrastructure was by no means the top concern for principals either in 2004 or 2007 (Table 4.9). However, the change in the importance of poor infrastructure between 2004 and 2007 is the greatest among all issues and points to a trend in the acuteness of schools' needs in this regard. Of the five problems listed, only one--heating of schools--seems clearly to have become less serious between 2004 and 2007, although that trend may simply be related to the severity of the winter in those two years. 4.39 In sum, the PETS findings regarding school infrastructure underline the current stresses in the system both in terms of capacity and maintenance, and the need 104 for significant attention on the part of the GOT but also the donor community for the foreseeable future. Extra-budgetary Transfers 4.40 The unequal expenditure per student described in an earlier section is confirmed by the PETS data, which further suggest, based on the pattern of funding inequality across regions, that school size in particular drives some of the differences (see Table 4.10). Thus, regions with larger schools spend less per student (with the exception of Sogd, which spends more per student than Khatlon although schools in Sogd are a little larger on average than schools in Khatlon). Table 4.10: Funding Per Student in General Education, by Region Region Average rayon per student funding Average school size (somoni per year) Dushanbe 105.31 1,598 RRS 145.01 302 Sogd 133.83 563 Khatlon 114.23 535 GBAO 267.88 215 Source: World Bank 2008a. 4.41 A second pattern Table 4.11: Per Student Funding in Tajik General observed is that, for any Schools, Unsubsidized versus Subsidized Rayons given school size, schools (somoni in 2006) in unsubsidized rayons are Average school size Unsubsidize Subsidized PCF better funded than schools d rayons rayons schools in subsidized rayons, as Under 150 pupils 268.84 272.58 shown in Table 4.11. From 150 to 225 pupils 168.91 Since unsubsidized rayons From 225 to 300 pupils 149.56 118.47 represent the more From 300 to 400 pupils 150.25 120.74 prosperous regions of the From 400 to 500 pupils 121.84 121.68 145.01 country, and although they From 500 to 600 pupils 138.36 93.12 146.69 pay net taxes that subsidize From 600 to 700 pupils 131.83 117.01 131.54 worse-off areas, it is to be From 700 to 1,000 pupils 119.18 110.03 119.39 expected that unsubsidized Over 1,000 pupils 97.54 113.42 rayons are not wholly Source: World Bank 2008a. deprived of their fiscal advantage and can afford more resources per student. A third pattern relates to PCF schools, which exhibit the highest levels of per student funding regardless of school size (with one exception--see Table 4.11). This finding probably resulted from the design of the per capita finance pilots, which sought to not immediately have an adverse impact on schools' budgets and create losers and thus opponents to the financing reform. However, the planned expansion of per capita finance of schools throughout Tajikistan presents a unique opportunity to address the issue of equity of public education expenditure in a phased and systematic manner. 105 Figure 4.11: Proportion of Schools that Received 4.42 While PCF schools Extrabudgetary Funding exhibit the highest per student public expenditures, 78% 94% extrabudgetary transfers are 76% 66% 68% 71% 69% especially frequent among 51% 57% 52% non-PCF schools. According to principals' declarations, over two thirds of all schools P dg l n lo have received at least one anbehsu RR nab lar F F at So PC Gbao Ur Ru To Khat such transfer in 2006 (Figure non-PC D 4.11). Schools under the PCF system generally receive the Source: World Bank 2008a. resources that are budgeted for them and are less in need of additional resources than schools under the traditional financing scheme. This might explain why only 52 percent of PCF schools received such a transfer compared to 71 percent of schools under the traditional financing scheme. Transfers are also less frequent in Dushanbe and Sogd than in the other oblasts, and especially frequent in GBAO. 4.43 Table 4.12 presents Table 4.12: Proportion of Schools that Received an schools that received an Extra-budgetary Transfer extrabudgetary transfer by geographic location and the (percent) source of the transfer. The Source of transfer Urban Rural Total most frequent providers are Government 10.3 7.5 7.8 students' parents followed by Private companies 1.8 0.1 0.3 Parents of pupils 41.6 30.6 31.9 international donors and Private sponsors/farmers 2.7 9.6 8.8 collective farms and state- School (income generating activity) 2.1 3.1 3.0 owned enterprises: respectively International donor 23.4 22.0 22.1 32, 22, and 20 percent of NGOs 0.0 6.5 5.7 schools reported that they have Collective farming/ 21.0 19.8 19.9 received a transfer from these state-owned enterprises sources. Transfers from Source: World Bank 2008a. students' parents are more frequent in urban areas than in rural areas. This might be related to a higher ability to pay in urban areas. By contrast, private sponsors and farmers are three times more likely to provide funds to schools in rural than in urban areas, and NGOs seem to support only rural schools. 4.44 The main purposes of the transfers are for school renovation (30 percent), construction materials (19 percent), and food products (15 percent). In terms of the providers of these transfers, students' parents are responsible for 50 percent of the transfers for school renovation, followed by international donors (15 percent). Construction materials are provided almost exclusively by collective farms and state- owned enterprises (54 percent) and students' parents (38 percent). Finally, international donors provide for fully 89 percent of the transfers for food products. 106 4.45 Table 4.13 reports the Table 4.13: Extrabudgetary Transfers per average transfers received per Student student by region and geographic (somonis) location: the sample average is Region Urban Rural Total SM 8 per student.42 There is a Dushanbe 17.6 17.6 great variation across regions. The RRS 117.5 13.4 24.4 average transfer per student is Sogd 10.4 2.9 3.8 higher in Dushanbe and the RRS Khatlon 16.3 1.9 3 than in the other regions: at the GBAO 0 1.2 1.1 extremes, schools in the RRS Total 33.7 4.7 8.2 receive on average SM 24 per Source: World Bank 2008a. student while schools in Khatlon receive only SM 3. There is also a variation in the amount received across geographic location: urban schools get on average SM 34 per student while rural schools get on average SM 5 per student. 4.46 In sum, therefore, in addition to observing disparities in per student funding across regions (albeit related to average school size in the region), better-off rayons, and especially PCF rayons, exhibit the highest spending per student. However, PCF schools receive less in terms of extrabudgetary transfers. The latter are slightly more prevalent in urban than in rural areas, and the amount per student varies greatly by region and geographic location, with urban areas and Dushanbe and the RRS clearly at an advantage. Students' parents are the most frequent source of extrabudgetary transfers and the lion's share of their contributions are dedicated to school renovation and construction materials. This is of course a further indication of the importance of improving school infrastructure particularly for the poorer segments of the population. D. SUMMARY 4.47 The present discussion has at its core the findings from the education PETS conducted in Tajikistan in 2007. Prior to implementation of the PETS, several important trends had been identified in the education sector: (i) rising student body size in an environment of already increasing education public expenditures, (ii) deteriorating infrastructure with further stresses expected as student numbers increase, and (iii) observed inequalities in public education funding. These trends portend further upward pressures on public education expenditures, particularly in the absence of efficiency gains. Thus, the PETS was conducted in an effort to assess potential efficiency gains, for example through identification of leakage of public funds. Since general education constitutes the bulk of public education expenditures, and within general education the lion's share goes to wages and salaries, a particular focus of the PETS was on the question of better deployment of staffing resources. In addition, the PETS questionnaire was designed to shed some light on the important issues of equity of expenditures and the state of school infrastructure in the country. 42 If we zero transfers are excluded from the sample, the mean reaches SM 17 per student. 107 4.48 Findings from the PETS indicate that the budget processes in the Tajik education sector are highly accurate down to the jamoat level. However, from the jamoat level down, the picture is less clear. The well-ordered system of recordkeeping and control of funds that exists, for example, in the rayons, does not exist at the stage of the budget process where goods and services are being distributed, in most cases by jamoats, to particular schools. Thus, jamoat accounting processes do not adequately carry financial accountability all the way down to the schools. There is an important exception to this rule: in the PCF rayons, the jamoat no longer plays a role, and schools know their own budgets and can report how they are executed. Thus, in PCF rayons the public expenditure tracking exercise was carried all the way down to the school level, and a high degree of consistency was found among budgets at all levels. 4.49 In terms of deployment of staffing resources, Tajik schools generally have reasonable student-teacher ratios and teacher workloads. On average, class sizes are probably smaller than optimal, but there is notable variation across schools and regions. However, decisions on class sizes do not follow strictly the standard rule. Instead, class splitting decisions are subject to local discretion, and sometimes result in "too few," sometimes "too many" classes per grade, relative to the rule. In total, however, the number of classes in the country does not differ greatly from the number that would be predicted if the standard rule were strictly applied. In addition, the PETS data provide statistical evidence that schools are responding to the PCF reform by consolidating classes. At the same time, although PCF schools consolidate classes, this has not led to a reduction in the wage bill as a share of the total budget. There is evidence, however, that PCF schools allocate less funding to utilities and more to goods and services. At the level of the school, the gains from the PCF reform may therefore not be in the realm of efficiency of expenditures but quality of education. 4.50 Certainly, as the PETS findings regarding school infrastructure underline, a high priority for general education in the country is the need to upgrade the capacity and quality of school infrastructure. In addition, particular attention must be paid, especially in the context of expanding the per capita finance reform to the entire country, to the issue of equity of public expenditures. The current state of affairs indicates disparities in public per student funding that are further exacerbated by the variation in extra-budgetary funding that schools receive. 108 Annex 4.1: Data Annex Table 4A.1.1: Number of Students, 1991­2006 (in '000s) Year Grades 1­4 Grades 5­9 Grades 10­11 Grades 1­11 1991/92 537.0 602.6 167.0 1,306.6 1992/93 519.7 604.3 133.1 1,257.1 1993/94 570.9 532.1 121.3 1,224.3 1994/95 593.5 562.6 116.6 1,272.7 1995/96 612.3 585.9 110.6 1,308.8 1996/97 638.7 581.8 106.4 1,326.9 1997/98 686.8 573.7 111.5 1,372.0 1998/99 690.3 659.1 85.2 1,434.6 1999/00 691.9 674.6 97.0 1,463.5 2000/01 680.1 691.3 131.6 1,503.0 2001/02 684.5 744.0 130.0 1,558.5 2002/03 694.9 798.6 124.2 1,617.7 2003/04 690.3 805.8 143.9 1,640.0 2004/05 693.1 802.0 158.2 1,653.3 2005/06 687.8 813.7 160.3 1,661.8 Source: National Statistics Committee of the Republic of Tajikistan 2006. Table 4A.1.2: Student-Teacher Ratio in Grades 1­11 by Oblast or RRS, 1991­2007 GBAO Khatlon Sogd Dushanbe RRS Total 1991/92 8.1 13.9 12.5 15.4 14.0 13.2 1992/93 7.5 13.2 12.2 15.2 13.6 12.7 1993/94 6.6 13.2 11.4 16.6 13.5 12.3 1994/95 7.5 14.5 11.7 19.5 15.0 13.3 1995/96 8.0 15.5 12.3 23.6 15.5 14.1 1996/97 8.1 15.5 12.1 25.3 16.9 14.3 1997/98 8.3 17.4 12.5 24.2 16.5 15.0 1998/99 8.3 17.8 12.6 24.9 16.0 15.1 1999/00 8.0 18.0 12.3 25.4 16.4 15.1 2000/01 8.0 18.6 12.4 25.7 16.1 15.3 2001/02 8.0 19.0 12.3 27.2 16.5 15.6 2002/03 7.7 19.8 12.4 27.0 17.2 16.0 2003/04 7.9 20.0 12.5 24.8 17.7 16.2 2004/05 9.1 19.6 13.4 27.3 18.3 16.9 2005/06 8.0 19.8 13.3 29.3 18.3 16.9 2007 (PETS) 10.8 18.7 12.4 26.0 17.6 16.4 Source: National Statistics Committee of the Republic of Tajikistan 2006; World Bank 2008a. Note: The ratios reported are based on numbers of teachers rather than full-time equivalents, and include schools for disabled children in the case of administrative data. 109 5. PUBLIC FINANCIAL MANAGEMENT REFORMS AND IMPLEMENTATION OF POLICY-BASED BUDGETING IN TAJIKISTAN A. INTRODUCTION 5.1 In the last few years, Tajikistan launched a number of important initiatives to better align public policies with budgets in efforts to promote efficiency of public expenditures. Public Administration Reform Strategy (PARS) approved by the Government in 2006 envisages important PFM-related reforms through the year 2015, that includes: (i) greater coordination between the budget and policies (ii) greater coherence in responsibilities for PFM (with a view to integrating investment planning and the budget process more closely; providing the MOF with overall responsibility for linking strategies and programs with the budget process); (iii) establishment of better procedures for financial decision-making in ministries; (iv) improvement of audit systems (strengthening internal and external audit; introducing new audit methods), (v) greater effectiveness of public procurement, and (vi) rationalization of revenue allocation mechanisms among levels of government. 5.2 These reform efforts are being supported by extensive donor assistance both lending as well as through technical assistance. Given the need to reform Tajikistan's public finance management system as well as past and ongoing lending and non-lending activities of various donors in broader PFM reform area, this chapter is structured in three main parts. The first part will review and highlight the status of PFM systems based on PEFA assessment conducted during 2007 and recent achievements and progress to date. The second part with cover issues related to Government's, the World Bank's and donors' efforts to develop an overarching strategy for the PFM. The third part will focus on non-lending and technical assistance, mainly focusing on the progress made and the next steps in the implementation of the policy-based budgeting (PBB) leading to meaningful Medium Term Expenditure Framework (MTEF). It will also cover associated next steps and linkages of ongoing PBB to a broader PFM work. Current Status of PFM system 5.3 The Public Expenditure and Financial Accountability Assessment for Tajikistan was completed in June 2007. The report identified key features of the PFM system, including the disjoint between capital and recurrent budgets (although some steps are taken towards closer integration), the relatively weak role of the PRSP, and the fragmentation of the system of government contribute to rather weak linkages between policies and budgeting in Tajikistan. It concluded that a number of areas covered by the assessment demonstrated good performance (the aggregate revenue outturn and comprehensive of information included in budget documentation). However, the assessment also highlighted serious weaknesses in the areas of predictability and control of budget execution (tax collection, control of commitments and payrolls, internal audit), accounting recording and reporting, external audit scrutiny and donor practices. 111 5.4 The government since has taken actions to improve the weaknesses of the PFM system identified by the PEFA reflecting its continued commitment to the PFM reform agenda. Due to a short time-frame between the last assessment and the time of this writing, this note highlights actions taken by the government to improve the PFM in the following areas: (a) credibility of the budget, (b) comprehensiveness and transparency, (c) policy-based budgeting, (d) predictability and control in budget execution, (e) accounting, recording, and reporting, (f) external scrutiny and audit, and (g) donor assistance. 5.5 Credibility of the Budget. Based on the preliminary data on budget outturns, limited progress has been made in this area. The deviations between actual and planned expenditure and between actual and planned revenues in 2007 significantly increased (see tables 3.1-3.3). Thus, the budget has yet to become a credible tool for policy planning. 5.6 Comprehensiveness and Transparency. To further improve the comprehensiveness and transparency of the budget, the Ministry of Finance (MOF) has established a Budget Classification Working Group on November 26th, 2007. It has already begun working on the administrative budget classification by defining the principles governing the segment. These principles are applied to the central government (republican) component of the administrative segment, fund segment, and geographic segment. However, the MOF has not yet begun work on applying the principles to local government budget organizations because it does not have information in Dushanbe on legal and operational relationships among them. It needs to gather information via local treasury offices and then consults with the local government budget organizations. 5.7 The Administrative Classification is Expected to Complete its Work is November 1st, 2008. Given the time required for design, approval, training and IT system changes as well as sufficient consultation with line ministries, the change in budget classification is envisaged to be introduced to the 2010 budget. The first draft of a new economic segment based on GFSM 2001 has been prepared and the Ministry of Finance is currently refining it. 5.8 Policy-based Budgeting. The government has pursued activities aiming at strengthening policy based budgeting that has been piloted in the education sector in 2007. The first MTEF document was disseminated to all sectoral ministries and local government institutions. In addition, the Ministry of Finance also published its first annual Budget Monitoring Report to increase budget transparency and ensure better reporting to donors and the general public. The government has issued a Presidential decree that integrates the MTEF cycle into the budget cycle for 2009-2011. As a result, the second policy-based budgeting paper covering 2009-2001 has been prepared and submitted to the Budget Commission and the Government by June 2008 as envisaged. In addition to education, three other sectors have been included on a pilot bases, i.e. health, social protection and agriculture. The education sector strategy including program costing is being updated with the assistance of UNICEF. Amendments to the Law on Public Finances have been approved by Lower Chamber of the Parliament and issued in May 2008. These amendments provide a solid legal framework for implementation of the MTEF and of the internal audit strategy (jointly with the European Commission). 112 5.9 Predictability and Control in Budget Execution. The Ministry of Justice registered the standard bidding documents for goods, works, and services and implementation regulations as referred to in the new public procurement law in January 2008. The new law will be gradually implemented to ensure "fully functional" procurement functions in six government agencies including Agriculture and Environmental Protection, Energy and Industry, Justice, Education, Health, and State Committee for Property and Foreign Investment. The fully functional procurement functions imply that (a) procurement units are established in the target agencies; (b) procurement officers in these agencies are identified, trained and accredited; and (c) the accredited procurement officers begin to procure goods, works, and services in accordance with the new procurement law. The Agency for Procurement of Goods, Works, Services and Organization of Stock-Exchange Activity with the support of the World Bank will conduct a training workshop on public procurement, thus enabling budget entities to conduct procurement according to the new law. Implementation of the new procurement law. 5.10 Accounting, Recording, and Reporting. The MOF has launched effort to modernize a public sector accounting system that is consistent with international standards with the support of the Public Financial Management Modernization Project. However, introduction of a modernized public sector accounting system will take time due to weak capacity. 5.11 External Scrutiny and Audit. A government working group was recently established to identify a strategy and action plan to establish an independent external audit in Tajikistan. The working group has now developed a strategy and action plan and it is being circulated and discussed in the government. It plans to complete the draft law on an independent external audit in the near future. Amendments to the Law on Public Finances have already been approved by Lower Chamber of the Parliament and is expected the amendments will be approved by Upper Chamber of the Parliament and signed by the President by end-March. These amendments provide a solid legal framework for implementation of the internal audit strategy. 5.12 Donor Coordination. Several bilateral and multilateral donors have been actively providing technical assistance to reform the PFM system in Tajikistan at both the Ministry of Finance and sector ministries in the social sectors (education, health and social protection). However, donors' activities both planned and ongoing are poorly coordinated due to lack of a PFM strategy that outline a road map and priority reform areas. Recognizing the need to improve donor coordination, the Government has contracted a local consultant to coordinate donors' activities beginning with developing a map of donors `activities in the PFM area during the interim period. 5.13 Looking Ahead. The government is certainly committed to the PFM reform as demonstrated by its effort to implement reform measures to address PFM weaknesses in some areas. However, the PFM reform agenda is enormous, compared to the government limited capacity and resources. There is an urgent need to prioritize the reform agenda and focus limited resource and effort on a high priority area. It is encouraging that the government will soon develop a public financial management reform strategy with the 113 support of development partners. The strategy is expected to provide a road map for the PFM reform in Tajikistan and thereby set a priority area that could be supported by development partners in the medium and long-term. Table 5.1: Comparison of Original Budgeted and Actual Expenditures, 2003-2007 2003 2004 2005 2006 2007 Budgeted primary expenditure (mln. tajik somoni) 571.4 859.0 1234.0 1501.0 1954.3 Actual primary expenditure (mln. tajik somoni) 668.7 984.0 1276.0 1619.0 2396.4 Difference between actual & budgeted primary 97.3 125.0 42.0 118.0 442.1 expenditure (mn. Somoni) Difference as % of budgeted primary expenditure (%) 17.0 14.6 3.4 7.8 22.6 Source: MOF/Treasury. Note: Primary expenditures exclude debt servicing payments and foreign and domestic funding of the PIP; funds shown include the CSIP, and the Social Protection Fund. B. PUBLIC FINANCIAL MANAGEMENT STRATEGY IN TAJIKISTAN 5.14 Effective Public Financial Management (PFM) systems are crucial to countries making progress in reducing poverty. Most Poverty Reduction Strategies (PRSs) that partner countries have developed recognize that sound PFM supports aggregate control, prioritization, accountability and efficiency in the management of public resources and delivery of services, which are critical to the achievement of public policy objectives, including achievement of the Millennium Development Goals (MDGs). In addition, sound public financial management systems are fundamental to the appropriate use and effectiveness of donor assistance since aid is increasingly provided through modalities that rely on well-functioning systems for budget development, execution and control43. 5.15 The "Strengthened Approach to Supporting PFM Reform" that has been developed by taking stock of the many years of donor-supported PFM reforms throughout the world places heavy emphasis on three pillars for ensuring their effectiveness: · A country-led agenda ­ a country-led PFM reform strategy and action plan; · A coordinated program of support ­ a coordinated IFI-donor integrated, multi- year program of PFM work that supports, and is aligned with, the government's PFM strategy; and · A shared information pool ­ a framework for measuring results that provides consistent information on country PFM performance, including progress over time. 43 Edited excerpt from Public Expenditure Working Group: "Supporting Better Country Public Financial Management Systems: Towards a Strengthened Approach to Supporting PFM Reform" (www.pefa.org) 114 5.16 The first pillar of the Strengthened Approach is therefore the establishment of a country-owned program of reform and a country-owned structure for managing the reform process. Country ownership and leadership are critical to successful reform, and an essential element for success is the development of an effective partnership between government and donors to support country ownership. In developing a country-led reform program, important considerations for all stakeholders are that the program is prioritized and feasible, and builds from the country's current circumstances and capacities44. 5.17 There is substantial consensus among the development community that PFM reforms need to be sensibly sequenced, with an effort to initially "get the basics right"45. In this regard, several methods exist for constructing PFM reform strategies based on a phased approach. In the case of Tajikistan, different donors have recommended a two-stage reform plan. The first stage would include further strengthening the budget preparation, developing the core treasury functions, and strengthening the public financial accountability structure through institutional and organizational measures. Once sufficient progress has been made on organizational structures, treasury functions, payment controls, expenditure tracking and the single treasury account, the second stage would be implemented, including introducing major accounting reforms to adopt international public sector accounting standards, developing an Integrated Financial Management Information System (IFMIS) in phases, and implementing tools for medium-term budgeting. 5.18 The development of a realistic and government-owned PFM strategy based on the principles outlined above and the establishment of a proper governance structure to implement it are of particular importance and urgency in the case of Tajikistan. Numerous diagnostic reports, such as the Public Expenditure and Financial Accountability Assessment (PEFA), the IMF's Report on The Observance of Standards and Codes ­ Fiscal Transparency Module (ROSC), and the Bank's Second Programmatic Public Expenditure Review, describe a situation with uneven progress across the areas of PFM. 5.19 Among PFM initiatives that have already shown some fruit are: (i) the establishment of the Treasury (1997) covering all components of the state budget; (ii) adoption of a new budget classification and chart of accounts based on IMF GFS 1986 methodology; and (iii) the implementation of tools to improve the sectoral allocation of expenditures under a medium-term expenditure framework. However, some relatively fundamental aspects of PFM lag considerably behind, such as the lack of commitment controls in budget execution, inadequate monitoring of payment arrears, a relatively high level of unreported government operations, the absence of an administrative classification of the budget, and inadequate internal and external control and audit, among others. 5.20 The inherited Soviet-style budgeting practices, characterized by ineffective macroeconomic and revenue forecasting tools, non-transparent budget allocation mechanisms, poor integration between capital and current expenditure planning and 44Ibid, para. 9. 45Schick (1997) as quoted in The World Bank PEM Handbook (1998) 115 budgeting based on input control, lower the efficiency of the newly introduced instruments for PFM. Although the basic budgetary legislation is in place, enforcement is not always adequate and regulations for budget execution are not comprehensive. The macroeconomic and medium-term forecasting functions of the MOF are not well developed, and the budget process does not effectively integrate the different allocation mechanisms, including PRSP and PIP. Basic treasury systems are still evolving and the computerization of the PFM treasury system is at an early stage and many budget and treasury operations, especially at the local level, are still manual. 5.21 Over time, the lack of a proper vision of reform sequencing, in particular those reforms that can be called "first generation" and "second generation", is likely to lead to poor outcomes for key reform initiatives, misunderstandings between the client and the donor community, wasted technical cooperation resources, and ultimately a loss of confidence on the very foundations of the reform effort. The urgent need for a PFM strategy in Tajikistan is not only to better understand the sequencing of the modernization efforts; it is also necessary for enabling the Government to understand the overall breadth of the effort required for PFM modernization and to make strategic choices based on it; in other words, the importance of seeing "all the pieces of the puzzle". For example, an internationally-accepted PFM practice is the existence of an independent external audit body with a broad mandate to promote probity and accountability, which does not exist in Tajikistan; a proper strategy would help visualize its importance in the overall PFM context and raise awareness towards such critical areas that have been relatively overlooked46. 5.22 Furthermore, there is a mismatch between the ambitious PFM Reform agenda being promoted by different donors and the Government's capacity to implement reforms. As donors have largely been focused on specific elements of PFM, the overall implementation capacity of the public sector have not been properly considered. The existence of a PFM Reform Strategy would enable all stakeholders to appreciate the magnitude of the combined effort and better gauge the absorptive capacity of the client. 5.23 The Ministry of Finance of Tajikistan is aware of the importance of undertaking a strategy development process. During the first quarter of 2008, consultancy services are being procured through the IDA-financed Public Sector Reform Project (PSRP) to support the Ministry in the preparation of the PFM Reform Strategy, following international good practices and in consultation with other development partners. The Minister of Finance has designated the Deputy Minister for Public Debt as the person in charge of guiding the preparation of the strategy. Once prepared, the Ministry would pursue its formal adoption by the appropriate legal instrument. 5.24 Despite challenges faced by Tajikistan in the PFM area, the MOF is moving ahead with the definition of the Strategy and has assigned a working group made up the managers of most of the related departments in the MOF. As part of the preparation of the Public Financial Management Project, the Bank has identified a preliminary set of areas of action that should be part of this strategy, together with their sequencing, 46 There is a current initiative supported by the Bank and SECO to conceptualize a new external audit function. 116 although both the content and sequencing will undoubtedly be enriched during the ongoing process. The preliminary benchmarks of the strategy are the following: 2008: · Implement budget classification revisions and adopt an organizational (administrative) classification · Continue pilots for medium-term planning and budgeting · Establish targets for adoption of IPSAS accounting and reporting standards · Establish an independent Supreme Audit Institution 2009: · Strengthen macroeconomic and fiscal forecasting · Develop linkages between macroeconomic and fiscal forecasting and the formulation of the budget · Adopt an integrated Budget Classification and Chart of Accounts · Establish the fundamentals of an internal auditing framework 2010: · Introduce modern and transparent methodologies for debt management · Develop and initiate implementation of a new PFM IT platform · Strengthen the budget execution process, including a commitment control system linked to the procurement process · Consolidate the Treasury Single Account and improve cash management · Adopt a formula-based transfer system to sub-national governments 2011: · Complete integration of current and capital budgets · Identify and address fiscal risks in the overall fiscal framework 2012: · Finalize implementation of IT platform · Introduce a program structure to the budget · Develop tools for dissemination of budget information to the public, including interactive web application 5.25 Broad international experience shows how important it is that a PFM reform program is owned and well steered by the Government. Without strong governance to shape and prioritize the reform program, the system easily becomes overloaded with work as the proponents of each project press for best practice methods to be applied to their area of interest. To promote government's ownership of the PFM reform, sharpen focus on achievement of results, improve transparency, timeliness and responsiveness of decisions related to PFM agenda, the Ministry of Finance is expected to establish a PFM Reform Council. The details of the composition of the Council will be known during mid-2008. 117 Box 5.1: IT Solutions to Support PFM Reforms At present, the Ministry of Finance operates a single-entry treasury system which, although reliable, is limited in the information it can produce and does not take advantage of the latest developments in information technology. With the support of the European Union, similar systems are being installed in regional treasuries. This is a temporary solution until a robust financial management information system is deployed, connecting the central treasury with the 73 local treasuries and generating all the information for accounting, monitoring, and decision-making purposes. The FMIS solution proposed for the PFMMP will be built as a core treasury system linked with existing/improved information systems of the MoF to improve the operational and allocative efficiency, and will be expanded over time as institutional capacity allows. All central key budget organizations and 73 territorial treasuries will be connected (around 300 system users, 400 user workstations, located at 103 connected sites) to the MoF through a web based system. Financial departments will have access to system functions through a web portal. Implementing the new FMIS necessitates the existence of an adequate information and communication technology infrastructure. The Government is aware of the significant shortcomings of the existing infrastructure at present and is committed to address them through measures such as the full deployment of a modern communications network, provision of electricity generators and fuel to all local treasuries, and any other measures needed to ensure the successful operation of the system. The Government of Tajikistan has made important progress towards the implementation of a TSA, resulting in substantial efficiency gains over the last few years. However, not all Somoni-denominated general government accounts have been incorporated yet. The MoF intends to expand the coverage of the TSA while respecting the levels of autonomy that different levels of government may be entitled to. In terms of cash management, the MoF established a Cash Management Unit in early 2006, but further technical assistance to this unit is necessary to achieve the recommendations made by several international organizations. Finally, the MoF intends to implement a robust commitment control process based on international good practices and linked to the decentralized procurement system, thereby safeguarding fiscal discipline and improving the availability of information for budget monitoring within the scope of FMIS. C. POLICY-BASED BUDGETING (PBB) Background 5.26 Based on the outcome of the 2004 Public Expenditure and Institutional Review (PEIR) as well as recommendations of the 2007 PPER on introduction of the MTEF and regional lessons learned from the development of the MTEFs, the GOT and the World Bank, supported by the U.K. Department for International Development (DFID) and the United Nations Children's Fund (UNICEF), agreed to undertake a series of initial budget reforms in the PFM area beginning with initial PBB exercise (phase 1 of the MTEF). As a response, an international resident consultant worked during six months with the MOF and developed a comprehensive reform plan and MTEF strategy note. Starting end-2006, with DFID/WB/UNICEF financial and technical resources, a group of international and local consultants assisted the GOT in the MTEF rollout with the pilot in education sector that resulted in drafting the first PBB paper (albeit called the MTEF paper) covering 2008­10. This paper was approved along with the draft state budget for 2008 in September 2007. The second policy-based budgeting paper, initiated by the 118 President's Decree in December 2007, has resulted in preparation on the PBB paper in May 2008 and submission to the Budget Commission in June 2008. It is expected that it will be approved by end-June 2008 and would serve as a solid base for preparation of the annual 2009 budget. What Has Been Achieved So Far? Legal foundations 5.27 In 2007, the World Bank and other international donors (notably DFID and UNICEF) continued to work together with the GOT on deepening public finance reforms and PBB in particular. Two Presidential Decrees of September 2006 (along with overall PBB guidelines) and a Decree of December 2007 provided a solid legal and conceptual basis for the PBB rollout in education and consequent strengthening the core PBB techniques and processes. In addition to this, a number of internal decrees and orders to this effect were passed both at the level of the MOF and the Ministry of Education (MOE) that helped streamline PBB work within ministries. Finally, the European Commission (EC) and the World Bank recently collaborated to ensure that changes to the public finance law on MTEF (making it part of the budget process) and internal control and audit were drafted and submitted to the GOT. This work was closely linked within the Programmatic Development Policy Grant. Amendments to the law have been adopted in May 2008. Procedural and administration aspects 5.28 In order to facilitate the ongoing PBB work, two working groups have been established within the MOF and the MOE to specifically undertake and supervise the PBB-related activities. Both working groups are led by senior technical staff of a respective ministry and report directly to the ministers. Internal ministerial decrees provide an outline of the roles and responsibilities of different units within the ministries as well as guide them through the required deliverables and timelines in order to ensure smooth process of drafting the PBB document and education sector policy note. This also includes clear guidelines for PBB preparation process, including macroeconomic framework, resource envelop, budget ceilings, standard templates, and so forth. 5.29 The MOF (macroeconomic and financial statistics unit) can now prepare initial macroeconomic framework as well as the sector resource envelope. The envelope is normally broken down into sectors, including details on gross state budget allocations, wages and salaries, capital investments, external financing, and extrabudgetary funds . This allows the MOE to prioritize its sectoral programs according to the expenditure ceiling (for example, for education) and formulate details of their ongoing and new programs in the form of a sector policy note. Sector working groups at the both ministries play an integral part in drafting and subsequent discussion of policy notes. Below is a generic action road map (key milestones) and a timeframe for the preparation of the PBB by respective ministries within the annual budget process. 119 5.30 As far as local governments are concerned, their interactions with sectoral ministries (such as the MOE) within MTEF and annual budget processes, as well as key input/output indicators and budget forms have been revised and outlined in detail in the comprehensive budget instructions for four sectors including education, health, social protection, and agriculture. In addition to sector-specific instructions issued by MOF at end-February 2008, general budget instructions for local government institutions have also been issued in the same period on formulation of the draft budget for 2009 and development of budget forecasts for 2010­11. These budget instructions also (i) serve as the basis for identifying all key budgetary organizations (KBO) in each respective sector, which should closely interact with MOF and respective LLM on all budget-related matters; and (ii) present a clear timeline for local budget submissions. Since about 82 percent of the education budget is at the discretion and supervision of local governments, these instructions are commonly viewed as an important element of timely and effective budget planning process and its subsequent monitoring at the end of the fiscal year. Table 5.2: Government's Key Milestones and Timeline for the Policy-based Budgeting Date / Period Activity / Milestone February 1, 2008 The macro-framework is prepared and issued by Ministry of Finance in coordination with Ministry of Economic Development and Trade (MoEDT) and National Bank of Tajikistan February 20, 2008 Revised MTEF guidelines (circular) with tentative expenditure ceilings (including PIP) for the preparation of the MTEF document are submitted to the Budget Working Group at the Ministry of Finance March 20, 2008 Draft sector policy notes and sector priorities are prepared by responsible lead line ministries (LLMs) and submitted to Ministry of Finance for follow-up bilateral discussions April 20, 2008 Final sector expenditure ceilings are issued by Ministry of Finance and sub- sector indicative ceilings are negotiated with LLMs May 10, 2008 Preparation and submission of medium-term sector programs for 2009-11 by lead line ministries to Ministry of Finance June 1, 2008 The comprehensive MTEF document for 2009-11 is drafted by Ministry of Finance and submitted to the Parliamentary Budget Commission of the Republic of Tajikistan end-June 2008 The MTEF document for 2009-11 is approved by the Government and disseminated among central and local government institutions through Ministry of Finance Source: MTEF, Ministry of Finance. 5.31 To improve transparency of budget execution and reporting mechanism, the World Bank consultants assisted the MOF in the development and issuance of the first 120 Budget Monitoring Report for 2006. The first report had been disseminated among all central government institutions and the general public. The report for 2007 will feature the GOT's quasi-fiscal stance and external debt situation, development assistance, and donors' budget support (external financing), in addition to a comprehensive state revenue and expenditure overview. The second issue of the annual Budget Monitoring Report is expected by June 2008. 5.32 Finally, relations with/between the Parliamentary Budget Commission and all other PBB key stakeholders have not been covered by any donor working in the PFM area. More specifically, since the Budget Commission is ultimately responsible for review of draft annual budgets and three-year PBB documents, it is imperative their active involvement and related capacity-building activity are ensured by ongoing and/or prospective donor initiatives. Capacity building for PBB 5.33 Given the fact the PBB is entirely a new paradigm for the public finance in Tajikistan (and many other developing countries), capacities for proper budget planning throughout the GOT, and notably in the MOF and more so in the MOE as a pilot, are extremely weak. Therefore initial capacity-building efforts were concentrated on the basic concepts of strategic budget formation and prioritization. 5.34 A series of introductory workshops for MOF and MOE staff were conducted during 2007 as well as two high-level workshops in May and September 2007. The latest high-level seminar held on September 18­19, 2007 in Khujand with the support of World Bank Institute (WBI) brought together half of the Tajik Cabinet, provided a road map for further PBB activities, and presented the GOT's Action Plan and a detailed PBB timeframe for 2009­11 adjusted to annual budget processes. At these two workshops, case studies and experiences from a number of transitional and former Soviet countries have been shared by guest international experts with the central government, key line ministries, and KBOs at the local oblast and rayon levels. In addition to this, study tours to Moldova, the Slovak Republic, and Armenia were organized for various levels of mid- level specialists and deputy ministers involved in the PBB process from MOF and MOE. The study tours were fully devoted to the financial system and budgetary relations between various levels of government. In particular, the Tajik delegation was introduced to budget planning system, including strategic planning, the medium-term expenditure program, and budget interdependence, as well as interlinks between PBB and PCF (such as in the Slovak Republic). 5.35 It is also important that proper attention is paid to capacity building of local governments and, specifically, rayon finance departments (RFDs), rayon education departments (REDs), technical staff in rayon and oblast level khukumats. For instance, UNICEF in conjunction with the World Bank conducted series of two-day introductory workshops for local governments (REDs) in October 2007, aimed at increasing their knowledge about ongoing institutional reforms in the state budget planning process, as well as general characteristics and mechanisms of PBB implementation in the education sector pilot. Other activities supported by donor agencies in this area are briefly described 121 in the next subsection. Therefore, institutional capacity and technical skills need to be developed among key government officials at the line ministry and local government levels to complement the shared international expertise on implementing PBB in transitional countries. International donors should provide continuous support to increase knowledge, qualifications, and general skills of public officials at the central level of MOE and MOF. 5.36 As part of the ongoing technical assistance, a comprehensive capacity-building program is being developed based on training needs as emerged from the first 20 months of PBB exercise with the first pilot in the education sector. At the present, donors are revising the capacity-building action plan targeted at increasing skills and competencies of specialists and mid-level officials at central and local government levels. Donor coordination and coherent approach to budgeting reforms 5.37 A joint April 2007 donor mission worked closely with the GOT and agreed to coordinate efforts between active and prospective donors within the offered technical assistance package in the medium term. It is of critical importance for all donors to work in the same direction and continue building up and capitalizing on the achievement so far. As an effort to meet this objective, a local consultant was hired in November 2007 under the World Bank's PHRD grant to help the MOF coordinate expanding and increasingly complex donor involvement in the PFM area. A PFM activities' mapping exercise has been conducted and a comprehensive matrix prepared that would serve as a starting point for better coordination and harmonization of donor assistance in this area. 5.38 Representatives of the World Bank, UNICEF, EC, and DFID have recently reinstated their pledges for continuing provision of technical and financial assistance for the roll-out of the first phase MTEF to other social sectors in Tajikistan. There are also other donors focusing specifically on local government capacity building regardless of the sector without officially being involved in initiatives under the general PFM umbrella. Nevertheless, further coordination of related institutional activities between donor agencies is necessary to effectively use donor funding under the current circumstances and Tajik context. Given the challenging task of building Tajikistan's capacity in moving towards a MTEF and PBB as well as the complexity and a range of issues involved, there is a need to give space to all interested and active donors and allow for a collective effort to reach the ultimate goal of building effective PFM system. Recognizing these needs, the GOT and international donors are also keen to establish solid linkages between PFM reforms and various strategic documents (for example, PRSP , National Development Strategy (NDS), National Strategy for Education Development (NSED), and so forth) that equally contribute to the prioritization process of medium-term programs and their subsequent financing from the state expenditure budget. Key Outcomes and Lessons Learned 5.39 A number of key outcomes have surfaced within the last 20 months of implementation of the initial PBB exercise. First, instilling ownership of the process by 122 the ministries has proved key to a successful implementation of the PBB so far. With continuous formal and hands-on on the job training and guidance, key stakeholders now feel comfortable leading the exercise. Having understood the purpose and importance of linking policy and budgets, the pilot ministry has the will and the needed understanding of the nature of linkages with sector policies as well as national and sector strategies. 5.40 As a result, the GOT has decided to expand the first-stage PBB expertise to other sectors, such as health, social protection, and agriculture. While this effort is commendable and shows a true commitment on part of the GOT, initial needs assessment needs to be conducted first; and for these sectors to be meaningfully included in the process, lessons from the education pilot need to be drawn to effectively synchronize the PBB action plan and milestones with formulation of the annual budget to avoid any delays and/or bottlenecks in the budgetary process. 5.41 GOT officials involved in the PBB process have initial understanding of the MTEF philosophy, process, and expected outcomes. Key technical staff have been trained in linkages to PFM reforms, PRSP/NDS, sector strategies (in education), and PIPs, and exposed to various experiences from the region, including from the Slovak Republic, Moldova, Armenia, Georgia, Russia, and Latvia. Technical and sector staff understand the mechanics of the PBB and have learned successful cases and, in fact, can continue with the PBB 2009­11 independently with some ongoing advice and guidance from the international consultants. 5.42 However, there is still lack of capacity and technical skills at all levels of the bureaucracy to undertake further MTEF reforms and move to more advanced stages of PBB. Thus, a comprehensive capacity-building program needs to be developed and funded to ensure proper technical staff at all levels (both national and regional) are adequately trained and prepared to further the reports. 5.43 It is imperative that more hands-on guidance and training be provided across all social sectors and levels of government in specific areas such as planning, prioritization, budgeting, costing, and expenditure efficiency analysis. During day-to-day interactions with officials at the MOF, some concerns were expressed about effective use of available international expertise and consultants; the result of which had been to prepare (train) local trainers who would eventually spread the knowledge and skills to the grassroots level (such as local finance departments and secondary educational institutions in rayons). 5.44 Three key lessons emerged as a result of the ongoing PBB pilot in Tajikistan: (1) In order to have a meaningful connection between policies and budgets, existing budget fragmentation needs to be addressed. (2) Adequate resource allocation and policy making can not go hand-in-hand unless current and capital budgets are synchronized. 123 (3) Any advance in PBB can not succeed unless technical capacities both at the national (ministry) and regional (oblast and rayon) level are developed and strengthened in all budget-planning and execution stages. Fragmentation of the budget 5.45 Budget fragmentation continues to be one of the major constraints in Tajikistan. From the PBB point of view and given existing given budget systems in Tajikistan, this poses major difficulties as there is no single institution to which a budget ceiling could be applied. In turn, that makes any link between policy and expenditure allocations a very vague concept. 5.46 Taking education as pilot, the first step was for the MOE to be formally approved as a LLM responsible for sector-wide budgeting and planning in education. Further, a country-wide list of all KBOs in education has been prepared and approved by the MOE and clear guidelines issued as to respective roles and responsibilities of KBOs and the LLM. However, this was only the first and the most necessary step towards addressing the issue of fragmentation in education budgeting process. Experience with the second cycle of the PBB revealed the underlying complexity of changing the relationships within the budget cycle and adopting PBB elements. Certain lessons could be drawn at this stage that might be useful to consider when introducing other sectors to this exercise. Even if these lessons learned might not necessarily be relevant for other sectors, they do provide essential knowledge about the potential obstacles and pitfalls. 5.47 Three major categories of budget fragmentation specific to education could be shown and all require own ways in addressing them. These patterns depend on the type of the KBOs involved and the schemes they use to communicate with the MOF when it comes to budgeting. Of the 40 KBOs in education 17 are local authorities, and 32 KBOs comprise the mix of education institutions and other ministries. 5.48 The first category involves independent education institutions (such as Tajik State National University and Tajik State Medical University) and their interactions with the MOE. Since these institutions are subject to all relevant education policies and strategies, the issue is how to bring them under the MOE umbrella and apply PBB principles in a single exercise and to put in place appropriate procedures and regulations. 5.49 The second category of budget fragmentation deals with sector ministries that have a number of education institutions that are immediate recipients of the funds from education budget. In this case the MOE education priorities and policies can not be easily applied due to contradictions with sector policies of the ministry they belong to. Examples of these institutions include quite a number of music and sport schools subordinated to the Ministry of Culture or State Committee on Sports respectively. Given what little influence the MOE has over policy formulation in these sectors, the mechanisms capable of bringing budget formulation for these institutions to MOE's hands are a subject for further elaboration. Failure to do this will significantly hamper success of PBB implementation in education sector since the MOE will still lack control over some public funds devoted to education. 124 5.50 The final category comprises the 17 KBOs--local authorities--demanding special relationships with the MOE. These local authorities receive capitation and investment grants by sector and are responsible for allocation of these funds according to the regional priorities. From this angle, implementation of the PBB mechanisms at the sub-national level turns into a very arduous task with multiple complex questions to address. In particular, how to address ad hoc priority setting of the sector policies at the regional level and, instead, synchronize them with national priorities and ensure allocation of funds accordingly. This issue calls for revision of existing concepts of budgeting and planning procedures between national and sub-national authorities. Further implementation of the policy-based elements is expected to be quite painful for the local authorities and therefore requires staged approaches, better coordination, and clear guidelines. 5.51 Another dimension of budget fragmentation lays in the way extra budgetary expenditures and unspent funds are handled and managed. It is quite normal for extra- budgetary funds to be allocated without proper policy, priorities, or rules established, or on an ad hoc basis, creating "dual" budgeting systems. These practices undermine credibility of budgets and unless addressed will continue creating perverse incentives through the budgeting process. Merger of current and capital budgets 5.52 Fragmentation of the budget into KBOs differs across social sectors and geographical regions and creates immense problems with sector budget policy formulation and constraints on the creation of fiscal space on a sector-by-sector basis. Capital and recurrent expenditure also appear to have diverged. They are presented in both the resource envelope and state budget as the Central State Investment Program (CSIP) and the PIP. The latter is mostly financed by the international donor community, while the former represents a small part of the state budget capital expenditure. Despite the fact that these diversions are common in most developing countries (especially in the FSU region), some developments have already been made by the GOT at the request of MOF. 5.53 In particular, the recent state budget law for 2008 presents an integrated budget where externally financed donor investment is included as part of the overall budget envelope along with budget support grants, public loans, extra-budgetary funds, and tax and nontax revenues (see Article 1 and Article 13 of the Law "On the State Budget for 2008" of the Republic of Tajikistan). Therefore, the integration process of all capital expenditure has certainly begun but needs strengthening at the sector level where most line ministries still do not practice regular recording and/or tracking and monitoring externally financed inflows of loans and grants, which complement state budget expenditure. 5.54 Discretion over distribution (and redistribution) of the state budget across current and capital expenditure categories is still not within the sector ministry's domain. Therefore, any ongoing PBB and further budgeting reform exercises should require integration of capital and current expenditure in the formulation of a common 125 resource envelope with predetermined tight sector expenditure ceilings, such as in the education sector. It is also imperative that all central level government institutions involved in the budgetary process have user-friendly electronic databases of all externally financed PIP resources, as well as technical tools, methodology, and competent staff, for tracking these expenditures on a regular basis. This will ultimately allow MOE to reallocate given sector-bound resources among economic budgetary categories depending on its sector priorities across new, ongoing, and completed programs. This will, in turn, give more freedom and flexibility to sector ministries in the budget formulation and execution process, and eventually allow for a more responsible and efficient decision making over public finances at the central government level. Capacity for budget planning and execution 5.55 International donor agencies including the World Bank, DFID, UNICEF, until quite recently the EU and others recently intensified their involvement in the PFM reforms arena. Many of these efforts focus on capacity building in one way or another as part of their technical assistance. While each individual activity is important, it is now vital to synchronize these efforts to address the most binding constraint to PBB--weak capacity in budget planning and execution. Therefore, it has been decided to create a unified roadmap of all donors' activities in the area of PFM with the help of a local consultant at the MOF. This exercise is currently ongoing and is expected to be completed by May 2008. Below is a snapshot of key undertakings in the area of capacity building and technical assistance in the PFM-related areas on the part of the donor community 5.56 In the past several years, the EC has been committed to continuous reforms in the public sector and currently implements the Sector Program for Social Protection (SPSP) 2007­09 (Action Program DCI 2007). The purpose of the SPSP is to assist the GOT in improving the design, management delivery, and effectiveness of national policies in the social protection sector in line with the NDS and PRS. The program also aims to strengthen PFM in the aforementioned sector on intervention. In addition, the Public Internal and Financial Control (PIFC) and Management Information System/Treasury System (Action Program, Tajikistan 2006). EC also launches a new three-year MTEF project to introduce all social sectors (health, education, and social protection) to the principles of PBB and proper budget execution practices through the team of ECORYS consultants. Finally, the Macroeconomic Modeling and Forecasting project aims to strengthen the macroeconomic and fiscal forecasting capacity of the Ministry of Economic Development and Trade (MoEDT) and MOF. 5.57 In addition to the preparation of the Fiduciary Risk Assessment in 2004, the DFID financed: (i) the World Bank Structural Adjustment Credit 2 (SAC2) critical evaluation report, which also assesses early lessons learned of the Institutional Building­ Technical Assistance: Phase 2 (IBTA2) budget administration reform project; (ii) a summary analytical note as part of the PARS PFM reform; and (iii) continuous consultancy services of international experts in relation to the introduction of MTEF reforms and its implementation in the education sector as a pilot through the MOE and MOF. 126 5.58 Since PBB was first piloted in the education sector, UNICEF has played a significant role in building capacities and providing oversight of the processes. In particular, technical assistance has been provided to address gaps in linking regional and national priorities, as well as replication of regional and area-specific issues in respective national strategies (such as the PRSP and the National Strategy for Education Development [NSED]). The UNICEF team currently reviews policy and capacity linkages against resource management processes within three sectoral ministries (MOH, MOE, and MoLSP), and supports international and local consultants for the MOE to help prepare and submit to MOF comprehensive education sector priorities and three-year development strategies, building in parallel much needed technical capacities of relevant mid-level staff. 5.59 In addition to core donors in the PBB process, other donors have also contributed to reforms in the area of overall PFM and local government budgeting: · The U.S. Agency for International Development (USAID) is addressing local governance and decentralization issues through the Local Governance and Citizen Participation project (started in 2002), which has been executed through the Urban Institute (UI); the project envisages introduction of participatory budgeting, and improvements in overall budget policy at the local government level and technical budget work by means of small grants, technical capacity building, and overall technical assistance. · The Swiss Cooperation Office (SCO/SDC) co-financed the issuance of the PEIR and several Public Expenditure Management Peer Assisted Learning Network (PEMPAL) initiatives, some of which pertained to capacity building of local government officials through study tours to other transitional and developing countries on areas of MTEF and PBB. · The Asian Development Bank (ADB) is providing technical assistance for Local Government Finance Capacity Building; the project is administered through the United Nations Development Program (UNDP) and the Open Society Institute (OSI) provides coordination of the respective activities on the ground. Experience with and Constraints to Strengthening Sector-Based Planning 5.60 As a legacy from the Soviet times, the MOE in Tajikistan (similar to other CIS countries) has broad responsibilities and powers in terms of setting overall education policy. However, direct engagement of the MOE in the budget formulation process is basically limited to subordinated budgetary organizations only. Moreover, the finance department of the MOE that is tasked with these functions has only very limited interaction with other departments dealing with education policy and is more focused on budget-level activities related to subordinated budget organization. Thus, it is not surprising that the MOE overall is not concerned with the budget implications of its education policy initiatives, and that the MOF and local departments of finance have much more control over allocations across sectors and line-item expenditures. As a result, it is often the case that even good education policies proposed and approved by the cabinet remain unfunded. Against this backdrop, the major challenge in introducing elements of the MTEF and PBB is to gradually transfer responsibility for budget 127 formulation to the MOE, and thus insure effective and efficient budget allocations (outcomes) within the sector. 5.61 Tajikistan's weak capacity and poor technical skill base is one major constraint for efficient PBB and sector-based planning. Therefore, the first and most critical task was to introduce strict budgetary ceilings for overall education sector public expenditures and allow the MOE to propose policies within the given budget constraint. From the recent experience with the education pilot it became evident that the MOE needs to be strengthened to carry out basic public expenditure analyses of existing programs as well as costing the proposed new policy initiatives. As a lead ministry responsible for education policy, the MOE should be given the power over budget allocations within the sector through setting priorities in pubic expenditures, budget allocation methodologies, and service delivery standards, regardless of institutional settings of service delivery (autonomous institutions, private providers, or locally subordinated schools). 5.62 A second major constraint associated with PBB is the lack of clearly defined and realistic goals within the priority programs. Therefore the immediate task is to insure that programs can be realistically designed and achieved, taking into account medium-term projections of the overall sector resource envelope and projected changes in main cost drivers. To achieve this at least in the medium term, all departments within the MOE dealing with policy need basic training in strategic policy planning and costing. It important to keep in mind that each level of education is very specific in terms of public policies and economics involved, so it will not be possible to consolidate all of these skills within one planning department. 5.63 One last major constraint to introducing an MTEF in the education sector has to do with the institutional settings of secondary education. Currently all policy-related decisions belong to central government while the governance and actual financing is assigned to the local budgets. It is clear that local authorities with no control over education policy will treat PBB tools with suspicion. On the other hand, central agencies, in particular the MOE, cannot control effectively the cost at local government levels within the existing framework. The solution to this problem can be seen in the ongoing efforts to introduce per capita financing and calculating budget appropriations by local authorities based on per capita allocations of public expenditures on secondary education. One of the fundamental solutions to this problem is to transfer responsibility for financing to the central government and decentralize actual implementation to the local authorities or even to the budget organizations. Moving Forward 5.64 Key issues moving forward with the PBB exercise and reasonable transfer of the responsibilities for the development of the education sector MTEF within the annual budget cycle include the following: (i) strengthen interactions between the MOF and the MOE as LLM responsible for the overall policy issues and corresponding distribution of public resources; (ii) strengthen interactions between the MOE, line ministries (with educational establishments under their supervision), and KBOs that interact with the MOF independently. 128 5.65 At present, both the MOF and MOE are prepared to perform the first type of interaction under MTEF that does not demand interaction among MOE, other line ministries, and education sector KBOs. This type of interaction is based mainly on the adequate understanding of the priorities of the education sector development and funding as well as a sound information base on the indicators and benchmarks of the sector development. 5.66 The most important immediate step for both ministries is to streamline their work to provide aggregate distribution of budget resources in education; and for the MOE to start interacting with other KBOs and regional authorities (local budgets) through roll-out of an intra-sectoral budget instruction that should be developed based on the general budget instruction issued by the MOF at the beginning of the 2008 exercise that sets the overall framework. 5.67 At the same time, the MOE currently is not prepared to fully take over the annual budgeting process in the sector--that is, build the annual budget on the base of the PBB that is fundamental for the successful implementation of the MTEF overall. It should be expected that in 2008, the annual budgeting exercise will be performed jointly with the MOF, to provide knowledge and information transfer to the MOE. Thus, full involvement of the MOE both within the sectoral MTEF and annual budgeting process is the main task for the 2010­13 PBB and this issue should be in the front line of the overall technical assistance provided by the donors. The key focus of the technical assistance should be on the following areas: (i) support the MOF in knowledge and procedures transfer to the MOE; (ii) help the MOE and sector KBOs achieve institutional preparedness (including through instruction and interaction with other line ministries and local KBOs) and build capacity in PBB by conveying knowledge of prioritization, costing, budgeting, revenue and expenditure analysis, and so forth. 5.68 In the medium term, key steps that need to be taken include the following: (i) merger of capital and recurrent budgets, and gradual integration of the PIP and the state budget; (ii) transfer responsibility for capital expenditures planning to the MOF,47 a step that would require strengthening capacity of the ministry to effectively coordinate ongoing and planned externally funded investment projects; (iii) ceasing the practice of not distributing additional budget resources between the sectors in the process of budgeting (both on PBB and annual budget levels) and keeping these resources as reserves to be distributed later on, usually after adoption of the budget;48 (iv) fundamental change in the revenue distribution structure to implement basic principles of fiscal discipline, replacing current annual adjustment of revenues according to centrally planned expenditures with a system of subsidies from the central budget to local budgets; (v) introduction of a legally binding, permanent, distributional scheme of revenues between central and local governments; (vi) fundamental attention given to improving tax forecasting and tax administration and corresponding capacity building in the relevant government bodies (including the Ministry of Finance). 47 Currently under Ministry of Economic Development and Trade. 48 The first attempt to do so by the MOF in March 2008, in the MTEF budget instruction sent to the MOE and other ministries participating in the MTEF pilot sectors. This practice needs to be continued and reflected in the annual budget expenditure ceilings. 129 D. SUMMARY 5.69 Since 2005 Government of Tajikistan committed itself to improving its public finance management systems and in particular budget processes and institutions. The current state of PFM system in Tajikistan is characterized by a number of strengths and weaknesses and the World Bank and development partners have been providing both lending and non-lending assistance to improve parts of PFM systems. In particular, Tajikistan progressed well over the last 20 months since the inception of the policy-based budgeting (PBB) exercise that focused on introduction of key elements of the PPB with education sector serving as a pilot. The process of establishing a working PFM Council is under way that is tasked with the development of the PFM strategy and action plan that would serve as a guiding framework for all efforts by the donors and the Government to improve its PFM functions. Authorities are committed to continue with the PBB pilots beyond education expanding to three other sectors, namely health, social protection and agriculture. It is clear that the challenges ahead are daunting as the country embarks at more difficult reform terrain in the area of public finance management reforms. Key challenges relate to the need to strengthen institutional capacity to undertake comprehensive PFM reforms at all levels; the need to better coordinate ongoing and planned activities within the framework of the overall PFM strategy (sequencing and prioritization); and, finally, the need to promote and ensure government's ownership of the PFM reform agenda. 130 6. THE IMPLICATIONS FOR PUBLIC FINANCES AND FISCAL SUSTAINABILITY OF ELECTRICITY SECTOR INVESTMENTS AND REFORMS A. INTRODUCTION 6.1 The purpose of this paper is to evaluate the long-term fiscal implications for Tajikistan of the capital investments associated with the Central Asia-South Asia (CASA) 1000 project, which consists of investments in major new hydroelectric and thermal power generating plants (HPPs and TPPs) and reforms in the domestic electricity sector. The CASA 1000 project involves generating electricity in Tajikistan and the Kyrgyz Republic and transmitting it via a 1,000 MW overhead transmission line (OTL) to Afghanistan and Pakistan. A smaller 220 kV OTL from Tajikistan to Afghanistan (AFG-220) is already under construction. Although the CASA 1000 OTL could transmit power to Afghanistan and Pakistan using the summer surplus power production from Tajikistan, eventually large new investments in power generation will be needed in Tajikistan, for three reasons: (i) to meet increasing domestic demand; (ii) to provide sufficient power for all year round exports to Afghanistan and Pakistan; and (iii) to bridge the large seasonal power shortfalls in winter (when demand is higher but supply from hydroelectric power plants is lower), for which investment in thermal power projects (TPPs) is required. 6.2 The Government of Tajikistan (GOT) will contribute to the investment costs of the CASA 1000 OTL, alongside the other states that are participating in the project. The GOT will almost certainly contribute to the investment costs of new power projects, alongside private investors, external public sector entities, and international financial institutions. The key reforms to the domestic electricity sector involve, among others, raising the average domestic tariff to the level required to cover the long run average incremental cost (LRAIC) of supply. 6.3 The issues addressed in this annex are threefold: i) How much will it cost the GOT to meet its share of the investment requirements of the CASA 1000 OTL and new power generation plants, and what will be the impact on the sustainability of Tajikistan's external public debt given that the GOT is likely to require external finance to fund its contribution to these investments? ii) How much additional budget revenue will be generated from the electricity investments and reforms? The major potential sources of budget revenue are higher collections of value-added tax arising from the increased domestic electricity tariffs, income (profits and debt payments) earned on the GOT's contribution to the capital investment in the power projects and a possible royalty levied on electricity production or exports. How much additional expenditure will government incur as a result of the investments and electricity sector reforms; from debt servicing, government utility bills, and energy compensation payments to low-income consumers? Hence what will be the net impact on the government budget? 131 iii) What are the potential contingent liabilities for the GOT entailed in these projects and reforms? Contingent liabilities could arise from two sources in particular: (i) power purchase agreements (PPAs) signed between power generation companies and public sector utilities, obliging the latter to take or pay for power; and (ii) the failure to eliminate the quasi-fiscal deficit if average domestic tariffs are not raised to the level required to cover the LRAIC of supply. 6.4 We evaluate the potential fiscal consequences under two scenarios. The first scenario entails the construction of the CASA 1000 OTL but no new power generation capacity other than the Sangtuda I HPP. Under this scenario, a limited volume of exports to Afghanistan and Pakistan would be supplied from the existing surplus summer production in Tajikistan and the Kyrgyz Republic. The second scenario entails the CASA 1000 OTL plus three new hydro power project projects (HPPs) and two new thermal power projects (TPPs) constructed in Tajikistan, which would enable much larger surpluses for export to be generated and exports to be supplied all year round to Afghanistan and Pakistan. Obviously the second scenario will involve much greater investment costs, revenues and risks than the first scenario. 6.5 Because the investments in generation and transmission capacity covered in this paper are long term in nature (with productive lifespans of 30 to 50 years), our analysis is also long term. We make projections out to 2030.The estimates and projections of the technical and financial aspects of the power projects--for example, the capital costs, operating costs and annual generation capacity--are taken from existing sources; notably the Regional Energy Export Potential Study (REEPS) (World Bank 2004) and the prefeasibility study for the CASA 1000 OTL conducted by SNC-Lavalin (SNC-Lavalin 2007). The objective of this paper is not to revisit the evaluation of the financial viability of these projects but to extend the existing analysis to cover the fiscal consequences in a comprehensive manner. 6.6 Our two scenarios correspond closely to two of the scenarios evaluated in the SNC-Lavalin Report. The SNC-Lavalin base case is similar to our second scenario, involving major new HPPs and TPPs, with two relatively small differences: we have excluded the Zerafshan HPP but assumed that the coal-fired TPP will have a larger capacity. Our first scenario is essentially the same as the no new generation scenario of SNC-Lavalin. Another difference with both scenarios of the SNC-Lavalin report relates to projections of domestic demand, which are explained in section 3. 6.7 The annex is organized as follows. Section 2 provides details of the proposed new investments in the electricity sector which are the main focus of this study. Section 3 sets out projections of electricity supply, domestic demand, and exports, which are necessary inputs into the fiscal projections. The impact on external debt stocks and external debt servicing, and hence debt sustainability, of the government borrowing required to finance the GOT's contribution to the investment projects is evaluated in section 4. In section 5 we make projections of the impact of the new investments and domestic tariff reforms on government revenues and expenditures, and hence also their net fiscal impact. Section 6 makes projections of the quasi-fiscal deficits in the electricity 132 sector arising from the failure to reform domestic tariffs and the potential contingent liabilities associated with new generation projects. Section 7 concludes. B. ENERGY INVESTMENTS 6.8 Table 6.1 lists all of the energy-related investment projects covered in this paper, which comprise one transmission line and five power generation projects. The total capital cost of the CASA 1000 OTL is estimated at US$455 million. We assume that GOT will fund the capital costs of that part of the OTL that is located in Tajikistan, which is estimated at 27 percent. Hence GOT's contribution to the capital cost will be US$123 million. 6.9 A financing proposal for the Sangtuda II project has already been prepared, which indicates that this project will be built on a build, operate, and transfer basis by the Government of Iran (GOI) (or public sector companies owned by the GOI). GOT will contribute US$40 million of the US$220 million total capital cost and will assume ownership of the project (and thereby acquire the revenues) after twelve and a half years of operation. Table 6.1: New Investments in Energy Projects and GOT Contribution (US$ million) Project Capital Cost GOT Share GOT Contribution Existing Assets Cash CASA 1000 Total 455 27% 123 Sangtuda 220 40 Coal Fired TPP 2,250 25% 563 Dushanbe Reconversion 309 25% 77 Rogun I 1,035 32% 250 79 Rogun II 2,220 32% 550 161 Total 4,489 800 1,042 Sources: SNC-Lavalin, World Bank. * Total cost of OTL including parts outside Tajikistan 6.10 The coal-fired TPP, which is essential to bridge the winter shortfalls in supply capacity, will cost approximately US$2.25 billion.49 The reconversion of the Dushanbe combined heat and power plant, to use coal as its primary fuel, will cost an estimated US$309 million. We assume that the GOT will contribute 25 percent of the capital costs of both these projects. The two phases of the Rogun HPP will cost a combined total of US$3.25 billion, which includes US$800 million of existing assets owned by the GOT.50 We assume that the GOT will contribute 10 percent of the new capital costs in addition to its existing assets, which will give it a 32 percent share of the capital of the Rogun 49 This is based on international cost comparisons of US$1200 to US$1500 per kW of installed capacity. The installed capacity will be 1,500 MW. We have taken the higher estimate of the cost range because cost pressures (in dollar values) are rising due to higher commodity prices and a depreciating dollar. 50 These assets were constructed during the Soviet period: their valuation is disputed. 133 projects. Summing up these requirements implies that the GOT will contribute a total of just over US$1 billion in 2007 prices, in addition to its existing assets, to these projects, of which more than half is attributable to the coal-fired TPP. This investment will be spread over the years 2008­2021: hence on average the GOT will fund about US$80 million per year, in 2007 prices, of new investment in electricity projects over the course of this period, which is the equivalent of about 2 percent of GDP. 6.11 We do not include the possible Zerafshan (also called Yova) HPP, which would cost US$260 million, because it is unclear whether the project will be implemented.51 We also exclude from our analysis the capital investments in the Sangtuda I HPP, the North- South OTL within Tajikistan, the AFG-220 OTL, or the ongoing rehabilitation works being undertaken by Barki Tajik (the government owned domestic electricity supply corporation), because the GOT is already committed to these projects, in most cases work on them has commenced (and in some cases is already near completion) and their financing implications are already incorporated in the debt sustainability analysis. We also do not include as GOT investment the urgent future investments by Barki Tajik (BT) to rehabilitate and upgrade the existing distribution system. Instead we assume that this will be funded from Barki Tajik's cash flow, once the domestic tariff has been increased to the LRAIC, given that this increase in the domestic tariff is designed to provide the resources needed for this essential investment. The scheduling of the investments in the projects listed in table 6.1, which is required for projecting indicators of debt sustainability and government revenues, is derived from the data in the SNC-Lavalin Report (in particular appendix 6.1).52 C. SUPPLY, DOMESTIC DEMAND, AND EXPORTS 6.12 To make projections of fiscal revenues and costs we require projections of potential supply capacity, domestic demand, and the surplus that will exist for exports. These projections are set out in table 6.2. The supply projections are based mainly on those in the SNC-Lavalin report. The existing annual capacity of about GWh 16,660 is assumed to be increased by GWh 622 as a result of the rehabilitation of the existing plant at Nurek (which will raise installed capacity by MW 150). This will be supplemented by the power generated from Sangtuda I (GWh 2,700 per year) and the HPPs and TPPs listed in table 6.1. We assume that the two TPPs will not be run at full capacity in the summer months once Rogun II starts to produce power, in 2019, in order to reduce the seasonal imbalance in power generation that will be exacerbated by additional HPP capacity. The marginal cost of producing power from the TPPs is higher than that of the HPPs, so if there is surplus seasonal power it is more efficient to reduce output from the TPPs rather than the HPPs. As a result of the power generated by these new investments, total annual power supply from plants in Tajikistan rises from about GWh 19,600 in 2010 to GWh 42,500 in 2030. 51 The Chinese government had proposed building this project financed by a loan to the GOT but this was rejected by the GOT. 52 We assume that the coal-fired TPP will have a capacity of 1,500 MW, rather than 1,000 MW as stated in the SNC- Lavalin report. The SNC-Lavalin report does not include Sangtuda II, therefore we have taken the technical and cost projections for this project from World Bank sources. 134 6.13 To project domestic demand, we utilize the projections made by the World Bank up to 2016.53 After 2016, we assume that the power requirements of the Tajik Aluminum Company (TALCO) will remain constant, that the requirements of the GOT will increase at the same rate as real GDP (because government services expand at this rate) and that demand from all other consumers, both commercial and residential, is determined by income growth (with an income elasticity of demand of 0.6) and real price changes (with a price elasticity of demand of -0.15).54 Consequently, domestic demand, including system losses, rises from just under GWh 14,000 in 2010 to just under GWh 20,000 in 2030. 6.14 These domestic demand projections are much lower than those made by SNC- Lavalin. The World Bank projects domestic demand in 2016 at GWh 15,146, compared to SNC-Lavalin's projection for that year of GWh 22,773 (that is, about one third higher). The consequence of lower domestic demand projections is that, for a given generating capacity, there will be a larger surplus available for export. It therefore improves the economic viability of the CASA 1000 OTL irrespective of whether new investments are made in generation capacity although it may make new investments in generating capacity less attractive, because they will be more heavily dependent on exports. 6.15 The excess of supply over demand is potentially available for export. The surplus rises from GWh 5,600 in 2010 to a peak of GWh 25,800 in 2022 before declining to GWh 22,800 in 2030 (the decline is due to rising domestic demand with supply constant after 2022). We do not take into account seasonal factors that could reduce the potential exportable surplus. Part of the surplus will be exported to Afghanistan through the AFG 220 OTL. These exports are projected to reach a maximum of GWh 2,600 per year by 2016 and then remain at that level. The remaining power surplus must be exported through the CASA 1000 OTL, if it is to be exported. In addition, there is likely to be an exportable surplus of power generated in the Kyrgyz Republic, which will also require transmission through the CASA 1000 OTL. The Kyrgyz power surplus is projected by SNC-Lavalin to rise from kWh 1,000 in 2010 to a peak of GWh 12,800 in 2022 before declining to about GWh 7,000 in 2030 (SNC-Lavalin report, chapter 3). The annual capacity of the planned CASA 1000 OTL will be about GWh 11,400. If exports through the CASA 1000 OTL are in line with those projected by SNC-Lavalin, they will account for about 43 percent on average of the combined power surpluses from Tajikistan and the Kyrgyz Republic, after deducting the power to be exported through the AFG 220 OTL. Hence if the generating capacity of the new HPPs and TPPs is to be fully utilized, it will be necessary to expand the transmission capacity of the CASA 1000 OTL (assuming a market will be available in Pakistan for additional power). 6.16 The SNC-Lavalin Report concludes that construction of the CASA 1000 OTL will be viable even if no new investment in generating capacity takes place in Tajikistan and the Kyrgyz Republic. In this case there is still an exportable surplus, albeit much lower and confined to the summer months. Without new generating capacity, the exports 53 These projections were made in a note setting out comments on the SNC-Lavalin report. 54 These are the same income and price elasticities of demand used by the World Bank in its comments on the SNC- Lavalin report. 135 through the CASA 1000 OTL average GWh 3,487 per year and those through the AFG 220 OTL GWh 1,151 per year (SNC-Lavalin, 3-22). 6.17 The viability of the new HPPs and TPPs in Tajikistan (listed in table 6.1) is clearly dependent on there being guaranteed export markets. As can be seen from table 6.2, domestic demand during the entire period 2010­30 can be met largely from the existing generating capacity plus the Sangtuda I HPP, although there would be some winter deficits. Hence there is no economic justification for building additional generating capacity just to serve the domestic market. D. EXTERNAL PUBLIC BORROWING 6.18 The GOT's total financial requirements to meet its contribution to the investment costs of all of the power related projects are estimated at US$1.04 billion in 2007 prices, spread over the period 2008 to 2021. We assume that all of these costs will be funded by external public borrowing, because there is very little scope in the state budget (which is funded from domestic revenues and a very limited amount of budget support from donors) to accommodate investment costs of this magnitude and almost no scope for domestic borrowing given the very shallow domestic financial sector. 6.19 We assume that the GOT will fund its contribution to the CASA 1000 OTL by borrowing on IDA terms. For all other projects, we assume that the external loans required by the GOT will be contracted from nontraditional lenders on terms that are similar to those applied to the loans from China which were contracted by the GOT in 2006; that is, a 2 percent interest rate, 20 year maturity and a five-year grace period for repayment of principle. This does not mean that we believe that China will provide all of the required finance; other potential lenders include the Russian Federation, Iran, and regional development banks. 6.20 To evaluate the impact of the borrowing requirements of power sector investments on external debt sustainability, we utilize the projections in the 2007 Debt Sustainability Analysis (DSA) baseline, adjusted for developments in 2007,55 and add to these projections the projected external debt stock and debt servicing requirements of (i) the borrowing on IDA terms for the CASA 1000 OTL. and (ii) the borrowing for all the power projects listed in table 6.1. These debt projections are presented in table 6.3, while figures A2.1 and A2.2 show two of the key debt indicators used to evaluate debt sustainability, the net present value (NPV) of external debt to GDP, and exports respectively. Note that the data in table 6.3 are in current prices. 6.21 The measure of external debt we focus on this analysis is public and publicly guaranteed (PPG) debt; hence it excludes private sector debt that is not guaranteed by the GOT or Tajik government agencies such as the National Bank of Tajikistan (NBT). It is likely that some of the non-GOT investment in the power projects will be funded by external debt, but we assume that this will not be guaranteed by the GOT or government 55 The adjustments we make for 2007 developments are (i) removal of US$400 million of additional semi-concessional loans from China which were included in the DSA baseline but not disbursed, and (ii) inclusion of US$328 million of private sector cotton finance debt which was guaranteed by the National Bank of Tajikistan (the central bank). 136 agencies and, therefore, we exclude it from our analysis of debt sustainability. While the private sector debt incurred by these projects could impose a burden on the Tajik economy, this is likely to be less of a problem than PPG debt, because the repayment of the former is dependent upon the energy projects that it finances being able to earn sufficient revenue to repay their debts, at least some of which will be earned in foreign exchange from the export of electricity. PPG debt, on the other hand, remains a liability of the GOT irrespective of whether sufficient foreign exchange or revenues can be earned by the projects to cover their debt servicing. 6.22 The adjusted DSA baseline projects an increase in the debt burden, in terms of the NPV of PPG external debt to GDP, which peaks in 2009 at 42 percent of GDP (compared to the threshold of 40 percent, which is deemed sustainable for a low- income country rated as a medium performer on the basis of Country Policy and Institutional Assessment [CPIA] ratings) and then gradually declines until the early 2020s when it flattens out at about 23 percent of GDP (the lower line in figure A2.1). The NPV of PPG external debt to exports exhibits a similar pattern: a peak at 180 percent (compared to the sustainable threshold of 150 percent) in 2008 followed by a decline until the ratio flattens out at about 134 percent in the early 2020s (see figure A2.2). Total debt servicing over the projection period (2008­30) averages US$138 million in 2007 prices, with a sharp peak in 2011 and 2012 because of the need to repay the NBT guaranteed cotton sector loans56 and the start of repaying the principle of the Chinese loans (these loans have a five-year grace period for repayment of principle). As a share of exports of goods and services, external debt service averages around 20 percent from 2008 to 2012 (compared to the sustainable threshold of 15 percent) but then declines and flattens out at around 7 percent by 2020. See table 6.3 for the details. 6.23 The borrowing for the CASA 1000 OTL makes little difference to these debt indicators. At most it adds 1.4 percentage points of GDP to the NPV of PPG external debt to GDP ratio (in 2010) and 5.6 percentage points of exports to the NPV of PPG external debt to export ratio (also in 2010). In terms of debt servicing, this borrowing adds less than US$1 million per year until 2019, when principle repayments start to fall due, and then rises to average just over US$3 million per year during the 2020s. Hence the borrowing requirement for the CASA 1000 OTL alone does not undermine external debt sustainability. 6.24 The borrowing required to fund the GOT's contribution to all of the power projects has a much more significant impact on debt indicators, for two reasons: first the amounts involved are much larger (eight times that of the CASA 1000 loan), and second, the borrowing for the power generation projects is on less concessional terms than that for the CASA 1000 OTL, and hence per nominal dollar of loan, the NPV of the debt stock and the debt servicing requirements is higher. The borrowing requirements for all of the power projects add 11.4 percentage points of GDP to the NPV of PPG external debt to GDP ratio (in 2014) and 53 percentage points of exports to the NPV of PPG external debt to export ratio (also in 2014). The addition of these loans delays the reduction of the NPV of PPG external debt to GDP ratio below the sustainable threshold 56 We assume that these loans will be repaid in equal instalments over five years beginning in 2008. 137 of 40 percent by five years (from 2010 in the adjusted DSA baseline to 2015) and that of the NPV of PPG external debt to exports to below the sustainable threshold of 150 percent by 15 years (from 2012 in the adjusted DSA baseline to 2027). The debt servicing requirements of all of the projects add an average of US$34 million in 2007 prices during the period 2008­30 (about 25 percent of the requirements in the DSA baseline), but this does not raise the debt service to exports ratio above the threshold for debt sustainability of 15 percent. 6.25 To summarize, the borrowing required for the CASA 1000 OTL alone does not have any significant impact on the sustainability or debt burden of Tajikistan's PPG external debt. However the GOT's borrowing requirements to fund its contributions to all of the power projects worsen long-term debt sustainability, which is already regarded as being fragile in the 2007 DSA. In particular, new borrowing for power projects keeps the debt to exports indicator above the sustainable threshold until the second half of the 2020s, thereby increasing the risk of debt distress. E. IMPACT ON GOVERNMENT REVENUE AND EXPENDITURES 6.26 The energy-related investments and electricity sector reforms will have a long- term impact on both government revenues and expenditures, though several channels. GOT revenue will be mobilized through the income earned by government investments in the energy projects, from possible taxes on electricity exports and the impact of higher domestic electricity tariffs on VAT. GOT expenditures will be affected by the debt servicing costs of borrowing to fund energy projects and the impact of increased electricity tariffs on government utility bills and compensation payments to low-income electricity consumers. 6.27 This section reports the projections for each of these components and then summarizes their net impact on the government budget. Tables A2.4 and A2.5 present the revenue and expenditure projections assuming that all of the new investments in power projects listed in table 6.1 are undertaken. Subsections 5.1 and 5.2 look at the impact of revenue and expenditure respectively, both under the scenario where all of the investment projects listed in table 6.1 are undertaken, while subsection 5.3 summarizes the net fiscal impact. The final part of this section, 5.4, examines the fiscal impact under the scenario in which the CASA 1000 OTL is constructed but no new investments in generation plants takes place and the projections pertaining to this scenario are set out in table 6.6. Government Revenues Income earned from power projects 6.28 The GOT should earn income on its own investments in each of the power projects (unless the projects are unable to earn sufficient revenue to cover their operating costs). For each project, with the exception of Sangtuda II, the GOT will earn income from debt (interest and repayment of principle) and equity in the project (for all 138 projects the debt equity ratio is assumed to be 70/30, with the GOT contribution to capital also divided in this ratio). 6.29 To make projections of this income, we have projected the earnings of each of the power generation projects, by multiplying projected annual output by the assumed generation tariff (the tariff charged by the project to the purchaser of electricity). We assume that the HPPs and TPPs will sell their power at 3 cents/kWh and 4 cents/kWh respectively in 2007 prices. We then deduct fixed and variable operating costs to compute operating revenue. The fixed and variable operating costs are taken from the SNC- Lavalin report. Operating revenue is then first allocated to debt servicing costs, determined by the debt/equity ratio of the project and an assumed interest rate of 8 percent, with the residual allocated to equity. The GOT share in the capital of the project (shown in column 2 of table 6.1) is then used to calculate the GOT's share of the returns to debt and equity. For the reasons given below, we assume that all of the power projects will be exempt from corporate income tax, hence the returns to equity holders are equal to operating revenue minus debt service payments. 6.30 The Sangtuda II HPP will be constructed on a build, operate, and transfer modality by the GOI. All revenues from the project will accrue to the GOI for the first 12.5 years of its operation, after which ownership and the full revenue stream will be transferred to the GOT. 6.31 For the CASA 1000 OTL, we assume that the revenues earned, minus the operating costs, will be divided between the four participating governments according to their respective shares of the capital cost of the project. Revenue is projected as the annual transmission of power, given in the SNC-Lavalin report 3-24 (which is based on the transmission capacity of the OTL), multiplied by a transmission tariff of 2 cents/kWh in 2007 prices. The transmission tariff of 2 cents/kWh is taken from the projections in the Fieldstone report (2007, page 43), which identifies this as the tariff required to cover operating expenses, debt service with an interest rate of 8 percent, and a return to equity of 16 percent. 6.32 The returns to GOT capital invested in the power projects in 2007 prices and as a percentage of GDP are shown in tables A2.4 and A2.5 respectively. In 2007 prices, government revenue climbs rapidly from US$16 million in 2011 to almost US$150 million in 2015 and then to US$200 million in 2020. It continues to rise before flattening out at around US$245 million in the second half of the 2020s. As a percent of GDP, this revenue rises to a peak of 3 percent in 2016 and then gradually subsides back to 1.3 percent by 2030. 6.33 These projected earnings should be regarded as a maximum that the GOT could mobilize from its investments in the energy projects, because they are based on all of the HPPs operating at maximum potential capacity (given a load factor of about 40 percent) and the TPPs operating at maximum capacity for at least 8 months of the 139 year, which would generate an average of GWh 40,000 per year from 2015 to 2030.57 As domestic demand in this period averages approximately GWh 20,000 per year, there will be an exportable surplus which averages almost GWh 20,000 per year. This compares with the transmission capacity of the CASA 1000 OTL of only GWh 11,388 per year and that of the AFG 220 OTL of GWh 2,628 per year. In addition, the power generation system in the Kyrgyz Republic will also generate an exportable surplus which would need to be transmitted through the CASA 1000 OTL. Hence unless additional capacity is added to the CASA 1000 OTL to transmit electricity exports, the new power generation projects will operate at much less than full capacity and the returns to capital will be substantially reduced as a result (see also section 6 which discusses potential contingent liabilities arising from underused capacity). Export royalty 6.34 We have made a projection of the revenue that could be earned by levying a royalty of 0.25 cents/kWh (in 2007 prices) on all exports of electricity generated in Tajikistan. Export volume is determined by the capacity of the CASA 1000 OTL and the AFG 220 OTL. We have assumed that two thirds of exports transmitted through the CASA 1000 OTL will be generated in Tajikistan alongside all of the power transmitted through the AFG 220 OTL. Hence Tajik-generated exports will amount to a maximum of about GWh 10,000 per year. At its maximum, from 2017 onwards, the export royalty will yield US$26 million in 2007 prices (a maximum of 0.4 percent of GDP). However, it is not certain that a royalty can be levied on exports without undermining their commercial viability. If the capacity of the CASA 1000 OTL were to be increased to unable the entire surplus generated in Tajikistan to be exported, the revenues from the export royalty would be much higher, by an average of about US$36 million per year during 2010­30. 6.35 We have assumed that a royalty will not be levied on domestic sales of electricity (which are already subject to a 20 percent VAT) because this will increase the domestic tariff which is required to cover the LRAIC of supply. If the GOT were to levy a royalty of 0.25 cents/kWh on supply to the domestic market, it would yield approximately double the amount of revenue mobilized from export sales. Corporate income tax 6.36 We have assumed that none of the power projects will be liable to pay corporate income tax (implying that they will be granted an exemption from the income tax law). We make this assumption because the financial projections of the costs of generation and transmission compared to the likely export purchase price of power suggest that the projects may only be viable, in terms of yielding a sufficiently high rate of return to investors, if they are exempt from corporate income tax. Investors are concerned with their after-tax rate of return, hence if the projects are liable to pay corporate income tax the tariff charged for generation must increase to maintain a given rate of after-tax return 57 We assume that, once Rogun II comes on stream, which will greatly increase generation capacity, the two TPPs will be closed down during the summer months (for a maximum of four months per year) when the supply-demand surplus will be at its largest, because the marginal cost of thermal generation is higher than that of hydropower. 140 to capital. Given that the hydropower projects must charge a tariff of 3­3.5 cents/kWh to generate pre-tax returns in the region of 15 percent, that the transmission tariff is assumed to be 2 cents/kWh, that the purchase price in Pakistan is unlikely to exceed 6 cents/kWh, and that we also incorporate an export levy of 0.25 cents/kWh (all in 2007 prices), the scope for levying corporate income tax without reducing after tax rates of return below the level needed to attract private capital, given the substantial risks involved for investors, would appear to be limited. This does not mean that we are necessarily in favor of granting income tax concessions to power sector investors; this should only be done if it can be justified following a comprehensive social cost benefit analysis of each project. We have made this assumption in order to ensure that the fiscal projections are not overly optimistic in terms of the potential revenue which can be mobilized. 6.37 BT is liable to pay income tax. Its revenues should increase substantially when the average domestic tariff is raised to the LRAIC. However we do not incorporate any additional income tax payments by BT into our projections, because BT will need this additional revenue to fund the necessary repair and maintenance of the transmission and distribution infrastructure, and therefore reduce the quasi-fiscal deficit.58 Value-added tax (VAT) 6.38 The increase in the average domestic tariff to the LRAIC (calculated as 2.3 cents/kWh in 2003 prices) will boost VAT paid on electricity sales. VAT is charged at 20 percent on all domestic sales of electricity. In gross terms the additional VAT arising from the tariff increase will generate an average of US$56 million per year from 2010 onwards. However, the net gain in VAT will be much smaller, because the increase in electricity tariffs will reduce the purchasing power of consumers for spending on other VATable products. We assume that consumers are budget constrained, and hence the marginal propensity to consume is constant with respect to the increase in electricity tariffs. Hence all of the increase in electricity costs to private consumers leads to an equivalent reduction in spending on other consumer products. We further assume that the reduction of spending on other products will fall exclusively on VATable products, rather than non-VATable products such as basic foods. Accordingly, most of the gain in VAT arising from higher electricity tariffs is offset by lower VAT payments on other consumer products. The net gain in VAT averages US$11 million in 2007 prices from 2008 to 2030. 58 Whether BT's taxable income (profits) increase with the higher tariff will depend on whether the additional repair, maintenance, and rehabilitation work undertaken by BT, and funded out of its higher revenues, is booked as current expenditure (and therefore reduces taxable income) or capital expenditure (which does not reduce taxable income). Irrespective of the accounting treatment of this expenditure, BT will need all of the additional revenue it earns from the increased domestic tariff to fund essential rehabilitation and maintenance of the supply network. Therefore, to the extent that its cash flow is diminished by tax liabilities, the quasi-fiscal deficit that arises from the failure to undertake essential rehabilitation and maintenance will be higher (see section 6). Hence there is no long-term net budgetary gain from imposing corporate income tax on BT. 141 Government Expenditures Debt servicing 6.39 The methodology and assumptions for projecting the debt servicing costs of energy projects to the budget are set out in section 4. In 2007 prices, debt servicing rises from US$3 million in 2010 to almost US$50 million in 2020 before flattening out at around US$55 million in the first half of the 2020s. It then begins to fall back in the second half of the 2020s, for two reasons: (i) debt service payments are set in nominal terms and so fall in real value over time because of inflation; and (ii) as the nominal value of debt is reduced by the repayment of principle, interest payments fall. Government electricity payments 6.40 The costs of the GOT's own purchases of electricity will rise substantially as a result of raising the average domestic tariff. We assume that tariffs for government consumption will be increased in line with the proposal made by the World Bank,59 which is that the tariff for government consumption should be raised to 4.58 cents/kWh, in constant 2006 prices, by 2010. We assume that government services will expand at least at a similar rate to that of GDP and that the government's income elasticity of demand is 0.6. Compared to the 2007 tariff (in real terms) which we use as a baseline, the higher tariff will add US$17 million in 2007 prices to government electricity bills in 2010 and this will rise to almost US$40 million by 2030. Payments under the Energy Compensation Mechanism 6.41 The energy compensation mechanism (ECM) provides subsidies through the budget to low-income electricity consumers. Under the current ECM arrangements, government pays for a fixed monthly volume of electricity consumption of eligible households whose aggregate household income falls below the average monthly salary for the applicable rayon. We assume that the ECM was applied to 327,000 households in 2007, that the number of households eligible will grow by 2 percent per year and that the ECM will cover annual electricity consumption of kWh 1,800 per household. At 2007 prices, the residential tariff for consumers who consume less than kWh 250 per month will rise from 0.8 cents/kWh in 2007 to 3 cents/kWh in 2010 and then remain constant in subsequent years. We assume that the full cost of this price increase will be absorbed by the ECM, up to the maximum annual consumption allowed per eligible household. This will increase the budgetary costs of the ECM by US$14 million in 2007 prices in 2010 and this will rise to US$21 million by 2030. Net Impact on the Budget 6.42 The net impact on the budget of the revenue and expenditure effects described above is negative in the first five years, and then turns positive for the rest of the period up to 2030. The negative impact in the first five years, up to 2012, arises because the 59 These are set out in the "World Bank Comments and Suggestions on the Proposal of the Government of Tajikistan to Adjust Electricity Tariffs to Reach Financial Viability Levels by 2010," May 2006. 142 impact of higher tariffs on government electricity bills and the ECM outweighs the net gains from higher VAT, while most of the income from the GOT's capital investments in power projects has not yet begun to flow. A large increase in income from projects (mainly the TPP and the CASA 1000 OTL) reverses the net budgetary impact in 2014. The net impact rises from US$66 million in 2007 prices in 2014 to US$176 million in 2025 and then flattens out. As a percentage of GDP, the net budgetary gains reach a maximum of 2.5 percent in 2016 but then gradually decline. 6.43 Over the period 2008­30, the net impact on the budget averages US$110 million in 2007 prices and 1.2 percent of GDP. This amounts to about 5 percent of projected government revenue over this period. Hence electricity investments and reforms make a modest net contribution to budget revenues but they do not represent a major source of fiscal space that could fund a large expansion of public expenditures. 6.44 We can also disaggregate the net fiscal impact into that arising from the GOT's contribution to new investments (returns to capital, royalties on exports, and debt service costs) and that arising from domestic tariff reform (higher VAT, government utility bills, and ECM payments). The net fiscal impact of new investments is positive in all years and averages US$143 million a year in 2007 prices during 2008­30. In contrast, the net fiscal impact of the domestic tariff reforms is negative in all years and averages negative US$33 million a year in 2007 prices during 2008­30. However the latter only includes the direct "on budget" impact of domestic tariff reforms, the primary motivation of which is to cut the quasi-fiscal deficit (see section 6). Hence it would be misleading to assume that the long-term fiscal position is harmed by domestic tariff reforms: the opposite is the case because the gain in reduction of the QFD outweighs the higher direct costs to the budget. Fiscal Impact with only CASA 1000 OTL Investment 6.45 The absence of investments in new generation plants (the HPPs and TPPs listed in table 6.1) affects three categories of GOT revenue and expenditure: (i) income from capital investments, (ii) royalties on exports, and (iii) debt service costs. 6.46 Under the no new generation investment scenario, income from capital investments is restricted to the GOT's share of the returns accruing to the CASA 1000 OTL. These returns themselves are lower than in the scenario with new generation capacity because the volume of power transmitted through the CASA 1000 OTL is much lower; while the CASA 1000 OTL can earn sufficient revenue to cover its debt service payments there is less revenue left over to pay equity holders. The GOT earns about US$11 million per year on average (in 2007 prices) from its capital investment in the CASA 1000 OTL compared to US$47 million a year on average when investment in new generation projects allows the transmission capacity of the OTL to be fully utilized. Royalty payments on electricity exports are also reduced because the volume of exports is lower, not just through the CASA 1000 OTL but also the AFG-220 OTL. In this scenario, royalty payments average about US$12 million per year in 2007 prices compared to US$20 million per year in the scenario, which includes new investment in generation. 143 6.47 Government debt servicing is much reduced in this scenario, as it includes only the servicing of the (highly concessional) CASA 1000 OTL debt. Debt servicing averages only US$2 million a year in 2007 prices compared to the US$35 million a year when the borrowing for new power generation projects has to be serviced. The net fiscal impact of the CASA 1000 OTL investment combined with the domestic tariff reforms is slightly negative, although the magnitude is small. On average during 2008­30 there is a net cost to the budget of US$14 million in 2007 prices or 0.2 percent of GDP. F. QUASI-FISCAL DEFICITS AND CONTINGENT LIABILITIES Quasi-Fiscal Deficits 6.48 The electricity industry incurred quasi-fiscal deficits (QFD) that were estimated at nearly 10 percent of GDP in 2006 and 7 percent of GDP in 2007. In essence, the electricity industry QFD is the deficit between the revenue that BT needs to earn to operate and properly maintain the electricity supply network and the revenue that it actually earns from the sale of power to domestic customers. The main consequence of incurring a QFD is that BT fails to devote adequate resources to the necessary maintenance and repair of the transmission and distribution network. 6.49 The causes of the QFD in the electricity sector are threefold: (i) average tariffs that are set at levels below the LRAIC of supply (a proxy for the marginal cost of supply), (ii) excessive technical and commercial losses, and (iii) losses attributable to noncash collection of electricity bills. The LRAIC was estimated at 2.3 cents per kWh in 2007 prices. Actual average tariffs were less than half this level; 0.9 cents per kWh in 2007. Setting tariffs below cost recovery levels contributed 82 percent of the estimated US$228 million QFD in 2007. Technical and commercial losses amounted to 16 percent of domestic supply in 2007, compared to an international norm of 10 percent, and this contributed 10 percent to the QFD. Finally noncash collections amounted to 18 percent of electricity bills in 2007 compared to the international norm of 2 percent and this contributed 8 percent to the QFD. 60 6.50 To eliminate the electricity sector QFD it is necessary to raise the average tariff to the LRAIC of supply (and maintain it at that level in real terms), reduce technical and commercial losses to the international norm of 10 percent, and reduce noncash collections to the international norm of 2 percent. Failure to eliminate the QFD will eventually create a burden on the government budget, because it will mean that government resources will have to be allocated to fund the essential rehabilitation and maintenance required to ensure that the electricity supply network can still function. The alternative would be a breakdown in electricity supply capacity, which would obviously have grave consequences for the economy. 6.51 BT has started to implement reforms to eliminate the QFD. Domestic tariffs were raised in 2006 and 2007. Investments in metering and billing systems have reduced 60 The reduction in the QFD as a percent of GDP in 2007 was attributable partly to a sharp increase in GDP measured in UD dollars (by 19 percent) and partly to a reduction in the GFD in absolute terms due to increased average domestic tariffs, lower technical and commercial losses and improved cash collections. 144 technical and commercial losses from 22 percent of supply in 2003 to 17 percent in 2007, and noncash collections have been reduced from 46 percent in 2003 to 18 percent in 2007. Nevertheless there is still some way to go to eliminate the QFD and some of the future reforms needed, especially further raising the average domestic tariff, are likely to prove politically contentious. 6.52 Figure 6.3 shows projections for the QFD based on two alternative scenarios. The first scenario is the full implementation on schedule by BT of the tariff reforms to raise the average tariff to the LRAIC and then maintain it at that level thereafter, together with the necessary improvements to distribution systems, billing, and collection that will bring technical and commercial losses and noncash collections down to international norms. Under this scenario the QFD is eliminated by 2010. 6.53 The second scenario differs from the first in that there are no further real increases in average domestic tariffs after 2007 (tariff increases merely keep pace with inflation). As a consequence the QFD remains a persistent problem, although it declines as a percentage of GDP to 2.6 percent by 2030. After the initial fall in the QFD as a percent of GDP, which is mainly attributable to the reduction in technical and commercial losses and noncash collections, the further falls arise because domestic electricity consumption grows more slowly than real GDP and hence the value of domestic electricity supply falls as a share of GDP.61 6.54 As noted in section 5, the increase in electricity tariffs to the LRAIC of supply involves a direct net cost to the budget, because the increase in government's electricity bills and ECM payments exceed additional VAT receipts on electricity sales, but these costs must be weighed against the benefits of reducing the QFD. The latter is far larger than the former. In 2007 prices, the average annual reduction in the QFD over 2010­30 as a result of raising electricity tariffs is US$206 million, compared to the average direct net cost to the budget of US$35 million. Hence raising the electricity tariffs improves the fiscal position, taking full account of "off budget" effects, by an average of US$170 million a year. Contingent Liabilities for Unused Power 6.55 As shown in table 6.1, the projected investment in new generation plants amounts to about US$6 billion, of which US$4.2 billion will be provided by investors other than the GOT (private investors, foreign publicly owned energy utilities, international financial institutions, and so forth). Because of the particular market structure of electricity markets, with the transmission and distribution system being a natural monopoly, investors in generation capacity face substantial market risk once their investments are "sunk" unless they have legally binding long-term agreements with power purchasers that can guarantee that the power generated will be sold at prices that are economically viable for the generators. As such, a precondition of new investment in power generation capacity will be the signing of long-term PPAs between the generator and future purchasers of its output. For example, Sangtuda I has signed a long-term PPA 61 This is mainly because of the assumed income elasticity of demand of 0.6 used to make long-term demand projections. 145 with BT under which the former commits to supplying a minimum volume of power and the latter commits to purchasing that power at predetermined prices. Such PPAs are often termed "take or pay" because the purchaser must pay for the power even if it cannot sell the power to final consumers. 6.56 Some of the power generated by the new HPPs and TPPs will be sold to BT for domestic consumption. This power will be covered by a take or pay PPA signed between BT and the generators. However, the arrangements for the power that is intended for export are potentially more complex. It is possible to envisage three possible arrangements: (i) the generator sells the power to BT (and signs a PPA with BT for this power) and BT then resells the power to a third party; (ii) the generator sells the power to an intermediary such as the company owning the CASA 1000 OTL and signs a PPA with this company (the intermediary would then seek to sign a PPA with the importer); or (iii) the generator sells power directly to the importer and signs a PPA with the importer. 6.57 The first of these arrangements (under which BT purchases power for export and then resells it) is the essence of the PPA signed between BT and Sangtuda I, under which BT purchases all of the power generated by Sangtuda I at predetermined prices, some of which will be sold on the domestic market and the rest exported. It appears that a similar agreement, covering all of the power produced by the generator, will be signed between BT and Sangtuda II. This means that BT has a liability to purchase power at predetermined prices even if it cannot resell this power to domestic consumers or to importers in Afghanistan and Pakistan. The Fieldstone Report (Fieldstone 2007, p. 75) states that PPAs will need to be signed between the appropriate government entities in Tajikistan and the Kyrgyz Republic as power sellers and Afghanistan and Pakistan as power purchasers as a precondition for the CASA 1000 project, which implies that all of the power exported from Tajikistan through the CASA 1000 OTL will be purchased by a government entity in Tajikistan and then sold on to the importer. This in turn implies that these government entities (presumably BT in Tajikistan) will have to sign PPAs with the generators. 6.58 The volume of power generated by the Rogun HPP and the coal-fired TPP is of a much larger magnitude than that from Sangtudas I and II. As most of the power produced by these projects will be exported, it is possible that investors will sign PPAs with both BT and a foreign electricity purchaser (in Pakistan), to diversify the risk. How the responsibility for purchasing power on a take or pay basis is split between BT and the foreign power utility is difficult to predict in advance. For the purposes of these projections, we assume that BT will commit itself to purchase all of the power that is intended for the domestic market, which comprises 500 MW of the power generated by the TPP and all of the power generated by the reconverted TPP in Dushanbe. We also assume (somewhat arbitrarily) that BT will commit to purchase half of all the power produced by Rogun and the TPPs that is intended for export, in which case it would also seek to sign legally binding agreements with importers to take or pay for the power. 6.59 These liabilities arising from the PPAs would then become a financial cost to BT in the event that the power purchased for domestic consumption cannot be sold to domestic consumers at a price that covers all costs, because demand is lower than 146 projected or because domestic tariffs are not raised to levels sufficient to ensure full cost recovery. One could envisage several plausible scenarios that would reduce domestic demand below projected levels or impede the raising of tariffs to full cost recovery levels. First, lower than projected real GDP growth will reduce domestic demand. Second, a sharp depreciation of the real exchange rate will raise the real value of the domestic tariff, denominated in somonis, that is needed to ensure full cost recovery, because the purchase price of power stipulated in PPAs will be denominated in U.S. dollars. Third, it may prove politically difficult to raise tariffs to the levels required to ensure full cost recovery. 6.60 They would also become a financial cost to BT in the event that power exports fall short of projected levels, because of disruption to the transmission system (for example from insurgent activity in Afghanistan or Pakistan) or because importers will not purchase the expected volumes of power. In the latter case BT may have legal recourse to the importer depending on the nature of any agreement signed between the two parties. These liabilities to take or pay for power on the part of BT are in effect contingent liabilities on the GOT budget, because BT is a government-owned entity and any impairment of its financial condition that undermines its capacity to supply the domestic market with electricity (for example, because it cannot fund essential maintenance of the supply network) will ultimately have a fiscal impact. For this reason, we treat any liabilities of BT to take or pay for power as contingent liabilities on the government budget. 6.61 Table 6.7 presents projections of the liabilities of BT arising from take or pay PPAs. The projections exclude the liabilities related to Sangtuda I (as this project is not covered in this paper for the reasons discussed in section 2). The total value of projected revenue from new power generation projects, and hence the total value of PPAs that generators will seek to sign, rises to around US$750 million per year in 2007 prices by 2022 and then flattens out at this level. Given the assumptions about BT's share of the PPAs discussed above, BT's liabilities will rise to US$450 million per year by 2020 and then flatten out at this level. As a percentage of GDP, BT's liability peaks at 7 percent in 2016 then declines to 3 percent by 2030. While it is unlikely that all of these liabilities will materialize as costs to BT (this would require both a collapse in domestic sales and exports), it is not implausible that a fraction of these liabilities translate into costs for BT, and ultimately for the government budget, if the domestic economy performs much more poorly than expected or if there is serious disruption to electricity exports after the sunk costs of new power projects have been incurred. G. CONCLUSIONS 6.62 From a policy perspective, it is optimal to evaluate the fiscal impact of planned or potential reforms and investments in the electricity sector under three distinct categories: (i) reforms to the domestic electricity market, (ii) investment in the CASA 1000 OTL, and (iii) investment in major new generation projects. 6.63 The domestic market electricity reforms analyzed here mainly comprise raising the average tariff to the LRAIC of supply (about 250 percent of the 2007 level). This is motivated by the need to eliminate the QFD of the electricity sector. Raising the average 147 tariff has a (relatively small) direct net cost to the budget, because higher budgetary costs will be incurred for the government's own electricity consumption and subsidies to low- income electricity consumers under the energy compensation mechanism, which exceed the gains from higher VAT payments. However, these direct budgetary costs, which average US$35 million a year during 2010­30 in 2007 prices, are far smaller than the indirect gains arising from the elimination of the QFD, which average US$206 million a year in this period. Hence raising domestic tariffs yields substantial net fiscal benefits. Indeed failure to eliminate the QFD would reduce fiscal space and present a threat to fiscal sustainability over the long term in Tajikistan. 6.64 Investment in the CASA 1000 OTL has relatively minor consequences for the fiscal position. This is because the GOT's financial requirements to contribute its share of this project are not large (US$127 million) and we have assumed that this can be financed on highly concessional terms by external lenders. Borrowing for the CASA 1000 OTL will have a negligible impact on long-term debt sustainability, with debt servicing costs averaging only US$2 million a year in 2007 prices during 2010­30. However, it is also the case that investment in the CASA 1000 OTL alone, without new investment in generation capacity, will not yield major fiscal benefits in terms of increased revenues. Revenues, from the income earned on the GOT's capital in the CASA 1000 OTL and royalties (assumed at 0.25 cents per kWh) levied on electricity exports (through both the CASA 1000 OTL and the AFG-220), average only US$24 million a year in 2007 prices during 2010­30, because there is insufficient surplus power to ensure that the transmission lines are operated at full capacity. 6.65 Investment in new HPP and TPPs, for which the GOT is expected to contribute about US$900 million in addition to its existing assets (which were constructed in the Soviet era), have a more profound fiscal impact: they increase the risk of debt distress and create contingent liabilities for the budget, but could also potentially mobilize significant budget revenues and make a positive net contribution to the budget. We assume that all of the GOT's investments in these projects, other than the existing assets, will be funded by external borrowing on semiconcessional terms and then compare the consequences of this borrowing with the projections in the DSA baseline. As a result, two of the key debt sustainability indicators remain above the threshold at which external debt is deemed sustainable for longer than is the case in the DSA baseline. In particular, the NPV of PPG external debt to exports remains above the threshold for sustainability until 2027, compared to 2012 in the DSA baseline. The new generation projects yield budget revenues in the form of income earned on GOT investments in these projects and export royalties, which average about US$190 million a year in 2007 prices over the 2010­30 period, compared to the debt servicing costs of these projects, which amount to an average of US$35 million. However, although the net fiscal impact is positive, revenues, net of debt servicing, average less than 2 percent of GDP, so these projects will not make a large contribution to the creation of fiscal space over the long term. 6.66 Finally the new power generation projects will entail liabilities for BT, in terms of guarantees it undertakes to purchase power from the generators on a take or pay basis. These liabilities of BT should be recognized as contingent liabilities of the government budget. The magnitude of these liabilities is, however, difficult to project as 148 it depends upon the nature of arrangements for the regional trading of power that have not yet been put in place. We have assumed that BT will sign take or pay power purchase agreements with generators for all of the power destined for the domestic market and for half of the power destined for export. This would imply that BT incurs liabilities amounting to an average of almost US$400 million per year in 2007 prices, or 4.5 percent of GDP, during 2012­30. How much, if any, of these liabilities would materialize as losses to BT and, therefore, costs to the government budget, will depend on the extent to which domestic demand is less than anticipated and/or electricity cannot be exported as planned. 149 0302 9202 8202 7202 0302 6202 L T O 5202 st eni 1000 2008- 4202 SA A projecl Basel C al A ht ht 3202 S D wi wi 2008-2030 ons 2202 ecti 1202 oj Projections Pr 0202 9102 GDP: P: to D G 8102 150 Debt to 7102 PPG of bte 6102 D NPV 5102 6.1: 4102 PPG 3102 Figure of V 2102 P N 1102 0102 9002 8002 50% 45% 40% 35% 30% 25% 20% 15% 0302 9202 1000 8202 ects SA CA baseline 7202 proj 2008-30 ll ith AS A w D 6202 2008-30 5202 4202 3202 Projections Projections 2202 1202 Exports: to 0202 Exports: 9102 Debt to 8102 151 PPG 7102 Debt 6102 External PPG 5102 of 4102 NPV 3102 6.2: External 2102 of Figure 1102 0102 NPV 9002 8002 200% 190% 180% 170% 160% 150% 140% 130% 120% 110% 100% 2 2 30 60 33 00 00 51 62 93 399 92 64 77 88 27 20 27 46 85 74 17 9307 16 28 88 26 166 424 196 178 228 140 113 9 0 3 0 0 1 226 329 399 2 0 1 9 2 0 6 8 8 202 6661 273 460 850 745 2494 9241 7491 174 3252 737 4011 262 1381 28 06 2 2 33 00 00 51 62 93 399 29 03 81 12 26 20 27 46 85 74 17 6838 61 28 88 26 166 424 188 171 236 140 113 2702 66061 226 3372 239 0064 0058 5147 993 49224 43481 75861 7661 05842 3247 01641 2862 38811 26 20 06661 2 2 33 00 00 51 62 93 399 27 46 85 74 29424 25081 11461 41 16 93442 67511 61041 28 26 88311 0252 66061 226 7332 239 6004 5008 4517 993 49224 68471 07661 6081 80842 6237 01641 6282 38811 4 0 3 0 0 1 3 2 8 2 5 4 3 6 8 8 226 329 99 202 6661 273 460 850 745 2494 7321 5751 157 5162 0681 4011 262 1381 2 2 23 60 33 00 00 51 62 93 939 92 84 40 44 07 59 16 28 88 GWh 20 27 46 85 74 15 26 166 424 169 154 255 106 140 113 2202 60661 226 3372 329 0064 0058 5147 993 92424 53661 39151 1451 39852 23821 16041 2862 88311 21 2010-2030 20 06661 2 2 33 00 75 82 18 62 93 27 46 63 83 11 22414 33361 84841 85 14 98052 97911 61041 28 26 88311 0202 66061 226 7332 239 6004 2504 3149 2421 35304 02461 56741 4571 32942 3998 01641 6282 38811 9 0 3 0 5 4 2 8 6 6 0 2 2 6 8 8 Projections 226 329 201 6661 273 460 212 931 124 8223 5721 4291 143 2502 598 4011 262 1381 8102 2 60 62 3372 2 00 45 93 46 6631 58 38 34 3041 20 7807 16 8262 88 166 102 371 154 140 217 140 113 Export 152 7 and 012 06661 3 226 732 329 0064 54201 6 361 85173 06151 28731 8 371 89912 3 646 61041 8 8 622 38 11 2 2 0162 66061 62 7332 93 4600 24501 3661 15873 14651 63131 5151 01222 7806 01641 6282 13881 Demand 5 012 226 3 0 6 5 3 7 6 66601 732 239 230 02451 361 48583 48451 33611 481 00132 972 21931 612 579 06 2 2 62 25 79 47 11 Supply, 62 93 2014 2733 5588 1490 1455 1718 2246 8065 166 280 145 130 134 103 6.2: 0132 66061 226 7332 239 0 4901 43722 26541 83921 4271 1728 3591 9585 8571 1014 2 0 Table 226 3 239 7 3 5 8 4 069 4 8 6 201 6661 273 0942 3981 2581 139 696 484 118 365 11 60 004 20 3372 93 02 22 80 13 1995 1421 9715 6711 3004 166 197 138 124 1002 660 002 0 7332 16 93591 941 13 07421 3451 6525 1601 8391 8391 oni sul itycapa noit rpu raen nda S C n metsys nversoc el m su ge 1 II ab Re De tioarene al gnitsix oni adut e rplu 0 dautg I II anbh citse portx S s 1000 E 22 A dit nga P S G E ad S San Rogun Rogun PT Dus latot selas esssol Dom wo wo kijaT rgyzy portx GF K E A CA 0302 0866 1456 3376 %4 %6 %3 % % % 7 7 1 % 0%.7 .17 %6 6 4 6 0 1.4 23. 23. 24. 138 139 143 07.3 12.3 80.3 8. 74.1 77.1 15.2 0292 5485 6085 8575 2%.32 %4. % % % 0 0 4 % % 6 5 9 4 23 4%.42 47. 136 138 144 71.2 76.2 52.3 .76 .86 7%.8 57.1 60.1 04.2 %9. 2%. %6. 3 4. 9 % % 9 9. 7 8 5% 6% 5% 2028 5047 5110 5420 49. 22 23 24 13 13 14 236. 241 319. .36 .46 5%.8 140. 143 190. 7 0 4 6 0. 1. 1. % % 5. 6. 4. 6% 9% 7% 202 458 464 501 2.2 2.2 4.2 33%1 35%1 45%1 207 212 292 .06 .16 5% 8. 9.15 126 129 178 6 9 5 7 202 415 422 465 %3.22 %6.22 %9.42 0. 1. 8. % % 2. 4. 3. 1 34%1 37%1 51%1 160 165 244 .25 .35 9% 7. 3.5 100 103 153 25 52 20 07 6% 0% 9% 4% 20 38 39 44 5.5 22. 23. 25. 134% 136% 153% 1.112 2.612 3.792 % % .47 .57 10. 5.531 8.831 8.091 4 24 74 43 86 %0 %4 %9 % % % 4 5 7 % % %5 2 6 3 2 20 35 36 41 56. 23. 23. 26. 134 136 157 96.1 01.2 81.2 .37 .57 10. 29.1 32.1 85.1 0232 3223 3933 9883 %4 %9 %1 % % % 8 0 3 % % .47 .67 %7 9 5 9 0 55. 23. 23. 28. 134 136 160 83.1 89.1 65.2 10. 23.1 27.1 78.1 0222 0983 1713 8113 %9. %5. %4. 1 3 8 % % 4% 7% 5% .47 .77 %8. 0 6 7 7 53. 23 24 29 13 13 16 172. 177. 249. 10 119. 122. 172. 1 8 2 2 8. 1. 2. % % 9. 7. 5. 5% 2% 9% 4% 8% 9% 202 289 297 365 4.2 5.2 0.3 13 13 16 160 166 239 .57 .77 1% 1.1 5.55 113 117 169 2007-2030 0 7 3 4 202 271 279 349 %2.52 %9.52 %4.23 0. 0. 9. % % 35%1 39%1 74%1 151 155 227 .57 .77 %3.11 7. 5. 5. 9.55 109 112 165 4 8 1 19 50 26 65 % % 0 0 6 6 9% 7% 2% 2. 3. 9. 2% 6. 7. 5. 20 25 26 33 25. 26. 34. 36%1 40%1 80%1 9.4 14 14 20 .67 .77 11. 10 10 15 Indicators: 18 92 66 09 % % % 7% 5% 8% 0% 20 23 24 32 3.4 26. 27. 35. 137 141 184 1.431 1.531 6.191 % % .77 .77 11. 3.201 1.301 2.641 9 Debt 0172 2422 3132 0223 %4 %3 %9 % % % 2 2 3 % % .87 .97 %6 5 3 5 0 99. 36. 27. 28. 36. 137 142 185 27.1 28.1 73.1 10. 00.1 35.1 % % % 6 6 1 % % 5 3 1 7 0162 1002 1692 8352 1%.82 0%.92 0%.83 153 138 142 186 26.1 27.1 61.1 .38 .48 6%.01 01.1 02.1 29.1 27. External %9. %9. %7. 3 3 2 % % 9 7 2 3 8% 3% 0% 2015 1969 2036 2703 of 28 29 39 13 14 19 142. 143. 168. .001 .101 %8. 21. 11 116. 117. 138. 4 2 6 2 0. 9. 8. 1% 1% 5% 201 187 193 258 0.3 1.3 1.4 41%1 45%1 94%1 166 166 181 %5.21 %5.21 8. 6. 1. 3. 7% 3.1 13 139 140 153 3 6 9 4 201 179 185 240 %7.13 %8.23 %4.24 6. 5. 3. % % 44%1 49%1 93%1 175 176 188 .141 .241 %1.51 5. 4. 5. 0.11 151 152 162 Projections 12 19 80 26 2% 4% 0% 20 17 17 22 33. 34. 43. 148% 153% 191% 2.842 2.942 8.752 %3.12 %4.12 1% 22. 6.912 5.022 1.822 5.8 6.3: 11 94 53 75 %4 %6 %6 % % % 2 1 2 20 16 17 20 36. 37. 44. 156 161 191 38.2 39.2 44.2 %9.12 %0.22 %4 0 8 4 5 5. 22. 16.2 16.2 21.2 Table 0102 6471 7031 8911 %3 %7 %1 % % % 5 2 7 39. 40. 45. 163 169 187 62.1 63.1 65.1 %1.61 %2.61 %4 0 7 0 0.3 16. 51.1 51.1 54.1 0092 6011 6391 7321 %1. 5 6 8 % 7 8 0 3 42 1%.34 %5. 9% 4% 4% 45 17 18 19 165. 165. 165. .581 %5.81 %6. 0. 18 157. 157. 158. 8 6 1 9 8% 200 144 145 145 1.4 %0.24 5. 5. 5. 2% 0% 1% 2% 2.4 18 18 18 178 178 178 %2.22 % 4. 4. 4. .222 2% 0.0 2.2 174 174 174 7 2 2 2 200 123 123 123 %7.63 7% 6.3 %7.63 7. 7. 7. % 68%1 68%1 68%1 srt 108 108 108 pox .841 %8.41 %8.41 7. 7. 7. secirp 0.0 108 108 108 nsoilli Efo 07 m d)et d)et s d)et )d 20 ent d)et ec $ us OTL us OTL portx us OTL nsoilli d)et us OTL rce rviest (adj 0 st GDP (adj 0 st E (adj 0 st m (adj 0 P st tesujda( LT O nsoilli us OTL st m (adj 0 st Debt G P P eniles ot 100 ec 100 ec 100 ec 100 ec 0001 $ deb ec 100 ec A S projl Debt eniles ot A S projl Debt eniles $ A eci S projl rve eniles sa A eci enli eci S projl rve eniles A ott S projl en of ba CA al of ba CA al of ba CA al S ba CA al S seab ASA projl rve C al S ba CA al V A hti hti V A hti hti V A hti hti A hti hti SA thi hti A hti hti NP DS w w NP DS w w NP DS w w Debt DS w w Debt D w w Debt DS w w emrcnI 5 5 6 5 6 9 7 0 9 9 1 2 7 1 3 2 2030 286. 5.1 5.2 245. 105. 3.4 0.2 1.4 180. 2030 1. 0. 0. 7.1 0. 3.0 0. 0. 1. 4 2 6 7 0 4 3 4 4 8 3 2029 286. 15. 25. 245. 110. 42. 20. 47. 176. 2029 1.2 1.0 2.0 1. 8.0 0. 1.0 3.0 3.1 3 9 6 9 6 9 9 8 8 2 1 2 8 2 4 3 2028 286. 14. 25. 245. 110. 40. 19. 49. 175. 2028 2. 0. 0. 9.1 0. 3.0 0. 0. 1. 2 6 6 1 9 5 5 9 4 2027 86.2 14. 25. 46.2 10.1 39. 19. 51. 75.1 2027 3.2 1.0 2.0 0 2. 9.0 3 0. 2.0 4.0 4.1 2 3 6 3 3 1 1 1 8 4 1 2 9 2 5 5 2026 286. 14. 25. 246. 110. 38. 19. 53. 175. 2026 2. 0. 0. 1.2 0. 3.0 0. 0. 1. 1 0 6 5 9 8 7 4 2 3 3 2025 286. 4.1 5.2 246. 110. 6.3 8.1 5.5 175. 2025 6.2 1.0 2.0 2. 0.1 0. 2.0 5.0 6.1 4 8 6 1 0 5 4 2 4 7 1 2 1 2 5 6 2024 277. 3.1 5.2 238. 110. 5.3 8.1 6.5 167. 2024 2. 0. 0. 3.2 1. 3.0 0. 0. 1. Prices 9 5 6 9 2 3 0 0 7 4 4 2023 269. 13. 25. 230. 107. 34. 18. 55. 162. 2023 8.2 1.0 3.0 2. 1.1 0. 2.0 6.0 7.1 GDP 3 3 6 5 4 1 6 7 9 2007 of 1 1 3 2 2 6 9 2022 3. 0. 0. 6.2 1. 4.0 0. 0. 1. 273. 13. 25. 234. 104. 33. 17. 53. 168. 2022 5 1 6 9 8 9 3 5 8 2021 76.2 13. 25. 37.2 04.1 31. 17. 55. 71.1 2021 3.3 2.0 3.0 8 2. 3.1 4 0. 2.0 7.0 1.2 millions, 9 Percent 8 6 5 6 8 0 9 3 1 2 3 3 2 7 7 2020 238. 12. 25. 200. 103. 30. 17. 55. 135. 2020 3. 0. 0. 6.2 1. 4.0 0. 0. 1. US$ 6 6 6 4 0 7 6 6 6 3 4 2019 208. 2.1 5.2 170. 6.9 9.2 6.1 9.4 112. 2019 9.2 2.0 3.0 2. 3.1 0. 2.0 7.0 5.1 0 4 6 0 9 7 3 9 1 2 2 4 3 2 6 9 2018 217. 12. 25. 179. 88. 28. 16. 43. 128. 2018 3. 0. 0. 6.2 1. 4.0 0. 0. 1. Expenditures: 0 2 6 2 7 7 0 0 3 8 4 2017 217. 12. 25. 179. 79. 27. 16. 36. 137. 2017 4.3 2.0 4.0 2. 2.1 0. 2.0 6.0 1.2 Expenditures: and 4 6 6 3 1 8 7 7 3 6 2 4 1 3 5 5 154 2016 9. 3. 0. 0. 1. 0. 0. 2. 214. 25. 179. 67. 23. 15. 27. 147. 2016 0.3 4.0 and 5 6 9 2015 79.1 4.9 5 6 0 4 3 22. 47.1 59. 23. 15. 21. 19.1 2015 2.3 2.0 4.0 6 2. 1.1 4 0. 3.0 4.0 1.2 Revenues 2 2 1 9 5 2 1 3 7 2 2 4 0 3 3 3 2014 9. 116. 19. 87. 50. 22. 15. 13. 65. 2014 2. 0. 0. 7.1 1. 4.0 0. 0. 1. Revenues 2 Related 5 4 2013 7.4 1.9 5 7 1 4 8 0 1.1 6.2 7.4 1.2 4.1 1.1 1.0 2013 0.1 2.0 2.0 0. 0.1 0. 3.0 2.0 0.0 Related 0 9 1 0 6 6 5 5 6 7 2 2 2012 32. 8. 9. 14. 43. 20. 14. 8. 11.- 2012 0. 0. 0. 3.0 0 1. 5.0 3 2 3 0. 0. 0.- Energy 4 2011 34. 7.8 6.9 0 3 7 2 16. 39. 19. 14. 5.5 0 5.- 2011 8.0 2.0 2.0 4 0. 9.0 5 0. 3.0 1.0 1 0.- Energy 6.5: 2 6 6 7 8 9 0 5 3 2 1 9 4 1 6 2010 13. 8. 4. 0.0 35. 18. 13. 3. 22.- 2010 0. 0. 0. 0.0 0. 5.0 0. 0. 6.4: 0.- Table 2009 1.5 1.5 0.0 0 6 9 0. 14. 7. 4.6 3.0 5 9.- 2009 1.0 1.0 0.0 0 0. 4.0 2 0. 2.0 0.0 3 0.- Table 5 5 0 2008 2. 2. 0. 0.0 0 5. 7.2 3 0 5 1 1 0 2. 0. 2.- 2008 0. 0. 0. 0.0 1 0. 1.0 1 0 1 0. 0. 0.- st st projec projec levy T rnsu levy energy T rnsu energy A port V ex rett s of A port V ex rett s of ec nte omrf gni ec nte gni omrf ojr P reu t omrf omrf ojr P reu t y paym pac y paym pac enue gain yt rvices M im enue gain yt rvices im nergy penditx ili C nergy penditx ili M C Rev Net Royalt E E Ut E Debt Net Rev Net Royalt E E Ut E Debt Net 8 5 4 0 4 9 7 8 6 2. 2% 2030 37. 15. 12. 10. 67. 43. 20. 29.- 0.- 7 2 4 2 5 4 3 9.2 8 %2 2029 37. 15. 12. 10. 65. 42. 20. 27.- -0. 6 9 4 4 8 9 9 0 1 3. 2% 2028 37. 14. 12. 10. 63. 40. 19. -26. 0.- 2007 5 6 4 6 0 5 5 5 2027 7.3 4.1 2.1 0.1 2.6 9.3 9.1 1.3 %2 24.- -0. 5 3 4 8 4 1 1 2 9 3. 2% 2026 37. 14. 12. 10. 60. 38. 19. 22.- millions, 0.- 4 0 4 0 8 8 7 3.3 4 %2 2025 37. 14. 12. 11. 58. 36. 18. US$ 21.- -0. 3 8 4 2 3 5 4 4 9 3. 2% 2024 37. 13. 12. 11. 57. 35. 18. 19.- 0.- 3 5 4 3 8 3 0 5.3 5 %2 2023 37. 13. 12. 11. 55. 34. 18. Capacity: 18.- -0. 2 3 4 5 3 1 6 6 1 3. 2% 2022 37. 13. 12. 11. 54. 33. 17. -17. 0.- 1 1 4 7 0 9 3 8 2021 7.3 3.1 2.1 1.1 3.5 1.3 7.1 7.3 %2 15.- -0. Generating 1 8 4 9 7 8 0 9 6 in 2. 2% 2020 37. 12. 12. 11. 50. 30. 17. 13.- 0.- 0 6 4 0 4 7 6 0.1 4 %1 2019 37. 12. 12. 12. 47. 29. 16. 10.- -0. 0 4 4 2 8 7 3 7 8 0. 1% 2018 37. 12. 12. 12. 45. 28. 16. 8.- 0.- Investment 0 2 4 4 5 7 0 8.0 5 %1 2017 37. 12. 12. 12. 44. 27. 16. 7.- New -0. 4 6 4 5 2 8 7 8 8 No 9. 0. 1% 2016 34. 12. 12. 40. 23. 15. 5.- 0.- 155 prices 4 4 7 1 0 4 7 with 2015 4.3 4.9 2.1 2.1 9.3 3.2 5.1 8.0 4.- %1 -0. 5 2 4 8 0 2 1 8 6 9. 0. 1% 2014 33. 11. 12. 38. 22. 15. 4.- 0.- 5 1.9 5 9 0 4 8 8.0 5 %1 2013 32. 10. 12. 37. 21. 14. 4.- -0. Expenditures 8 9 8 1 0 6 5 9 2 8. 8. 0. 1% 2012 30. 13. 36. 20. 14. 5.- 0.- and 7 7.8 8.8 2 8 7 2 9.0 1 %1 2011 30. 13. 34. 19. 14. 4.- -0. 2 6 6 3 8 9 6 2 8. 4. 2010 13. 3.3 8.1 0. 5% 13. 20.- 0.- Revenues 1.5 1.5 4 9 3 2009 4.1 7. 4.6 1.0 9.- %3 -0. 5 5 0 2. 2. 5. 7.2 3 0 5 2. 0. 2.- 1% Related 2008 0.- 1000 GDP Energy A of levy S T CA ent 6.6: A port V rnsu s of ex ret nte omrf gni perc omrf reu t ast y 1000 paym pac gain A rvices pac enue S yt Table im penditx ili M mi C Rev Net Royalt CA E Ut E Debt Net net 0302 9202 e 8202 tariff easr in 7202 e inc 6202 LRAIC ffi AIC to LR increas 5202 tar at real 4202 tariff no Increase thouti thi thi 3202 Tariff w W W 2202 and 1202 without thi 0202 and CI w 9102 with LRA 8102 GDP 156 GDP to of of 7102 6102 cent percent 5102 as per 4102 QFD as 3102 QFD 2102 Projected 1102 6.3: ectedj 0102 Figure Pro 9002 8002 7002 8% 7% 6% 5% 4% 3% 2% 1% 0% 2030 497 534 1.5 1.3 2029 749 453 .45 3 3. 0282 497 534 7.5 4.3 prices, 0272 749 453 .06 7 2007 3. 0262 749 453 4.6 9.3 millions 2025 749 453 .86 1 4. US$ 2024 497 534 3.7 4.4 Tajik, 2023 750 454 .87 7 4. Barki 0222 507 544 4.8 1.5 of 0212 729 453 .78 4 5. 0 Liabilities 022 708 451 0.9 8.5 2019 645 420 .88 7 5. ntingent 2018 246 184 1.9 1.6 Co 157 5 and 2017 624 419 .89 6. 2010-2030 0162 256 194 4.01 0.7 0152 556 385 9.9 9 6. Agreements 0142 307 232 .95 4.4 84 48 Purchase 2013 7.1 7 1. 25 52 2012 .50 5.0 Power of 0 0 0 0 2011 0. Value 0 0 0 2010 0.0 GDP P of D Projected s A P P 6.7: of ue ytil Gfotnecrep nte text. perc as See: valla abiil sal ytil Table T Tot B taoT abiil T B Sources REFERENCES Avanesyan, Vahram. 2005. "Per Capita Budget Allocation for Education." World Bank, Washington, D.C. Fieldstone, Chadbourne, and Parke LLP and Nexant. 2007. CASA 1000 Revised Phase 1 Report, Commercial Assessment. Herrera, S. and Gaobao Pang 2005. Efficiency of Public Spending in Developing Countries: An Efficiency Frontier Approach, World Bank, Manuscript. M.J. Farrell 1957. "The Measurement of Productive Efficiency," Journal of the Royal Statistical Society, Series A (General), Part III. Ministry of Finance of the Republic of Tajikistan. 2007. "Medium-term Public Expenditure Framework (MTEF) for 2008­10." National Statistics Committee of the Republic of Tajikistan 2006. "Education in the Republic of Tajikistan." Organization for Economic Co-operation and Development (OECD). 2008. Education at a Glance 2007--OECD Indicators. OECD, Paris. SNC-Lavalin. 2007. Draft Pre-Feasibility Phase 1 Report: Techno-Economic Pre- Feasibility Study for the 1000 MW Transmission Line (CASA ­ 1000). UNDP. 2005. "Investing in Sustainable Development: Millennium Development Goals Needs Assessment in Tajikistan," Dushanbe, Tajikistan. United Nations Children's Fund (UNICEF). 2007. TransMONEE 2007 Database, Innocenti Research Centre, Florence. World Bank. 1994. "Tajikistan: Health Sector Policy Note,". ________. 2002. "Tajikistan Education Sector Review--Meeting the Education Challenge: Policy Options for the Next Generation." World Bank, Washington, D.C. ________. 2004. "Central Asia: Regional Electricity Export Potential Study." World Bank, Washington, D.C. ________. 2005. "Millennium Development Goals: Progress and Prospects in Europe and Central Asia Region," Washington, DC. ________. 2005. "Tajikistan Energy Utility Reform Review". 159 ________. 2005. "Public Expenditure and Institutional Review Washington, D.C. ________. 2007. Welfare Dynamics and the State of Service Delivery in Education and Health Sector. ________. 2007. "Tajikistan: First Programmatic Public Expenditure Review (PPER 1), Report no: 39771-TJ, Washington, D.C. June ________. 2007. "Tajikistan: Public Expenditure and Financial Accountability Assessment--Performance Report," SECO (Swiss Economic Cooperation Office). June ________. 2008a. "Republic of Tajikistan: Education Sector Public Expenditure Tracking Survey Washington, D.C. ________. 2008b. "Welfare Dynamics and State of Service Delivery in the Health and Education Sectors. Washington, D.C." 160