Tajikistan Programmatic Public Expenditure Review Public Disclosure Authorizedpublic Disclosure Authorized

eotN.371 J aiitn Programmatic PublicExpenditure Review Tajikistan Report No.39771- TJ Report No. 39771- TJ Tajikistan Programmatic Public Expenditure Review Public Disclosure AuthorizedPublic Disclosure Authorized

June 22, 2007

Poverty Reduction and Economic Management Unit Europe and Central Asia Region Public Disclosure AuthorizedPublic Disclosure Authorized Public Disclosure AuthorizedPublic Disclosure Authorized

Document of the World Bank Public Disclosure AuthorizedPublic Disclosure Authorized

NCPS Net Credit to the Private Sector NDS National Development Strategy NFA Net Foreign Assets NPV Net Present Value OIN Other Items Net OTL Overhead Transmission Lines PARS Public Administration Reform Strategy PDPG Programmatic Development Policy Grant PEFA Public Expenditure and Financial Assessment PEIR Public Expenditure and Institutional Review PEP Private Enterprise Partnership PETS Public Expenditure Tracking Surveys PFM Public Financial Management System PFMMP Public Financial Management Modernization Project PIP Public Investment Program PIT Personal Income Tax 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 PSRP Public Sector Reform Project QFD Quasi fiscal deficit RM Reserve Money SAC Structural Adjustment Credit SECO Office of the Secretary of State for Economics SFCC State Financial Control Commission SME Small and Medium Enterprises SURE Seemingly Unrelated Regression Equations SWAps Sector-wide Approaches TA Technical Assistance TFP Total Factor Productivity TIN Taxpayer IdentificationNumber TOE Tons of Oil Equivalent UNICEF United Nations Children’s Education Fund USAID United States Agency for International Development VAT Value added tax

Country Director: Annette Dixon Table of Contents

PREFACE ...... 1 EXECUTIVE SUMMARY ...... 1 1. INTRODUCTION ...... 1 A . BACKGROUND...... 1 B . THE DEVELOPMENTCONTEXT IN TAJIKISTAN...... 2 c . MAINFINDINGS AND RECOMMENDATIONS OF THE 2003/04 PEIR ...... 4 D. KEYCHALLENGES FOR PUBLIC EXPENDITUREREFORM ...... 6 E. THE THEMESAND CONTENTS OF THE PROGRAMMATIC PUBLIC EXPENDITUREREVIEW ...... 8 F. SEQUENCING THE WORK OF THE PPER ...... 12 2 . ECONOMIC GROWTH: CONSTRAINTS, CHALLENGES AND IMPLICATIONS FOR PUBLIC EXPENDITURE...... 13 A . INTRODUCTION...... 13 B . EXPLAINING TAJIKISTAN’SECONOMIC GROWTH PERFORMANCE ...... 14 c . CHALLENGES AND CONSTRAINTS TO SUSTAINABLE FUTURE GROWTH ...... 19 D. COMPARATIVEADVANTAGE AND POSSIBLE GROWTHTRAJECTORIES ...... 24 E. MODELING THE GROWTH EFFECTS OF PUBLIC EXPENDITURE ...... 28 F. CONCLUSIONS AND IMPLICATIONS FOR PUBLIC EXPENDITUREPOLICIES ...... 30 3. FISCAL SPACE IN TAJIKISTAN ...... 34 A . INTRODUCTION...... 34 B . FISCAL POLICY SINCE INDEPENDENCE ...... 35 c . FISCAL SPACE FOR PRIORITY EXPENDITURES...... 39 D. DOMESTIC REVENUES ...... 39 E. BUOYANCYAND INCOME ELASTICITY OF REVENUES ...... 41 F. CHANGES IN TAXPOLICY ...... 42 G. TAX ADMINISTRATION...... 44 H. GOVERNMENTBORROWING ...... 49 I. FISCAL IMPLICATIONS OF REFORMS IN THE ELECTRICITYSECTOR ...... 56 J . RESOURCE ENVELOPEPROJECTIONS ...... 64 K. THE EFFICIENCY OF GOVERNMENTEXPENDITURE ...... 66 L. CONCLUSIONS ...... 67 4 . HOW CAN THE MTEF BE INTRODUCED IN TAJIKISTAN? ...... 72 A . INTRODUCTION...... 72 B. MOTIVATIONFOR INTRODUCING THE MTEF ...... 73 c . WHAT HAS BEEN ACHIEVED SO FAR? ...... 74 D. CONSTRAINTS TO INTRODUCING SECTOR BASEDPLANNING ...... 75 E. WHAT CHANGES ARE REQUIRED TO SUPPORT THE INTRODUCTION OF THE MTEF? ...... 79 F. HOW SHOULD THE INTRODUCTION OF THE MTEF BE SEQUENCED OVER TIME? ...... 86 G. STRATEGY FOR PHASED IMPLEMENTATION OF THE MTEF ...... 90 H. CONCLUSION ...... 93 APPENDIX 1 MEDIUM-TERMBUDGET FRAMEWORK ...... 94 APPENDIX 2: ACTION PLAN OF STEP TO IMPLEMENT MTEF IN PILOT SECTOR 2006-2007 ...... 96 APPENDIX 3: REGIONAL LESSONS FROM THE DEVELOPMENT OF MTEFS ...... 98 5. PUBLIC EXPENDITURE AND FINANCIAL ACCOUNTABILITY (PEFA) ASSESSMENT ...... 105 A . INTRODUCTION...... 105 B. INTEGRATED ASSESSMENT OF PFM PERFORMANCE ...... 106 c . ASSESSMENTOF THE IMPACT OF PFM WEAKNESSES ...... 110 D. PROSPECTS FOR REFORM PLANNING AND IMPLEMENTATION...... 1 12 Annexes

ANNEX1. THEIMPLICATIONS OF PUBLIC EXPENDITUREFOR ECONOMICGROWTH IN TAJIKISTAN...... 119 ANNEX11. EXPENDITURESOF THE STATE BUDGET OF THE REPUBLIC OF TAJIKISTAN 2005-2010 ...... 156 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 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 UK Department for International Development (DFID) and the Office of the Secretary of State for Economics (SECO) of Switzerland.

The PPER is intended to be 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 these PFM reforms through analytical and diagnostic work.

The PPER has been designed to cover five themes which represent the key challenges for fiscal policy and PFM reforms in Tajikistan. The themes are as follows: Theme 1: Public Expenditure, Fiscal Space and Growth; Theme 2: Policy-based Budgeting; Theme 3: Efficiency of Public Expenditure in the Social Sectors; Theme 4: Fiduciary Risks; and Theme 5: Inter-governmentalfiscal relations.

The PPER will be undertaken over three fiscal years (2006/07-2008/09) and published in three volumes, of which this represents the first. The work in Volume 1 covers two of the issues under Theme 1 - the implications of public expenditures for economic growth and long-term fiscal space for priority expenditures - which respond to the need to analyze how scarce budget resources can best be allocated to meet the Government’s developmental objectives for growth and poverty reduction. Volume 1 also includes an outline of the strategy for implementing the MTEF, which responds to a decision taken by the Government in 2006 to begin implementing the MTEF in 2007, and a summary of the PEFA results.

This report was written by Sudharshan Canagarajah (TTL) and Martin Brownbridge (Consultant). The PPER team include: Sudharshan Canagarajah, William Dillinger, Jariya Hoffman, Utkir Umarov, Shuhrat Mirzoev (ECSPE); Ernest0 Cuadra, Dina Abu-Ghaida, Anne Bakilana, Peyvand Khaleghian, Saodat Bazarova, Vlaidmir Kolchin (ECSHD); Santiago Cornejo, Logan Brenzel (HDNHE); Waly Wane (DECRG); Pascale Kervyn (ECSPS); Aziz Khaidarov (ECCTJ); Martin Brownbridge, Nathan Smith, Taras, Pushak, and Victor Murinde (consultants). The peer reviewers are Anand Rajaram (PRMPS), Mary Betley (external - DFID) and Elena Nikulina (ECSPE). Overall guidance for this report was provided by Carlos Felipe Jaramillo and Cheryl Gray. Takhmina Jumaeva and Damika Somasundaram assisted the team. EXECUTIVE SUMMARY

Chapter 1: Introduction

1. The Programmatic Public Expenditure Review 2007-2009 (PPER) aims to provide analysis, diagnosis and policy recommendations in respect of the key medium and long term challenges facing fiscal policy and Public Financial Management (PFM) in Tajikistan. The PPER is programmatic in the sense that its work program extends over three years and encompasses a planned three volumes, of which this is the first. The extended time scale of the work program reflects the broad scope and depth of the PFM issues which need to be covered as well as the intention to align the work program with ongoing technical assistance (TA) to support PFM reforms and, therefore, to sequence the work appropriately.

2. The PPER was launched in May 2007 at a workshop in Dushanbe with senior Government of Tajikistan (GOT) officials, where the substantive themes to be covered in the PPER were discussed and agreed upon. A public expenditure working group, chaired by the Minister of Finance, was set up and provides the formal link between the Bank (and other donors) and the GOT for purposes of planning the PPER. Development partners, notably DFID (UK) and SECO (Switzerland) have contributed to the financing of the PPER. The PPER builds on the Public Expenditure and Institutional Review 2003/04 (PEIR) which provided the first comprehensive overview of public expenditure issues in Tajikistan (World Bank, 2005D).

3. The PPER is structured around five themes which represent the main challenges for PFM reform in Tajikistan and which cover the main PFM and fiscal policy issues which have been prioritized by the GOT and by the donors who are assisting the GOT with PFM reforms. These themes are: Theme 1: Public Expenditure, Fiscal Space and Growth; Theme 2: Policy-based Budgeting; Theme 3: Efficiency of Public Expenditure in the Social Sectors; Theme 4: Fiduciary Risks; and Theme 5: Inter-governmental fiscal relations.

4. Theme 1 responds to questions of fiscal policy design, especially related to the composition of expenditure and the magnitude of external borrowing, which are becoming increasingly important in Tajikistan and which are likely to have a profound impact on the long term development of the economy. The PPER takes a public finance perspective to evaluate how fiscal policy can best contribute to long term sustainable economic growth and how fiscal space can be created for priority public expenditures, taking account of the scope to mobilize domestic revenues, attract donor grants, access sustainable borrowing and improve expenditure efficiency. The analysis of growth oriented fiscal policy, presented under theme 1, is an approach which was recommended by the Development Committee ofthe Bank in April 2006.

i 5. Theme 2 responds to the decision taken by the GOTin 2006 to reinvigorate efforts to implement the Medium-term Expenditure Framework (MTEF) - which had begun in 2002 under the Bank’s IBA2 project - and to seek financial and technical support from development partners for this endeavor. The PEIR had identified the lack of linkages between budget allocations and strategic policy formulation as a major weakness of the budget process; a weakness which the introduction of the MTEF aims to address. The analysis undertaken under Theme 2 is primarily intended to support the TA for implementing the MTEF, by diagnosing the main constraints and making feasible recommendations which can facilitate the gradual introduction ofthe MTEF.

6. The social sectors, notably, education, health and social protection, will play a key role in achieving the Millennium Development Goals (MDGs), but the quality of social services provided to the population is impeded not just by shortages of budgetary resources but also by serious weaknesses in the allocative and technical efficiency of expenditures in these sectors. Theme 3 focuses on expenditure efficiency in the social sectors, an issue which has not hitherto been subject to rigorous analysis, using methodologies which include efficiency frontier analysis and Public Expenditure Tracking Surveys (PETS).

7. Previous diagnostic studies of the PFM system, including the Bank’s 2003 Country Financial Accounting Assessment (CFAA), the International Monetary Fund’s (IMF’s) 2003 Safeguards Assessment and the DFID’s 2005 Fiduciary Risk Assessment, identified serious weaknesses in fiduciary controls. These weaknesses also deter donors, who could potentially provide more aid to the government budget, including budget support, from doing so (currently a large share of donor aid is actually disbursed off budget, for example to Community Based Organizations). Theme 4 focuses on fiduciary issues and includes a comprehensive evaluation of the PFM system using the internationally accepted best practice methodology of the PEFA assessment.

8. Inter-governmental fiscal relations, between the various layers of Government, are both complex (as explained in the PEIR) and evolving in Tajikistan. They impact on various aspects of the PFM system, notably through the Local Budget (the budget for local governments) which accounts for the bulk of social sphere expenditures and in the collection of tax revenues. Reform of the PFM system, including the introduction of the MTEF, must therefore take account of inter-governmental fiscal relations. Theme 5 will provide advice on how the structure of inter-governmental fiscal relations can be reformed to best accommodate the requirements of the MTEF, other PFM reforms and the GOT’SPublic Administration Reform Strategy (PARS).

9. This volume of the PPER focuses mainly on Themes 1 and 2, with chapters on how public expenditure can best support economic growth, and on fiscal space for priority expenditures (under Theme 1) and the introduction of the MTEF (Theme 2). It also includes a summary of the PEFA, the main report of which will be published separately. The work under Themes 3 and 5 requires thore time to complete and will be published in subsequent volumes. It is anticipated that further work under Theme 2, as

11 which the economy needs most to generate higher growth rates is not investment in public capital such as infrastructure, because the economy is already quite well endowed with this type of capital, but investment for the production of marketable goods and services, for which the private sector is much better suited than the public sector. Third, the private sector, driven by the profit motive and operating in competitive markets, faces much stronger incentives to generate improvements in factor productivity than does the public sector.

15. Private sector investment rates are very low, at 5-6% of the Gross Domestic Product (GDP), despite the recovery of the economy. One of the main reasons for the lack of private investment, as found in surveys of the private sector. such as the 2005 EBRD-World Bank Business Environment and Enterprise Performance Survey (BEEPS), is the poor institutional climate for investors, characterized by multiple complex and untransparent regulations administered by many different Government agencies which give scope for discretion on the part of public officials and offer opportunities for corruption. In reality, there is no level playing field for private investors; instead each firm makes its own individual arrangements with officials, a situation which is extremely damaging for efficient resource allocation and is anathema to a predictable and stable business climate.

16. Consequently, a major policy priority should be the urgent implementation of measures to improve the institutional climate for private investors, by reforming the systems for permits and licenses and tax administration, so that a more simple, predictable and transparent regulatory framework is put in place, to create a clear set of rules which apply to all investors. The Government has begun to implement reforms in this area, by streamlining licensing procedures, for example, but much more needs to be done to strengthen the investment climate.

17. Tajikistan appears ta have a comparative advantage in the generation of hydroelectric power for export to neighboring countries, and a number of hydropower projects (HPPs) are currently under construction or being planned. The PPER argues that the Government should allow private investors to play the leading role in owning and managing all of the new HPPs for two reasons. First, given the huge scale of investment resources required (several billions of dollars for all of these projects combined), the Government does not have the financial resources or borrowing capacity to fund these investments without jeopardizing external debt sustainability. Second, energy projects can be operated on a commercial basis and are, therefore, more suitable for private sector investment 'which can provide more efficient management of these projects than can the public sector. However, while the generation and export of hydropower is potentially profitable, if long term export contracts can be secured, it will create very few jobs once the construction phase of these projects has been completed. It should not comprise the sole engine of growth for the economy; and while HPPs can contribute to long-term development of the Tajik economy, they should not be viewed as a panacea.

18. Given that budgetary resources are scarce and that public sector borrowing capacity is very constrained, the careful prioritization of Government expenditures is

iv essential. The PPER cautions against large public new investments in infrastructure, even if these have the attributes of public goods (e.g., roads) and will not, therefore, be supplied by the private sector. Tajikistan already has a large stock of public capital, relative to the size of its economy, which is a legacy of heavy investment during the Soviet era, but the Government lacks the budgetary resources to maintain this adequately. Moreover, surveys of the private sector do not indicate that deficiencies in the public infrastructure are a major constraint to private sector business. A World Bank regional study of transport and constraints to trade in Central Asia, including Tajikistan, concluded that the road networks in the region are relatively extensive and largely sufficient to meet the needs of users, although roads are in poor condition because of inadequate maintenance (World Bank, 2004B). Consequently it is unlikely that investment in new public capital assets is a priority for the economy. Instead, priority should be given within the budget to funding adequate maintenance of the core networks of the existing public infrastructure; expenditures which are likely to offer higher rates of return than investment in new capital assets.

19. Labor productivity is low but can be improved over the long term through better public education and health services. Therefore, expanding access to, and improving the quality of, public education and health services should be a priority for budget policy, with larger budget allocations to these sectors alongside measures taken to improve expenditure efficiency. Public provision of these basic social services is essential because, given low per capita incomes, private provision is not affordable for most of the population.

20. To summarize, this chapter of the PPER advocates a development strategy in which priority in the Government budget is placed on expanding and improving the provision of the education and health services needed to boost human capital development, alongside the essential maintenance of the existing core infrastructure networks. New capital projects should not be a priority for the budget, because it is not evident that an expansion of the public capital stock is really essential to support economic growth and because they would inevitably crowd out funding for essential public services and maintenance. Government should leave the private sector to invest in commercially viable infrastructure projects as well as a broad range of other, more labour intensive, goods and services, but increased private investment will require a marked improvement in the institutional climate for business. Such a strategy offers the prospect of sustainable economic growth with employment creation, rising labor productivity which will raise real wages, and improvements in public services which will enhance the non income aspects of human welfare.

21. There is a danger that poor public policy decisions, influenced by an excessively optimistic view of the benefits to be gained from large infrastructure projects, lock Tajikistan into an inferior development trajectory. Such policy choices would include heavy public investment in infrastructure, funded by external borrowing, the neglect of education and health services and the failure to improve the institutional climate for private investment. The outcome of such policies is likely to be capital intensive growth

V which generates few jobs and, therefore, does little to reduce poverty, and in the long run may be derailed by an unsustainable public debt burden.

Chapter 3: Fiscal Space in Tajikistan

22. The concept of fiscal space refers to the additional budgetary resources which could be mobilised for priority public expenditures. This chapter adopts the fiscal diamond framework, which was proposed in a paper presented to the Bank’s Development Committee Board in April 2006, to evaluate fiscal space over a 10 year time frame. The fiscal diamond entails analysis of the scope for mobilising resources from four sources: domestic revenues, donor grants, sustainable borrowing and expenditure efficiency improvements, while ensuring that fiscal policy is consistent with macroeconomic stability. This chapter also includes analysis of the budgetary implications of reforms in the electricity sector to reduce the QFD (estimated at 9% of GDP in 2006) because this is critical for long term fiscal sustainability.

23. The fiscal space projections offer grounds for guarded optimism, but also raise important warnings about fiscal sustainability if key policy reforms are not implemented. Fiscal space equivalent to 8% of GDP per annum could potentially be mobilized by 20 16, but this is contingent on reforms to tax administration and PFM systems. Without these institutional reforms, very little fiscal space will be created. The dangers to fiscal sustainability arise from excessive external borrowing and the large QFD.

24. Domestic revenues could potentially be increased by 5-6% of GDP (from nearly 19% of GDP in 2006 to over 24% of GDP in 2016). This is the most important potential source of fiscal space. Tax administration reforms (discussed in paras 51-53 below) contribute the bulk of the increase. Given that there is substantial scope for improving . the efficiency of the existing tax administration, the PPER projects that, if tax administration reforms recommended by the IMF and the Asian Development Bank (ADB) are implemented effectively (which will take time), tax revenues could rise by about 20% over 10 years, which would raise additional tax revenue of more than 3% of GDP per annum after 10 years.

25. There is only limited scope to raise more revenue through tax policy reforms for two reasons. First because rates are already quite high by international standards for two of the most important taxes; 20% for value added tax (VAT) and 25% for corporate income tax, and so further increases in these rates is not advisable to avoid distortions. Second, although customs duty rates are relatively low, most imports enter Tajikistan duty free under regional trading arrangements with the Former Soviet Union (FSU) countries; raising customs tariffs will not generate much additional revenue. The most promising option for increasing tax rates is provided by the personal income tax (PIT), because current rates are low, with the top marginal rate being only 13%. Less than 5% of labor income was paid as PIT in 2004. The PPER recommends raising PIT rates, including raising the top marginal PIT rate to 25%, which is the same level as that of the corporate income tax (which would remove a distortion and an avenue for tax avoidance). Higher PIT rates could potentially double the share of labour income paid as tax and

vi generate revenue of about 1% of GDP. Higher PIT rates are also justified on the grounds that they would make the tax system more progressive.

26. Revenues from new HPPs, comprising both taxes or royalties on the export of electricity and returns to the GOT equity stake in these projects, could contribute around 1% of GDP to revenue by 2016, although this will depend on factors which are not yet certain, such as the price paid by neighboring countries ,for electricity and the cost of transmission to export markets.

27. Donor grants currently constitute a relatively small share (about 5%) of the budget resource envelope. The bulk of donor grants are disbursed off-budget, mainly to CBOs, which partly reflects donors’ concerns about fiduciary weaknesses in the Government budget. In 2005, donors disbursed at least US$109 million in grants off-budget, compared to only US$18 million in the form of on-budget grants. Fiscal space would be created by an increase in donor grants to the budget, arising from an overall increase in grant aid to Tajikistan, in line with commitments made by the G-8 countries to increase their aid to the poorest countries, and/or by donors shifting a larger proportion of their aid to the Government budget. The latter will require improvements in public financial management to strengthen and make more transparent the budget process and to improve the allocation of expenditures, especially in those sectors which can contribute to meeting the MDGs.

28. It is very hard to predict how donor aid to Tajikistan will evolve over a 10 year period, and the range of possible outcomes is large. Tajikistan might receive no increase in grants, because donors remain unconvinced that fiduciary standards are improving sufficiently and/or because Tajikistan contracts more loan finance from non traditional lenders, thereby reducing the perceived need for more grant aid. At the other extreme, grants could increase several fold as donors both increase grant aid to Tajikistan as part of a general expansion of aid to low income countries and channel a larger share of their grant aid to the government budget. The projection made in the PPER falls between these two extreme scenarios. Grants to the budget are projected to rise by 10% a year in nominal dollar terms, using the 2007 budget projection as a base. This will raise grants from 1.2% of GDP in 2006 to 2.2% of GDP in 2016.

29. Tajikistan’s external public and publicly guaranteed (PPG) debt was reduced to sustainable levels as a result of debt relief from bilateral creditors and the IMF and by the Government restricting borrowing for the Public Investment Programme (PIP) to 4% of GDP per annum during the 2000s. However, in 2006, GOT contracted new loans of US$604 million from China, as a result of which external loan disbursements will jump to an average of 14% of GDP during 2007-2009 (see Chart 1 below). Consequently, the Debt Sustainability Analysis (DSA), undertaken in early 2007 by the Fund and Bank, found that Tajikistan is at a high risk of debt distress and will breach two of the thresholds for external debt sustainability stipulated in the IMF/IDA Low Income Countries DSA Framework.

vii Chart 1: External loan disbursements as percent of GDP and the implied path of the NPV of PPG external debt to exports, 2005-2016

18%

180% 16%

14% 150% n 8 12% z Exp 120% Y W 10% 9 +Loan c f Disbursements to GDP - RHS E g 90% 8% E 0 P z 6% 60% \ -16 4 % 30% J 2% 0% 0% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2018 30. The approach taken in the PPER to calculate the amount of fiscal space that could be created by external borrowing is to assume that, after the peak in new loan disbursements during 2007-2009, further borrowing will be constrained so as to reduce the NPV of Public and Publicly Guaranteed (PPG) external debt to the thresholds of 150% of exports and 40% of GDP which are deemed sustainable under the IMFIIDA Low Income Countries DSA Framework by 2016. This implies that after 2009, external loan disbursements will have to fall back to an average of around 5% of GDP if the external debt is to be sustainable, as shown in Chart 1, which projects loan disbursements as a percent of GDP and the associated path of NPV of PPG external debt to exports from 2005 to 2016.

3 1. Over the long term, average annual disbursements of external public debt should be restrained to 56% of GDP at most if external debt sustainability is not to be jeopardized. Although this is higher than the level in 2006 (by 2.5% of GDP), it is much lower than the amount of loan disbursements projected over the medium term (2007- 2009), a level which is clearly not sustainable over the long term. These projections do not include non-PPG debt such as cotton debt. If the GOTwere to assume the cotton debt or borrow to fund debt relief for the cotton farmers, the ratios of NPV of external debt to exports and GDP would be higher and so the scope for new borrowing after 2009 would be correspondingly lower if the external debt is to remain sustainable.There is no room for domestic financing of the budget because the domestic financial market is very shallow; is too heavily dollarized, and is undiversified (the non-banking sector is very small).If Government were to borrow domestically, control of the money supply would be jeopardized (as happened in the 1990s) leading to high inflation, and private sector borrowers would be crowded out of credit markets. The long term sustainable rates of domestic borrowing should be slightly negative to allow monetary growth to be

... Vlll controlled, the National Bank of Tajikistan (NBT) to accumulate foreign exchange reserves and for private sector credit to grow.

32. The scope for efficiency improvements of public expenditure is hard to quantify because of data limitations. However, major allocative and technical inefficiencies characterize public expenditures, so there must be the potential for improving both types of efficiency through reforms to PFM systems, such as the introduction of the MTEF, which over the long term should create fiscal space. Given that Government final consumption and investment expenditures (i.e., excluding transfer payments and interest payments) amounted to 17% of GDP in 2006, even a relatively modest 10% improvement in expenditure efficiency over the long term would generate fiscal space amounting to 1.7% of GDP.

33. Electricity sector reforms will have a profound impact on the budget over the long term. The sector incurs a QFD estimated at 9.3% of GDP in 2006, which is measured as the difference between the revenue which Barki Tajik, the state owned electricity utility, requires to operate and properly maintain the existing infrastructure for the supply of electricity, and the revenue which it actually collects in cash. Sub-optimal electricity tariffs, which are only one quarter of the level needed to cover the long run average incremental cost (LRAIC) of domestic supply, contribute the largest part of the QFD, supplemented by technical and commercial losses and poor cash collection of electricity bills. Although the QFD is not currently a cash burden on the Government budget, it will become so if no measures are taken to eliminate it, because the consequences of the Government not funding the QFD will be that essential maintenance and rehabilitation of the electricity transmission and distribution system will be neglected, and hence electricity supply will deteriorate, with adverse consequences for the economy.

34. The most important reform needed to reduce the QFD is an increase in the average domestic tariff to the LRAIC of supply. Higher electricity tariffs will have a negative impact on some components of the budget, reducing tax revenue paid by the aluminum smelter, and increasing the costs to the budget of Government’s own electricity bills and of transfer payments to compensate low income electricity consumers under the energy compensation mechanism (ECM). The aggregate fiscal impact of electricity sector reforms, including the revenues mobilized from HPPs, is overwhelmingly positive, generating net gains of almost 10% of GDP, most of which constitutes the reduction of the QFD.

35. Table 1 below summarizes the fiscal space which could be created in 2016, compared to 2006, if the reforms identified above, tax administration, tax policy, PFM and the electricity sector, are implemented. Additional budgetary resources amounting to 8.6% of GDP could be mobilized, mainly from domestic revenues, but additional expenditures amounting to 1.9% of GDP would be required to fund higher interest payments (because of higher external debt), higher Government utility bills and ECM payments and to fund the remaining QFD in the electricity sector (reduced to 0.6% of GDP). Expenditure efficiency gains would add 1.7% of GDP to fiscal space; hence net fiscal space would amount to 8.4% of GDP per annum by 2016.

ix Table 1. Fiscal Space in 2016 compared to 2006; Yo of GDP

2016, change over 2006 Budget Resource Envelope 8.6 Domestic revenues 5.6 Grants 1.o I Financing I 2.0 I Net External Financing 2.5 Net Domestic Financing -0.5

Additional resources required for: 1.9 Interest Davments 0.4 Government electricity bills 0.5 Energy Compensation Mechanism 0.4 Funding remaining QFD 0.6

Gains in Expenditure Efficiency 1.7 (assumes 10% efficiency gain)

X Chapter 4: How can the MTEF be introduced in Tajikistan?

36. This chapter of the PPER responds to the GOT’Sdecision taken in 2006 to implement a MTEF and to seek technical assistance for this from donors. The introduction of the MTEF also reflects the desire of Tajikistan’s development partners to develop meaningful sector budget strategies, especially in the social sphere, around which they can provide aid to the budget. The chapter provides a broad strategy for implementing the MTEF in a gradual and phased basis, building on the strategy presented at a roundtable organised by the MOF in May 2006.

37. The motivation for introducing the MTEF is to improve budget resource allocation and to link this with the strategic policy objectives of the government, including those set out in the National Development Strategy (NDS) and the Poverty Reduction Strategy Paper (PRSP). Previous efforts to introduce an MTEF under the Bank’s IBTA2 project, had not gone further than the preparation of a rudimentary fiscal framework which provides a hard budget constraint for aggregate expenditure and, therefore, is a useful tool for macroeconomic management, but does not play any meaningful role as a budget planning tool at a lower level of expenditure than that of aggregate expenditure.

38. The major constraints to introducing an MTEF which can play a meaningful role as a tool for medium-term sector planning are two-fold. First, the State budget is very fragmented, with more than 100 Key Budget Organizations (KBOs); each prepare budgets and deal separately with the Ministry of Finance (MOF) in the budget process. In some of the major sectors, such as education, responsibility for preparing and implementing budgets is disbursed among multiple KBOs, including local governments and various types of central spending agencies (there are more than 40 KBOs in the education sector). The line ministries have the responsibility for preparing sector policies but are effectively marginalized from preparing the budgets for their sectors; hence there is a disconnect between policy formulation and budgeting, especially in the social sectors where the bulk of expenditures are channelled through the Local Budget and bypass the line ministries. In addition, the current and capital expenditures are treated separately in the budget preparation process, with the former being the responsibility of the Ministry of Economy and Trade.

39. The second major constraint to introducing an MTEF is the very weak capacities for budget planning in the MOF, line ministries and other KBOs. Familiarity with modern budget techniques and concepts is very scarce among public officials. The line ministries have very small budget planning departments because they have not previously had to undertake these functions. Budgets are essentially prepared by KBOs in a purely mechanistic way, following rules and formulae stipulated by the MOF in the budget circulars. Introducing a system of policy based budgeting such as the MTEF requires putting in place the skills needed for forecasting demand for public services, prioritising and costing expenditures and making choices between competing expenditure demands; these skills are virtually non-existent at the moment.

40. The MTEF strategy involves giving the line ministries a much greater role in the budget preparation process, so that budget allocations can be better aligned with

xi sector strategies. This will require changes to the budget process and a major strengthening of capacities for budgeting in the line ministries. In each sector, the line ministry responsible for policy formulation will be designated the Lead Line Ministry (LLM). At the start of the budget process, the LLMs will provide guidance to all KBOs in their sectors to ensure that budget submissions accord with sector policy. The KBOs will then channel their budget submissions for expenditures in each sector to the relevant LLM, each of which will collate all the submissions for its sector and prepare a consolidated sector budget to be submitted to the MOF. In addition, the MTEF strategy will integrate the capital and current budgets.

41. The severity of the capacity constraints means that implementation of the MTEF must of necessity proceed cautiously, with the introduction of new budget processes being matched to the building of technical capacities for budgeting in the MOF and line ministries, supported by effective donor TA projects. Consequently, the MTEF will be implemented in phases, beginning with a pilot sector; education. Education has been chosen as a pilot sector in part because donors are willing to provide a significant amount of TA to the sector. The pilot phase, which is expected to last for two years, will involve:

a) Strengthening the capacities of the MOF to produce a medium-term fiscal framework, through TA for revenue forecasting and macroeconomic forecasting;

b) Building a genuine sector budget process in the education sector, with the support of TA for the Ministry of Education from the Bank, the European Union (EU) and the United Nations Children’s Education Fund (UNICEF).

42. Additional sectors will be brought into the MTEF after the pilot phase has been completed, and the lessons learned in the pilot phase should help expedite this. As noted above, capacity building will be a prerequisite for success of the MTEF. Fortunately donors have prioritised TA for capacity building so funding for these activities should not be a serious constraint, but effective coordination and harmonisation of donor TA efforts will be essential. In this regard, a joint donor mission on supporting introduction ofthe MTEF, which was held in April 2007, was a useful step forward.

xii Chapter 5 PEFA Assessment

43. This chapter provides a summary of the PEFA assessment carried out between October 2006 and April 2007. The purpose of the PEFA assessment is to provide the GOT with an objective assessment of the country’s PFM systems, support a better understanding of the overall fiduciary environment of the budget and assist in identifying those parts of the PFM systems most in need of reform. The PEFA assessment following the standard PEFA methodology which grades 28 different aspects of the PFM system and 3 related to donor practises, with each indicator graded on a scale from A (highest) to D (lowest). Table 2 presents the grades for each indicator as evaluated by the PEFA assessment.

44. Tajikistan’s overall performance against the PEFA indicators is poor. Out of the 28 PFM indicators, only two received A grades (aggregate revenue outturn [which has consistently over-performed against the budget] and the comprehensiveness of budget information); there were only four B grades, compared to 11 C grades and 7 D grades. Four indicators were not assessed because of lack of information. Each of the three indicators for donor practices received D grades, implying that the performance of donors is also poor.

45. Hence there is considerable room for improvement in PFM systems as well as with donor practices. Areas which scored badly included: classification of the budget; public access to information; multi year perspective for budget planning; predictability of the availability of funds for expenditure commitments; quality and timeliness of financial statements, external audit and the proportion of aid which is managed through national budget procedures. However, some of these deficiencies are being tackled through ongoing PFM reforms supported by TA from donors, such as the introduction of the MTEF, support for the external audit agency and reform of the internal audit system.

Policy Recommendations of the PPER

46. This section brings together the policy recommendations from each of the different chapters of the PPER as well as the preliminary findings of a just concluded Bank mission to Tajikistan to analyse the sequencing and content of reforms to PFM following on from the PEFA assessment. The policy recommendations of this volume of the PPER are discussed under three headings (which are not mutually exclusive); (i)institutional reforms; (ii)composition of public expenditures; and (iii) threats to the sustainability of public finances.

xiii Table 2: Overview of PEFA Assessment Results Indicator I Issue covered I Rating I A. PFM Out-Turns: Budget Credibility I

PI-17 Recording and management of cash balances, debt and guarantees C+ [pending finalization] PI-18 Effectiveness of payroll controls not assessable/further data required PI- 19 Competition, value for money and controls in procurement not assessablelfurther data reauired

47. Major institutional reforms are essential, to create fiscal space and are also a prerequisite for translating this fiscal space into realizable benefits for the economy, public services and poverty reduction. Institutional reforms are needed in three areas in particular; PFM systems, tax administration and the business climate for private investment.

xiv 48. PFM reforms will create fiscal space, by improving the efficiency of expenditures and thereby freeing up resources for other budget priorities and by enhancing the credibility of the budget as a tool for poverty reduction, which will help the Government to mobilize more donor grants for the budget. In addition PFM reforms to improve the allocative and technical efficiency of expenditure are needed to strengthen the delivery of public services. Key PFM reforms which are currently being implemented or planned are: (i) the introduction of the MTEF, beginning on a pilot basis in the education sector; (ii)the implementation of wage reform in the civil service and health sectors, to strengthen incentives for critical staff; (iii) the introduction of automated Treasury systems at the local level, which are linked to the MOF and which will eventually be followed by the introduction of a computerised Financial Management Information System (FMIS; (iv) the development of a modern internal control and audit system; (v) the alignment of accounting methodologies with international public sector accounting standards; and (vi) the introduction of a new budget classification system.

49. The PFM reform agenda is ambitious, especially given the severe capacity constraints in the public service. Hence the PFM reforms should not be hurried and should be sequenced to evolve in parallel with the design of methodology, enactment of new legislation and capacity building. A comprehensive capacity building program is required to train staff throughout the public service, and especially in the line ministries, in modern budgeting and accounting methodologies.

50. The success of PFM reforms will depend to a large extent on the details of their implementation. Therefore it is essential that each of the PFM reforms is well planned and realistic. Implementation should focus on key aspects of the reforms which are essential for improving PFM systems in Tajikistan, rather than attempting to raise the whole of the PFM system to international standards, which is not feasible. A system-wide perspective should be taken to help determine the best program of reforms in terms of outcomes and risks. This means attending to the weak links in the PFM system and focusing resources on the areas of highest potential gain.

51. Tax administration reforms are needed for two reasons. First, to create fiscal space by strengthening revenue mobilisation: tax administration reforms could potentially generate tax revenue equivalent to more than 3% of GDP per annum by 2016. Second, to improve the business climate for the private sector by making the tax system more transparent, fair and less open to abuse: the BEEPS reported that 80% of businesses pay bribes to tax inspectors (World Bank, 2005A: 16-17).

52. A strategy for tax administration reform has been recommended by the IMF (Harrison et al, 2005) and the ADB. The priority for reform is to introduce modern tax administration practices facilitated by computerization. Self assessment by taxpayers, backed up by selective risk based audits, should replace the current system whereby tax officials attempt to verify manually every tax return. Over the longer term, the structure of the tax department should be re-organized to reduce the number of small tax offices in the regions which play a marginal role in tax collection. Large taxpayers should be handled by the large Taxpayer Unit and model tax inspectorates should be established in each of the three regions to handle medium sized taxpayers. The tax police should be abolished, as this unit is not compatible with modem tax administration practices.

xv 53. Over time, the closure of small offices, together with computerization and changes in working practices, will allow staff numbers in tax administration to be reduced. This will create resources for raising staff salaries substantially, so as to strengthen incentives to attract and retain higher caliber staff and reduce incentives for corruption. Tax officials also require substantial training, especially in auditing.

54. As well as tax administration reform, improving the institutional climate for the private sector to attract more private investment will require cleaning up the system for business permits and licences. The aim should be to reduce the absolute number of permits and licences required and the number of different agencies involved in processing them and to put in place a system which is transparent, predictable and does not offer opportunities for arbitrary decision making on the part of public officials.

55. The importance of the institutional reforms discussed above cannot be over- emphasised. These reforms are essential to improve economic governance. World Bank research on the transition economies of Europe and Central Asia (ECA) found that the impact on economic growth of the size of the Government budget depends on the quality of governance (Varoudakis, 2006). In countries with poor governance, an increase in the size of the Government budget impedes growth, because a large tax system characterised by poor governance seriously distorts resource allocation while poor governance reduces the positive benefits of potentially productive public expenditures. It is only in countries with good governance that creating fiscal space for productive expenditures will actually be translated into higher growth. Tajikistan scores badly on most measures of governance; it is ranked second worse among 27 ECA countries in terms of government effectiveness. Consequently, improvements in governance are a prerequisite if increased Government expenditures in the priority sectors are to bring about higher economic growth rates; improved governance must precede an expansion of the budget.

56. The PPER makes clear recommendations about how the allocation of expenditures within the Government budget should be shifted to support sustainable economic growth and poverty reduction. Increased public investment, whether in commercial energy projects or in infrastructure projects which have the character of public goods, should not be a priority for the budget. Commercial energy projects can be owned and financed by the private sector, without the (or with minimal) need for direct Government funding and economic growth does not depend on new public infrastructure projects. Moreover, the huge funding requirements for public investment projects will inevitably crowd out resources for priority current expenditures and/or undermine external debt sustainability.

57. Over the medium-term (2007-2009), slightly over 50% of the entire Government budget will be allocated to capital expenditures, leaving virtually no room for any real increases in the funding for current expenditures. Given the constraints on current budget expenditures, it is not even clear how the new public assets created by this surge in capital spending can be properly maintained over the long term. Once this surge in capital projects is completed (in 2009), the budget should be rebalanced towards recurrent expenditures, which is where more funds are needed the most.

xvi 58. Instead of new capital projects, priority in the Government budget should be accorded to the key social sectors, health and education, which can strengthen human capital formation. These are services which will not be provided optimally by the private sector, if only because most people in Tajikistan cannot afford private provision, so the Government must ensure comprehensive provision of basic health and education services. Allocating more budgetary resources to health and education, alongside PFM reforms to improve the efficiency of expenditures in these sectors, is necessary to expand access to and improve the quality of health and education services. In turn this will enable the quality of the workforce to improve and thereby boost labour productivity, which will be an important source of future economic growth.

59. There are problems with the existing public infrastructure, not least because a lack of funds has prevented most of it from being properly maintained since independence. Therefore the priority in the budget should be to ensure that sufficient funds are made available for maintaining adequately the core networks of the existing infrastructure. This will generate higher returns to spending than investment in new infrastructure assets.

60. The potential benefits to be derived from institutional reforms and the creation of fiscal space for priority expenditures will be undermined if the long-term sustainability of public finances is put in jeopardy. There are two main threats to fiscal sustainability in Tajikistan, excessive external borrowing and the QFD of the electricity sector.

61. Tajikistan faces a high risk of external debt distress as a result of large new loans contracted in 2006 for infrastructure projects. External loan disbursements during 2007-09 will average 14% of GDP, a level which is not sustainable. To ensure long term external debt sustainability it will be necessary to restrict new loan disbursements after 2009 to an average of around 5% of GDP per year. The argument sometimes made to justify heavy external borrowing, that the projects financed by the loans are economically viable and will generate large benefits, are not tenable. External debt sustainability is a macroeconomic concept and the external debt burden must be evaluated in relation to macroeconomic variables, such as the size of GDP, exports and Government revenues, irrespective of the benefits yielded by individual projects. Key internationally recognized thresholds for external debt sustainability, such as the Net Present Value (NPV) of debt to exports and GDP, will be breached by very wide margins if external borrowing is not reduced sharply after 2009.

62. The policy recommendation of the PPER is that the GOT should set clear annual limits on the level of new external borrowing for the medium- and long-term, which are consistent with bringing the NPV of external public debt back into line with internationally accepted thresholds for debt sustainability in a low income country, such as those defined in the IMF/IDA Low Income Countries DSA Framework. This will give a clear signal to the private sector and Tajikistan’s development partners that the Government will implement sound debt management and ensure that a future debt crisis is avoided.

63. The electricity sector’s QFD will inevitably become a burden on the budget over the long term unless it is eliminated through electricity sector reforms, because the

xvii consequences for the economy of not properly maintaining and rehabilitating the transmission and distribution network will be too damaging. Therefore, the Government budget will have to fund these essential expenditures if Barki Tajik cannot afford to do so. Furthermore, the large size of the QFD, estimated at more than 9% of GDP in 2006, means that it will severely squeeze the fiscal space available for priority expenditures and/or force Government to incur unsustainable borrowing.

64. The electricity sector reforms needed to eliminate the QFD are not in dispute. Average electricity tariffs must be increased to around four times the current level. In addition, Barki Tajik must reduce its commercial losses by improving its billing and collection systems and reduce technical losses through proper maintenance and repair of the distribution system. These reforms will however be politically painful, as consumers who currently enjoy subsidised electricity will have to pay much higher tariffs.

65. As noted above, this is the first of a planned three volumes of the PPER. Subsequent volumes will make further policy recommendations. There will be further recommendations related to the implementation of the MTEF, based on the practical experience of those working on the ground to provide TA and build capacity in the MOF and line ministries. Recommendations will also be made to improve expenditure efficiency in the health and education sectors. In addition, a comprehensive PFM reform strategy, which is agreed with the GOT and can be supported by development partners, will be set out.

xviii 1. INTRODUCTION

A. BACKGROUND

1. The PPER aims to deepen understanding of key analytical issues related to fiscal policy and PFM in Tajikistan and to support the ongoing PFM reform program in the country. The fiscal policy and PFM issues addressed in the PPER include the need to focus fiscal policy on supporting sustainable economic growth and poverty reduction, the introduction of policy based budgeting, improving the efficiency of public expenditures and strengthening fiduciary controls. The analysis and diagnostic work in the PPER will provide the basis for the preparation of a government owned PFM reform strategy and action plan which will guide the policy dialogue and technical assistance on PFM reforms.

2. The PPER adopts a programmatic approach with the work taking place over three years (2006/7-2008/9). The PFM reform agenda in Tajikistan is wide and challenging and the volume of work to be covered is accordingly large; hence the three-year time-frame of the PPER. The sequencing of the work program of the PPER is guided by dialogue with the GOT with the aim of supporting the ongoing PFM reforms being undertaken in the country. This volume represents the findings from the first year’s work of the PPER.

3. A workshop to launch the PPER and discuss the three year agenda was held in Dushanbe on 15 May 2006 and attended by senior GOT officials. The views expressed at the workshop guided the design of the PPER and the associated work program. To facilitate Government ownership of the PPER, a Public Expenditure Working Group, which includes representatives from the MOF, other key government agencies and donors supporting PFM reforms, was established to coordinate the work program for the PPER and meets regularly.

4. A Public Expenditure and Institutional Review (PEIR), undertaken in 2003/04, provided the first comprehensive overview of public expenditure issues in Tajikistan (World Bank, 2005D). Together with a 2003 Country Financial Accountability Assessment (CFAA) and 2003 Country Procurement Assessment Report (CPAR), the PEIR identified critical weaknesses in the processes for allocating budget resources, in budget execution and PFM. However, implementation to date of PEM and PFM reforms has been at best slow and uneven. It is evident that neither the GOT nor the donors has fully comprehended the complexities of implementing these reforms in the Tajik context: on the GOT side it is unlikely that the reform steps have been fully owned or supported, while the donors have not collectively analyzed the institutional characteristics of the budget processes in Tajikistan and the difficulties which this has posed for the implementation of reforms. However, in 2006, the GOT took a number of steps to move the PFM reform agenda forward, including the introduction of a

1 MTEF and the drafting of a PARS. The PPER responds to, and supports, the GOT’S ommitment to its own PFM reform agenda.

5. The PPER is organized around five themes which represent the critical challenges for budget system reform in Tajikistan. The rationale for these five themes is that they cover the main areas of fiscal policy and PFM reforms which have been prioritized by the GoTand by the donors who are assisting the Go with PFM reforms. These five themes are: (i)strategic fiscal policy issues: aligning fiscal policy and public expenditures with the requirements to support sustainable economic growth and poverty reduction; (ii)reform of the budget process to support the introduction of policy based budgeting; (iii)efficiency of public spending in the key social sectors of health, education and social protection; (iv) public financial management, especially internal controls to prevent corruption, reporting of budget outturns, the internal and external audit functions and procurement; and (v) intergovernmental fiscal relations, especially relations between the center and local governments (local hukumats). These themes are very similar to the issues that have been identified as critical for improving PFM in other Commonwealth of Independent States (CIS) countries

6. The recommendations arising from the PPER will be brought together to comprise a PFM reform strategy and action plan. This action plan will provide the main policy document of the GOT to guide PEM and PFM reforms and will also serve to guide the technical assistance programs of interested donors (e.g., World Bank, European Commission (EC), DFID, ADB, and the United States Agency for International Development (USAID). The Public Expenditure Working Group will oversee implementationof the PFM reform strategy and action plan.

7. The rest of this introduction is organized as follows. Section 2 briefly reviews the developmental context in Tajikistan. Section 3 summarizes the findings and recommendations of the 2003/04 PEIR. Section 4 outlines the most critical challenges for public expenditure reforms. Section 5 outlines the objectives, activities and outputs of each of the five themes of the proposed PPER. Section 5 also notes which of the reports under each of the themes are included in this volume of the PPER and which will be included in subsequent volumes.

B. THE DEVELOPMENTCONTEXT IN TAJIKISTAN

8. Tajikistan is a low income country which is recovering from civil war and economic collapse and where economic and budget management is impeded by severe institutional weaknesses. The World Bank, alongside other donors, is providing support to policy reforms through a series of Programmatic Development Policy Grants (PDPG) and technical assistance projects such as the Public Sector Reform Project (PSRP) and Public Financial Management Modernization Project (PFMMP). Reforms to public financial management (PFM) and public administration reform (PAR) are important components of the overall economic reform program of the GOT, supported by the Bank in the Country Partnership Strategy (CPS) and through the PDPG series.

9. Tajikistan gained independence from the Soviet Union in 1991. During the Soviet era, what is now Tajikistan was a state-controlled command economy dominated by the cotton and aluminum industries and heavily subsidized by transfers

2 from the rest of the Soviet Union. The Soviet era left two important legacies for the independent state of Tajikistan. First, much of the production structure was very rigid and ill suited to the demands of a market economy. Second, Tajikistan inherited a relatively well developed economic and social infrastructure.

10. Following the break up of the Soviet Union, Tajikistan suffered economic collapse as a result of three factors: the disruption to integrated trade and production systems within what had been the Soviet Union, the ending of subsidies from the Soviet Union and a civil war which raged from 1992 to 1997. Real GDP fell by more than 70% before it began to recover at the end of the 1990s. The restoration of peace and security throughout the country combined with the bringing back into production of underutilized capacity allowed for a rapid recovery of output, with real GDP growth averaging just under 10% per annum during 2000-2005. Nevertheless, aggregate real output is still about 30 percent lower than its level in 1990.

11. Moreover the living standards of the Tajik population suffered both from the fall in real GDP and the ending of subsidies from the Soviet Union. Tajikistan now has the status of a low income country, with a per capita GDP of around US$360 per annum. With employment opportunities severely limited in Tajikistan, approximately one third of the working age population is estimated to migrate abroad for work, either on a seasonal or more permanent basis, with the majority of migrants working in Russia. The incidence of income poverty was estimated at 64 percent in 2004, although the country fares better with respect to some of the indicators of human development. '

12. The economic collapse and ending of Soviet subsidies also had profound implications for the real value of government's budget resources, which fell dramatically. Prior to independence, about 40% of the budget for what is now Tajikistan, was subsidized by transfers from other parts of the Soviet Union. The Government sector inherited by the independent Tajikistan is far too large to be funded adequately by the much reduced budget resources available to the Government, even though these budget resources are now supplemented by donor aid, mainly in the form of project support, amounting to around 4% of GDP.

13. Because there was no real effort to cut the size of public employment, real wages in the Government sector have plummeted to levels far below that which is required to provide even a minimum subsistence level of income; while funding available for the operational and maintenance expenditures of government agencies is only a fraction of that required to deliver public services or maintain the public infrastructure. The very low salary levels are an inducement to corruption among public servants.

14. Since the end of the 1990s, the GOT has implemented economic reforms which have begun to yield positive results. Macroeconomic management in the 2000s, supported by a Poverty Reduction and Growth Facility (PRGF) program agreed with the IMF, has improved, with inflation reduced to single digit levels in 2004 and 2005. As noted above, the economy has recovered more than half of the fall in production

' Human development indicators are generally better than the average for low income countries. For example, life expectancy at birth is 66 years, compared to the low income country average of 59 years. Gross secondary school enrolment is 82% compared to an average of 46% for low income countries.

3 suffered in the 1990s. The external public debt of Tajikistan has been reduced as a result of debt relief agreements negotiated with Russia and other bilateral creditors and more recently the Multilateral Debt Relief Initiative (MDRI) with IMF.2 Tajikistan has accumulated balance of payments surpluses in the last few years, enabling the country’s foreign exchange reserves to be gradually rebuilt. Hence the economy does not suffer from serious domestic or external macroeconomic imbalances.

15. Structural and institutional reforms are now the major challenges for Tajikistan as it seeks to develop its economy and reduce poverty. The priorities for reform efforts are: firstly, the promotion of private investment to provide employment and gradually shift the economy away from the rigid capital intensive production structures built up during the Soviet era and secondly, the restructuring of the public sector so that it can focus scarce public resources on the efficient provision of essential public goods and services. The latter issue is discussed in more detail in the following sections. c. MAINFINDINGS AND RECOMMENDATIONSOF THE 2003/04 PEIR

16. The 2004 PEIR focused on the following issues: Strategic Policy Formulation; Budget Management and Execution; the Centre-Sub-National Relations; and the Public Investment Planning and Process. Since the 2004 PEIR was the first such activity for Tajikistan, it focused on providing an overview of all the PFM issues. The main findings of the PEIR are as follows.

The budget is very poorly aligned to strategic policy formulation, mainly because the public administration is still organised along the lines of the old Soviet model, in which the budget was subordinated to central planning. The roles and functions of different institutions within public administration are often unclear or overlap. Fiscal relationships between central and local governments are highly complex and impede rational budget preparation linked to clear national priorities. Public administration is hampered by a severe lack of capacity, very low wages and corruption. Capital expenditures are fragmented and not integrated with the current budget. The delivery of public services has deteriorated because of budget resource constraints exacerbated by the weaknesses in budget management.

17. Some of the recommendations of the 2004 PEIR have led to important reform measures being implemented such as the establishment of a National Treasury, a budget audit agency and the enactment of a public finance law. However, the overall record of implementing budget reforms has been poor due to the very weak capacities

In 2000, external public debt had risen to 128% of GDP but this was reduced to just under 40% of

GDP in 2004 as a result of debt cancellation by creditors and the growth in GDP (IMF, 2005). In 2006, , the stock of debt increased when the GOT contracted loans of US604 million from China.

4 within the public service for understanding and managing reforms, a lack of strong ownership from the GOTand a lack of political will and specific technical know-how with respect to how to implement budget reforms. The sections below provide further details on the actions implemented under each category reviewed in the PEIR. The remaining challenges for budget system reform are discussed in detail in Section 4: these are areas which require additional analysis which is undertaken as part of the PPER.

(a) Strategic Policy Formulation and Policy Management

18. Tajikistan has retained most of the organizational features of government from the Soviet era, which are ill suited to the effective formulation and implementation of policy in a market economy. The policymaking and budget processes are highly fragmented, without a clear demarcation of responsibilities for different agencies and with several parallel resource allocation mechanisms. Although some steps to address these problems have been taken, notably the endorsement of the PARS, the establishment of a civil service register, the submission to Parliament of a civil service law and the endorsement of a civil service wage reform strategy, the major institutional reforms required to facilitate policy based budgeting, such as the implementation of the MTEF, have yet to be implemented.

(b) Budget Management and Execution

19. Given the institutional weaknesses referred to in the above paragraph, it is very difficult to translate strategic policies into coherent budget allocations. This has had profound consequences for the budget, because with the budget suffering a substantial decline in real resources, a strategic re-allocation of budget resources is essential. Instead Government has attempted to maintain all existing budget structures even though there are insufficient budget resources to fund them adequately. Some important budget reforms have been implemented, such as the enactment of a Public Finance Law, the establishment of a National Treasury and establishment of an external audit agency. However the major reforms to the organizational structure of the budget which are necessary to facilitate policy based budgeting have yet to be undertaken. These reforms, focused around introduction of the MTEF, are at center of the budget reform strategy to be implemented over the medium-term and are supported under Theme 2 of the PPER

20. Public investment expenditures are fragmented between the externally funded PIP and the Central State Investment Program (CSIP) funded by the State budget, are not properly integrated into the rest of the budget and are not guided by a single development strategy. Also, there is no formalized methodology for the selection of investment -projects according to transparent criteria. The Government has taken some actions to improve the PIP, which is now prepared on a three-year rolling basis and updated annually, while the external borrowing requirement associated with the PIP must be submitted to Parliament for approval. The priorities for reform in this area are to integrate planning of investment expenditures with the rest of the budget and to adopt rigorous formal procedures for evaluating and selecting projects for inclusion in the budget.

5 (d) Intergovernmental Fiscal Relations

21, Sub national units of administration play a major role in the provision of public services, including education, health care, and housing and community services. The system used to finance these services continues to bear the imprint of the Soviet era, requiring lengthy negotiations between tiers of government over spending needs and revenue prospects. As a result, sectoral ministries have little influence over the allocation of resources at the local level. The PEIR discussed several options for reform. At the strategic level, these included a shift of budget making authority from sub national units to the sectoral ministries. The Government has opted for this approach, which is a prerequisite for introducing a sector-based MTEF (see Chapter 4), but is still working out the procedural details of how to do this. The PEIR also recommended the introduction of more transparent criteria into the allocation of fiscal transfers. With the exception of the experiment in capitation-based education financing, no progress has been made on this front.

D. KEY CHALLENGES FOR PUBLIC EXPENDITUREREFORM

22. The main challenges for reform of public expenditures, which are addressed in the PPER, relate to the allocation of scarce budget resources to meet strategic policy objectives, the efficiency with which budget resources are used for service delivery, a range of fiduciary issues and the budgetary relationships between different tiers of government. The identification of these challenges was guided by the priorities identified in strategic planning documents of the GOT, such as the NDS, the views expressed by GOT officials, the findings of the PEIR and other diagnostic work undertaken by the Bank and other donors (e.g., the CFAA). They also reflect lessons drawn from fiscal reform programs among other CIS countries which have faced similar problems in reforming Soviet model public administration systems.

23. While there has been a marked improvement in the macroeconomic aspects of fiscal policy management since the late 1990s, important challenges are emerging which may threaten long term fiscal sustainability and macroeconomic stability. There is pressure from both within and outside Tajikistan to increase government spending on MDG related programs, although the budgetary resources that is available to fund an expansion of such programs, from domestic revenues or donor aid, are likely to be quite limited. The Government has also prioritized large infrastructure projects in the energy and transport sectors and is seeking to fund these with external borrowing. It contracted loans of US$604 million from China in 2006 to fund road rehabilitation and the construction of an overheard power line. It seems probable that Tajikistan may be able to access further large external credits to fund infrastructure projects, but the Debt Sustainability Analysis (DSA) undertaken by the Bank and Fund in early 2007 (IMF and World Bank, 2007) found that Tajikistan faces a high risk of debt distress because of the sharp rise in new borrowing.

24. Any substantial increase in Government expenditures also raise concerns over the absorptive capacity of Government, the burden imposed on the state budget to fund counterpart contributions to foreign financed capital projects and the impact on inflation through increased aggregate demand. After having been brought down into single figures through prudent fiscal and monetary policies in 2004 and 2005,

6 inflation rose to more than 12% in 2006, buoyed by a surge in aggregate demand as well as supply side shocks.

25. The allocation of budget resources involves two distinct but related problems: first, the process by which budget resources are allocated (the institutional relations of the budget process) and secondly, the substance of the budget allocations (i-e., how much does Government spend and what does it spend its resources on). The main developmental priority is to promote rapid and sustainable economic growth so that real per capita incomes can be raised. This implies that a key priority for government spending should be to support rapid economic growth. The robust economic growth recorded since 1998 was largely driven by the rebound from economic collapse during the 1990s, with existing productive capacity (e.g., in the cotton and aluminum sectors) brought back into production. But, as the economy moves back towards its production possibility frontier, it does not seem likely that these high rates of economic growth can be sustained for much longer without attracting new private investment in commercial enterprises and without public expenditures which can prevent the deterioration of existing infrastructures and/or build new infrastructure which complements private investment. In this context, how best to allocate budget resources to support economic growth is a key policy question.

26. The size of public sector - 385,000 employees - and the scope of its activities are too large given the very limited budget resources available to efficiently fund these activities. To avoid drastic cuts in public employment, real salaries have been cut to levels that are often too low to meet subsistence needs. The Government has also inherited from the Soviet period an extensive portfolio of capital assets (including economic and social infrastructure) which it cannot afford to operate or maintain properly.

27. Therefore, the public sector needs to scale back its activities and focus scarce budgetary resources on areas where market provision of goods and services is not feasible. However, even within the general area of public goods and services, rigorous prioritization will be necessary because legitimate demands far outstrip potential budget resources. There are important choices to be made between competing expenditure demands; for example, between human resource development, social protection through transfer payments to vulnerable individuals, maintaining the existing infrastructure and constructing new infrastructure. The trade offs in these situations are not always clear or easy to evaluate. The analytical work of this PPER is intended to facilitate understanding of these tough choices.

28. Fiduciary issues have moved to the center of policy discussions on financing government programs for economic growth and poverty reduction. Although Tajikistan moved from “post conflict” to a “gradual developer” three years ago, donor projects are heavily focused on social service delivery, and a majority of these projects are funded off budget and bypass local institutions due to capacity constraints and fiduciary concerns. This has led to a lack of coherence between the Government budget and donor projects resulting in a fragmentation and duplication of local service delivery. Unfortunately, independent assessments have identified serious concerns related to fiduciary risk in the budget and for this reason donors are reluctant to provide more of their aid in the form of budget support. In order to address these donor concerns, Government needs to undertake periodic reviews of fiduciary risks

7 which will assess the progress and recommend corrective measures to address these challenges.

29. Tajikistan’s intergovernmental fiscal relations are an important challenge for improving PFM. Because a majority of social expenditures are implemented through the local government budgets, it will not be possible to implement an MTEF based on sector planning in the key sectors without reforming intergovernmental fiscal relationships. As such, the recently adopted MTEF action plan includes changes in budget planning in terms of the role of line ministries and local Hukumats while per capita financing requires changes in national-sub national relations and resource allocation mechanisms.

E. THE THEMESAND CONTENTS OF THE PROGRAMMATIC PUBLIC EXPENDITURE REVIEW

30. Given the challenges outlined above, the PPER provides analysis to assist the Government to design and implement reforms under the following five themes.

Theme 1: Public Expenditure, Fiscal Space and Growth

3 1. How best to allocate budget resources to support economic growth is addressed in Theme 1 of the PPER, through a DFID-financed study on the determinants of economic growth and their implications for public expenditure. The study examines the factors behind Tajikistan’s recent strong economic recovery and assesses what will be required to sustain economic growth over the long term. It identifies which sectors of the economy are likely to be the drivers of growth over the long term, in light of the country’s potential comparative advantage, and reports the results of an endogenous growth model to examine econometrically the causal links between different types of public expenditure and growth in Tajikistan and other CIS countries. The findings of this study are presented in Chapter 2 of this volume of the PPER. Further work under this theme will examine the implications for poverty reduction of different trajectories of economic growth.

32. There are important questions pertaining to the potential size of the budget resource envelope and therefore the scope for expanding expenditures in priority areas to support growth and poverty reduction: for example, what scope is there for growth of domestic revenues?; what is a sustainable level of public sector borrowing?; how much more grants from foreign aid donors can be mobilized?; can resources be freed up through improvements in expenditure efficiency? These questions are covered in Chapter 3 of this volume of the PPER which includes long term projections of fiscal space utilizing the analytical framework proposed in a recent joint IMFBank report to the Development C~mrnittee.~ Theme 2: Policy-Based Budgeting

33. Improving the allocation of budget resources will require institutional changes to the budget process. The existing budget process has been shaped by the structures

“Fiscal Policy for Growth and Development: An Interim Report”, Background report by the IMF and World Bank for the Development Committee Meeting, April 2006.

8 inherited from the Soviet era, in which the Government budget was a component of the broader system of economic central planning rather than having its own separate functions and objectives, and by the decentralization which took place after the end of the civil war, which established a system of dual subordination which left the role of line ministries unclear. The budget process is extremely fragmented and complex, with over 100 KBOs each having their own budget and dealing directly with the MPF in the budget process. Many of these KBOs have responsibility for expenditures covering multiple functions. The line ministries have responsibility for policy formulation but their direct control over budgets in their respective sectors is highly circumscribed by the fragmentation of the budget. The structure of the existing budget system provides no institutional mechanisms for allocating budget resources in a coherent and transparent manner to meet strategic policy objectives, hence it is very difficult to link the strategy outlined in the NDS or PRSP to the budget. Instead, annual budget allocations are determined by a combination of rigid formulae linking allocations to existing structures and bargaining between the central institutions (such as the MOF) and KBOs. In addition, the current and capital budgets are essentially determined separately, with the capital budget controlled by the Ministry of Economy and Trade.

34. Introducing policy based budgeting in Tajikistan will require major institutional changes to the budget process. These changes will entail reorganizing the budget on a sectoral basis and strengthening the role of line ministries in the budget allocations for their respective sectors together with the integration of the current and capital budgets. These issues are covered in Theme 2 of the PPER. The studies prepared under this theme include an analysis of the major changes in the budget process and in relations between different budget organizations which will be necessary to enable the MTEF to be implemented in Tajikistan, which is presented in chapter 4 of this volume of the PPER. Subsequent volumes of the PPER will include analytical work related to the specific needs of the education sector, which is pilot sector which will spearhead the implementation of the MTEF. This will include analysis of how the line ministry will relate to the myriad budget organizations which implement education expenditures.

Theme 3: Efficiency of Public Expenditure in Social Sectors

35. Theme 3 examines the efficiency of expenditures in the social sectors. The focus is on the social sectors, rather than the budget, because these sectors will play a crucial role in the country’s efforts to achieve the MDGs. Donors are keen to provide additional financing for these sectors (e.g., The Education for All-Fast Track Initiative), but are concerned about how efficiently funds will be spent, and also because it’is not feasible, in terms of resources, to employ the methodologies proposed to evaluate efficiency (e.g., Expenditure Tracking Surveys) across the whole budget.

36. The social sectors (mainly health, education and social protection) comprise approximately half of all non-PIP government expenditures. As noted above, relatively well developed but expensive and structurally rigid educational, health and social protection systems were put in place during the Soviet era but the sustainability of these systems and the quality of service delivery has been severely undermined by the contraction of the Government’s budget resource envelope over the last decade

9 and a half. These problems are compounded because the limited evidence which is available indicates that the resources which are channeled to the social sectors are not allocated in an optimal manner and that operational efficiency is poor because of weak management systems, lack of appropriate incentives and corruption. Hence, improving allocative and technical efficiency in the social sectors is essential if strategic public expenditure objectives in these sectors are to be achieved. However more detailed diagnostic analysis of the allocation and use of resources in the social sectors are required to inform recommendations for policy reform in this area.

37. Work carried out under this theme will include an examination of the technical efficiency of expenditures in the education and health sectors which utilizes the efficiency frontier methodology.

38. PETS in the education and health sectors will be conducted in collaboration with the Ministries of Education and Health and local government^.^ The findings of the PETS will be reported in subsequent volumes of the PPER. The PETS will provide a complete picture of resource flows in the education and health sectors (e.g., how much is spent and on what? how do different KBOs obtain their funds? how are budget decisions made? and how are funds managed in the KBOs?). They will identify the supply side factors that break the chain between the spending of resources and its transformation into service delivery. Equity of access to educational and health services will also be covered by PETS and household survey analysis. The PETS will establish base line indicators for monitoring the use of resources in the key social sectors. In addition, the findings of the PETS will help to improve transparency and accountability in budget execution and the resource utilization in both sectors. Furthermore, it will provide important inputs and recommendations for policy reform necessary for the development of sector wide approaches (SWAPS)in the education and health sectors. Although the PETS will be launched in FY 2007, the analysis and dialogue on the recommendations is scheduled through FY 2008 and FY 2009. This theme will also include a study of social protection which will be launched in FY 2008.

Theme 4: Fiduciary Risks

39. The Bank and other donors have supported several diagnostic studies - of the PFM system, including the Bank’s 2003 Country Financial Accounting Assessment (CFAA), IMF’s 2003 Safeguards Assessment and DFID’s 2005 Fiduciary Risk Assessment, which have concluded that serious weaknesses in the system of PFMs are a major impediment to improving the efficiency of budget resource use in Tajikistan and deter aid donors from providing a greater share of their aid in the form of budget support rather than the less efficient project aid (although there are circumstances where project aid is a more suitable vehicle for delivering aid, for example when technical assistance or external knowledge is required, the composition of donors’ aid to Tajikistan is very heavily weighted towards projects, which comprise 96% of projected aid to the Government in 2007, with budget support comprising only 4%).

See Annex 2 for an expanded concept note, ‘Public Expenditure Tracking Surveys in Education and Health’ which examines the rationale for, and objectives of, the proposed PETS in more detail

10 40. The complexity and scale of problems in fiduciary systems has meant that implementation of the PFM reform agenda has so far been slow and difficult. Nevertheless it is imperative that PFM reforms do not allow faltering, hence Theme 4 of the proposed PPER, covers the PFM issues, using the PEFA methodology. The rationale for including PFM issues in the PPER is not to simply add further diagnoses to those which have already been undertaken, but to specifically focus on what needs to be done to facilitate implementation of PFM reforms.

41. The major objectives of Theme 4 are to review the progress made in implementing the PFM reforms recommended by previous diagnostic studies such as the CFAA and CPAR, to assess whether, and to what extent, the recommended reforms have been implemented under on-going World Bank, IDF and other donor TA projects, to identify the obstacles to implementing these reforms, to set out the remaining agenda for PFM reforms and to provide guidance for its implementation. The accepted framework of PEFA is used as the methodology for benchmarking the 2003 CFAA, CPAR and PEIR assessments and then to review the fiduciary status of public finances in 2006. T he PEFA framework includes a government self assessment of the PFM system which is intended to ensure maximum government participation in and ownership of the process. Key issues which are examined under Theme 4 include: the adequacy of PFM systems, measures to prevent corruption and misuse of funds, the reporting of fiscal data to the public and the quality and transparency of these data, the adequacy of internal and external audit systems, the effectiveness of the State Financial Control Commission (SFCC) and procurement reform’. A summary of the PEFA Report, which includes all 3 1 of the PEFA ratings, is presented in Chapter 5 of this volume of the PPER.

Theme 5: Inter-Governmental Fiscal Relations

42. Reform of intergovernmental relations cannot be avoided if the Government is to implement a sector-based MTEF, as proposed by the MOF at its MTEF Round Table in May 2006. Hence theme 5 of the PPER covers intergovernmental fiscal relations. It examine the process of allocating budget resources and the roles played by central ministries, line ministries and local governments (oblast, rayon, jamoat), the arrangements for revenue sharing between the center and local governments, and the implications which these arrangements have for resource allocation. I t also examines the implications for local government budgets of introducing per capita financing in the health and education sectors, and more broadly the PAR strategy. The objective of this analysis is to generate recommendations for streamlining the local government system consistent with the PAR strategy and integrating it into a budget system which can support policy based budgeting. The work in Theme 5 is closely linked to the MTEF implementation action plan and the recommendations in the PAR strategy: as such, the priority in the first year of the PPER is to provide guidance on restructuring intergovernmental relations in the budget preparation process for the education sector, which will be the pilot sector for the MTEF. Further work under this theme will be determined by the needs of the other sectors which are subsequently brought into the MTEF as well as by the lessons learned from the

httD:liwww.Defa.org provides a description of the methodology and the steps involved in carrying out a PEFA assessment,

11 education sector beginning FY08. The work under this theme will be reported in subsequent volumes of the PPER.

F. SEQUENCING THE WORK OF THE PPER

43. The PPER will be carried out over the course of three fiscal years (2006/07- 2008/09) and published in three volumes, of which this is the first. The sequencing of the work has been guided in part by the requirements for supporting the ongoing program of PFM reforms which is being implemented by the GOTand by the need to respond to pressing demands for analytic work on how best to align the limited resources of the budget with the Government’s strategic objectives for growth and poverty reduction. Hence Chapters 2 and 3 of this report, which address the issue of how best government expenditures can support economic growth and the potential fiscal space available to fund priority expenditures, were identified as priority areas for analysis and have been included in the first volume of the PPER.

44. The analysis in Chapter 4 of the first volume, outlining a strategy for implementing the MTEF, responds to the decision taken by the Government in 2006 to implement an MTEF, beginning in 2007, and to seek TA from donors to support this. Further work under Theme 2 will build on this strategy to support the MTEF reforms in the pilot sector and other sectors. This work has already begun in the education sector and will be included in subsequent volumes. In addition to the sector work, the next set of issues which are pertinent to the implementation of the MTEF involve the inter-governmentalfiscal relations which are covered under Theme 5 and these will also be addressed in subsequent volumes of the PPER.

45. The PEFA assessment is also a priority for the PPER for two reasons. First, because it provides essential diagnostic analysis on which to begin developing a PFM reform strategy and secondly, because a comprehensive fiduciary assessment is a prerequisite for many donors to consider providing budget support to Tajikistan, which is essential to create more fiscal space for priority expenditures. Hence, the PEFA assessment was conducted in the first year of the PPER and a summary is included in this volume of the PPER.

46. The work under Theme 4, which examines the efficiency of public expenditures in the social sectors, was started in the first year of the PPER but, because of the time required to design, prepare and carry out the PETS and analyze the data as well to complete the frontier efficiency analysis, the results will be reported in subsequent volumes of the PPER.

47. Drawing on the analysis of the PPER, and especially the findings of the PEFA, a PFM reform strategy and action plan will be drawn. This will identify the priorities for PFM and set out a sequenced action plan for achieving these reforms. The PFM reform strategy and action plan will be included in a subsequent volume of the PPER.

12 2. ECONOMIC GROWTH: CONSTRAINTS, CHALLENGES AND IMPLICATIONS FOR PUBLIC EXPENDITURE

A. INTRODUCTION

48. Sustained and rapid economic growth will be a necessary, but not sufficient, condition for the eradication of mass poverty in Tajikistan. Although the recent growth performance of the economy has been very strong, this largely represents a recovery from the very steep economic contraction which took place in the first half of the 1990s, which was made possible by bringing back into production underutilized factors of production. Growth will inevitably subside once the scope for further increases in the capacity utilization of the existing capital stock is used up, unless a combination of new capital investments and improvements in factor productivity provides new sources of growth by expanding aggregate supply. The role of fiscal policy in supporting the recovery of the economy has primarily involved its role in macroeconomic stabilization, but future growth will depend far more on a strengthening of the supply side of the economy, critical elements of which, notably public infrastructure and human capital, will depend on budgetary policies.

49. The contention of this chapter is that fiscal policy, and especially public expenditure policies, will not only be important for supporting sustainable long term economic growth, but will also shape the pattern of economic growth and development in Tajikistan. It is possible to envisage alternative long-term growth trajectories for Tajikistan, distinguished in part by whether public expenditure priorities are focussed on large infrastructure investments, or on social services and human capital development, and the extent to which government improves the investment climate for the private sector and economic governance more generally. These trajectories have markedly different implications for poverty reduction and human development.

50. This chapter argues for a balanced development strategy in which the priorities for public expenditure are: (i) the strengthening of human capital, mainly through the improvement of basic education and health services, and (ii) the maintenance of the core networks of the existing infrastructure (especially, those components of the infrastructure which comprise public goods which are unlikely to attract private investment). . Such a strategy offers the prospect of economic growth combined with employment creation, poverty reduction and improvements in human welfare.

51. One of the motivations for this analysis is that developing countries, which have followed a strategy of heavy public sector-led investments in infrastructure and capital intensive projects funded by external borrowing often failed to achieve sustainable long-term growth or poverty reduction and eventually ran into an external

13 debt crisis which had extremely damaging economic and social consequences. It is clearly important for Tajikistan to avoid making these mistakes.

52. The chapter is organized as follows. Section 2 reviews the performance of the economy since independence, analyzes the factors which explain the trends in growth and assesses the implications of this for future growth. Section 3 evaluates the constraints to growth in Tajikistan, drawing on the findings of several recent World Bank studies. Section 4 examines the potential comparative advantage of Tajikistan, what this implies for possible long-term economic growth trajectories and how such trajectories will be shaped by economic and budgetary policies. Section 5 provides some econometric evidence on the links between different types of public expenditure and growth in a sample of transition economies which includes Tajikistan. Section 6 concludes and examines the implications for public expenditure policies of the analysis of the constraints to, and prospects for economic growth in Tajikistan.

B. EXPLAINING TAJIKISTAN’S ECONOMIC GROWTH PERFORMANCE

53. Tajikistan’s economy contracted steeply in the six years following independence in 1991. The low point in output occurred in 1996, by which time the cumulative fall in GDP was 68%. In the next few years the economy began to recover, slowly at first, with growth accelerating in the 2000s, averaging nearly 10% per annum during 2000-2004, before subsiding to 6.7% in 2005. By 2005, the Tajik economy had recovered 30% of the fall in GDP which had taken place in the 1990s, so that real GDP in 2005 was about 60% of the level in 1990. The decline in agriculture was steeper than that of the GDP and its recovery was not as rapid. The agriculture’s share in GDP fell by 15 percentage points over the 1991-2005 period, whereas the share of industry in the economy is now at approximately the same level as it was at independence.

54. Tajikistan’s economy followed a similar path to those of the other CIS countries, all of which suffered a decline after the break up of the Soviet Union at the start of the 1990s (dubbed the “great contractions” by De Broeck and Koen 2000) followed by a recovery beginning later in the 1990s. However, the magnitude of the initial drop in GDP, its duration and the pace of recovery, differed among these economies. Tables 2.1 and 2.2 compare Tajikistan’s GDP growth with that of other low and middle income CIS economies since the break up of the Soviet Union. In five countries (Armenia, Azerbaijan, Kazakhstan, Uzbekistan and possibly Turkmenistan, [although data after 2001 are not available], real GDP had recovered to more than the 1990 level by 2005. The economy of the Kyrgyz Republic had only attained 80% of its 1990 level by 2005, while those of Moldova, Georgia and Tajikistan were still between 40% and 50% lower than in 1990. Those countries whose output has now surpassed its 1990 level all suffered a smaller initial contraction of GDP than Tajikistan, and began their recovery earlier than Tajikistan.

55. In the decades prior to the break-up of the Soviet Union, the economy of the USSR had undergone extensive economic growth; i.e., growth was generated by increasing the supply of factor inputs, especially capital. Consequently, the economy was characterized by high capitaVoutput ratios. The capital/output ratio was officially estimated at around three in 1986 but some Western estimates put it higher, at around five (Easterly and Fischer, 1994: 32). The elasticity of substitution between capital

14 and labor also appears to have been low, which together with the high capital/output ratio, implies diminishing returns to capital, which can explain the long term slowdown in economic growth in the Soviet Union (Easterly and Fischer, 1994).6 In the Tajikistan Republic of the USSR during 1971-1990, the average growth in the capital stock was 5.2% per annum, similar to the average of 5% for the USSR as a whole. Output growth in Tajikistan during 1971-90 was 2.6% per annum, similar to the rate of 2.3% for the whole of the USSR. However, the growth in the labor force in the Tajikistan Republic was 3% per annum, much higher than the average of 1% for the USSR (De Broeck and Koen, 2000: 15). This would imply that the capital/output ratio in Tajikistan was similar to that of the average for the USSR but that its capital/labor ratio was lower than the average for the USSR.

56. The problems in the economy were evident long before the break-up of the Soviet Union. In the 20 years before the break-up (1970-1990), its growth of output per worker was worse than that of the Soviet Union as a whole, which was itself on a downward path. In this period, annual growth per worker in Tajikistan was only 1%, compared to 2.8% for the USSRq7Growth in all of the Central Asian republics was below the average for the USSR in this period. The worst performing sector in Tajikistan was agriculture, with average growth per worker during 1970-1990 of negative 1.8% per year (Easterly and Fischer, 1994: 44, 45). The poor performance of agriculture characterized all of the Central Asian republics in this period, possibly because of environmental problems caused by over-investment in cotton.

57. Several researchers have used a growth accounting framework to analyze the sources of economic contraction and growth in the CIS countries (De Broeck and Koen, 2000), low income CIS countries (Loukoianova and Unigovskaya, 2004) and Tajikistan (Matovu, 2005), following their independence from the Soviet Union. The growth accounting framework disaggregates GDP growth into three components: those attributable to the growth in the capital stock, the growth in the labor force (and labor productivity), and the growth in total factor productivity (TFP). The latter is essentially a residual which picks up all of the influences on the productivity of factors of production, as well as measurement errors: TFP does not just reflect exogenous technological progress (De Broeck and Koen, 2000). These growth accounting exercises assume a constant return to scale production function with elasticities of output with respect to labor and capital of 0.7 and 0.3 respectively.

Note the negative TFP growth rates for the periods 1981 -85 and 1986-90 shown in Table 2.4. ’ These growth rates refer to Net Material Product (NMP), which was the definition of output in the Soviet Union. The growth rates of NMP per worker reported in Easterly and Fischer (1 994: 44) are higher than those implied by the data in De Broeck and Koen (2000: 15); the latter imply negative growth rates of output per worker in Tajikistan, given that annual rates of labour force growth (3 percent%) were higher than output growth (2.6 percent%), but both sources indicate that growth rates of output/NMP per worker were lower in Tajikistan than the average for the USSR.

15 P 0 0 hl m 0 0 hl hl 0 0 hl

7 0 0 hl

0 0 0 hl

Q) Q) Q) 7

03 Q) Q) r

b Q) Q) 7 to z0)

In Q) Q) 7 t zQ) m Q) Q) 7 3

9) fi 3m m 4

L 9) E

.IE E 0 .I c,

Ih 2: I

.Id sa

v1 E u.. z s9) F 58. Matovu (2005) found that TFP growth was the main cause of the collapse and then recovery in output in Tajikistan. Matovu’s results are shown in Table 2.4. TFP growth was strongly negative in 1991-95, slightly positive during 1996-2000, and very strong during 2001-04, mirroring the trends in output. The capital stock was virtually stagnant throughout the period since independence, although there was a slight recovery (growth of 1.1% per annum) in 200 1-04. The labor force declined in the 1990s by about 1% per annum, but then grew by 3% per annum in 2001-04. The results for Tajikistan in the 1990s reported by Loukoianova and Unigovskaya and De Broeck and Koen are consistent with those of Matovu. The recovery in GDP in Tajikistan was, therefore, driven mainly by a recovery in TFP, and to a lesser extent, growth in the labor force. Matovu estimates that TFP growth contributed about three quarters of the GDP growth during 2001-2004, boosting GDP growth by an average of 7.2% per annum.

59. The trend in TFP in Tajikistan was similar to that of most other CIS countries. Throughout the CIS region, there was a decline in TFP associated with the contraction in output, while TFP growth was the main contributor to the subsequent recovery in output. Except in Azerbaijan, capital investment made no more than a very minor contribution to the output recovery in the low income CIS countries (Loukoianova and Unigovskaya (2004).

1981-85 1986-90 1991-95 . 1996-2000 2000-04 Capital 4.6 2.9 0.2 -0.1 1.1 Labor 3.1 2.9 -0.9 -1.1 3.0 TOP -2.4 -1.6 -15.5 1.3 7.2 output 1.2 1.3 -16.1 0.5 9.7

60. The contraction in output in Tajikistan was caused by three factors. First, the Tajik economy was integrated closely into that of the USSR, but the break-up of the Soviet Union severely disrupted trade patterns and production, and led to higher prices of imported materials and fuel which had previously been subsidized. Secondly, the economy suffered a huge fiscal shock when transfers from the Soviet Union, which had accounted for almost half of the Government budget during the Soviet era, ceased at independence (IMF, 2006B: 4). Thirdly, the civil war which afflicted the country between 1992 and 1997 severely disrupted the economy. As a consequence of the output collapse, much of the country’s capital stock was underutilized in the 1990s. As output fell much faster than capital (which can only adjust very gradually), TFP also fell. As such, the economy was operating far below its Production Possibility Frontier (PPF) by the mid 1990s. The output collapse was accompanied by severe macroeconomic imbalances, with huge fiscal deficits largely financed through money creation, hyperinflation which exceeded 2000% in 1995 and

18 a rapid build up of external debt, mainly in the form of trade credits from other CIS countries which exported goods to Tajikistan.*

61. The economy began to recover in 1997, with growth accelerating in 2000. Growth appears to have been driven mainly by a recovery in aggregate demand, especially private consumption, which in turn was boosted by an expansion of official capital inflows and workers’ remittances. Given that the economy was operating well within its PPF in the mid 1990s, an expansion of supply capacity was not required for a recovery of output; all that was required was to bring back into production the existing under or un-utilized productive capacity, as occurred in the aluminium industry, for example. T he movement back towards the PPF is reflected in the recovery of TFP which is found in the growth accounting estimates.

62. The economic recovery was boosted by two factors. First, the restoration of peace and security in 1997, and secondly, the fiscal adjustment which began in the second half of the 1990s and allowed the macroeconomy to be stabilized. As noted above, the pattern of economic contraction and recovery in Tajikistan was similar to that of other CIS#economies. Segura-Ubiergo, Simone and Gupta (2006) analyze the relationship between fiscal adjustment and growth in the CIS countries during the period 1992-2001 and argue that fiscal adjustment boosted economic growth because the reduction in the monetization of fiscal deficits and the enhanced credibility of macroeconomic management helped to stimulate an inflow of donor inflows and private capital (as well as remittances) which contributed to the recovery of aggregate demand.

c. CHALLENGES AND CONSTRAINTS TO SUSTAINABLE FUTURE GROWTH

63. It is unlikely that Tajikistan can rely on the same sources of growth which dominated the recovery of the economy during the 2000s, notably the high rates of TFP growth made possible by a shift back towards the economy’s PPF, driven by an expansion of aggregate demand, for much longer. This is because most of the spare capacity which characterized the economy in the mid 1990s has now been brought back into production and very little new capacity has been added, because rates of investment have been low (Table 2.3). In addition, it is possible that the PPF has actually shifted inwards since independence because of a decline in the size and quality of the labor force, caused by the migration of workers to seek work outside Tajikistan, although we do not have accurate and up to date statistics to substantiate this.

64. To explore whether the economy has now returned to close to its PPF, Chart 2.1 displays two estimates of capacity utilization which differ in the assumed average depreciation rate of the capital stock. To estimate capacity utilization, we apply to an initial estimate of the capital stock the reported rates of gross fixed capital formation (GFCF) and assumptions of the depreciation rate. Tajikistan’s peak output was in 1988. In the absence of data on the capital output ratio for Tajikistan, we make the assumption that the capital output ratio in the peak year for output was three, which is in line with the official estimate for the USSR as a whole in 1986, but lower than

Tajikistan had no external debt at independence, but by 1996, its public and publicly guaranteed external debt had reached US$867 million, equivalent to 84% of GDP.

19 Western estimates. It is not clear whether the Tajik economy was more or less capital-intensive than that of the USSR as a whole, given that the former comprised sectors which were probably much more capital intensive than the average for the USSR (hydropower and aluminum smelting) and sectors which were less capital intensive than average (agriculture).

65. We assume that there was full capacity utilization in the peak year of output, and that the real value of the capital stock was the same in 1990 as in 1988. Real GDP in 1990 was 94% of the peak year level (De Broeck and Koen, 2000: 9), which implies that capacity utilization in 1990 was 94%. For depreciation of the capital stock, we make two separate assumptions. T he first is the 3% estimate used by Loukoianova and Unigovskaya (2004) for all low income CIS countries. However, this estimate of a 3% depreciation rate only includes the physical deterioration of capital; it does not take account of obsolescence of capital (also called economic deterioration). To incorporate economic deterioration of capital, we also include a higher, 5% depreciation rate. We use estimates of GFCF in constant prices, from the World Bank country database for Tajikistan.

66. As shown in Chart 2.1, if the depreciation rate was only 3% throughout the 1990-2005 period, the real level of the capital stock would be at virtually the same level in 2005 as it was in 1990; Le., gross fixed investment matched depreciation on average throughout this period. In that case, capacity utilization in 2005 would still be only just under 70%, despite it having doubled since the mid 1990s. Hence there will still be room for further increases in capacity utilization which could generate more economic growth. However, if the depreciation rate, including economic deterioration, is assumed to be 5%, then the real value of the capital stock in 2005 would be only 80% of the 1990 level (Le., gross investment has been insufficient to offset depreciation) and capacity utilization would be almost 90% in 2005. In that case, the scope for further increases in capacity utilization will run out quite quickly.

20 Chart 2.1. Estimated Utilisation of the Capital Stock with 3 percent and 5 percent Depreciation Rates, 1990-2005

30% - --i -5% depreciation 20% -. -3% Depreciation 10% -

0% 7.

Source: see paras 64-66.

67. The higher rate of depreciation is probably more realistic than the lower one, for two reasons. First, it is likely that physical depreciation rates in Tajikistan were higher than the average for low income CIS countries because some of the capital stock in Tajikistan was damaged in the civil war in the 1990s. Secondly, economic deterioration cannot be ignored because some of the public enterprises established during the Soviet era are not economically viable in a market economy, and the capital stock of these enterprises cannot readily be used in other sectors of the economy.’ Consequently it appears likely that the economy is close to the point where capacity utilization cannot be increased any further and is, therefore, close to its PPF.

68. If the economy is now operating close to its PPF, future growth in Tajikistan will require an expansion of aggregate supply (a rightward shift in the PPF), through higher rates of capital investment and improvements in TFP, including labor productivity. Factor productivity could be improved by strengthening competition in the economy, by attracting more foreign direct investment which could provide modern management skills and, over the long term, by strengthening human capital through better education and training which would enhance labor productivity. Given that average per capita incomes are very low in Tajikistan, the scope for mass private sector provision of education is limited and hence the public sector must provide the bulk of education services. Labor productivity could also be raised if the health status of the workforce could be improved through better health care. Compared to other

For example, many of the machine producing factories in Tajikistan set up during the Soviet era relied on inputs brought in from elsewhere in the Soviet Union. These factories became unprofitable after the break up of the Soviet Union because of increases in the prices of their inputs and the cost of transportation (Umarov and Repkine, 2004).

21 CIS countries, labor productivity in Tajikistan is very low (only 70% of the level in the Kyrgyz Republic and half that in Armenia and Georgia for example) although this cannot be attributable solely to the quality of the labor force; the high costs of doing business make an important contribution to low labor productivity (World Bank, 2005B: 3).

69. Increased private investment will be crucial to an expansion of the supply side of the economy for three reasons. First, given budgetary constraints, the scope for funding much higher levels of public investment on a sustainable basis, without pre- empting budgetary resources for essential current expenditures, is very limited. Secondly, the type of investment which the economy needs most to generate higher growth rates is not investment in public capital such as infrastructure, but investment for the production of marketable goods and services, for which the private sector is much better suited than the public sector. Thirdly, private sector investors, operating under competitive conditions and facing the incentives to make profits, are more likely to generate improvements in factor productivity than would be the public sector which lacks the same incentives.

70. Private investment has been very low in Tajikistan in recent years, averaging only just over 5% of GDP in the 2000s (see Table 2.5). Given the rapid recovery in aggregate demand since the late 1990s, together with the restoration of macroeconomic stability, one might have expected a stronger response from private investors to take advantage of the expanding domestic market, as well as possibly to export to neighboring countries, some of which are also enjoying buoyant growth.

Table 2.5: Private Investment, Percent of GDP: 2001-2006

2001 2002 2003 2004 2005 2006* Private 4.2 5.4 5.5 5.4 6.0 6.0 investment

7 1. One of the reasons for the paucity of private investment is that the institutional climate for the private sector is poor. Surveys by the World Bank have found that the business climate is characterized by unpredictability, a lack of transparency, an uneven playing field for businesses and corruption. Most damagingly, instead of a clear set of rules for all businesses to ensure fair competition, individual firms make their own special arrangements with Government officials (World Bank, 2005B: 22-24). There is a plethora of complex regulations which act as a barrier to entry by start-up businesses and are a burden for existing firms. Businesses need to obtain many permits from different government agencies; the procedure for obtaining permits is complex, inefficient, not transparent, and costs businesses both money and time. Corruption is pervasive; the findings of the BEEPS indicate that private firms pay bribes to tax and customs inspectors, to local authorities, to obtain permits and to access utilities. The BEEPS reported that 80% of businesses paid bribes to tax inspectors (World Bank, 2005A: 16-17). The operation of'tax administration, which involves the tax liabilities of companies being computed by tax officials rather than through self assessment by taxpayers backed up by risk based audit, is also a

22 hindrance to the private sector, because it allows too much scope for discretionary decisions and predatory behavior on the part of tax officials and the tax police.

72. Improving the institutional climate for private investors, by radically curtailing the number of licenses and permits required, enhancing transparency in the implementation of regulations, ensuring a level playing field for all investors, curbing opportunities for corruption, and introducing modern concepts of tax administration, will be a prerequisite for an increase in genuine private sector investment. Some measures have already been taken to address these problems. Parliament enacted and amended legislation in 2005 and 2006 to streamline licensing procedures and reduce the number of activities which are subject to licensing, but there is still much to be done to improve the investment climate.

73. While higher rates of private investment in marketable goods and services is clearly essential for long term growth, whether the economy requires more public investment is less certain. The country inherited a large stock of public capital from the Soviet era, to the extent that the capacity to supply utilities is “over-dimensioned’’ for the size of the economy (World Bank, 2006C), although some of the public capital assets are not especially well suited to a market economy (for example in the transport sector, investment in the Soviet era favoured the railways over the road network) and some assets are now in a poor state of repair because adequate maintenance has not been carried out since the country became independent.

74. Two sectors are particular pertinent for any discussion of the country’s public capital requirements: electricity and transport infrastructure. As discussed below, Tajikistan has a comparative advantage in the production of hydroelectricity, but this does not mean that the supply of electricity for domestic consumption needs to be expanded. Tajikistan already consumes a much higher level of electricity per capita than the average for a developing country.” Moreover, domestic electricity demand is likely to stagnate when electricity tariffs are raised to cost recovery levels as planned by the Government.” Therefore, it is unlikely that an expansion of electricity supply to the domestic market is critical to support future growth of the economy. Expenditures are required to rehabilitate and maintain the transmission and distribution system, but when electricity tariffs are raised to cost recovery levels it should be possible for Barki Tajik, the electricity utility, to fund these expenditures without recourse to the Government budget.

75. Investment in hydro electricity generation for export to neighboring countries is potentially commercially viable and offers major investment opportunities, but these opportunities will be best exploited by private investors with their own capital at risk, provided that Government can provide credible guarantees against expropriation and other non-commercial risks. Public investment should not be required to develop hydroelectricity generation for export, and indeed is not desirable because this will

10 Tajikistan’s annual per capita consumption of electricity is 2,236 KWh, compared to the average for a middle income country of 1,422 KWh, even though its GDP per capita is only around one fifth of the average for a middle income country. Excluding the electricity consumption (about 40 percent% of the total), Tajikistan still consumes around 1,300 KWh per annum (World Bank, 2006C: 36). ” The current average level of electricity tariffs is around 0.6 cents per KWh. The Government intends to raise tariffs to the level required to cover the long run average incremental costs of supply, which is 2.3 cents per KWh in current prices.

23 pre-empt budgetary resources which are required for essential public services such as education.

76. It is also unlikely that major investments in the transport infrastructure are necessary to support long term economic growth, because there is little evidence to indicate that the inadequacies of the transport infrastructure are a major constraint to private investment and growth. In the 2005 BEEPS, only 2% of businesses in Tajikistan ,cited transportation as a major obstacle for the growth and operation of their business, while only 11% of businesses in Tajikistan cited transportation as a moderate obstacle (World Bank, 2006C: Annex 1). A World Bank regional study of Transport and Constraints to Trade in Central Asia, including Tajikistan, concluded that the road networks in the region are relatively extensive and largely sufficient to meet the needs of users, although roads are in poor condition because of inadequate maintenance. It argued that Central Asian countries should focus on the maintenance of existing roads, focusing on the core road networks that can be maintained on a sustainable basis within the resources available in each country (World Bank, 2004B).

77. To summarize, Tajikistan economy’s investment requirements to support sustainable long-term growth consist mainly of investment in the production of marketable goods and services, including electricity for export, rather than investments in infrastructure to provide public goods and services. Hence the private sector rather than the public sector is much better suited to take responsibility for financing, owning and managing these investments.

D. COMPARATIVE ADVANTAGEAND POSSIBLE GROWTHTRAJECTORIES

78. Tajikistan’s future development will be shaped to some extent by its comparative advantage, as well as by government policies and regional developments. A country’s comparative advantage on world markets is closely linked to its resource endowments, or at least those resource endowments which are immobile between countries, in line with Heckscher-Ohlin trade theory. Wood and Mayer (200 1) demonstrate empirically that relative resource endowments explain 40-60%of the composition of exports in a large sample of 111 countries, in which exports are classified as manufactures, processed primary products and unprocessed primary products and resource endowments comprise land, unskilled labor and skilled labor.

79. On the basis of its resource endowments, where should Tajikistan’s comparative advantage lie? The methodology employed by Wood and Mayer (2001) involves comparing two indicators of resource endowments: land area per worker and years of schooling per worker. Countries with a higher than average land area per worker and with an average low (high) years of schooling per worker export primary (processed) commodities. Those countries with low land area per worker and with low (high) average years of schooling per worker export labor (capital) intensive manufactures. Alongside land area per worker we also consider arable land per worker, because only a small share of Tajikistan’s total land area is arable. We also include energy endowments, because these are important for Tajikistan. Energy endowments are measured as tones of oil equivalent (TOE) per worker per annum obtainable from proved recoverable reserves of crude oil and natural gas and from hydropower potential. Hydropower potential is defined as the economically

24 exploitable hydropower capacity.’* Hydropower potential is expressed in Terawatt hours per annum and then converted into TOEs per annum.I3 Proved recoverable reserves of crude oil and natural gas are the volume occurring in known natural reservoirs which can be recovered in the future under present and expected economic conditions with existing available technology. To convert recoverable reserves into a flow of energy per year, it is assumed that these reserves will be recovered over a 20 year time horizon: hence the annual recoverable volume is the total recoverable reserves divided by 20. Table 2.6 shows Tajikistan’s resource endowments compared to those of seven major regional groups of countries and two individual large countries.

80. What do the data in Table 2.6 suggest about the nature of Tajikistan’s comparative advantage? It is unlikely that Tajikistan has a comparative advantage in agricultural exports, because its arable land area per worker is no higher than the world average. The sectoral patterns of growth in Tajikistan since 1970 also suggests that the country’s comparative advantage does not lie in agriculture because, despite the fact that the Soviet Government had invested heavily in agriculture in the Tajik Republic, the growth rate of the sector has consistently lagged behind that of the rest of the economy for the last three and a half decades. Nevertheless, there may be exceptions to this; for example, Tajikistan may have a comparative advantage in cotton production given that this crop has historically been economically viable. Tajikistan does not have the relative resource endowments of an exporter of labor intensive manufactures - low land area per worker and low skill levels - and so is unlikely to have a comparative advantage in these products. Tajikistan’s landlocked geography also impedes its ability to export labor intensive manufactures, because transport costs are more important for manufactured exports as a result of their import content generally being higher than that of other types of exports. This does not mean that no agriculture or manufacturing can be viable in Tajikistan. There are parts of the country, such as the Ferghana Valley, which are very fertile and can support economically viable agriculture, such as cotton or food crops for the domestic market. It is also possible that some manufacturing activities, especially those involving the processing of locally produced raw materials, such as cotton, can be viable. But neither agriculture nor manufacturing, with the exception of the energy intensive aluminum smelting, are likely to be the leading sectors of growth in the traded goods sectors.

I2 This is the amount of gross theoretical capacity that can be exploited within the limits of current technology under present and expected local economic conditions. Gross theoretical capacity is the maximum power which could be generated if all natural water flows were turbined with 100 percent efficiency to sea level or the country’s borders. The definition is from the WEC Survey of Energy Resources. ’’One Megawatt hour is equal to 0.086 TOEs.

25 Table 2.6: Resource endowments of Tajikistan and other countries and groups of countries

Land area Arable land Years of Energy sq kilometres sq kilometres schooling TOE per per 100 workers per 100 workers worker p .aI

MENA 4.0 0.3 5.5 15.7 SSA 6.0 0.4 3.0 0.9 LAC 5.6 0.4 6.1 2.9 EA&P 1.1 0.1 6.4 0.3 SA 0.6 0.2 4.6 0.1 AdEcon 4.2 0.4 8.8 0.7 EE&FSU 8.7 0.9 9.5 2.1 China 1.o 0.1 6.4 0.4 USA 4.3 0.8 12.1 1.2 Tajikistan 4.8 0.3 10.1 9.1

World 3.2 0.3 6.7 1.4

MENA is Middle East and North Africa; SSA is Sub-Saharan Africa; LAC is Latin America and the Caribbean; EA & PC is East Asia and the Pacific, excluding China; SA is South Asia; Adcom is Advanced Economies, excluding USA and EE&FSU is Eastern Europe and the Former Soviet Union. Sources: Years of schooling; Barro and Lee (2001) and Wood and Mayer (2001) Population; World Development Indicators Land area and arable land area data are from the Nationmaster.com website. Energy data are from the website of the WEC Survey of Energy Resources.

81. The resources which Tajikistan is relatively well endowed with are energy supplies and skilled labor, defined in terms of the average years of schooling of the workforce, although this indicator may overstate the skill levels of the workforce because of both a deterioration in the quality of education since the break up of the Soviet Union and migration of workers. Tajikistan is very well endowed with energy resources, with 9 TOES of energy per worker per annum, compared to the world average of 1.4. Almost all of Tajikistan's energy resource endowment derives from its hydropower potential. Only one country in the rest of the world, Iceland, has a greater hydropower potential per worker than Tajikistan and there are only five countries in the world which have larger potential hydropower resources even in absolute terms, all of which are much larger countries than Tajiki~tan.'~Moreover, Tajikistan is geographically close to countries which are both large and energy scarce, notably China and the countries of South Asia, which can provide potential export markets for electricity. New hydropower projects in Tajikistan could generate electricity at a cost of only 2.5-3 cents per KWh, and transmit power at around 0.5

14 These are the Democratic Republic of Congo, Brazil, Canada, the USA and the Russian Federation.

26 cents per KWh to export markets, compared to the costs in potential export markets of around 5 cents per KWh (World Bank, 2004C). Hence Tajikistan has a strong comparative advantage in the energy sector, specifically the generation of electricity from hydropower.

82. The generation of hydroelectricity for export is, therefore, likely to be a major growth industry in Tajikistan over the long term. However, this is not an unmitigated blessing. Countries endowed with natural resources, including those among the transition economies, have tended to experience slower growth than those countries which are less well endowed. Kronenberg (2004) argues that the prime reasons for this “resource curse” in the transition economies were corruption and the neglect of basic education. In addition, hydropower generation is very capital intensive ands will create very few jobs, other than in the construction phase of the projects. Hence Tajikistan cannot depend upon hydropower generation as the sole engine of growth if that growth is to contribute to employment creation and poverty reduction. Whether or not the exploitation of Tajikistan’s hydropower potential turns out to be a curse or a blessing for the country will depend on the nature of its economic and fiscal policies. It is possible to envisage two quite different, and admittedly extreme, long term growth trajectories in Tajikistan, with the differences between these two trajectories essentially attributable to differences in the Government’s budgetary policies and its policies towards private investment. We set out each of these two possible trajectories below.

Growth Trajectory I: Growth with jobs and human development

83. The first of these two possible growth trajectories would lead to a beneficial outcome in terms of human development and welfare. It would involve private sector investment in hydropower generation, mainly for export, with government restricting its own role in the industry to providing a sound regulatory framework, reforming the domestic electricity market to ensure the long run commercial viability of domestic supply, and providing credible guarantees for private investors against non- commercial risk such as expropriation or major changes in the tax regime.

84. Government would also implement fundamental reforms to improve the institutional climate for private investment outside the electricity sector. This is essential, because if the economy is to combine growth with employment creation and therefore poverty reduction, it must attract private investment into sectors other than electricity generation which are more labor intensive. Given that the growth of export earnings is likely to be dominated by electricity exports, the domestic market for goods and services will provide the main focus for private investment outside of the electricity industry.

85. Under this trajectory, Government would focus its budgetary resources on strengthening the provision of basic public services to enhance human capital and fight poverty, especially by improving the public provision of education, and providing other essential public services such as the maintenance of law and order, protection of property rights and regulation of markets where this essential to protect the public. Over the long term, this will facilitate an improvement in labor productivity and thereby complement increased private investment in boosting sustainable economic growth. Improved labor productivity will also enable real

27 wages to rise. Consequently this growth trajectory offers the prospects of three important development gains: (1) rapid and sustainable economic growth, driven by private investment in hydropower generation for export and in other more labor intensive industries mainly serving the domestic market, combined with increased labor productivity arising from better education and training and from factor productivity improvements brought about by more competitive markets; (2) income poverty reduction brought about by employment creation and rising real wages as a result of private investment in the non-electricity sectors and better education and training; and (3) reduction in the non-income aspects of poverty as a result of improvements in essential public services such as education and health care.

Growth Trajectory 2: Growth without jobs and human development

86. Investment in hydropower generation for export also plays a major role in the second growth trajectory but unlike in the first trajectory, the Government itself would undertake a large share 'of the investment, using external loans to supplement its own resources. The second difference with the first trajectory is that the overall investment climate remains fundamentally unreformed with the consequence that private investment does not increase and, therefore, job creation is minimal.

87. With the capital investment requirements of the energy sector, and perhaps large capital projects in other sectors such as transport, taking up a major share of the Government budget, the fiscal space for funding basic public services will be squeezed, foreclosing the possibility of achieving major improvements in the provision of education and health services. In addition, the external borrowing required to fund the Government's capital investments may threaten external debt sustainability, which would further reduce the fiscal space for funding basic public services. Consequently, this trajectory offers the prospects of economic growth, although at lower rates than the first trajectory because of lower private investment in the non electricity sector and lower investment in human capital, but it will create few jobs and do little to reduce poverty, whether income poverty or the non income aspects of poverty such as health.

88. The basic differences in terms of public policy between the two trajectories are twofold. First, the priorities for public expenditure in the first trajectory are basic social services, especially education, while in the second trajectory they are capital investments. Secondly, fundamental reforms are enacted to strengthen the investment climate in the first trajectory but not in the second.

E. MODELINGTHE GROWTH EFFECTS OF PUBLIC EXPENDITURE

Public Expenditure and Economic Growth: What Does Endogenous Growth Theory Tell Us?

89. There is now a large literature dealing with endogenous growth theories and the empirical modelling and testing of these theories. The causal factors of economic growth in exogenous growth theory are the accumulation of physical and human capital, alongside various institutional factors (macroeconomic stability, trade policy, etc) which, inter alia, can affect the efficiency with which capital is used. Exogenous growth models, therefore, allow for a possible role for fiscal policy (as well as other

28 economic policies) in the growth process. Fiscal policy could boost economic growth through the impact of public expenditures on physical or human capital, whereas it might depress growth if fiscal deficits crowd out private investment or lead to macroeconomic instability. Endogenous growth theory suggests that the type of public expenditures are likely to be important for growth; not all public expenditures may be beneficial and some may have negative affects, depending on the specific circumstances in individual countries.

90. There have been many empirical studies analysing the impact of public expenditures on economic growth, but the findings are not unanimous. The research of Aschauer (1989), Barro (1991) and Easterly and Rebelo (1993) support the argument that government expenditures on physical infrastructure have a significant positive impact on growth through their affects on private-sector productivity. Other researchers have argued that the relationship between government expenditure and economic growth tends to be sector-specific. For example, Bose, Haque and Osborn (2003) examine the growth effects of government expenditure for a panel of 30 developing countries in the 1970s and 1980s, with a particular focus on sectoral expenditures, finding that the share of government capital expenditure in GDP is positively and significantly correlated with economic growth, but that current expenditure is insignificant. They also find that, at the sectoral level, government investment and total expenditures on education are the only outlays that are significantly associated with growth once the budget constraint and omitted variables are taken into consideration.

91. Because the circumstances of each country differ, the results obtained from a large sample of countries may not necessarily be applicable to every individual country. To support the analysis of the links between fiscal policy and economic growth in Tajikistan, an endogenous growth model, similar to that of Bose, Haque and Osborn (2003), which included different types of government expenditures, including capital and current expenditures, and education and health sector expenditures, alongside other explanatory variables such as indicators of the level of human capital, the initial level of GDP, and the fiscal surplus and private investment, was estimated with data from a sample of eight transition economies in Central Asia and the Caucasus, including Tajikistan. This study was carried out with DFID financing as a background paper to this chapter and is attached as an annex.

92. The methodology was applied in two steps. The first step was to use the model to study the growth effects of government expenditure in the eight sample countries for the period 1996 - 2004 and identify the patterns of government expenditure that are associated with economic growth. The second step was to use Seemingly Unrelated Regression Equations (SURE) techniques to isolate the coefficients for Tajikistan specifically and hence identify the types of public expenditure which are most likely to support sustained economic growth in Tajikistan.

93. Although there are a number of drawbacks with this type of econometric analysis - the sample size is small, the growth relationships across all of the sample countries may not be identical, the growth effects of expenditures may have long gestation periods and may not be readily captured in annual data, reverse causation, from growth to public expenditure, cannot be ruled out, etc - the results are in line with those of many other empirical studies and do provide useful insights into the

29 type of relationships between fiscal policy and economic growth which exist in Tajikistan.

94. The evidence from this study supports the prevalent view in modern growth theory that education is an important factor in economic growth, with the level of government expenditures on education having a significant and robust correlation with growth. The study also found that the share of government capital expenditure in GDP is positively and significantly correlated with economic growth. This result provides the basis for government investments in sectors such as transport and communication, where it may be difficult to attract the optimal levels of private investment because of the public good characteristics of these investments. However, the impact of public investment seems to be conditional on the non-government investment variable (which is a proxy for private investment) in Tajikistan. Non- government investment is associated with economic growth in a significant and positive manner, especially in the SURE estimates for Tajikistan. There is also strong evidence to support human capital variables in the eight transition countries and in Tajikistan. In particular, the initial human capital variable, expressed as enrolment in primary and secondary schools is consistently significant in the regression estimates, as is life expectancy. Finally, the Government budget surplus has a significant and positive impact on economic growth (or alternatively, a larger fiscal deficit depresses growth); moreover the magnitude of the impact is quite large: on average, a one percentage point increase in the budget surplus is associated with an increase in real per capita growth rate of about 0.6 percentage points.

F. CONCLUSIONS AND IMPLICATIONS FOR PUBLIC EXPENDITUREPOLICIES

95. Tajikistan suffered a dramatic contraction of its economy in the first half of the 1990s as a result of the disruption to trade and production patterns caused by the break up of the Soviet Union, the loss of fiscal transfers from the Soviet Union and the civil war. The economy began to recover in the second half of the 1990s with strong rates of growth, averaging 9% per annum, being achieved in the 2000s. The recovery was largely the result of a rebound in aggregate demand, facilitated by the gradual stabilization of the economy and inflows of remittances and donor aid, which enabled under-utilized production capacity to be brought back into production. Although by 2005, real output was still approximately 40% below the level at the start of the 1990s, before the break-up of the Soviet Union, the limits to further rapid growth based solely on increasing the utilization of existing capacity will soon be reached. The economy’s aggregate supply capacity has probably fallen since independence, because gross investment rates have not been sufficient to offset the depreciation of existing capital and because many skilled workers have left the country, either permanently or as migrant workers.

96. Sustainable future growth will depend on expanding the supply side of the economy, through higher rates of capital investment and improvements in factor productivity, rather than on simply boosting aggregate demand. The critical requirements for long term future growth are a major rise in private investment in marketable goods and services, investment in human capital to raise labor productivity and in increase in competition and the removal of market distortions to enhance incentives for strengthening total factor productivity. To attract more private investment it will be necessary to improve the poor institutional environment for

30 private sector business, which is characterized by excessive, complex and untransparent regulations and corruption.

97. An expansion of public investment is not so pressing a need for the economy, because it is already well endowed with public capital assets as a result of high investment rates during the Soviet era. Some of the existing infrastructure systems, such as electricity supply, are over dimensioned for a low income economy. The responses to business surveys do not suggest that infrastructural inadequacies are the major constraint to private sector businesses. Moreover, the budget lacks adequate funds to maintain properly even the existing public capital assets, and hence it is likely that allocating more funds to maintenance will generate higher social rates of return than investment in new public capital assets.

98. Tajikistan has enormous hydropower generation potential and borders large energy scarce countries which can provide future export markets for electricity generated in Tajikistan. The capital costs of new hydropower developments are enormous (the combined costs of two projects currently under consideration are in excess of 100% of GDP) and hence it is essential to attract private capital into these investments. Private investment in hydropower generation will also help to ensure that they are managed efficiently and profitably, but to attract private investment it will be necessary for government to reform the domestic electricity market so as to make electricity supply commercially viable in the long run, and to provide a secure regulatory and tax structure for the electricity sector which can provide private investors with guarantees against non commercial risk.

99. Tajikistan cannot rely solely on hydropower generation as the engine of growth if it is to benefit from equitable growth with job creation and poverty reduction, because hydropower generation is very capital intensive and hence private investment in more labor intensive industries will be necessary to create jobs. The exploitation of natural resources has not always benefited developing and transition economies. Transition economies endowed with natural resources have experienced slower rates of growth than those transition economies that are less well endowed. The relatively poor performance of natural resource abundant transmission economies has been attributed to the neglect of education and corruption.

100. To examine the causal links between different types of public expenditure and economic growth, we used an endogenous growth model on a sample of eight transition economies including Tajikistan. The regressors included a vector of budget variables alongside other variables which are hypothesised to affect growth in endogenous growth theory, such as human capital. Because there were insufficient observations to estimate the model on Tajikistan alone, we used the Zellner procedure of seemingly unrelated regression equations (SURE) to isolate the coefficients for each country.

101. The evidence from the regressions supports the prevalent view in modem growth theory that education is an important driver of economic growth. We also found that the share of government capital expenditure in GDP is positively and significantly correlated with economic growth. Non-government investment is also associated with economic growth in a significant and positive manner, especially in the Zellner estimates for Tajikistan. Also, we find strong evidence for the role of

31 human capital variables in the growth process; in particular, the initial human capital variable, expressed as enrolment in primary and secondary schools was consistently significant, as was life expectancy.

102. What are the Government expenditure implications of the analysis in this chapter for sustainable economic growth in Tajikistan? Government expenditures on investment and human capital development, especially education, would boost economic growth in the long run. However, if Tajikistan is to follow a development strategy, which combines economic growth with human development, it should prioritize expenditures on education, health and other forms of human capital development, rather than investment, for two reasons. First, because a shortage of public capital is not a major constraint to economic growth, given that the public infrastructure is already relatively well developed, and second, because the scale of investments required to develop major infrastructure projects, such as hydropower generation for export, are far greater than can be accommodated within the Government budget unless the Government were to embark on heavy external borrowing which would jeopardize future external debt sustainability.

103. Given that budgetary resources are inevitably constrained by the imperative of maintaining debt sustainability, an increase in government capital expenditures will squeeze the fiscal space available for current spending on education and other forms of human capital development and on the essential maintenance needs of the existing infrastructure. A growth strategy focussed on heavy public investment in energy and other large infrastructure projects would be detrimental to the balanced long run development of the country, especially if it were combined with a failure to reform the institutional climate for private investment, because hydropower generation will create very few jobs and because of the adverse budgetary consequences for education spending.

104. Instead it would be preferable to shift the emphasis of budgetary allocations towards human capital development, especially education and health, and to the essential maintenance requirements of the core assets of the existing public infrastructure, while leaving investments in large commercially viable infrastructure projects such as hydropower generation to the private sector. Improving the education and health systems and maintaining the existing infrastructure assets are clearly public goods which will not be undertaken by the private sector, and hence government must take the lead in providing these goods. In contrast, the private sector can, in principle, provide the capital needed to develop major infrastructure projects provided that the risks involved can be controlled by putting in place the appropriate regulatory framework. Tajikistan does not spend enough on public health and education services. Aggregate budgetary spending on health and education is low, at 1.3% and 4% of GDP respectively in 2006. Health expenditures command just 7% of the State budget, less than half of the international target of 15%.”

105. A development strategy in which Government focuses on expanding and improving the provision of the key public services needed for human capital

IsCompared to other CIS countries, Tajikistan spends about the same proportion of GDP on education but only just over half as much on health (Lorie, 2003: 26), although it is arguable that most of the CIS countries need to spend more on these sectors to reverse the decline in human capital which has occurred since the break up of the Soviet Union.

32 development, offers the prospect of sustainable growth with employment creation. Rising labor productivity will raise real wages, and improvements in public services will enhance the non-income aspects of human welfare.

33 3. FISCAL SPACE IN TAJIKISTAN’~

A. INTRODUCTION

106. The design of fiscal policy in Tajikistan poses major challenges for policymakers. The real value of budget resources available to fund government expenditure has contracted sharply since independence, because of the economic collapse and the ending of transfers from the Soviet Union which subsidized service provision during the Soviet era, and as result. the quality of public services has deteriorated. An expansion in the budgetary resources allocated to public services such as education, health care and the maintenance of public infrastructure, combined with improvements in the efficiency of government expenditures, is essential if Tajikistan is to achieve its developmental objectives and especially its goals for poverty reduction. However, if an expansion of the budget resource envelope is to be consistent with the promotion of sustainable economic growth, it is crucial that any increases in the tax burden do not distort incentives for saving, investment and trade, and that Government borrowing does not jeopardize macroeconomic stability, crowd private sector borrowers out of credit markets, or threaten government solvency. Moreover, even if an expansion of the Government budget allows expenditures on priority sectors to be increased, improvements in economic governance, such as public financial management, will also be needed to ensure that this generates higher growth and poverty reduction (Varoudakis, 2006).

107. This chapter examines the fiscal space which is potentially available to Tajikistan over the next 10 years. The concept of fiscal space refers to the additional budgetary resources which can be mobilized for priority public expenditures through increases in the budget resource envelope and improvements in the allocative and/or technical efficiency of government expenditures. We employ the framework of the fiscal diamond to analyze fiscal space; the four points of the fiscal diamond represent the four potential sources of additional resources for priority public expenditures: domestic revenues, grants from donors, sustainable borrowing and improvements in the efficiency of public expenditure^.'^ The quasi fiscal deficits in the energy sector are also included in the analysis, because they will drastically reduce fiscal space available for priority expenditures unless they are eliminated through energy sector reforms, and these energy sector reforms will also affect the budget.

108. The chapter is organized as follows. Section 3.2 reviews fiscal policy in Tajikistan since the 1990s, highlighting the progress made in recovering from the dire fiscal position of the mid 1990s but also the remaining challenges. Sections 3.3 to 3.7 utilize the framework of the fiscal diamond to analyze the scope for expanding the l6This chapter draws heavily on a Tajikistan case study done for the Development Committee Paper on “Fiscal Policy for Growth and Poverty Reduction” presented to the World Bank board in April 2007. 17 The concept of the fiscal diamond as a tool for analyzing how the scope for fiscal space differs across countries was recommended in a paper prepared by the Bank for the Development Committee of the Bank and the Fund; “Fiscal Policy for Growth and Development: An Interim Report”, April 2006.

34 budget resource envelope in Tajikistan without jeopardizing fiscal sustainability or macroeconomic stability. This involves analyzing the potential to mobilize additional budget resources from domestic revenues (section 3.4), grant aid (section 3.5) and borrowing (section 3.6). The measures needed to tackle the quasi fiscal deficits in the energy sector, and the fiscal impact of these measures, are discussed in section 3.7. Section 3.8 summarizes the impact on the aggregate budget resource envelope of the projections for domestic revenue, grants, borrowing and the quasi fiscal deficit. Section 3.9 examines the scope to create fiscal space by improving the efficiency of government expenditures. Section 3.10 provides a conclusion.

109. The main conclusion of this chapter is that Tajikistan could expand its fiscal space substantially in 10 years time, potentially by around 8 percent of GDP annually, but only if the Government implements major reforms to tax administration, public expenditure management and electricity pricing. Reforms to tax administration could potentially add almost 3.2% of GDP to the fiscal resource envelope. Even more important are electricity sector reforms needed to cut the sector's quasi fiscal deficit that amounted to 9.3% of GDP in 2006. If these institutional and electricity sector reforms are not implemented, not only will there be very little growth in the budget resource envelope and no gains from improved expenditure efficiency, but the Government will eventually have to allocate substantial fiscal resources to the electricity sector to fund essential rehabilitation and maintenance of the transmission and distribution system, which will drastically cut the fiscal space available for priority expenditures. Moreover, even if additional budget resources could be mobilized without reforms to public expenditure management, especially to improve the allocation and technical efficiency of expenditures, it is doubtful whether additional government expenditure would have a positive impact on economic growth,

110. Hence the main policy message of this chapter is the urgency of reforms to improve governance in public financial management, in tax administration and in the electricity sector. These reforms are a prerequisite for both mobilizing fiscal space for priority expenditures and for ensuring that any expansion of the budget made possible by fiscal space translates into higher growth and poverty reduction.

B. FISCALPOLICY SINCE INDEPENDENCE

11 1. Tajikistan gained independence in 1991, but the country suffered a civil war between 1992 and 1997 and an economic collapse. The country also suffered a massive fiscal shock in the early 1990s, with a dramatic shrinking of budget revenue which may have been by as much as 80% in real terms. The fiscal shock arose from two sources. First, during the Soviet era, Tajikistan had been a recipient of transfers from the Soviet Union which accounted for almost half of the Government budget, but these transfers ceased at independence (IMF, 2006B34). Secondly, the tax base contracted with the economic collapse. The fiscal shocks led to huge fiscal deficits which worsened the macroeconomic instability.

112. Although the fiscal data for the first half of the 1990s, when the country still used the Russian ruble, are not comparable with data for subsequent years. It is clear that very large fiscal deficits were incurred in this period, in the region of 20% of GDP on average. The large fiscal deficits incurred in the first half of the 1990s were

35 almost entirely financed by credits from the central bank, as government’s domestic borrowing requirement could not be met by the very shallow domestic financial system. Central bank financing of the fiscal deficit, combined with the collapse of the economy, fuelled hyperinflation, which reached 2,100% in 1995.

113. In addition, quasi fiscal deficits led to a sharp build up of government and government guaranteed external debt, caused mainly by the Government financing imports of commodities from Russia and other former Soviet Union (FSU) countries. Tajikistan, which at independence had no external public debt, had by 1995 accumulated over US$SOO million of external debt (equivalent to 134% of GDP), of which the major share was bilateral debt contracted from FSU countries. By 1994, Tajikistan was already experiencing serious difficulties in servicing its debt and rescheduling agreements were signed with three FSU countries in the mid 1990s.

114. The Government began implementing an IMF ESAF program in 1998, a key element of which was the reduction in the fiscal deficit and the elimination of central bank financing of the deficit. The overall fiscal deficit was reduced from 11.2%of GDP in 1995, to an average of just under 3% of GDP in the final four years of the 1990s, and domestic financing of the fiscal deficit to an average of 1% of GDP in this period (see Chart 2.1).‘* Consequently the broad money growth rate was brought down in the second half of the 1990s and inflation was cut to 30% by 1999. However, Government and government guaranteed external debt rose by a further US$lOO million in the second half of the 1990s, to just under US$900 million in 1999.

’* These deficits are on a cash basis. There were substantial expenditure and revenue arrears, but the lack of a proper treasury system impeded the accurate reporting of these arrears.

36 Chart 2.1

Fiscal Aggregates as Percent of GDP: 1995-2005 30%

25%

20%

15%

10%

-5%

-10%

-15% 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 [-Expenditure - ‘Domestic Revenue - - Domestic Financing -Fiscal Deficit] Source: MF

115. The cut in the fiscal deficit in the second half of the 1990s was achieved by cutting total government expenditures sharply in real terms: between 1995 and 1999 the real value of expenditures contracted by 40%. Cuts in bread subsidies, which had amounted to nearly 10% of GDP in 1995, accounted for a large part of the cut in expenditures.” Expenditures on salaries and wages also fell sharply, from 7.4% to 3.6% of GDP between 1995 and 1999, because of both cuts in the real value of wages and in the size of the payroll. Revenues actually declined by over 10% in real terms in this period, mainly because the Government rationalized and reduced the heavy rates of taxation which had been imposed on the cotton and aluminum export sectors, and also because these sectors suffered declines in output.

116. Real GDP growth, which had begun to recover in 1997, accelerated in the 2000s, averaging 9% per annum during 2000-2005. Strong economic growth and better tax administration boosted government revenues, which rose by more than 6% of GDP from 1999 to 2005 (see Section 3.4). This allowed government expenditures to increase while maintaining the overall fiscal deficit at between 2% and 4% of GDP, with the exception of 2001 when it temporarily jumped to 5.6% of GDP, and avoiding virtually any resort to domestic financing of the deficit. Without the need to finance the fiscal deficit, the Central Bank was able to exert tighter control over the money supply and inflation; inflation was brought down to below 15% in 2001 and to below 10% by 2004, although it subsequently jumped to over 12% in 2006.

19 Bread subsidies were replaced by cash transfers, but substantial arrears were accumulated on these transfer payments.

37 Chart 2.2 External Debt, Percent of GDP: 1993-2005

140 -c

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Source: IMF

117. The recovery in government expenditure was very pronounced. Primary government expenditures rose from 13% of GDP in 1998 (the lowest level since independence) to 21.8% of GDP in 2005. Given that GDP also recovered strongly over this period, this implies an increase in the real value of government expenditure (using the GDP deflator as a proxy for the price index for government services) of approximately 180% between 1998 and 2005. Basic public services have been among the major beneficiaries of the recovery of the Government budget. For example, expenditures on education in the State Budget (Le., excluding the PIP) rose from 2% of GDP in 2000 to 4% of GDP in 2005, while expenditures on health in the State Budget rose from 0.7% of GDP to 1.5% of GDP in the same period.

118. The external debt burden has been reduced since 2000, when the nominal dollar value of external debt (including non government guaranteed debt) peaked at US$1,227 million (128% of GDP). External debt management was also tightened after 2000: new borrowing was restricted to concessional loans and a ceiling of 4% of GDP (approximately US$30 million in 2001) was imposed on new external borrowing to fund the Public Investment Program (PIP). Public and publicly guaranteed (PPG) external debt continued to rise, albeit slowly, until 2004 when debt cancellation, debt for asset swaps and debt restructuring agreements were reached with Russia and Pakistan. In 2006, Tajikistan was granted debt relief on its IMF debt under the Multilateral Debt Relief Initiative (MDRI) worth US$97.5 million in nominal terms. By 2005, the nominal value of PPG external debt had fallen below 40% of GDP. However, the external debt indicators began to deteriorate again in 2006, because Government contracted US$600 million of loans from China to finance infrastructure projects (see Section 3.6 for more details).

38 119. Despite the progress made in strengthening fiscal policy, serious problems remain. The Government has inherited from the Soviet era a structure of public institutions and physical assets which is large too large to be operated and maintained properly given the available budgetary resources. Consequently the inputs into public services cannot be funded adequately and are spread too thinly over many different activities. For example, salaries of government employees are far too low to provide even a subsistence wage (the average monthly wage was around 60 Somonis, equivalent to US$20 in 2004). There are insufficient funds for critical non-wage operational inputs, such as medicines in the health service: government expenditure on health amounts to the equivalent of only US$5 per capita. The physical infrastructure cannot be maintained properly. Moreover, while the Government’s own finances are relatively sound, there are serious problems elsewhere in the public sector, in particular in the electricity sector, which incurred a quasi fiscal deficit (QFD) of 9.3% of GDP in 2006.

120. To summarize, important progress has been made in several key aspects of fiscal policy since the late 1990s. First, there has been a major fiscal consolidation which has been a critical factor in stabilizing the macro economy, with the virtual elimination of domestic financing of the fiscal deficit. Secondly, there has been a recovery in domestic revenues which in turn has boosted the budget resource envelope. Thirdly, the expansion in the budget resource envelope has enabled the funding of social services to be increased substantially. Fourthly, the external public debt burden has been reduced from unsustainable levels through bilateral debt relief and the MDRI, although the risk to future debt sustainability is high because of the recent contracting of large new loans from China. c. FISCALSPACE FOR PRIORITY EXPENDITURES

121. The following sections of this Chapter examine the scope for creating fiscal space for additional priority public expenditures over the next 10 years from the four points of the fiscal diamond; domestic revenues, grants, sustainable borrowing and improvements in the efficiency of expenditure. We also analyze the budgetary implications of reforms to the electricity sector which are needed to eliminate the QFD of the sector and the potential budgetary impact of hydropower investments. The base from which our long term fiscal projections are made is the preliminary fiscal outturn for 2006. The budgetary resource envelope for primary expenditures amounted to 21.3% of GDP in 2006. This was composed of domestic revenues of 18.8% of GDP, grants of 1.2% of GDP (this excludes the MDRI grant of US$97.5 million which has also been excluded from net domestic financing), and financing of 1.8% of GDP. The latter consisted of net external financing of 1.6% of GDP, net domestic borrowing of -0.3% of GDP and privatization receipts of 0.6% of GDP.

D. DOMESTICREVENUES

122. There are three potential sources of growth in domestic revenues as a share of GDP; first, the income elasticity of some taxes, secondly, changes in tax policy (tax rates, thresholds, exemptions etc), and thirdly, improvements in tax administration. This section explores the scope for raising domestic revenues from each of these three sources over the next 10 years.

39 123. Table 3.1 compares the tax revenue collected in Tajikistan with that of other CIS countries.20 Although Tajikistan had the second lowest revenue to GDP ratio among the CIS countries, this is at least partly explained by Tajikistan having the lowest GDP per capita income in the CIS. Lorie (2003), investigating tax performance in CIS countries, regressed the tax revenue to GDP ratio on PPP GDP per capita for a large sample of developed, transitional and emerging market economies, with an additional regression which included a dummy variable for the OECD, graduating transition and CIS countries. The predicted tax revenue to GDP ratio for Tajikistan was 13.6% and 18.1% in the regressions with, and without, a dummy respectively. Tajikistan’s tax revenue of 16.6% of GDP in 2006 was not, therefore, out of line with what would be expected based on the country’s per capita income.

Table 3.1. Tax Revenue/GDP Ratios in CIS Countries, 2004

GDP per capita US Tax Revenue as percent dollars of GDP Armenia 1060 17.2 Azerbai ian 940 16.9

Unweighted average 1270 24.1 Unweighted average of 70 1 20.2 low income CIS countries

124. Domestic revenues amounted to 18.8% of GDP in 2006, of which 2.2% of GDP was non tax revenue (see Table 3.2).21 Domestic revenue relies heavily on income and profit taxes, which include a payroll-based social tax, which contributed 25% of revenue, and VAT, which contributed almost 40% of revenue in 2006. Customs duties contribute a relatively small share of revenue (6%) mainly because large shares. of imports are accorded duty free entry under bilateral trade agreements with CIS countries.

’O Data for 2004 for Turkmenistan are missing. Lorie (2003: 37) reports data on tax revenue for Turkmenistan of 23 percent of GDP in 1998. ” This includes extra budgetary funds of 1.3 percent of GDP reported in IMF fiscal data.

40 Table 3.2. Domestic Revenues as % of GDP

2000 2001 2002 2003 2004 2005 2006

Domestic revenues 13.9 14.6 16.6 16.2 16.8 19.3 18.8

Tax revenues 12.9 14.0 14.8 15.0 15.2 16.6 16.6 personal income tax 1.2 1.3 1.2 1.1 1.1 1.1 1.2 profit tax from corporations 0.6 0.6 0.6 0.5 0.6 0.5 0.6 minimum tax on enterprise profits 0.0 0.0 0.0 0.0 0.0 0.3 0.4 social taxes 1.6 1.6 1.8 1.7 1.8 2.1 2.0 tax on real estate 0.0 0.0 0.0 0.0 0.2 0.1 0.2 land tax 0.4 0.4 0.3 0.3 0.2 0.1 0.1 single tax from agric producers 0.0 0.0 0.0 0.0 0.0 0.8 0.5 sales tax on exports 3.3 2.5 1.9 1.9 1.4 1.5 0.6 cotton fibre 2.1 1.7 1.2 1.3 1.1 0.8 0.6 aluminum 1.2 0.8 0.6 0.6 0.3 0.7 0.0 tax on retail sales 0.0 0.0 0.0 0.0 0.3 0.4 0.4 VAT 2.5 3.6 4.8 5.2 5.7 6.3 7.4 domestic VAT 1.5 1.5 1.3 1.3 1.3 1.4 1.9 external VAT 1.o 2.1 3.4 4.0 4.4 4.9 5.5 road user tax 0.4 0.5 0.6 0.7 0.6 0.5 0.7 simplified tax 0.0 0.0 0.0 0.0 0.1 0.1 0.1 excise tax 0.5 0.6 0.9 1.o 0.9 0.8 0.7 customs duties 1.4 1.8 1.7 1.5 1.6 1.4 0.9 processed product tax 0.4 other taxes 0.9 1.1 1.1 1.2 0.6 0.6 0.6

Non tax revenues 1.o 0.6 1.8 1.2 1.7 2.7 2.2 Sources: Ministry ofFinance, MF

E. BUOYANCYAND INCOME ELASTICITY OF REVENUES

125. Domestic revenues have been quite buoyant since 2000, rising by 4.9% of GDP in six years (Table 3.2). Because there have been many changes in tax policy since 2000, with some taxes abolished, new ones introduced and the coverage and/or rates of other taxes changed, it is difficult to disentangle precisely all of the reasons for the buoyancy of tax revenues between 2000 and 2006. However, VAT on imports has played the major role, adding 4.5% of GDP to tax collections. The growth in VAT was largely attributable to a change in the methodology, effected in the early 2000s, for calculating the VAT from the origin to the destination principle for trade with CIS countries (IMF, 2006: 10); Le., the VAT on goods which are traded between Tajikistan and other CIS countries is now levied by the importer (destination) instead of the exporter (origin) which was previously the case. As Tajikistan has a large trade deficit with other CIS countries, the change in methodology for applying the VAT served to greatly expand the VAT base.22 In 2006 VAT was further boosted by very buoyant import growth of 38%, which was at least partly driven by the strong growth of remittances.

22 Levying VAT on the basis of the origin of imports was common to all CIS countries for their intra- CIS trade, but most have now switched to the destination basis.

41 126. Growth in social taxes and excise taxes also made modest contributions to the buoyancy of tax revenues during the 2000-06 period. In 2005, tax collections were boosted by a number of tax policy changes made that year, such as the increase in the aluminum tax rate (from 1% to 3% of exports, which was then replaced by the processed product tax in 2006) and the introduction of a minimum tax on enterprises and a single agricultural tax.

127. While it is difficult to assess the income elasticity of the major tax handles based on their historical performance because of the changes in tax policy, their observed buoyancy does provide some guide as to whether or not these tax handles are income elastic. In fact, as noted .above, only VAT on imports has displayed much evidence of buoyancy during the 2000s. It does not appear likely that income or profit taxes will be income elastic in the future. In the case of profit taxes the formal enterprise sector of the economy largely comprises enterprises set up in the Soviet era which are not very dynamic, and hence are unlikely to expand rapidly as a share of GDP, while the bulk of formal sector employment is in the Government sector, and hence the scope for increasing the tax base is also highly constrained, given that any increase in the Government wage bill faces budget constraints. Because of the structure of the economy and employment, it is not surprising that neither individual income tax nor profit taxes have displayed much buoyancy over the last five years.

128. Will VAT on imports continue to increase as a share of GDP? This seems unlikely. Imports are already high as a share of GDP (48% in 2006), especially for a landlocked country, and so will probably not rise further. The main factor that led to the buoyancy displayed by VAT on imports during the 2000s, the change in the methodology for computing the VAT, is a one off effect which will not be repeated. Consequently it seems prudent to conclude that natural income elasticity of the tax system over the medium term will not be a significant source of additional revenue.

E”. CHANGES IN TAXPOLICY

129. What scope is there for raising more revenue through tax policy changes without creating distortions to resource allocation which would impede growth, which effectively means that tax rates should be moderate and not out of line with international levels. This subsection discusses the scope for generating additional revenues by raising the tax rates on the tax handles which are major contributors to tax revenues.

Corporate Profits Tax

130. Corporate profits are currently taxed at 25% (the rate was reduced from 30% in 2005). The current rate is in line with the unweighted average corporate profit tax for CIS countries and the transition countries of Eastern Europe (Lorrie, 2003: 38) which implies that if Tajikistan wants to be competitive in creating a business friendly tax system, it should not reverse the 2005 tax rate reduction. Moreover, the tax base for corporate income tax is small, given the rate of 25%, taxable corporate income is only about 2.5% of GDP. Raising the corporate profit tax rate would not generate much additional tax revenue.

42 Personal Income Tax

13 1. Personal income tax (PIT) rates are low. There are three rates; O%, 8% AND 13%. The current top marginal rate of 13% is much lower than the equivalent in other CIS countries, which ranges from 20% to 35% (Summers and Baer, 2003: 7). Less than 5 percent of labor income was collected in personal income taxes in 2004, which is attributable to both the schedule of low rates and the fact that the incomes of a large share of workers fall below the threshold at which tax begins to be paid. Nevertheless it should be possible to increase the average PIT collection to about 10% of labor income by raising the top marginal rate to 25% and levying a rate of 10-15% on average incomes, which would generate additional revenue of about 1% of GDP. Raising PIT rates could also make the personal income tax more income elastic, because the schedule of taxes will become more progressive. The top marginal PIT should be set at the same level as that of the corporate income tax rate, in line with international best practice, to remove the opportunity for the owners of small businesses to reduce their tax liabilities by shifting income from profits to their own salary. There is also a case on equity grounds for raising the top marginal rate of PIT.

Social Security Taxes

132. Social security taxes make a substantial contribution to revenue but the applicable rates are already high - enterprises pay 25% of their monthly wage bill - and this is, in effect, a tax on the hiring of labor. Hence raising the tax rate would discourage employment. Moreover, these taxes are earmarked for pension and social security benefits so there is little point in raising the contributions unless this is needed to meet a gap in the funding of benefits.

VAT

133. The VAT rate is already 20% (and is effectively 22% because of the road user tax which is also levied on the VAT base at a rate of 2%). This is at the high end of VAT rates worldwide, so an increase in the rate is unadvisable.

Customs Duties

134. Customs duties rates are relatively low: there are seven rates with a range of between 0% and 15%. However, although imports are high as a share of GDP, the customs tax base is much narrower because around three quarters of all imports originate in the CIS countries, with many of which, including Russia, Tajikistan has signed free trade agreements. Although the average customs duty rate is about 7%, the effective tariff rate during 2001-2005, calculated as customs duties collected as a percentage of the value of total imports, was only 2.5%, implying that about 65% of imports enter Tajikistan tariff free. Hence any increase in customs tariffs would apply to a minority of imports which would limit its potential for revenue generation, as well as distorting resource allocation.

43 Excise Duties

135. There could be scope for raising excise duty rates, but this is also probably limited. Almost two thirds of excise duties comprise taxes on petroleum products. Petrol is already taxed at a high rate (about US$0.5 per liter) which limits the scope for further increases, although diesel is taxed much more lightly. There may also be scope for raising taxes on alcohol.

Summary of Tax Policy Measures

136. Raising the PIT rate (with the top marginal rate aligned with the corporate income tax rate) offers the most promising avenue for mobilizing more revenue through tax policy changes without adversely affecting incentives for economic growth. T his would also improve the progressivity of the tax system and make it more income elastic. Apart from higher excise duty rates on alcohol and possibly fuel, there is limited scope for raising more revenue from indirect taxes, either because tax rates are already high (e.g., VAT) or because the tax base is narrow (customs duties).

G. TAXADMINISTRATION

137. Improvements in the tax administration, which is implemented by the Ministry of State Revenues and Duties (MSRD) through its tax and customs departments, offer the greatest scope for enhancing domestic revenue. Reforms to tax administration have been underway for several years, and some progress has been made. The tax department has been restructured along functional lines and a Large Taxpayer Unit (LTU) set up. A new tax and customs codes were introduced in 2005. Nevertheless, major weaknesses in tax administration have still to be rectified.

138. Although the tax code is sound in most respects and in line with international standards (Summers and Baer, 2003: 8), it has not yet been fully implemented. Working practices do not accord with the norms of a modern tax administration, notably in the lack of self assessment by taxpayers. The tax department is badly organized, with resources spread over far too many offices (74) geographically, of which the majority deals with relatively few taxpayers whose contribution to total tax is marginal. The setting of monthly targets for each tax district also impedes effective tax collection at the national level. Although all taxpayers have a Taxpayer Identification Number (TIN), the law does not require that the TIN be quoted on all invoices, which is an impediment to preventing fraud in the payment of indirect taxes. The registration of taxpayers is chaotic. There is virtually no computerization of tax administration. The quality of tax audits is very poor because there are too few tax officers with the requisite skills. Tax administration suffers from a shortage of funds for operational activities and office buildings are poorly maintained and lack basic equipment. Tax collection is also seriously hampered by corruption, which is partly the result of the low salaries paid to tax officers.

139. The collections of some indirect taxes appears to be too low given the size of the tax base, which would indicate that substantial scope exists for raising tax revenue by improving compliance through better tax administration or by tightening up on exemptions. For example, the amount of VAT collected in 2006 is consistent with

44 VAT having been levied on only about a third of private consumption expenditures, which implies that either compliance is poor or that a large share of consumption items are exempt from VAT.

140. The priority for tax administration reforms is to introduce modern tax administration practices backed up by comp~terization.~~Tax officials should stop attempting to verify manually every tax return, and instead adopt the principle under which taxpayers are expected to comply voluntarily, through self assessment, with their returns being subject to audits on a selective basis. The auditing system needs to be strengthened, with better trained auditors undertaking audits selected on the basis of greatest risk. The law should be revised to enforce the quoting of a TIN on all invoices. Much work needs to be done to enable the tax code to be fully implemented, such as drafting and ratifying supporting regulations. Efficiency in tax administration can be enhanced by the introduction of IT systems for taxpayer registration and accounting and core administration functions.

141. Over the longer term, the structure of the tax department should be re- organized to reduce the number of small tax offices. in the regions which play a marginal role in tax c~llection.~~Large taxpayers should be handled by the LTU, which needs strengthening. Model tax inspectorates should be established in each of the three regions to handle medium sized taxpayers. The closure of small offices, computerization, plus a change in working practices, will then allow the number of officials employed by the tax department to be reduced by at least 50% (Le., from the current level of around 1,600 to 750). This would enable salaries for the remaining staff to be raised substantially which would provide better incentives to attract and retain higher caliber staff and reduce incentives for corruption. In addition the tax police, which is a separate tax enforcement agency within the MSRD comprising 20% of its staff, should be abolished and the resources which are thus saved used to strengthen tax administration. The tax police are not compatible with modern tax administration practices which are based on self assessment: instead the existence of a separate tax agency which has close contact with taxpayers encourages corruption (Summers and Baer, 2003: 24).

142. The Customs Department can also be strengthened through the upgrading of facilities and infrastructure at border posts, developing application systems and procedures to support modern customs practices and training customs officers (Harrison et al, 2005).

143. How much additional revenue could tax administration reforms generate? Although it is difficult to be precise, based on the experience of other developing countries which implemented major reforms to tax administration, an increase in tax revenue of around 20% should be attainable over the long term; i.e, tax revenue could be raised from the current level of around 16% to 19-20% ofGDP.

23 These paragraphs are based on the recommendations for reform of an ADB funded tax administration reform project and the recommendations in Harrison et al (2005). 24 The geographical fragmentation of the tax system reflects the system of financing local government budgets, whereby a proportion of taxes collected locally are retained by local authorities (and usually supplemented by transfers from the center). The rationalization of tax offices, therefore, needs to be linked to changes in the financing of local government budgets.

45 Long-term Revenue Projections

144. It does not appear likely that the income elasticity of the tax system will make a significant contribution to revenue growth in the future, for the reasons discussed in Section 3.4.1 above. The scope for raising tax rates is also limited because the rates of several of the most important tax handles, such as corporate income taxes and VAT, are already at levels which are in line with comparable rates in similar economies, while the scope for raising more revenue from customs duties is constrained by Tajikistan’s free trade agreements with its main trading partners. The exceptions are personal income taxes and excise duties, where raising tax rates could mobilize at least another 1% of GDP in tax revenue. Improving tax administration offers the best prospects for enhancing tax revenue. The tax system is characterized by serious weaknesses which, in principle, can be rectified. Introducing the practices of modern tax administration could generate additional tax revenue of over 3% of GDP. Although it will not be easy to reform tax administration, it should be possible to implement major reforms over the 10 year time horizon of this paper.

145. Table 3.3 provides 10 year projections for domestic revenues as a share of GDP. The projections incorporate the following assumptions. All existing revenue handles growth in line with nominal GDP except for the following. Based on the assumption that marginal PIT rates are raised, with the top PIT rate set at 25%, the tax collected from PIT as a share of labor income rises from the current level of over 4% to 8% by 2010, and by a further 0.25 percentage points of labor income per year. . Formal sector labor income is assumed to be a constant 25% of GDP. The VAT projections incorporate the gains from VAT levied on the higher cost of electricity, which is explained in section 3.7. However, we also make an offsetting adjustment to take into account the likelihood that higher electricity tariffs will depress demand for spending on consumer goods other than electricity, and hence depress the VAT collected from sales of these goods. The assumption is that total private consumption expenditure will be inelastic with respect to increases in electricity tariffs; Le., the marginal propensity to consume is not affected. Therefore, consumption of goods and services, other than electricity, falls by the same absolute amount as the increase in utility bills for private consumers, which we proxy by taking the total increase in electricity bills minus that of government and TALCO. Consequently, almost three quarters of the gain in VAT from higher electricity bills is offset by lower VAT payments on goods and services other than electricity. The processed product tax, which is levied on TALCO, is reduced by the amount of subsidy required to offset the impact of higher electricity tariffs on TALCO’s input costs, so that by 2009 TALCO no longer pays this tax, as explained in section 3.7.

146. To project the cotton export sales tax, we make use of the modeling of cotton sector development scenarios made in World Bank (2005B: 56) which examines the potential impact of reforms to raise efficiency in the sector, including reforms to establish secure land rights, resolve the problem of farm debt, abandon the setting of production targets, introduce competition in input supply, establish alternative sources of financing for cotton farms and liberalize trade in raw cotton and cotton fiber. Efficiency enhancing reforms could raise cotton yields per hectare by a third, to 2.5 tons per hectare, which was the lower bound of yields obtained during the Soviet period, and could raise the efficiency of ginning by 19%. Total cotton incomes would therefore rise by just over 50%, to almost US$330 million a year. Assuming that the export tax rate remains at lo%, export tax revenue could be raised to US$33 million, compared to the tax collection of US$16 million in 2006 (when production was adversely affected by drought). Our projections assume that this level of export taxes will be reached at the end of the 10 year forecast period, in 2016, after a series of equal increases in each of the preceding years.

147. The gains from tax administration reform are projected at a cumulative 2 percent of total tax revenue per year for 10 years, but only beginning in 2008 because time will be required to strengthen the institutional capacity of tax administration; hence by 2016 tax administration reform boosts tax revenues by 18%. The revenue projections also include the projected revenue earned from government’s capital investment in the hydropower projects, and the surplus in the form of taxes or royalties which can be earned from electricity exports, which is explained in section 3.7.

148. Domestic revenues increase by 5.6% of GDP, from 18.8% of GDP in 2006 to 24.4% in 2016, as a result of the following factors. The increase in individual income tax, together with the elasticity of that tax handle, generates gains of 1.2% of GDP. VAT increases by 0.3% of GDP as a result of the net impact of electricity tariff rises. However, the processed product tax revenues amounting to 0.4% of GDP in 2006 are foregone because of the impact of higher electricity tariffs on the cost structure of TALCO. Tax administration reforms yield additional revenue of 3.2% of GDP by 2016, contributing almost 60% of the growth in revenues over the next 10 years. A further 1.3% of GDP in revenue is provided by income from capital investments in HPPs and royalties and/or taxes from exports of electricity.

Grants

149. Donor grants to the Government budget have been low, although project grants may have been under-reported. Excluding the MDRI debt relief received in 2006 (which is recorded as a grant in fiscal data) grants to the budget amounted to 0.7% of GDP in 2005 and 1.2% of GDP on 2006, but grants should jump to around 2% of GDP in 2007, boosted by the receipt of US$9.2 million from the Education for All- Fast Track Initiative. The majority of donor grants are disbursed to agencies outside the Government budget, such as community based organization^.^^ Future levels of

25 An estimated US$l09 million of grant aid from donor organizations was disbursed outside the budget in 2005 compared to US$17 million of grants disbursed to the budget in the form of both budget support and project aid. The estimate for the non budget grants is a minimum because data were not available from all donors.

48 grant aid to the budget, in the form of budget support and/or project aid, could be boosted if donor governments increase their overall aid to Tajikistan, in line with commitments made at international fora to expand aid in real terms to the poorest countries, and/or if a larger share of donor aid were to be channeled to the Government budget. The latter will depend on the perceived quality of public financial management, especially with regard to aid in the form of budget support. Concerns about fiduciary weaknesses have deterred donors from providing more aid in the form of budget support (the summary of the PEFA assessment in Chapter 5 gives details of fiduciary weaknesses).

150. It is very difficult to make any quantitative projections of the volume of budget grants over the long term, because the range of possible outcomes is large. At the low end of the range, Tajikistan might receive no increase in grants, because donors remain unconvinced that fiduciary standards are improving sufficiently and/or because Tajikistan contracts more loan finance from non traditional lenders, thereby reducing the perceived need for more grant aid. At the other end of the range, grants could increase several fold, possibly to around 5% of GDP, as donors both increase grant aid to Tajikistan as part of a more general expansion of development assistance to low income countries and channel a much larger share of their grant aid to the Government budget. The latter will clearly require improvements in public financial management to strengthen and make more transparent the budget process and to improve the allocation of expenditures, especially in those sectors which can contribute to meeting the MDGs. The assumption we have used here is that donor grants will increase by 10% a year in nominal dollar terms, using the 2007 budget projection as a base, which is consistent with the commitment made by the G8 countries in 2005 to increase their aid to poor countries by 60% by 2010. We assume that the 10% growth in grants will be sustained through to 20 16.

15 1, Projected donor grants for the budget as a percent of GDP are shown at the bottom of Table 3.5. They rise from 1.2% in 2006 to 2.0% of GDP in 2007, and then slowly increase to 2.2% of GDP by 2016. Consequently, the projected increase in grants makes only a very modest contribution to fiscal space over the long term. However, these projections are unavoidably highly speculative and depend to a large extent on decisions which are outside the control of the Tajikistan government.

H. GOVERNMENTBORROWING

Recent Trends in Borrowing

152. Net borrowing to fund the Government budget has been relatively modest in the last six years. During 2001-2006, net external financing averaged 2.8% of GDP and net domestic borrowing averaged negative 1% of GDP. The former was constrained by the need to restore external debt sustainability, which led to government imposing a constraint on external borrowing to fund the PIP of 4% of GDP per annum. Domestic borrowing was constrained by government’s policy priority to stabilize the macro economy which required that Government should avoid financing its deficit by borrowing from, and instead accumulate savings with, the Central Bank, to allow for a deceleration of money supply growth and the accumulation of foreign exchange reserves by the Central Bank. The very shallow nature of the domestic financial system offers very little scope for Government to

49 borrow either from the commercial banks or from non bank financial institutions. Consequently, the macroeconomic imperatives dictated that domestic financing should be negative. Privatization receipts added an average of 0.5% to the budget’s financing resources during 2001-2006. Table 3.4 below provides details of financing for the Government budget during 2001-2006.

Table 3.4: Government Financing 2001-2006 % of GDP

Other external financing includes arrears and debt rescheduling Net borrowing from the NBT excludes the receipt of the MDRI debt relief in 2006 (equivalent to 3.5 percent of GDP). Other domestic borrowing includes non bank borrowing and the discrepancy Source: Ministry of Finance, LMF

153. The following subsections evaluate the magnitude of government borrowing which could be incurred in a sustainable manner over the next 10 years.

External Borrowing

154. Government’s net external financing will increase dramatically over the medium term as a result of its contracting of a US$604 million loan for infrastructure projects from China and its intention to contract a loan of US$400 million from China for a HPP. Net external financing is projected to rise to around 14% of GDP in 2007 and 2008 before falling to 11% of GDP in 2009. The magnitude of external financing over the next three years will lift the NPV of Tajikistan’s PPG external debt over at least two of the five thresholds at which debt is deemed sustainable under the IMF/IDA Low Income Country Debt Sustainability Assessment Framework (see Tables 3.5 and 3.6).

50 Table 3.5: IMF/IDA LIC DSA External Debt Sustainability Thresholds for a Medium CPIA Performer

Indicator Threshold (%) NPV of external debt to GDP 40 NPV of external debt to exports 150 NPV of external debt to government revenue 250 I External debt service to exPorts I 20 I

51 2 3 U

2 3 U

E 3 U

1 3 U

1? 3 U

- 3 U

2 3 U

D 3 3 U

0 3 3 U

c 3 3 U

3s U

m 155. Our projections for external financing are derived using the following methodology. For the next five years (2007-201 l),we adopt the projections made in the most recent DSA, which was completed in January 2007 (IMF & World Bank, 2007). External loan disbursements and amortization in this period mainly reflect the implementation of loan agreements already signed, such as the US$604 million Chinese infrastructure loan. We then make our own projections for 2012-16. After 201 1, there is more scope for loan disbursements to vary, as these will involve loan contracts not yet signed. As the purpose of this chapter is to assess the scope for fiscal space which is consistent with sustainable borrowing, our projections are based on the assumption that the Government will aim to bring back, over the long term, its external debt indicators to levels which do not exceed the thresholds at which external debt is deemed unsustainable, by restricting new external borrowing. Hence, given the projected loan repayments and the grant element of the loans, we construct a profile of loan disbursements which ensures that all sustainability indicators are met by 20 16.

156. The MF/IDA Low Income Country Debt Sustainability Framework identifies five thresholds for external debt sustainability, which are conditional on the CPIA scores of the debtor. Higher CPIA scores imply stronger economic and fiscal management and this allows the sustainable level of external debt to be higher. Tajikistan was judged a medium CPIA performer in 2006 (after being previously scored as a low performer). We assume that Tajikistan will be able to maintain its status as a medium CPIA performer over the long term, and as such, its external debt sustainability will be evaluated on the basis of a medium CPIA performer, the external debt thresholds for which are set out in Table 3.5

157. The projected net external financing over the medium term will lift the NPV of debt to both GDP and exports above the sustainability thresholds, although the indicators will start to decline after 2009. The NPV of debt to exports is the most problematic, as this exceeds the threshold by the largest margin. Under our projections, this indicator falls back to the threshold of 150% by 2016. The NPV of debt to GDP returns to the threshold earlier, by 2013. We assume that all external PPG debt is used to finance the Government budget,26 that the average grant element of the outstanding debt stock in 2007 (about 23%) will remain the same throughout the 10 year period of the projections and that average interest payment on external debt will rise to 1.75%. Repayments after 201 1 are projected on the assumption that the average loan maturity is 25 years with an average grace period of five years, and that, on average, one third of the portfolio in each year is still under its grace period.

158. Table 3.6 presents the profile of external loan disbursements and repayments, together with the implied trends in the debt indicators, over the next 10 years. As can be seen from the table, to restore external debt sustainability, it is necessary for external loan disbursements to fall sharply after 2009, by which time the bulk of the Chinese loans will have been disbursed. Disbursements must average not more than about 4% of GDP during 2009-13 and then about 5% of GDP in the following three years.

26 This may not necessarily be valid, for two reasons. m First, if Tajikis tan contracts new credits from the IMF under the PRGF, the financing will not be used for the budget. Secondly, some of the external loans may be used by public enterprises, or by the Central Bank.

53 159. Consequently, once the projected increase in external loan disbursements during 2007-09 has been absorbed, there is very little scope left for any further increase in external borrowing without breaching the debt sustainability ceilings. The sustainable level of gross public external borrowing is only 4-5% of GDP, and even this level of borrowing would be unsustainable if the country were to suffer a major shock to GDP growth or export growth. This is one of the key policy messages of this paper: the Government cannot solve its budgetary resource constraints through higher external borrowing on a sustainable basis, even if external lenders are willing to extend it more credit in the medium term, because that will jeopardize long term debt sustainability.

Domestic Borrowing

160. The main constraint on government domestic borrowing is the need to maintain macroeconomic stability without crowding out private sector borrowing from the banking system. Government’s domestic debt stock is small, at around 4% of GDP, and this debt carries a fixed interest rate of about 5%; hence the sustainability of the domestic debt stock is not the binding constraint on domestic borrowing. Given the low level of domestic debt, Government could issue more domestic debt and remain solvent, but the issuance of more debt faces three constraints: the very narrow domestic monetary system (broad money was only 8.5% of GDP at the end of 2006), the high level of polarization of the banking system (foreign currency deposits were 69% of total bank deposits at the end of 2006), and the lack of significant non-bank financial institutions. The poorly developed domestic financial sector also constrains the use of alternative macroeconomic instruments to create fiscal space. If Government were to fund a more expansionary fiscal policy by borrowing from the Central Bank, the latter would then face the need to sterilize the reserve money thus created, but its ability to do so is constrained by the same characteristics of the domestic financial system which restrict the scope for government borrowing from the domestic market.

161. To derive projections of domestic borrowing consistent with macroeconomic stability, we assume that only borrowing from the Central Bank and commercial banks is possible. This is because the non bank financial sector is too small to hold any significant amount of government debt. Sustainable borrowing from the Central Bank is determined with reference to its balance sheet; whereby the increase in reserve money (RM) is identical to the sum of the changes in its net foreign assets (NFA), net credit to government (NCG), net credit to the private sector (NCPS) and other items net (OIN).

ARM ANFA + ANCG + ANCPS + AOIN

162. We assume that the velocity of circulation of reserve money will fall at 2.5% per year; hence reserve money will grow at the annual rate of nominal GDP growth plus 2.5%, NFA is the sum of foreign exchange reserves and foreign exchange liabilities. We assume that the former increase at the same rate as imports (which grow at the same rate as GDP), thereby maintaining the foreign reserve to import cover constant, while foreign exchange liabilities, valued in current dollars, remain constant. We assume no net lending by the Central Bank to the private sector and a zero change in its other items net. With these assumptions, the Central Bank’s lending

54 to the Government can be derived as the residual. After 2007 (for which we take the projection from the Government budget) this is negative up to 2016 and averages 0.4% per annum (Table 3.6).

163. The constraints on government borrowing from the commercial banks are the growth of their liabilities and the extent to which government wants to avoid competing with the private sector and public enterprises for the very limited pool of commercial bank credit. Commercial bank liabilities are mainly deposits, but more than two thirds of deposits are foreign currency denominated deposits and we assume that commercial banks will not use foreign currency deposits to fund purchases of government debt because of the currency mismatch that this would entail. Consequently, the resources available for funding government borrowing are restricted to domestic currency deposits. We assume that Government will restrict its net borrowing from commercial banks to 30% of the increase of domestic currency deposits, thereby leaving the remaining 70% to fund lending to the private sector and public enterprises, as well as holdings of cash reserves and other assets. We assume that the velocity of circulation of Somoni deposits will fall by 5% per annum. As can be seen from Table 3.6, government borrowing from the commercial banks is positive but very small, averaging only 0.1% of GDP per year during 2007-2016. Net domestic borrowing, the sum of net borrowing from the Central Bank and commercial banks, is slightly negative over the next 10 years, averaging 0.2% of GDP per annum. This is slightly larger than the average so far during the 2000s (Le,, the savings are slightly less) but the amount of fiscal space created is very small.

Privatization Receipts

164. The Government budget received modest inflows from privatization receipts, averaging 0.5% GDP during 2000-2006 (see Table 3.4). This was the result of a program of privatizing mainly small scale public enterprises, which is now almost complete. AI though there are still several large-scale enterprises remaining in the public sector, such as TALCO, the aluminum smelter, we have not made projections of privatization receipts for the next 10 years because it is not yet clear what further plans the Government has for privatizing these public enterprises.

Interest Payments

165. To make projections for external interest payments, we use the DSA projections for the period 2007-2011 and in subsequent years we assume that the average interest rate on the outstanding stock of debt is 1.75%. For domestic debt, the interest rate on the 300 million Somonis of existing government debt, which is non- marketable, is 5%. We assume that new debt will be issued on the money market, through a competitive auction, and that it will, therefore, pay a market interest rate. It is likely that new marketable debt will be short term, given that the money market is not liquid. Hence we assume that the interest rate paid on this debt will be equal to that paid on short term time deposits. To project this interest rate, we take the average deposit rate for time deposits with a maturity of 1 month to 3 months over the last five years, which was 11.5%.

166. The projected external and domestic interest payments, as a percent of GDP, are shown in Table 3.6. As can be seen, they increase by 0.4% of GDP during the 10

55 year period of the projection, from 0.5% of GDP in 2006 to 0.9%of GDP in 2016, with all of the increase attributable to higher interest payments on external debt. . Interest costs remain moderate because the bulk of the Government’s debt stock is external debt contracted at concessional rates of interest.

I. FISCALIMPLICATIONS OF REFORMSIN THE ELECTRICITY SECTOR

167. Government’s fiscal position over the long term will be affected profoundly by developments in the electricity industry. The electricity industry, which is almost entirely in public ownership, incurs substantial QFDs in serving the domestic market, mainly because of selling electricity at below its long run marginal cost. However, the future prospects for the industry and for the Government budget are brighter if Government fully implements agreed reforms. It should be possible to reduce the QFDs through electricity tariff reform and investments in system rehabilitation, while major new investments in hydropower generation will create the potential for substantial export earnings and government revenue, although there will also be costs for Government. This section examines the long-term fiscal implications of reforms and investments in the electricity sector. It begins with an explanation of the QFDs, briefly outlines the main planned and potential investments in the sector, and then makes 10 year projections of the fiscal impact of electricity sector investments and reforms.

Quasi Fiscal Deficits

168. The electricity sector is the main source of QFDs in Tajikistan. World Bank (2005E) estimated the QFD in the sector as the difference between the actual revenue collected in cash by Barki Tajik, the state owned electricity utility, and the revenue which would be collected if it set electricity tariffs at levels sufficient to cover the long run average incremental cost (LRAIC)27and reduced its technical and collection losses to levels which are in line with international norms. The LRAIC is computed as the sum of the costs required for capital investment in rehabilitation of the electricity system plus the incremental operational and maintenance expenditures on that system, divided by incremental sales of electricity, all evaluated over a 20 year time horizon.

169. The electricity sector QFD for 2003 was estimated in World Bank (2005E) as US$273 million. We have used this methodology to update these estimates to 2006.28 The QFD arises from three sources. First, Barki Tajik incurs technical and commercial losses which in 2006 amount to 17% of electricity supplied to the domestic market, as a result of loss of electricity in the distribution system, theft of electricity, defective metering and inappropriate use of norm-based billing for customers without meters. The norm for technical and commercial losses is set at 10% of electricity generated, hence Barki Tajik’s excess technical and commercial losses, which contribute to the QFD, amount to 7% of electricity supplied in 2006. Evaluated at the LRAIC, the excess technical and commercial losses amount to US$25 million in 2006. The second source of QFDs is the difference between the average tariff charged by Barki Tajik and the LRAIC. The LRAIC is estimated at 2.3

21 The LRAIC is used as a proxy for long run marginal cost. 28 This methodology is also employed in Ebinger (2006) where the QFDs are referred to as hidden costs.

56 cents per KWh in 2006 prices (2.1 centsKWh in 2003 prices), compared to the average tariff charged by Barki Tajik of 0.6 cents per KWh in 2006. The losses arising from charging average tariffs which are not much more than a quarter of the LRAIC are projected to reach US$210 million in 2006. Thirdly, Barki Tajik incurs losses because only 67 percent of the electricity bills it issues to its customers are actually paid in cash, compared to the international norm of 98 percent. These losses amount to a further US$27 million. The total QFD, therefore, is estimated at US$262 million, or 9.3% of GDP, in 2006.

170. What are the fiscal implications of the QFDs in the electricity sector? In most years, the QFDs are not subsidized directly through the Government budget; an exception was 2001 when Barki Tajik defaulted on tax payments because of liquidity constraints. However, some form of government subsidy will eventually become inevitable unless the QFDs are eliminated, because Barki Tajik will not be able to fund the investment and associated operational and maintenance costs which are required to rehabilitate the electricity system and keep it functioning at an adequate standard. Consequently either the costs will have to be borne from the Government budget, thereby reducing fiscal space by an equivalent amount, or the electricity system will deteriorate, reducing the supply and reliability of electricity to customers, which will inevitably have a negative impact on economic growth. To eliminate the QFD in the electricity sector, it will be necessary for Barki Tajik to raise the average tariff to the LRAIC and to reduce its technical and commercial losses and its billing losses, through better metering and enforcement of payment of bills, for example.

171. The gas utility, Tajikgas, also incurred QFDs as a result of technical and commercial losses (especially incorrect billing), but these were much lower in magnitude than those incurred in the electricity industry; in the region of US$13 million, or 0.8% of GDP in 2003. Tariff reforms implemented since 2003 have largely eliminated the QFD in the gas industry. It is not clear whether there are significant QFDs outside of the energy sector because no comprehensive investigation of this issue has been undertaken,. There are potential QFDs in the cotton sector, where cotton farmers have debts which have not been fully verified but which could amount to as much as US$300 million. These debts are unpayable, in that they greatly exceed the capacity of cotton farmers to service them out of their income from cotton, and Government may face pressure to provide some form of bail-out for the farmers.

Future Hydropower Investments

172. Major investments in the electricity sector will have an impact on the Government budget over the long term, through their capital costs if undertaken with public funds and through the fiscal revenues which could be generated by taxing exports of electricity. Table 3.7 lists major investments which are ongoing or proposed. These include rehabilitation of the existing generation, transmission and distribution system, for which investments amount to US$199 million and investments in overhead transmission lines (OTLs) of US$244 million. Most of the investments in rehabilitation and OTLs (US$223 million) will be made by Government, largely through on lending foreign loans to Barki Tajikq2’ If these

29 One of the Chinese loans referred to in section 3.6 will fund the construction of an OTL.

57 investments are completed over a five year period (2006-20 lo), Government's contribution will amount to an average of 3% of GDP per year.

Table 3.7: Capital Costs of Major Ongoing, Planned and Potential Electricity Sector Investments

Millions of US dolllars I Projects and Nature of I Total capital '1 GOT 1 Source of GOT Investments cost contribution contribution Rehabilitation of generation, 199 199 Mainly loans transmission and distribution from ADB, Kwf, systems IDA Overhead Transmission Lines 244 224 Chinese loans

Sangtuda 1 HPP 600 100 From existing assets Sangtuda 2 HPP 220 Rogun 2900 800 From existing assets Other HPP 54

Total 4317 1331

173. At least two large investments in HPPs are currently being planned or under consideration. The most advanced in terms of project preparation are the Sangtuda I and I1 projects, with a total cost of US$820 million, of which Government's contribution is US$lOO million in the form of existing assets.30 An even larger project is the Rogun HPP, which will cost about US$3 billion, and for which Government will contribute existing assets which it claims are worth US$800 million. The HPPs will expand the generating capacity of Tajikistan and thereby provide power surpluses for export to countries such as Pakistan, Afghanistan, Iran and Russia. The Sangtuda HPP could be commissioned in 2009 and attain full production of TWh 2.67 per annum in 2013.3' Rogun could be commissioned in 2014 and reach full capacity of TWh 14.2 per annum in 2020. Together with the loss reductions attained through rehabilitation of the transmission and distribution network, these investments could add more than TWh 17 per annum by 2020 to the current capacity of TWh 12 million. By 2016, the time horizon for this paper, power sector investments could add almost TWh 8 per annum to the existing capacity.

174. The Government should leave the bulk of the investment in new hydropower projects to the' private sector, except for the assets which it already owns (the construction which took place during the Soviet period). The combined investment costs of the Sangtuda and Rogun projects amounts to around 100% of GDP, which far outstrips government's borrowing capacity (World Bank, 2005B: 46). Moreover, the generation and export of electricity is not a public good which needs to be provided

30 These are assets built during the Soviet era, when the projects were started but not completed. 3' A TeraWatt hour (TWh) is one billion Kilowatt hours (KWh).

58 by the Government; rather it is a commercial enterprise which can be undertaken by the private sector which is more likely to provide efficient management than Government.

175. If the HPP investments listed in Table 3.7 are implemented, the Government budget will benefit through two channels. First, government will acquire a share in the capital structure of the projects, in the form of equity and debt because it will contribute to the projects the value of assets constructed during the Soviet era. Hence Government will earn a return on capital in the form of interest and dividends. Secondly, the export of electricity could generate a surplus over and above the financial costs of generation and transmission, part or all of which could accrue to Government in the form of taxes or royalty payments.

Projections of the Fiscal Impact of Electricity Sector Investments and Reforms

176. The reforms to the electricity sector include electricity tariff reforms, to raise average domestic tariffs to the level of the LRAIC by 2010, which together with the investments outlined above, will affect the fiscal position through seven channels: (i) reduction of the QFD; (ii) increase in the cost of government’s own electricity consumption; (iii) increase in VAT paid on electricity consumption; (iv) higher payments through the Energy Compensation Mechanism; (v) a return on Government’s capital investment in the Sangtuda and Rogun HPPs; and (vi) a share of the surplus earned by exports generated by the Sangtuda and Rogun HPPs. Table 3.8 provides 10 year projections of the fiscal impact of the electricity sector reforms. The projections for the dollar values of the variables are all expressed in 2006 prices, unless otherwise stated.

QFDs

177. The projected QFDs take account of the scheduled reduction in technical and commercial losses by Barki Tajik from 17% of gross supply to the domestic market in 2006 to 10% by 2016, and a rise in cash collections from 67% of payments in 2006 to 90% in 2016. The main source of QFD reduction is the proposed increase in the average tariff from the 2006 level of 0.6 centsKWh to the LRAIC level of 2.27 centsKWh, in constant 2006 prices in 2010.32Together these measures cut the QFD from US$262 million (9.3% of GDP) in 2006 to US$30 million (0.7% of GDP) in 2016.

Government Utility Payments

178. Government is projected to pay US$2.8 million for its own electricity consumption in 2006. We assume that tariffs for government consumption will be increased in line with the proposal made by the World Bank33,tariffs should be raised to 4.58 cents/KWh, in constant 2006 prices, by 2010. This will push up

32 The tariffs on which the projections in this sub-section are made are those 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 201O”, May 2006. 33 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.

59 Government’s electricity bills to US$29 million by 2016, compared to US$3.5 million if tariffs were to remain at the 2006 levels. Hence an increase of US$26 million, or 0.6% of GDP by 2016, can be attributed to the proposed tariff increases.

VAT on Electricity

179. Electricity is subject to a 20% VAT. With the proposed tariff increases and the projected electricity demand, VAT paid on all electricity consumption is projected to rise to US$82 million in 2016, compared to a projected US$22 million if tariffs were to remain at 2006 levels.34 The gain in VAT payments as a result of the tariff increase is US$62 million, or 1.2% of GDP, in 20 16.

34 The projected US$20 million in VAT payments in 2016 which assumes no tariff increase does not take account of the higher domestic demand that would result from lower prices. This is because it is very unlikely that higher demand could actually be met given the levels of domestic supply, especially as the Sangtuda HPP, which will contribute to domestic supply from 2009 onwards, will not be economically viable if tariffs do not increase.

60 Table 3.8: Fiscal Impact of Electricity Reforms and Investments 2006-2016; % of GDP

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

1, Quasi Fiscal Deficit -9.3 -7.4 -5.6 -3.3 -0.4 -0.Q -0.8 -0.7 -0.7 -0.7 -0.6 1.a from excess technical -0.B -0.7 -0.6 -0.4 -0.3 -0.2 -0.1 0.0 0.0 0.0 0.0 and commercial losses 1.b from pricing below LMIC -7.5 -5.5 -3.8 -1.8 0.7 0.0 0.0 0.0 0.0 0.0 0.0 1.cfrom excess non cash payments -0.Q -1.1 -1.2 -1.1 -0.8 -0.8 -0.7 -0.7 -0.7 -0.7 -0.6

2. Increase in govi electricity payments 0.1 0.2 0.3 0.6 0.6 0.6 0.6 0.6 0.5 0.5 2.1 Govt payments at 2006 tariffs 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 2.2 Govi payments at projected tariffs 0.1 0.2 0.3 0.4 0.7 0.7 0.7 0.7 0.6 0.6 0.6

3. Gain in VAT 0.3 0.8 1.0 1.4 1.4 1.3 1.3 1.3 1.2 1.2 3.1 VAT at 2006 tariffs 0.6 0.6 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.4 0.4 3.2 VAT at projected tariffs 0.6 0.9 1.2 1.5 1.9 1.Q 1.8 1.8 1.7 1.7 1.6

4. Required TADAZ subsidy 0.0 0.2, 0.4 0.6 0.6 0.5 0.5 0.5 0.5 0.4

5. ECM 0.1 0.1 0.3 0.5 0.5 0.5 0.5 0.4 0.4 0.4 5.1 ECM at 2006 tariffs 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 5.1 ECM at projected tariffs 0.1 0.2 0.3 0.4 0.6 0.6 0.6 0.6 0.5 0.5 0.5

6. Return on HPP investments 0.1 0.1 0.2 0.2 0.2 0.5 0.7 6.1 Sangtuda 0.1 0.1 0.2 0.2 0.2 0.2 0.2 6.2 Rogun 0.0 0.0 0.0 0.0 0.1 0.3 0.5

7. Share of HPP surplus 0.1 0.2 0.2 0.4 0.6 7.1 Sangtuda 0.1 0.2 0.1 0.0 0.0 7.2 Rogun 0.0 0.0 0.1 0.4 0.6

8. Aggregate Fiscal Impact -9.3 -7.2 -5.4 -3.3 -0.6 -1.1 -0.8 -0.6 -0.5 0.1 0.5 Source: See Text Item “4. Required TADAZ subsidy” is factored in for illustration purposes. This may be true for low case scenario.

TALCO

180. Raising the electricity tariff will have a negative impact on the finances of TALCO, which consumes around 6.5 TWh of electricity per year. We assume that the tariff paid by TALCO will rise in line with the proposal made by the World Bank, which is that it will increase to 1.36 centsKWh, in 2006 prices, by 2010. With the electricity tariff raised to this level, TALCO electricity costs will rise to US$lll million a year (inclusive of VAT) by 2016, an increase of US$42 million over its electricity costs if the tariff was to remain at the 2006 level.

18 1. It is not clear whether TALCO could absorb such an increase in input costs and remains solvent, given that its current profits amount to only US$20 million per year; lower profitability of TALCO is mostly driven by tolling arrangement that prevents the company to benefit high world market prices for aluminum. The financial viability of TALCO is also constrained by the very high costs of transporting its material input (alumina) and output (aluminum) overland: these costs are around 4.5 times higher, per ton of aluminum, than comparable costs for a competitive developed country aluminum smelter, which are only partly offset by lower labor costs (World Bank, 2005B: 48). A more detailed analysis of the long term commercial viability of TALCO is required to enable a more definitive judgment to be made about the financial capacity of TALCO to absorb higher electricity costs. Given that the purpose

61 of this paper is to make long term fiscal projections, rather than recommendations specific to TALCO, we think it is justified to adopt a conservative view, for the purposes of fiscal planning, of the capacity of TALCO to absorb higher electricity prices, with the proviso that we are not expressing the view that TALCO should receive subsidies from the Government budget; it would clearly be preferable if the efficiency of TALCO could be improved sufficiently to allow it to absorb higher electricity costs without recourse to government subsidies.

182. Consequently, we assume that TADAZ can only absorb electricity cost increases up to a maximum of US$20 million per year, and that all cost increases above this level will have to be absorbed by the budget, through, for example, government foregoing taxes owed by TADAZ.35Therefore the cost to the budget of subsidizing TADAZ’s electricity consumption in 2016 will be US$22 million, or 0.4 percent of GDP. We assume that this subsidy will be affected mainly through lower tax payments by TADAZ.

Energy Compensation Mechanism

183. Under the ECM, Government makes payments through the budget to cover a fixed monthly volume of electricity consumption of eligible households whose aggregate household income falls below the average monthly salary for the applicable rayon. After implementing planned revisions to the ECM, it is expected that it will cover annual electricity consumption of KWh 1,800 per household and that 312,000 households will benefit.36 We assume that the number of households covered by the ECM will rise in line with population growth of 2% per year. At 2006 tariffs for residential consumption of less than 250 KWh per month, this would imply that the budgetary costs of electricity subsidies under the ECM would be US$4.2 million in 2016, but the costs will rise to US$24.5 million if these tariffs rise from the current level of 0.5 cents/KWh to 2.93 centsKWh (in 2006 prices), in line with the proposal of the World Bank. Hence electricity tariff reform will raise the costs of the ECM by US$20 million, or 0.4% of GDP, by 2016.

35 The main source of taxation of TADAZ was the export sales tax, but this was replaced in 2006 by the processed products tax. TADAZ paid US$lO million in processed product tax in 2006, much less than the US$29 million projected in the Government budget. 36 These data are from World Bank (2006B).

62 Return on Government Investment in Sangtuda and Rogun

184. To project the return which will accrue to Government from its share of the capital in the HPPs, we multiply Government’s percentage share of the capital of each HPP by the estimated financial cost of the HPP expressed in centsKWh of output and by the projected annual output of the project. The financial costs of the projects and their projected outputs are taken from World Bank (2004C: 31, 57 and 58). The financial costs of Sangtuda and Rogun Phase 1 (which is relevant for the time horizon considered here) are estimated at 2.44 centsKWh and 2.93 KWh respectively. Government’s share of the capital of Sangtuda 1 & 2 is assumed to be 12%, while that of Rogun is assumed to be 20%..37 Revenues do not begin to accrue until 2010, when Sangtuda is projected to begin generating electricity. By 2016, by which time the Sangtuda HPP will have reached its maximum output, it is projected that it will earn an annual return of US$8 million for Government, while Rogun, which will have reached only a third of its maximum output, will earn an annual return of US$27 million. The combined returns from the two HPPs will amount to 0.7% of GDP in 20 16. urplus from Electricity Exports

185. Electricity could be exported at prices above the financial cost of generation and transmission, thereby earning a surplus, all or part of which could be captured by the Government in the form of taxes or royalties. T o project the potential budget revenue from electricity exports, we assume that exports will equal the amount to which domestic supply of electricity exceeds domestic demand, and that the supply of exports will be met first by Rogun, if available and sufficient, and then by Sangtuda. We assume that the export price will be 5 centsKWh. Transmission costs are estimated at 0.53 cents per KWh.38 We assume that half of the surplus earned by the Sangtuda and Rogun HPPs can be captured by Government; the distribution of the surplus will be determined through a process of strategic bargaining between government and its co-investors. The share of the surplus which accrues to Government rises from US$l million in 2012, the first year of exports, to US$29 million in 2016, equivalent to 0.6% of GDP. However, the surplus is very sensitive to the export price. I f this were lower at 4 centsKWh, exports would still be viable because the price would still exceed the financial costs of generation and transmission (although only just in the case of exports from Rogun), but the surplus would be drastically cut.

Net Fiscal Impact of Electricity Reforms

186. The net fiscal impact of the electricity sector is derived by summing together all the components discussed above. We are concerned primarily with the change in the net impact over a 10 year time horizon which amounts to a positive gain of 9.8%

37 The investment cost of Sangtuda 1 and 2 is USIS820 million, of which government will contribute US$IOO million. The investment cost of the Rogun project is estimated at about US$3 billion. Government claims that the value .of existing assets, which will comprise government’s contribution, is US$800 million, but as this may be disputed by its investment partners we have assumed that government’s share will be valued at around US$600 million.

This is the average of the transmission costs to Iran, Pakistan and Afghanistan given in World Bank (2004C: 36).

63 of GDP in 2016 compared to 2006. The QFD reduction provides by far the largest component; a gain of 8.7% of GDP. The tariff rises which yield most of the QFD reduction also boost VAT payments by 1.2% of GDP, but raise Government electricity bills by 0.5% of GDP and ECM subsidies by 0.4% of GDP, while forcing Government to subsidize TALCO, probably by foregoing taxes, by 0.4% of GDP. Government earns a return on its investments in the HPPs of 0.7% of GDP and a surplus on their electricity exports of 0.6% of GDP. while the net impact of electricity sector reforms is positive, the main benefit of the reforms is to eliminate most of the QFD, rather than to generate significant amounts of additional budgetary resources which could be used to fund an expansion of public services. The on- budget impact of the reforms and investments in the electricity sector is a net gain of 1.1% of GDP in 2016 versus 2006; tax and non-tax revenues increase by 2.5% of GDP while Government expenditures on utility bills, ECM and the TALCO subsidy rise by 1.3% of GDP.

J. RESOURCEENVELOPE PROJECTIONS

187. This section brings together the projections in the previous four sections to derive a budget resource envelope for the next 10 years. The budget resource envelope is shown in Table 3.9 above and comprises all of the budget resources which are available to fund government expenditures, under the State Budget and the PIP. It is the sum of domestic revenues, grants, net external financing (disbursements of loans minus amortization) and net domestic borrowing. We also show the budget resource envelope for primary expenditures; i.e., excluding interest payments, and the projected additional requirements for government electricity bills and the ECM. The quasi fiscal deficit for the electricity sector is also included in Table 3.9.

188. From a level of just under 22% of GDP in 2006, the budget resource envelope expands rapidly in the medium term, to over 35% of GDP in 2007 and 2008, due to the large disbursements of external loans for infrastructure projects. However, as discussed in section H above, this level of borrowing is not sustainable and hence the budget resource envelope falls back sharply to around 26% of GDP during 2010- 2013, before climbing to around 30% of GDP in 2016. By 2016, the budget resource envelope will have risen by 8.6% of GDP compared to the level in 2006. The rise in domestic revenues contributes just under a third. Grants contribute an additional 1% of GDP to the budget resource envelope but the contribution of net domestic financing declines by 0.5% of GDP, as we have not made any projections for privatization receipts.

189. Interest payments increase by only 0.4 percentage points of GDP between 2006 and 2016, hence the primary budget resource envelope expands by just over 8% of GDP in this period. However, budget resources amounting to a further 1% of GDP will be required to fund higher government electricity payments and the ECM, and there will also still be a small quasi fiscal deficit of 0.6% of GDP remaining in the electricity sector in 2016, which may have to be covered from budget resources. Therefore the increase in budget resources for priority expenditures, after taking account of higher interest payments, government electricity bills, the ECM and the need to cover the remaining QFD in the electricity sector, is around 6.5% of GDP.

64 5N

$ 3 N

2N n 5N

N r 3 N

r 5N

2 3 N

VI W m 3 3N

x) 3 3N

3Ns K. THEEFFICIENCY OF GOVERNMENTEXPENDITURE

190. In principle, fiscal space can be created by improving the allocative and technical efficiency of government expenditures, thereby allowing more outputs in terms of public goods and services to be produced for a given amount of budgetary resources. However it is very difficult to quantify in a rigorous manner the magnitude of expenditure inefficiency and the potential scope for creating fiscal space through efficiency improvements.

191. Allocative efficiency of government expenditures in Tajikistan is likely to be low because budget allocations are still heavily influenced by those which prevailed during the Soviet era. These allocations were heavily biased towards capital intensive technologies, highly centralized service provision and very rigid expenditure structures. For example, the health service is overly specialized, segmented and hierarchical with a bias in favor of secondary and tertiary health care and against primary health care. The number of health facilities in the country is around 2,800, which is excessive, and as a result many of these facilities cannot be adequately staffed, supplied with necessary inputs such as medicines, or properly maintained. In the education sector, employee compensation consumes 83% of the recurrent expenditure in the local budgets, from which most schools are funded, leaving insufficient funds for teaching materials and essential classroom maintenance. Many of the schools are in a very poor state of repair and are often unusable in winter because heating systems have broken down. Expenditures on repair and maintenance of all government assets from the State budget amount to only 1.5% of GDP (World Bank, 2005D: 43),39 which given that the value of the public capital stock must be in excess of 100% of GDP, cannot be anywhere near sufficient. The weaknesses in the budget process for allocating expenditures, discussed in Chapter 4, are a constraint to improving the allocative efficiency of government expenditures.

192. It is also likely that technical efficiency is low in Tajikistan, given that many aspects of the country’s public expenditure management system are weak. For example, accountability for the spending of public funds is weak. In principle, spending units face a dual accountability, to both the local authority and the applicable line ministry, but in practice this simply means that there is no effective accountability for the spending of funds. Control of the payroll is impeded because there is no complete civil service register with data on all staff in all budget units. Internal audit is deficient as it focuses mainly on the legal compliance of expenditures, its operating principles fall short of international standards and there are no internal audit functions in spending agencies. There are weaknesses in public procurement arising from the lack of a clear legal framework for open competitive bidding and potential conflicts of interest in the roles of the Procurement Agency. (Radev et al, 2005)

193. Methodologies exist for quantifying the technical efficiency of government expenditures. Herrera and Pang (2005) estimate the technical efficiency of education and health expenditures in a sample of 140 countries, using the Free Disposable Hull (FDH) and the Data Envelopment Analysis (DEA) methodologies: these

39 This may underestimate total government expenditure on repair and maintenance because it is possible that some of the PIP expenditures, although classified as investment, are actually for repair and maintenance.

66 methodologies involve estimating an efficiency frontier, which represents the level of efficiency reached by the most efficient country in the sample, from a production function linking inputs (sectoral government expenditures) to sectoral outputs (indicators of health status and educational attainments). Herrera and Pang report results for Tajikistan which indicate that the technical efficiency of education spending is between 65% and 78%, and that of health expenditure is between 65% and 88% of the respective efficiency frontiers. However, there are serious weaknesses in data which impede any comparison of expenditure efficiencies in Tajikistan with those of other countries. In particular, the measures of outcomes in key areas such as health and education are unreliable. For example, while official government statistics show an infant mortality rate (IMR) of 27.9 deaths per 1000 live births, estimates of the IMR from living standards and demographic surveys range between 78 and 94.5 (World Bank, 2005C). In addition, the production frontier methodology takes no account of factors of production other than government expenditure which could have a significant impact on the output indicator. Government expenditures comprise only a small share of total expenditure in the health sector, approximately 16% in 2003; it is difficult to distinguish the impact of government health spending on health outcomes from that of health expenditures from other sources. Furthermore, the efficiency frontier methodologies only apply to technical inefficiency, not to allocative inefficiency. Given the serious deficiencies in the budget process which impede the alignment of budget allocations with strategic policy objectives, together with the dearth of basic skills for budget preparation, such as the prioritization and costing of expenditure proposals in spending agencies (Chapter 4), allocative inefficiencies are probably at least as, if not more, important as a source of expenditure inefficiency than technical inefficiencies.

194. As such, it is probably not realistic to make any quantification of the magnitude of expenditure inefficiencies in Tajikistan. At best we can acknowledge that major allocative and technical inefficiencies exist and that there is potential for improving both types of efficiency through reforms to the PEM systems, which over the long term should create fiscal space. Given that government final consumption and investment expenditures (i.e., excluding transfer payments and interest payments) amounted to 17% of GDP in 2006, even a relatively modest 10% improvement in expenditure efficiency over the long term would generate fiscal space amounting to 1.7% of GDP by 2016. L. CONCLUSIONS

195. The major long term challenges for fiscal policy in Tajikistan are to restructure government expenditures to improve the quality and efficiency of public services in a manner which can support economic growth and provide poverty reducing public services, while maintaining macroeconomic stability and reducing the large QFD in the electricity sector.

196. Fortunately, the findings of this study suggest that there is potentially large fiscal space for expanding government expenditures, but realizing this fiscal space will require major reforms by Government. Moreover, failure to implement effective reforms to public financial management, tax administration and the electricity sector would pose serious risks to fiscal sustainability and macroeconomic stability over the medium- to long-term.

67 197. Even before taking into account the scope for improving the efficiency of expenditures, fiscal space amounting to 6.6% of GDP could potentially be realized over the next 10 years, comprising an increase in the budget resource envelope of 8.6% of GDP, minus the additional budget resources required for higher interest payments (0.4% of GDP), higher government electricity payments (0.5% of GDP), higher payments to the energy compensation mechanism for low income electricity consumers (0.4% of GDP) and the funding of the remaining QFD in the electricity sector (0.6% of GDP). If the implementation of public expenditure management reforms could raise the efficiency of government expenditure on the provision of final consumption and investment expenditures by 10% over 10 years (which is a relatively modest target), additional fiscal space of about 1.7% of GDP could be created, raising the combined fiscal space from budget resource mobilization and expenditure efficiency improvements to more than 8%of GDP by 2016. Table 3.10 summarizes the projections of fiscal space in 2016, the final year for the projections in this study, compared to 2006.

68 2016, change over 2006 Budget Resource Envelope 8.6

Gains in Expenditure Efficiency 1.7 (assumes 10% efficiency gain)

Net fiscal space 8.4

198. Increased domestic revenues, potentially amounting to 5.6% of GDP make the largest contribution to fiscal space over the next 10 years; however most of these gains are contingent upon major institutional reforms to the tax and customs administration. Tax administration reforms, which include internal restructuring of the tax department, introduction of modern systems of tax administration focused on voluntary taxpayer compliance enforced by selective risk based audits, computerization and better training and incentives for tax and customs officials, could potentially generate additional tax revenue of 3.2% of GDP. Most of the additional gains to domestic revenue, on top of those generated by tax administration reforms, are also contingent on policy reforms. Increasing the marginal rates of personal income tax could contribute gains of up to 1.2% of GDP, while revenues amounting to 1.3% of GDP could be generated from the HPP investments, in the form of returns to government’s own investment in the HPPs and the tax or royalties levied on their earnings. The viability of the HPPs is also dependent upon reforms in the electricity sector, particularly increasing domestic tariffs to commercially viable levels.

199. The scope for expanding the budget resource envelope through additional borrowing is very limited. Disbursements during 2007-2009 of the external loans, which have already been contracted, will raise the NPV of PPG external debt above the GDP and export thresholds at which it is considered sustainable under the IDA/IMF Low Income Countries Debt Sustainability Framework. Consequently, while external debt disbursements are projected to rise sharply during 2007-09 because of the disbursement of Chinese loans, they must subsequently fall back to around 4-5% of GDP if the NPV of external debt is to be brought back in line with the sustainability thresholds.

200. The main constraint to higher domestic borrowing is not the sustainability of government’s domestic debt, which is very small, but the macroeconomic

69 consequences of domestic borrowing from very shallow and highly dollarized domestic financial markets. Broad money is the equivalent of only 8% of GDP, foreign currency deposits comprise more than two thirds of all bank deposits and the non bank financial system barely exists. Consequently there is very little scope for issuing government debt to the commercial banks without crowding out private sector borrowers from credit markets. There are no non-bank financial institutions which can purchase debt, and any significant amount of government borrowing from the Central Bank would jeopardize the latter’s efforts to control reserve money. The improvement in the economy achieved since the late 1990s has been made possible in part because of the restoration of macroeconomic stability, which in turn is attributable to the reduction of the fiscal deficit. The macroeconomic stability achieved over the last few years will be put in jeopardy if Government resorts to domestic borrowing or if external debt again becomes unsustainable because of excessive external borrowing.

201. Public expenditure reforms will also play a role in mobilizing more donor grants for the budget. The Government budget has benefited from only a small fraction of the donor grants disbursed, although this situation improved in 2006. Increased grant inflows, driven by both higher levels of donor aid to low income countries in general and a greater willingness on the part of donors to channel their aid through the Government budget, which is contingent upon reforms which can assure donors that their funds are efficiently and transparently spent, could potentially yield additional budget resources in the form of grants equivalent to 1% of GDP by 20 16.

202. The largest single threat facing the budget is the QFD of the electricity sector, estimated at 9.3% of GDP in 2006. The QFD is the difference between the revenue which the state owned electricity utility requires operating and properly maintaining the existing infrastructure for the generation and supply of electricity and the revenue which it actually collects in cash. The main cause of the QFD is suboptimal electricity tariffs, supplemented by technical and commercial losses and poor cash collection of electricity bills. Although the QFD is not currently a cash burden on the Government budget, it will become so if no measures are taken to drastically reduce it. The consequences of the Government not funding the QFD will be that essential maintenance and rehabilitation of the electricity transmission and distribution system will not be undertaken, and hence electricity supply will deteriorate. Fortunately there are measures, albeit politically painful ones, which can be taken to slash the QFD, principally by raising electricity tariffs to the level equivalent to the long run average incremental cost of electricity supply (which requires an almost fourfold increase in average tariff levels from those prevailing in 2006). Higher electricity tariffs will have a negative impact on some components of the budget, reducing tax revenue paid by TALCO, the aluminum smelter, and increasing the costs to the budget of government’s own electricity bills and transfer payments to compensate low income electricity consumers. But the aggregate fiscal impact of electricity sector reforms is overwhelmingly positive, generating net gains of almost 105 of GDP.

203. The fiscal space detailed in these projections, if realized, would allow government expenditure to be raised to around 30% of GDP over the long term, compared to 22% of GDP in 2006. While this would make the funding of priority expenditures much more feasible, it will not be sufficient to generate higher growth

70 and poverty reduction without substantial improvements in the quality of governance, and especially in PFM systems. The preliminary findings of the World Bank’s ECA Regional Fiscal Study indicate that, in countries with poor governance (which include Tajikistan which is ranked second worse among 27 Eastern Europe and Central Asian countries in terms of government effectiveness), potentially productive public expenditures (defined as education, health, housing and economic affairs, including transport) have little positive impact on economic growth, whereas unproductive public expenditures have a negative impact. Consequently, in countries with poor governance, large total public expenditures reduce growth, probably because they require high levels of taxation which distort incentives for work, investment and saving; hence growth is maximized by keeping the size of government small (Varoudakis, 2006).

204. The implications for Tajikistan are that major improvements in the quality of governance will be a prerequisite if increased government expenditures in the priority sectors are to have a positive impact on economic growth. Improved governance must precede an expansion of the budget. Mobilizing more tax revenues to fund an expansion of the budget in the context of poor economic governance, including an untransparent tax administration system, may depress economic growth through distortions to resource allocation, while the benefits of an increase in government expenditure will be limited if governance is not improved. Moreover, many of the measures identified in this chapter to expand the budget resource envelope, such as tax administration reforms, will take many years to generate more resources, whereas there is an urgent need to ensure that the existing budgetary resources are used more efficiently to deliver better quality public services.

205. Consequently the message that emerges clearly from this Chapter, as well as other Chapters,, is that institutional reforms to improve the quality of public expenditure management, government fiduciary systems and tax administration, and reforms in the electricity sector, as well as reforms to the business climate for private investors, are essential if Tajikistan is to both generate fiscal space over the long term and to benefit, in terms of economic growth and poverty reduction, from the higher government expenditures made possible by the fiscal space.

71 4. HOW CAN THE MTEF BE INTRODUCED IN TAJIKISTAN?40

A. INTRODUCTION

206. The MTEF is a tool for sector-based medium-term budget planning which has been introduced in a wide range of developing and transitional countries,‘ often with the support of donor agencies who view reforms to public expenditure management (PEM) as critical to the achievement of the MDGs. The major objective of the MTEF is to facilitate policy-based budgeting (Le., linking budget allocations to strategic policy objectives) within a sound fiscal framework. The MTEF is now regarded as “the sine qua non of good PEM” (Le Houerou and Taliercio, 2002).

207. In May 2006, the Ministry of Finance (MOF) organized a Round Table, attended by government officials and representatives of donor agencies, to present a proposal for introducing an MTEF in Tajikistan on a pilot basis, beginning with the 2007 budget process. The MOF subsequently circulated a Government Resolution to obtain political authority for the introduction of the MTEF and the radical changes to the budget process that this will entail. This is not the first effort to implement an MTEF in Tajikistan. The MOF had begun work on an MTEF (termed the Medium Term Budget Framework [MTBF]) in 2000 and actually presented the MTBF at the 2002 Consultative Group Meeting. However, this previous effort, while providing some useful information on projected medium term expenditures, broken down by functional classification, had no more than a marginal impact on budget planning.

208. The problems experienced by Tajikistan in introducing an MTEF (and in implementing PEM reforms in general) are not unique to this country especially in comparison with the other CIS which share a similar history and institutional structures. A recent World Bank study (Andrews and Moon, 2003) explored the implications of weak PEM capacity for developing the PRSP approach (for which policy based budgeting is integral) in the PRSP countries of Eastern Europe and Central Asia (ECA). One of the main findings of this study is that PEM capacities are even weaker in this group of countries than in those low income countries which qualify for HIPC debt relief; that capacity for budget formulation is the weakest aspect of PEM, and that efforts to introduce PRSPs in ECA countries have run far ahead of the PEM capacity required to support effective policy based budgeting. Moreover, of the 11 ECA countries evaluated in this study, Tajikistan was judged to have the weakest PEM capacity.

40 This chapter is based on the non lending technical assistance for launching the MTEF in Tajikistan which began in January 2006 and is managed by the World Bank and DFID.DffD. The strategy for implementing the MTEF was presented at a Round Table organized by the Ministry of Finance in Dushanbe in May 2006 and subsequently endorsed through a Government Resolution in October 2006.

72 209. The purpose of this chapter is to understand why the MTEF proved so difficult to introduce in Tajikistan, examine the obstacles to its introduction and to map out a set of feasible measures which will allow for a gradual introduction of the MTEF over the next few years. While there is now strong political will to implement an MTEF in Tajikistan, there are major constraints especially in relation to the existing institutional relationships of the budget process and the very weak (almost nonexistent in some agencies) capacities for implementing modern budget process in the public service. Hence it is necessary to be realistic about the pace at which an MTEF can be introduced and to be cautious in implementing reforms, especially in relation to aligning the introduction of new budget processes to the building of technical capacities in the relevant agencies, as well as promoting a broader understanding of the rationale for the MTEF within the public service and in civil society. A dogmatic approach to MTEF implementation should be avoided; there is no blue print for the MTEF which can be applied uniformly to every country. Instead the introduction of MTEF reforms must be done in a pragmatic matter, with each step in the reform process carefully designed to ensure that it is compatible with the technical and institutional capacities which are necessary for its effective implementation.

210. The chapter is organized as follows. Section 2 explains the motivation for introducing the MTEF. Section 3 outlines the extent of progress made so far in developing the MOTIF. Section 4 discusses the critical constraints and obstacles to introducing an MTEF in Tajikistan. Section 5 outlines a strategy for the introduction of an MTEF, in a phased manner, and Section 6 sets out the measures which will be required to implement this strategy.

B. MOTIVATIONFOR INTRODUCING THE MTEF

21 1. The primary motivation for introducing the MTEF is to improve budget resource allocation and to link this with the strategic policy objectives of the Government, including those set out in national development strategies such as the PRSP. Rational budget allocations, in which expenditures are prioritised to best meet the budget’s objectives is not possible under the existing budget system for two reasons. The first is the fragmentation of the budget among more than a hundred KBOs and the separate determination of current and capital budgets, which is discussed in Section 4, and the second is the norm-based system of allocating budget resources, whereby expenditures are determined by formulas linked to structures in place (such as schools or hospitals).

212. The existing structure of facilities and staffing is highly suboptimal and inefficient: it was inherited from the Soviet era when budget resources were much larger, and even if it was optimal then (which is debatable) it is certainly no longer so given the steep decline in available budget resources. For example, there are approximately 2,800 health facilities in the country, including hospitals in villages, but many of these facilities are little more than crumbling shells: there are no patients in the beds, no drugs and very few staff in attendance. There are also large numbers of authorized staff posts, reflecting the norms and the facilities in place, but only a fraction are filled because the salaries are far below even subsistence level. Rationalising health facilities and staffing levels is essential to improve allocative efficiency but the existing system provides neither the institutional mechanisms nor the incentives for doing this. The most appropriate institutions for allocating budget

73 resources in line with strategic policy objectives are the line ministries, which have responsibility for formulating sector policies, but under the existing budget system the line ministries have only limited influence over sectoral budget allocations because they directly control only a small share of these allocations; hence policy formulation is divorced from budget allocations. The most important change to the budget process, with the introduction of the MTEF. will bring about is the centralization of responsibility for formulating sector budgets in the relevant line ministry.

2 13. Although the main motivation for reforming the budget allocation system is to improve the allocative efficiency of the budget, it will also provide additional benefits, not least because by linking budget allocations more clearly to policy objectives in a transparent manner, it will encourage donors to channel more of their aid directly to the Government budget in the form of budget support. This is essential if the severe constraints on funding for essential current expenditures, such as operational and maintenance expenditures in the health and education sectors are to be alleviated. c. WHAT HAS BEEN ACHIEVED SO FAR?

214. The MTBF prepared by the MOF consists of a set of three year fiscal projections covering the main revenue heads, expenditures and financing. The 2006 MTBF, which includes the 2006 budget plus the projections for 2007 and 2008, is shown in Appendix Table 1. The Government budget consists of two separate components: the state budget and the PIP. The MTBF breaks down the state budget expenditures into sectors, which correspond to broad functional classifications: e.g., law enforcement structures, education and health. There is no breakdown of the PIP into sectors. The state budget is composed of three main institutional components: the Republican budget (comprising national institutions such as line ministries); the local government budgets and special funds (such as pensions). This breakdown is not shown in the MTBF.

215. The main value of the existing MTBF is that it provides a medium-term aggregate fiscal framework which is both internally consistent (there are no financing gaps) and consistent with macroeconomic objectives, in particular Government’s projected domestic borrowing requirement is consistent with the Bank of Tajikistan’s planned monetary program. Up to 2006, the MTBF closely reflected the three year PRGF programme which the Government negotiated with the IMF (the PRGF ended in 2006 and no successor programme has been implemented) , The fiscal framework of the MTBF provides a hard budget constraint for aggregate expenditure and, therefore, is a useful tool for macroeconomic management. However, the existing MTBF does not play any meaningful role as a budget planning tool at a lower level of expenditure than that of aggregate expenditure.

216. The projections of sector expenditures contained in the MTBF have the appearance of sector expenditure ceilings, but they do not play this role in the budget process. This is because no budget institution actually receives the ceiling for each sector and has the responsibility for drawing up a sector budget within that ceiling. As discussed in the following section, responsibility for budgeting is fragmented among a huge number of KBOs, both within, and spanning, different sectors. In fact, because of the fragmentation of budget institutions, there is no effective sector

74 planning linked to budget allocations in Tajikistan. Instead of acting as a hard budget constraint for sector expenditure planning, the sector expenditure projections are simply the aggregation, across all KBOs, of their projected functional expenditures related to each sector: e.g., the projected education sector expenditure is the sum of all expenditures on education by each of the 40 plus KBOs which undertake education expenditures. At best these projections provide some limited data on projected expenditures by functional classification, but even that is somewhat tenuous, because in the absence of hard budget constraints at the sector level, there is no mechanism within the budget process to ensure that, once future budget estimates are prepared, they will not deviate from the sectoral breakdown currently shown in the MTBF.

217. It is useful to view an MTEF as consisting of budget planning at three levels, differentiated by level of aggregation.

218. Level 1 is the determination of a budget resource envelope derived from a medium term fiscal framework, consistent with macroeconomic objectives. The budget resource envelope defines how much Government can allocate to all of its expenditures in aggregate.

219. Level 2 is the division of the budget resource envelope into a set of sector ceilings. A sector ceiling should represent a hard budget constraint under which a sector should plan and prioritize its expenditures: Le., the sector ceiling is not simply a forecast or a projection of a sector’s expenditures; it is a crucial planning tool of the MTEF. T he allocation of budget resources into sector ceilings should be determined by inter-sectoral budget priorities which are set out in national planning documents, such as the PRSP and NDS.

220. Level 3 is the allocation of the expenditure ceiling for each sector into a sector wide budget. This allocation should be driven by the intra-sectoral priorities set out in the medium-term sector expenditure plans of each sector.

221. Viewed from this perspective, the MTBF in Tajikistan fulfills the requirements of Level 1 of MTEF budget planning, but not Levels 2 and 3. This is also reflected in the respective roles which different budget institutions have played in preparing the MTBF. T he line ministries make almost no input into preparation of the MTBF, which instead is prepared almost entirely by the MOF.

222. The priority for PEM reform should be to try and gradually extend the MTEF to take on board levels 1 and 2, consisting of the inter and intra sectoral allocations of expenditures, without jeopardizing the successes achieved in the area of macroeconomic management brought about largely through a sound fiscal framework. The following section examines the constraints and obstacles to extending the MTEF to encompass sector expenditure planning.

D. CONSTRAINTS TO INTRODUCING SECTOR BASEDPLANNING

223. Sector-based budget planning faces three major constraints in Tajikistan: the fragmentation of the budget, the separation of current and capital budgets and the lack of capacity for budget planning within budget institutions.

75 Fragmentation of the Budget

224. There are hundreds of different budget organizations (Le,, organizations which receive and spend funds from the State budget). A critical distinction is between KBOs and other budget organizations (non-KBOs). KBOs deal directly with the MOF in preparing budget submissions and negotiating with the MOF during the budget process. There are 118 KBOs comprising line ministries, The Presidential Administration, the Majlisi Oli, the courts, sundry commissions, committees, academies and universities and local governments (three Oblasts, 13 Rayons of Republican Subordination (RSS) and Dushanbe. Non-KBOs do not deal directly with the MOF; instead they are subordinated to a KBO. For example, the rayons, except for the RSS, are subordinated to the Oblast in which they are situated. There are several hundred non-KBOs. The fragmentation of the State budget is partly a legacy of the Soviet era, when Republican budgets in the Soviet Union were typically fragmented among many budget organizations without there being dominant line ministries, and partly the consequence of the decentralization implemented after the end of civil war in the 1990s.

225. The problem for sector-based budget planning is not the large number of non- KBOs - in most countries there is a plethora of budget organizations which actually implement expenditures - but the large number of KBOs. With 118 KBOs, many of which, including all of the 17 local governments, have responsibilities which span several sectors, the number of KBOs in most sectors runs into double digits. In the education sector, there are around 40 KBOs. In most of the sectors, there is a line ministry which has responsibility for formulating policies, but that line ministry only controls a fraction of the expenditures within its sector. For example, the Ministry of Education directly controls only 9% of the total education budget: the remaining 91% is split among the 40 plus KBOs in the sector, with the 17 local governments accounting for almost all of this. The Ministry of Health controls only about 10% of the health sector budget. As each KBO submits its budget requests directly to the MOF and then negotiates with the MOF to finalize its budget estimate, the line ministries are effectively bypassed in the determination of expenditure allocations. The line ministries can influence the content of the budget instructions which are sent by the MOF to KBOs at the start of the budget process, but they have very little influence over the actual allocation of expenditures except for their own expenditures (which are largely their administrative expenditures and the expenditures of those non KBOs which are subordinated to them).

226. From the standpoint of sector budget planning, the fragmentation of the budget among over 100 KBOs has several adverse consequences. First, there is no single budget institution in each sector to which the sector ceiling can be applied and which can take responsibility for prioritizing sector-wide expenditures according to sector policies. Secondly, the connection between sector policy formulation and budget planning is very tenuous, because there is no mechanism through which policies formulated by the line ministries can be translated directly into budget allocations. Thirdly, in terms of the volume of expenditures, the most important KBOs are the local governments, which control about 30% of all of the expenditures in the State budget, excluding the special funds. The focus of the local authorities is not meeting nationally determined sectoral objectives; it is to meet the specific needs of their communities. There are legitimate reasons for giving local authorities control over

76 budgets, but for them to control the major part of the budgets of key sectors such as health and education is not compatible with a budget based on nationally formulated sector budget policies. It would still be possible to develop an MTEF in a decentralized system, in the sense that medium-term expenditure plans could be drawn up, but it would be an MTEF devoid of sector-based expenditure allocations and sector-based expenditure plans and, therefore, an MTEF which would be very difficult to align with strategic policy objectives which are also based around sectoral objectives, such as a typical PRSP.

Separation of Current and Capital Budgets

227. In addition to the fragmentation of the budget among the numerous KBOs, the budget is fragmented between the current budget and two separate capital budgets; the Central State Investment Program (CSIP), which is part of the State budget, and the PIP which is fully funded by external donors. The CSIP and the PIP fall under the aegis of the Ministry of Economy and Trade (MOET). The selection of projects for the CSIP and PIP involves a process of negotiation between the KE30 promoting the project, usually in collaboration with the external funding agency if the project is in the PIP, and the MOET. Although the CSIP, unlike the PIP, is formally integrated into the State budget, what matters for budget planning is that the process of making allocative decisions in relation to the CSIP is distinct from that relating to the current expenditures in the State budget.

228. It is not clear how the selection of projects relates to expenditure ceilings, if it does at all, and there is no relationship between ceilings given to KBOs for current expenditures and capital expenditures. Therefore, sectors or KBOs cannot trade off capital expenditures for current expenditures, or vice versa, and so cannot formulate a sector-wide budget which is fully comprehensive and covers current and capital expenditures. In fact it would appear that KBOs have every incentive to try and maximize their share of the CSIP and PIP, because this has no negative impact on the allocation of budget resources that they will receive for current expenditures, but may boost their allocation in the future because these allocations are largely driven by spending norms linked to structures (therefore the more structures that are built now, the larger will be the current budget allocation in the future).

229. The separation of the PIP from the current budget is, of course, typical in many developing countries mainly because PIP projects are driven by the availability of funds from donors, and hence decisions to undertake these projects are not easy to integrate within the normal budget process. However, it is more unusual to also treat separately that part of the capital budget which is funded from the Government’s own budget resources.

230. The consequence of the fragmentation of the budget among over 100 KBOs and the separation of the current and capital budgets in the budget process is that meaningful sector planning, whereby sector wide expenditures are prioritised and allocated according to a coherent medium term sector expenditure plan, is rendered impossible. The sector expenditures shown in the MBTF are not the product of a coherent sector planning process, they are simply the amalgamation of numerous atomised allocative decisions taken largely in isolation from each other. Furthermore, the “sector ceilings” shown in the MBTF can play no meaningful role in the budget

77 planning process; their only value is to show, once the budget process is complete, how much resources have been allocated to each sector.

Capacities for Budget Planning

231. Capacities for budget planning throughout government are very weak. Even basic concepts of strategic budget formulation and prioritization are not well understood, partly because the budget system which was inherited from the Soviet era, and which is still largely intact, was very different from that of a modern western budget system, in which a set of clearly defined government activities are costed and prioritized. The Soviet budget system was essentially a mechanical exercise in which the budget was subordinated to a broader five year economic plan. The methodology for determining budget allocations involved the use of input norms, whereby allocations reflected structures in place, rather than any systematic attempt to match resources to policy objectives. Normative budgeting is still the dominant methodology for determining budget allocations in Tajikistan. In addition, the Soviet budget system did not entail a central role for strong line ministries in budget management. Therefore these ministries did not develop a body of expertise in managing the budget in their particular area. Instead, as was noted above, the budget system was highly fragmented with numerous small budget organizations having responsibility for expenditures, which prevents economies of scale in activities such as budget planning being realised.

232. Line ministries have very few staff, if any, with expertise in formulating budgets, such as expertise on costing activities, undertaking quantitative project appraisals, or prioritising expenditures. Within line ministries, only a small share of the staff is devoted to budget formulation, which is not surprising given that the line ministries themselves have such a limited role in the budget process. For example, in the Ministry of Education, the planning and economy department which deals with budget formulation (in the 44 budget organizations directly subordinated to the MOE) has only seven staff members. Tajikistan’s systemic weaknesses in budget formulation are shared with most other CIS countries (Andrews and Moon, 2003).

Summary

233. The analysis above suggests that the lack of progress achieved so far in implementing the MTEF in Tajikistan is not simply the result of a lack of “political will”. Rather it reflects major constraints arising from the structure of the budget process and severe capacity weaknesses. The budget system inherited from the Soviet era could hardly be less conducive to the development of policy based budgeting focussed on sectoral budget allocations and sector planning, which is the essence of an MTEF. Yet this budget system has changed in few fundamental respects in the 15 years since independence, and the major change which has occurred, the introduction of “ad hoc” decentralization processes and the shifting of most social expenditures into the local budget, is a further barrier to budget formulation driven by sector budget policies formulated at the national level.

78 E. WHAT CHANGES ARE REQUIRED TO SUPPORT THE INTRODUCTION OF THE MTEF?

234. To implement fully the MTEF in Tajikistan, it will be necessary to effect radical restructuring of the relationships between different budget institutions in the budget process, to restructure line ministries and build up their capacity for budget formulation, to integrate the state capital investment program and, at a later date, the PIP, with the current budget, to lengthen the budget timetable and to centralize tax payments. Radical institutional restructuring of budget processes would be difficult even in public administrations characterized with strong administrative and technical capacities and experience of implementing institutional reforms, neither of which characterizes Tajikistan. Consequently, it will be necessary to proceed with the reforms cautiously, with the introduction of new budget processes being matched to the building of technical capacities for budgeting in the MOF and line ministries, supported by TA from donors.

Chart 1

Existing Budget Institutional Arrangements in Each Sector

KBOs Ministry

uDepartment 1 OtherKBO I

Budget instructions from MOF to KBOs Expenditure requests from KBOs to MOF Bilateral negotiations between MOF and KBOs Final budget allocations from MOF to KBOs

Note: Some KBOs budget for expenditures in more than one sector Restructuring Budget Relations

235. The most important changes required to implement the MTEF relates to the organization of the budget into sectors. E ach sector, includes many KBOs, but one

79 budget organization in each sector must take the lead in coordinating the KBOs in the sector and formulating budget policies if a comprehensive sector-wide budget is to be prepared. The only feasible KBOs which can assume this task are the line ministries. Although some sectors contain one or more line ministries, only one of these has responsibility for policy formulation in that particular sector (e.g., the Ministry of Education has responsibility for education policy, although there are other line ministries which operate educational establishments). We will refer to a line ministry with policy responsibility as a lead line ministry (LLM). The critical reform needed to enable effective sector budgeting to be carried out is to make the LLMs responsible for formulating the entire budget in their respective sectors, so that budget allocations can be aligned with sector policies.

236. The required restructuring will entail all KBOs in a given sector channelling their budget submissions through the relevant LLM for that sector, with the latter having the authority to revise the budget submissions of KBOs to ensure that they accord with the sector’s priorities and can be accommodated within the overall sector expenditure ceiling. This will mean that most KBOs will no longer make their budget submissions directly to the MOF. Instead the MOF will receive a consolidated budget submission covering the entire sector from the LLM and will negotiate with that LLM to finalize the sector’s budget allocation. In effect KBOs which are not LLMs will be subordinated to the LLMs for the purpose of budget planning and will become more like non-key budget organisations. In the Republican budget, this switch will be straightforward, at least in theory, because most KBOs other than line ministries are relatively small and have a single area of responsibility: hence they will be subordinated to one LLM for all of their expenditures. The local governments, however, have expenditure responsibilities which span several sectors, and hence they will be subordinated to several different LLMs.

Chart 2

MTEF Budget Institutional Relationships in Each Sector

Other KBOs

Other .’,’

4 b ...... b Budget instructions, including sector ceilings Allocation of sector expenditures from MOF to LLMs Negotiations between LLM and Submission of sector budget estimates from other KBOs LLMs to MOF Bilateral negotiations between MOF and L%s to finalise budget allocations 237. Chart 1 show the budget relationships under the existing budget system: each individual KBO deals directly with the MOF, receiving budget instructions from the MOF, making a budget submission to the MOF and then negotiating with the MOF. Chart 2 shows the budget relationships required for a sector-based MTEF. In each sector, the LLMs intermediate between the MOF and all other KBOs in the sector.

238. It will not be possible to restructure the entire budget around sectors led by a LLM, for two reasons. First it is unlikely to be politically feasible to subordinate the budgets of politically powerful KBOs such as the Apparatus of the President or the Majlisi Oli to a LLM. These KBOs will continue to deal directly with the MOF outside of any sectoral structures which are put in place to cover public administration. Secondly, while it may be possible to create a sector covering domestic law enforcement institutions, it is likely that the defense will be treated separately, for security reasons. This is not a fatal omission for the MTEF however. In most countries where the MTEF has been implemented, there are budget organizations which still operate outside of sectors, for political or security reasons. What is important is to organize as large a share of the budget as possible into sectors, and especially those components of the budget which are most closely linked to the policy objectives in the PRSP, such as the social sectors.

239. Table 4.1 shows the five best candidates for creating sectors, together with the LLM for each sector and some of the other KBOs which will become subordinate to the LLM. As can be seen from the table below, these five sectors comprise almost half of all expenditures in the State budget (and slightly more than half of primary expenditures). The main omissions are public administration bodies and foreign economic relations, and law enforcement structures, which together account for 35% of the State budget. As mentioned above, it would be desirable to create a sector for domestic security, encompassing the Ministry of Internal Affairs, police, prisons and the law courts.

Sector Lead Line Ministry Subordinated KBOs Share of State budget (%) Education Ministry of Education Committee on Sport and 21.8 Physical Education, univ., Local governments Health Ministry of Health Local governments 6.9 Social Protection Ministry ofLabour and Social Protection Fund 12.4 Social Welfare Agriculture Ministry ofAgriculture Ministry of Environment 3.2 Protection and Forest Economy, Ministry of Land Improvement and Water Economy Academy of Agricultural Sciences Forest Economic Production Association

Transport I Ministry ofTransport I Local governments 4.9

81 240. In addition to the KBOs, there are hundreds of budget organizations which are non-key budget organizations and which are subordinated to KBOs in the existing budget system. In most cases, these relationships will not change when the budget system is restructured around sectors.

241. The required restructuring of the budget process to move towards a more centralized structure for preparing a sector-based budget will clearly not be easy to implement and will take several years to complete; both to reorganize the budget process and build capacities for budgeting in line ministries. Hence the strategy recommended in this Chapter is to move gradually, beginning with one pilot sector, and then bring other sectors into the MTEF at a later date. The pilot sector is the education sector, hence in the short term (1-2 years), the aim should be to create a sector budget process in this sector, with the Ministry of Education as LLM. In the medium term (3-5 years), another two sectors can be brought into the MTEF. The lessons learned from introducing the MTEF in the pilot sector should prove valuable when the MTEF is introduced in other sectors.

242. With a sector-based MTEF, the role of the LLM will be central to budget planning in each sector. The LLM will receive a single expenditure ceiling for its sector from the MOF at the start of the budget process. The LLM will take responsibility for allocating that ceiling among all the KBOs in its sector and for ensuring that each KBO’s expenditure is allocated according to the sector’s priorities. It will send budget instructions to all the KBOs in its sector, receive budget requests, and hold negotiations with them as part of the budget process. The sector itself, led by the LLM, should have the discretion to allocate expenditures among the main economic classifications of the budget provided that the overall sector ceiling is not breached and that statutory requirements, unavoidable expenditures such as utility payments and government directives are fully taken into account. Having prepared a sector-wide budget estimate in consultation with the KBOs in its sector, the LLM should then submit this estimate to the MOF, the aggregate of which should not exceed the sector ceiling. There will then be negotiations between the LLM and the MOF to finalize the sector’s budget allocation. The MOF itself should not revise the sector’s estimates provided that they comply with some basic budget requirements, such as fully providing for unavoidable expenditures. In effect, the role of the MOF in the sector’s budget is to ensure that the sector ceiling is not breached and that budget implementation will not be jeopardized because insufficient budget allocations have been made for unavoidable expenditures, such as statutory expenditures and utilities. Responsibility for budgeting within the sector should be delegated to the sector itself, under the direction of the LLM.

82 The Role of Local Governments in the Budget Process

243. The local authorities will continue to implement expenditures in various sectors but the introduction of an MTEF will entail major changes in their role in the budget process. In the current system, the local authorities, which are KBOs, are responsible for receiving budget submissions from subordinated budget organizations and then making a budget submission to the MOF. To be consistent with a sector- based MTEF, the local authorities, which are KBOs, should make their budget submissions to the relevant LLMs and consequently they will no longer deal directly with the MOF in the budget process, at least for those areas of expenditure which are organized into sectors.

244. Centralizing budget formulation in line ministries will clearly impinge on the authority of local governments, although the discretion of local governments is heavily circumscribed under the existing budget system. It will be desirable, especially from the standpoint of promoting greater local accountability and autonomy, for Government to allocate a portion of the local budget to local authorities to be spent entirely at the discretion of the local authorities themselves. For example, these expenditures could cover local authority administration costs and specific projects important to local needs. This should be accompanied by measures to build capacity in local governments for budget planning, so that the budget funds which are available to be spent at their discretion are utilised in an optimal manner to meet local needs. In the medium-term, each Oblast and RSS should be encouraged to prepare a Local Government Budget Framework Paper, setting out the local priorities for public expenditure and how local government budget allocations will be used to meet these priorities.

Integration of Capital and Current Expenditures in Sector Budgeting

245. With the sectors themselves having responsibility for all of their expenditures, including the allocation between the different economic classifications of expenditures, there will be no need for a separate Central State Investment Program. All capital investment projects will instead be fully incorporated into the sectoral budgets, and planned entirely at the sector level, and the decision to undertake a proposed capital investment project will be made by the relevant sector, taking into account whether the project is in accord with its expenditure priorities, whether it is affordable under the medium-term sector ceiling and the implications of the project with respect to current budget costs in the future, once the investment is completed. Each sector, led by its LLM, will have the discretion to determine how much is allocated to current expenditures and how much to capital expenditures, within the overall sector expenditure ceiling. The aggregate amount of capital expenditure in the budget will, therefore, not be determined centrally: it will simply be the outcome of the sum of decisions taken at the sector level. Therefore, the role of the MOET in planning capital expenditures in the State budget will become largely redundant.

246. Eventually the PIP should also be incorporated into sector ceilings, although this is more complicated because of the separate funding arrangements for the PIP and also because the level of external borrowing should be a budget constraint.

83 Strengthening Capacities in Line Ministries

247. The greatly enhanced role of line ministries in the budget process will not be possible without a major strengthening of their capacities for budget planning; including capacities for costing expenditure proposals, linking activities to policy objectives, cost benefit appraisal of projects, and the long term forecasting of demand for services. Line ministries will need to be restructured to allocate many more staff to budget planning functions. They will also require substantial technical assistance in budget planning, especially in critical sectors such as the social sectors, and this should be a priority for the TA programs of donors.

248. Sector Working Groups (SWGs) should be set up in each of the sectors to act as a forum for all of the major KBOs to discuss both sector policy and budget allocations. The SWGs should be chaired by a senior official of the LLM of that sector. Their principle tasks will be to: (i)prepare a medium term expenditure plan, setting out the policy objectives of the sector, the prioritization of expenditures to best meet those objectives, and the costing of expenditures; (ii) provide a forum for discussions within the sector on the allocation of that sector’s budget resources; and (iii)monitor the implementation of the sector’s expenditures and progress towards meeting the sector’s policy objectives.

The Process for Determining Sector Expenditure Ceilings

249. The allocation of the budget resource envelope into sector expenditure ceilings is not simply a technical exercise; it has a political dimension because the allocation of budget resources involves political choices between competing strategic demands. Therefore to ensure that the MTEF can command widespread political support, there must be a political mechanism for determining inter-sectoral budget allocations. In effect, this means that the final decision over these allocations should be made by a body comprising key political actors, such as Ministers, even though such a body will require technical advice from the MOF to help it reach decisions. The National Budget Commission (NBC) could play this role. Whichever body is given this role will need to draw up criteria for guiding inter-sectoral budget allocations.

The Budget Timetable

250. The institutional arrangements required to implement an MTEF also have important implications for the timetable of the budget process, because some steps in the budget process must logically precede others if the MTEF is to be implemented in a meaningful way. In particular, resource envelopes must be determined before resources are allocated; this applies both at the level of aggregate budget and at the sector level. In the existing budget process, the determination of the resource envelope and the allocation of expenditures are conducted simultaneously, because the budget submissions from KBOs include both expenditure requests and revenue projections.

25 1. There are five distinct steps in the budget process which must be followed in sequence if the MTEF is to be implemented properly.

84 Step 1: The MOF determines the budget resource envelope.

Step 2: The budget resource envelope is allocated into sector expenditure ceilings and the sectors are informed of their respective ceilings.

Step 3: The sectors, led by the LLM, determine detailed budget allocations within their ceilings and then, through the LLM, submit budget requests to the MOF.

Step 4: MOF holds bilateral negotiations with the LLM in each sector to finalize sectoral budgets.

Step 5: MOF submits the budget to Cabinet for approval, followed by submission to the Majlisi Namoyanagon.

252. The current budget process starts towards the end of April or the beginning of May. Introducing the MTEF will require a longer budget process, hence an earlier starting date, in order to allow sufficient time for stages 2 and 3 in the above sequence; and especially stage 3 which is likely to be time consuming.

Distribution of Taxes

253. Under the existing budget system, the collection of taxes is decentralized to offices of the Ministry of State Revenues and Duties in the local government areas. During the budget process, revenue projections are made for each local government area. The expenditures under the local budget are then funded through a combination of the retention of a predetermined percentage of the taxes (such as VAT and excise tax) collected in the local area and transfers from the central treasury. Under this system, the local governments bear the risks and benefits of any deviation in taxes collected at the local level relative to the budgeted levels. This system provides perverse incentives for local governments to under budget for revenues; because this allows them to retain more of the revenue which is actually collected and to spend this without proper accountability (e.g., it is spent on items which have never been appropriated by the National Parliament).

254. Because an MTEF requires that most expenditure allocations should be determined centrally, there is no reason for taxes collected by tax offices in the local authorities to be retained by the local authorities. Instead all taxes should be remitted to the national Treasury which should then distribute funds to the local authorities, in line with the budget estimates and the dictates of cash management. The exception would be those taxes specifically earmarked for spending at the discretion of local authorities (such as land taxes) which would be retained in full by the local authorities. Centralizing tax collections would also mean that the risk of revenues under performing would be transferred to the center, because expenditures on the local budgets would no longer be tied directly to taxes'collected in the same local area. Consequently there would be no need for separate revenue projections for each local government area. Instead tax administration could be carried out in a more centralized manner, with many of the local offices, which collect a negligible share of total tax revenue, closed down, which would enhance efficiency in tax administration.

85 This should be implemented over the medium-term, in line with the reforms to strengthen tax administration.

F. HOW SHOULD THE INTRODUCTION OF THE MTEF BE SEQUENCED OVER TIME?

Basic Principles

255. Experience shows that there is a need to avoid being too dogmatic about MTEFs; in practice they can take a variety of forms aimed at strengthening the link between government policies and budgetary allocations over time. At the same time, building such a link can take significant time and the steps should be built organically. Stakeholders should beware of overly ambitious expectations. Often, the beginnings of a framework can be put in place with the substance for the framework being improved over time.

256. In terms of sequencing of the development of the MTEF process itself, experience in the region4’ indicates:

0 Focussing initially on the top-down (macro-fiscal framework and initial sector ceilings) rather than on the bottom-up process (the costing and content of detailed sector expenditure strategies) has tended to work most effectively.

0 In the design of the strategic budgeting process, simplicity in programmes works most effectively. It is important that information requested from sector ministries is not overly detailed, time-consuming or unlikely to be credible. In particular, three-year allocations for detailed activities are unhelpful when the first priority should be to get broad resource allocations more in line with sector strategies.

0 Once there is a basic strategic budgeting framework in place, Government should work to ensure that existing elements of the strategic budgeting process are made stronger before introducing new elements (such as performance indicators).

0 Finally, Governments should program the link with wider budget systems (particularly, Treasury systems) explicitly into the reform programme.

257. The essential prerequisite for meaningful sector work is that the Ministry of Finance has established its central budget disciplining role and can control aggregate fiscal di~cipline.~~Moreover detailed sector analyses are wasteful of staff time if basic expenditure control does not exist and if reliable expenditure information is not available.

41 See Appendix 3 of this chapter. 42 This is not a hard-and-fast rule as sector expenditure analyses can help sector ministries to prioritize their intra-sectoral expenditures and thereby assist with sector reforms. In general, however, if the Ministry of Finance has not established fiscal discipline in relation to sector ministries then there is the risk of detailed and well-planned expenditure plans being undermined.at the Ministry of Finance.

86 Phases of MTEF Development

258. Lessons from experiences of introducing MTEFs have highlighted a number of phases of relative development of MTEFs (see Table 4.2). The process of working through the phases is not necessarily linear, and movement across phases should be expected to take considerable time and sufficient institutional capacity development.

259. During the preparatory phase for introducing an MTEF, the basics of the legal and institutional framework (e.g. ensuring that a modern Budget Systems Law43 is in place, establishment of a Government MTEF Steering responsibility for coordinating the MTEF given to a department in the Ministry of Finance such as the Budget Department, etc.) have been established or are underway. In some countries, the legal and institutional framework is the foundation for the beginning of a two- stage strategic budgeting process. What is important in the legislation is reference to a two-stage process (rather than necessarily specifically introducing an MTEF per se) whereby there is a stage prior to the start of the annual budget process where Government reviews the fiscal framework. Whilst not having the legal and institutional framework in place does not necessarily prevent countries from beginning such a process, it appears to be an important legitimising first step.

260. In the primary MTEF process, once the basic legal and institutional framework has been established, countries starting implementation of a two-stage strategic budgeting process begin with the development of a basic multi-year macro/fiscal framework, largely based around an externally-developed, e.g. financial programming, framework. At this stage, there is hited policy analysis of the overall resource framework, and there are no, or only limited, sector expenditure strategies in place, such as an initial sector review, setting out the sector’s policy objectives and classifying sectoral activities into broad programme areas to meet the objectives. Sector ceilings are based on overall policy priorities and limited analyses of the main economic categories of expenditures.

261, In the next phase of development, an intermediate MTEF process, countries undertake more advanced analyses for the macro/fiscal framework and sector expenditure strategies. The setting of overall expenditure ceilings is based on some cross-sectoral analyses (overall government priorities e.g., poverty reduction, investment/recurrent ratio etc.). On the sector side, sector ministries (perhaps initially at the pilot stage) have begun to analyse their existing activities by broad programme and activity in order to prioritise existing activities and identify those which are no longer deemed to be priorities. The budget allocation mechanism has been changed to enable the budget to reflect more clearly government policy priorities by program/activity, as reflected in sector expenditure strategies, in addition to the items of expenditure. Initial changes to the budget classification to reflect sectoral and ministry programmes are being put in place.

43 In Tajikistan this is the Law on State Finances. 44 While it is understood that the same agencies should be involved in both the strategic and detailed estimate preparation phases of an integrated MTEFhudget process (Le. that the MTEF process should precisely not be the responsibility of separate groups or bodies), it is often useful to initiate the process of developing an MTEF by entrusting a steering body with the task of ensuring that this new way of doing business is harmoniously mainstreamed with existing processes and that those are adjusted as and when need be.

87 262. When a country has a more advanced MTEF process, the macro/fiscal framework includes all resources, including external project finance, extra-budgetary finance and off-budget resources. Macro/fiscal projections are based on a more sophisticated forecasting model and on detailed policy analyses of proposed revenue policy changes and macro feedbacks. Sector ceilings are based on detailed analyses of cross-sectoral issues. With respect to sector expenditure strategies, initial or detailed costing of existing activities (both recurrent and capital) have been carried out, areas of cost savings and improvements to programme efficiency and operations have been identified, and broad expenditure implications of the new or restructured programme activities have been calculated. In this level, new prioritized activities to achieve policy objectives have been identified, a clear mechanism has been established for prioritising against policy objectives, and a mechanism is in place for allocating resources based on cutting or phasing out non-priority activities and funding resources to the highest priority and most cost-efficient activities.

263. At the sector level, the output is a detailed, fully-costed and prioritized medium-term sector expenditure strategy, with medium-term expenditure plans which contain detailed expenditure (recurrent and investment) implications for programs/activities and which are consistent with sector ceilings. In the most advanced cases, performance measurement indicators and analyses of sector performance are carried and sector budgets have been restructured to realize cost and efficiency savings. Finally, a mechanism to cost out and prioritize new proposals and integrate them into the MTEF has been established. At this level, changes to the budget classification and Chart of Accounts are fully in place, and budgets are executed in accordance with programmes and activities linked to policy priorities.

264. Tajikistan is effectively still at the preparation stage although it has implemented one of the main components of the primary stage (the macro/fiscal projections). The main impediment to further progress in the primary stage and to progression onto the intermediate stage is the fragmentation of the budge process discussed in Section 4.4, which prevents meaningful sector ceilings being used as budget planning tools.

Development of MTEFs within Broader Measures to Strengthen PFM

265. As part of the process for introducing a more strategic phase to the budget, Governments should pay attention to the wider budget systems on which budget plans depend. The MTEF is an integral part ofthe budget/PFM cycle; weaknesses in any of the links in the chain potentially undermine the ability of the MTEF to be used to achieve Government policy objectives.

4s There is a continuum with respect to including performance elements in the budget process, from merely introducing performance indicators alongside budgets, for sectors to monitor and report on in their sector expenditure strategies, to a process of linking budget allocations not to broad programmesprograms of activities but rather to specific outputs or performance targets. Countries at IevelLevel 3 in the development of an MTEF process are likely to start at the simpler end of the continuum. The more advanced along this continuum a country moves the greater are the analytical capacity requirements.

88 Firstly, if the broad strategic allocations in the MTEF are not translated into approved allocations in the appropriated annual budget, then the resources employed preparing a strategic budgeting phase are wasted. Secondly, well-made budget plans which are ignored or changed significantly during budget implementation undermine the credibility of the budget process and any strategic decisions that are made. More sophisticated planning of expenditures is futile if there are not also mechanisms to ensure that budget implementation is in line with plans, and if good information is not available on budget performance. Specifically, if the supporting budget systems (e.g., the budget classification system) are unable to allocate resources to the priority policy areas in line with the strategic budget plans, then there could be a mismatch between the use of budgetary resources and the achievement of budget policies. If accounting and reporting systems do not allow the recording and monitoring of budget programs, then central agencies, Governments and Parliaments will be unable to monitor if budget objectives are being met. Finally, in terms of scrutinizing the MTEFhudget, if the executive, the legislature, and civil society do not understand the budget in its strategic form, then it is more difficult to achieve an appropriate match between budget policies and resources.

Table 4.2: Broad phases of the development of an MTEF process Pre-MTEF Primary Intermediate Advanced MTEF IMTEF process MTEF Drocess Drocess Macro-jscal Based on Simple Inclusive of all pamework Establishment simple (e.g., domestic resources; more (top-down) and/or financial model; stronger sophisticated model; strengthening programming) macro/ fiscal alternative scenarios of basic legal framework; analysis; may inform decision- and focus on inclusion of making institutional domestic external finance parameters, revenues Setting of including Overall policy Some analysis Deepedmore sector ensuring that priorities, of inter-sectoral comprehensive analysis ceilings (top- a modern littleho analysis priorities incl. of inter-sectoral down) Organic of these basic policy priorities Budget Law is priorities or of priorities and in place. cross-sectoral cross-cutting issues issues Sector Limited sector Intra-sectoral Costing of activities expenditure expenditure broadly and stricter strategies strategies prioritised prioritization (bottom-up) strategic/ mechanisms inform programme intra-sectoral strategic framework in expenditure

I place framework.

266. Governments frequently introduce strategic elements to the budget as part of a wider PFM reform program. The decision of when best to introduce a strategic budgeting process within such a reform program is an important one. In practice, as

89 discussed above, when best to undertake the development of an MTEF depends on the basic state of public financial management. If a PEFA46or other diagnostic analysis of PFM indicates a large number of low scores, then the basic requirements of a PFM system are unlikely to be in place.

G. STRATEGY FOR PHASED IMPLEMENTATION OF THE MTEF

267. Restructuring budget relations along the lines discussed in Section 3 will be difficult and involve risks. Although the existing budget process is far from optimal as a process for allocating resources in line with policy objectives, it does actually allow a budget to be prepared and implemented: at least as a bureaucratic exercise it is effective. Even this could be threatened by an ill planned restructuring of the budget process, which could be highly disruptive to the functioning of Government. To minimize the risks, it will be necessary to move gradually, with the MTEF introduced in phases over many years. The strategy recommended here involves a twin track approach. The existing budget process will remain in place but reforms will gradually be introduced in the way in which the MOF handles the centralized functions of the budget process, while simultaneously sector budgeting will be introduced in a pilot sector. The next step will then be to bring more sectors of the budget into the MTEF, so that the existing budget processes are gradually replaced by new ones.

268. The full introduction of the MTEF will take many years, and will need to be accompanied by complementary reforms to public financial management, such as strengthening of treasury systems, auditing, monitoring and evaluation and payroll management. One of the important findings of the World Bank study on the implications of weak PEM capacity for the PRSP approach, is that PEM reform should be undertaken in a holistic manner: improving one link in the PEM system will not generate better budget outcomes if other links in the system are unreformed and remain weak (Andrews and Moon, 2003). The Government will need assistance to implement the MTEF, and complementary PFM reforms, from donors; in particular, it will need technical assistance for the MOF and the line ministries in the pilot sectors. Because the budget systems in CIS countries are markedly different from those in most other developing countries, the former face unique challenges in implementing PEM reforms. It will, therefore, be important that TA provided by donors can draw on the expertise gained in implementing PEM reforms elsewhere in the CIS region.

269. In terms of the sequencing discussed in Section 4.6, the three phases set out below are intended to take Tajikistan from the preparatory stage (with some components of the primary stage) through into the intermediary stage.

Preparatory Phase

270. The earliest that implementation of the MTEF can start will be the 2007 budget process, which will produce budget estimates for 2008 and budget projections for the following two years. Before the 2007 budget process begins, several preparatory

46 See PEFA Secretariat, Public Financial Management: Performance Measurement Framework, June 2005

90 measures will have to be taken. These are: (i)obtaining political approval for the introduction of the MTEF through a Government Re~olution;~’(ii) reviewing all relevant legislation (e.g., the Law on State Finances) to determine if any revisions to legislation are needed; (iii) select one (or more) pilot sectors; and (iv) draw up a new budget timetable with an earlier start date to allow for a longer budget process.

27 1. The best candidate for a pilot sector is the education sector for several reasons. It is already benefiting from TA designed to strengthen budgeting allocations in the sector and, under the Education for All Initiative, the sector will receive grants, a condition of which is the preparation of a sector strategy. T he preparatory phase should prepare the Ministry of Education (MOE) for the expanded role it will play in the budget process. A study should be carried out to assess the best way to restructure the budget process within the education sector so that the MOE is given the responsibility of coordinating all budget requests from the KBOs in its sector and preparing a single consolidated submission for the MOF. A chart should be prepared by the MOE showing all of the KBOs in the education sector and the expenditures which they control. The MOE should draw up budget instructions to be sent to all KBOs in the education sector, which request details of budget requests and give guidelines on the form and magnitude that these should take. Internal restructuring of the MOE will also be necessary so that more staff is allocated to working on the budget process.

Phase 1

272. Phase 1 should start with the beginning of the 2007 budget process, which should be no later than March 2007, and last for two years.

273. The key tasks for the MOF will be to prepare a budget resource envelope at the start of the budget process which can provide the basis for allocating sector ceilings. The European Union (EU) is planning to provide technical assistance in the area of macroeconomic and tax revenue forecasting which will assist the MOF to prepare the budget resource envelope. Other tasks which should be implemented by the MOF in phase 1 are the setting up of a process for allocating the aggregate budget resource envelope into sector ceilings, involving drawing up criteria for making these allocations (which in principle should be linked to policy objectives in the PRSP) and an intra-governmental process for reaching a political consensus on the allocations, so that they are not merely a technical exercise carried out by the MOF.

274. In the education sector, the minimum requirement of phase 1 will be to establish functioning relationships between the MOE and all other KBOs in the sector to allow the education sector budget submission to be prepared on schedule. The MOE should also begin to prepare a medium-term expenditure plan for the sector, which can guide the sector’s budget allocations. This will require capacity building in the MOE and probably long term TA. n education sector working group should also be established, comprising key stakeholders among the KBOs in the sector.

275. Appendix 2 sets out the main actions required to implement the MTEF in both the MOF and the pilot sectors in the preparatory phase and phase 1.

47 The Government Resolution endorsing the MTEF was passed in October 2006.

91 Phase 2

276. Phase 2 should begin in the first quarter of 2009. The key tasks of phase 2 will be fourfold.

277. First, the MOF should consolidate the process for preparing its budget resource envelope and the allocation of this into sector ceilings. In particular, a formal process for agreeing the sector ceilings within government should be established: One option would be to use the National Budget Commission.

278. Secondly, the education sector, having restructured the bureaucratic links in the budget process to channel budget submissions through the MOE, should then focus on improving the quality of budget allocations by applying technical expertise in the MOE to evaluate expenditure requests, cost new proposals and assess the links between expenditures, inputs, outputs and outcomes in the sector. Obviously these qualitative improvements in the budget process will take time to be realized. The SWG should also be strengthened so that it can play an effective role in the budget process.

279. Thirdly, responsibility for determining all education related capital expenditures in the CSIP should be transferred to the MOE from the MOET. The education sector expenditure ceiling will, therefore, cover both current and capital expenditures.

280. Fourthly, two other sectors should begin implementing sector budgeting in phase 2, undertaking similar reforms to those carried out in the education sector in phase 1. Health would be a good candidate for a second pilot sector, not least because it is so critical for achieving the MDGs. The social protection sector can also be included in this phase. The restructuring of the budget process in the social protection sector should be more straightforward than in other social sectors because the local governments play only a minor role in this sector.

Phase 3

281. The third phase will probably begin in 201 1 and is likely to last for more than one year, as it will involve bringing the rest of the budget into the MTEF. The MOF should prepare new budget instructions for all LLMs and the KBOs which remain outside of sectors. These instructions should include the three year expenditure ceilings. All of the remaining sectors should be brought into the MTEF, drawing on the experiences and lessons learned by the pilot sectors to reorganize the budget process within the sectors. All sectors will take over responsibility for the relevant capital projects under the CSIP.

282. Tax collections should be centralized in phase 3, with all collections other than those taxes specifically earmarked for local authority funding being submitted to the central treasury. Local authorities will, therefore, receive their funds in the form of transfers from the central treasury, instead of retaining taxes when they are collected.

283. The final step in the implementation of the MTEF will be the integration of the PIP projects into the sector ceilings of each sector.

92 H. CONCLUSION

284. The existing budget system, characterized by a very high level of fragmentation among over one hundred KBOs and a very limited role played by line ministries is a major constraint to the introduction of a sector-based MTEF in Tajikistan. Reforming the budget system is, therefore, a prerequisite for implementation of a meaningful MTEF. The key reforms involve restructuring the budget process into sectors, so that a sector expenditure ceiling can be allocated in a coherent manner among competing expenditure demands in the sector, according to the sector’s expenditure priorities. In each sector, a lead line ministry should be put in charge of coordinating the budget process in that sector.

285. It is necessary to be realistic about the pace at which the reforms recommended here can be implemented. The institutional reforms to the budget system will be difficult to implement and take considerable time; hence they should be introduced cautiously, in a phased manner, to minimize the risks of severely disrupting the budget process. The creation of functioning sectors, with sector wide expenditure ceilings and budget allocations, should be implemented first on a pilot basis beginning in the education sector. The pilot phase is likely to last two years before other sectors can be brought into the MTEF. The MTEF will place much greater responsibilities on line ministries to coordinate and lead budget planning in their respective sectors, but capacities for budget planning in line ministries are very weak, and so strengthening these capacities through training and technical assistance will be imperative if the MTEF is to succeed. Consequently, substantial financial and technical assistance from donors will be essential to support the implementation of the MTEF.

93 APPENDIX 1 MEDIUM-TERMBUDGET FRAMEWORK

Table 1

Budget for 2006 and Forecast for 2007 and 2008; thousands of Somoni

tevenue and Grants 2006

ievenues and Grants to state budget 150000C If which VAT repaid -1 oooc ievenues and Grants 151000C tax revenues 1383736 income individuals 9027E 1 10279 123512 profit tax from corporations 425% 45410 50859 minimum tax on enterprise profits 22847 23192 25975 social taxes 17200C 197000 225000 tax on real estate 14972 17217 19283 land tax 8312 8338 9340 single tax from agric producers 66203 66865 71546 sales tax inc 58837 131125 146860 cotton fibre 581 12 74125 83020 aluminum 725 57000 63840 tax on reta.il sales 301 34 33748 37798 VAT 505532 552 134 618390 domestic 124990 145210 162635 external 380542 406924 455755 road user tax 47173 47941 53694 simplified tax 5835 8158 9137 excise tax 81901 91 105 10428 1 domestic 11331 13886 15840 external 70570 7721 9 8844 1 customs duties 108121 132768 147364 patent fee 12072 15202 17026 vehicle owner tax 8700 10596 1 1867 bonus and royalty 6262 7189 8051 state duty 16468 17621 19736 processed product tax 85531 ion tax revenues 88264 95606 11431 1 grants 38000 40000 40000 )we: Ministry of Finance

94 Table 1 (continued)

Medium Term Budget Framework: Budget for 2006 and Forecast for 2007 and 2008; thousands of Somoni ixpenditures 2006 200; 2001

,tal expenditures 154200C 1696991 190453( Iublic admin bodies & foreign econ relations 184656 20321; 22806: -aw enforcement structures 2387 1E 262706 29483 Social Sphere total 7631 1E 84043: 94453( education 33607C 37035( 41608’ health 106482 11730: 13251; social protection 19165C 210911 236701 compensations 3724C 4098: 4599! culture and sports 5741 1 631 8: 7090s other social services 34264 37701 4232( interest payments 5108E 5510; 5770: externa I 3508E 3749! 3794; domestic 1600C 17601 1976; rota1 economic services 23382€ 25782: 292 19( agriculture and agro-indistrial complex 49691 5493! 6237~ transport and communications 74842 8236f 9282: mining and construction 1153E 1269f 14241 fuel and energy 25575 28296 32581 utilities 67246 74 1O! 8405f other economic services 4935 543’ 609! ither expenses 7060C 7769f 87191 ‘IP spending from external sources 33600C 36400( 40400( ‘otal expenditures including external PIP 187800C 2060991 230853( ludget Deficit and Financing ludget Deficit excluding PIP -4200C -4550’ -5050’ nancing 4200C lomestic 150151 privatisation receipts changes in govt deposits proceeds from securities domestic debt repayment ixternal 22784s Foreign debt repayment -108151 -101 60f -11501( External financing to PIP 33600C 36400( 40400( leficit including PIP -37800C -40950’ -45450‘ iurce: Ministry of Finance

95 APPENDIX 2: ACTION PLAN OF STEP TO IMPLEMENT MTEF IN PILOT SECTOR 2006-2007

Time Period Ministry of Finance Lead Line Ministry for Pilot Sector (MOE) October 2006 Designate education sector as pilot sector and Ministry of Education (MOE) as Lead Line Ministry for that sector October 2006 Send out budget instructions for implementing MTEF in education sector November- Draw up new budget timetable December for 2007, with earlier start date 2006 for budget process November- Meet with MOE to discuss Meet with MOF to discuss plans December plans for implementing- MTEF for implementing MTEF in 2007 2006 in 2007 November- Meet with MOE to discuss TA Meet with MOF to discuss TA December requirements from donors requirements from donors: make 2006 requests to donors for TA December Assist MOE with budget data Draw up list of all KBOs in 2006 on KBOs education sector and obtain from MOF budget estimates for 2005 and 2006 for each of these KBOs December Organise training workshops for Attend training workshops 2006 -January senior officials and budget 2007 officials in MOE and selected KBOs in education sectors January 2007 Strengthen budget departments by re-allocating staff to them January 2007 Write to all KBOs in education sector informing them of new budget procedures for implementing MTEF January - Work with MOE to draft new Work with MOF to draft budget February 2007 budget instructions to be sent by instructions to be sent by MOE to MOE to all KBOs in education all KBOs in education sector sector January - Make projections of budget March 2007 resource envelope for State Budget using forecasts of revenues, grants, debt service, borrowing, privatisation receipts January-March Consult with National Budget 2007 Committee on criteria for allocating budget resource

96 envelope among sectors February- Meet with joint Public Finance March 2007 Management donor mission to discuss and agree medium term assistance for MTEF and PFM reforms March 2007 Determine sector ceilings for education sector: send sector ceilings to MOE April 2007 Advise MOE on dividing Divide sector ceiling from MOF consolidated sector ceilings into into ceilings for each KBO, based ceiling for each KBO on projections for protected items, sector policies, etc. April 2007 Send budget instructions, including ceiling for each KBO, to each KBO in sector May-June Meet with each KBO in sector to 2007 discuss budget allocation July 2007 Receive budget requests from KBOs in sector; review each request and revise if necessary August 2007 Submit consolidated sector budget submission to MOF August 2007 Meet with MOE to finalise Meet with MOF to finalise sector sector budget submission budget submission September Amalgamate consolidated 2007 budget estimates from education sector with budget estimates for all non pilot sector KBOs October - Begin drafting medium term December sector expenditure plans 2007 November Review implementationof 2007 Review implementation of 2007 2007 budget process in pilot sector budget process in pilot sectors with MOE; identify problems with MOF; identify problems and and necessary adjustments for necessary adjustments for 2008 2008 budget process; identify budget process, identify where where more training, TA is more training, TA is required required November - Training of senior officials and Attend training organized by December budget officers in MOF, MOE, MOF 2007 LLMs for future MTEF sectors and selected KBOs

97 APPENDIX 3: REGIONAL LESSONS FROM THE DEVELOPMENT OF MTEFS

Experience of developing MTEFs in other CIS countries indicates that, whilst there are a number of common lessons, there is no standardized methodology or blueprint for introducing greater strategic elements into the budget process. The development of an MTEF process needs to be adapted to suit the particular circumstances of each country, including the extent to which fiscal discipline has been established, the relative strength of the role of the MOF in the budget process, relative capacities in the MOF compared with sector ministries, and the degree of political commitment by the Government and by MOF.

The introduction of MTEFs in the CIS region has led to a number of achievements, including providing a tool for the Government to maintain fiscal stability and discipline through publicizing and committing itself to the overall budget framework. In addition, MTEFs have provided GovernmentParliament, sector ministries, and sub-national governments with a more comprehensive source of analytical information on the budget (i-e., trends and key issues); a common problem in many countries is the lack of basic information and analysis on where funds are being spent. This is particularly important when central agencies do not have influence, or little or indirect, over the budgets of sub-national entities. MTEFs have provided a forum for communicating these budget parameters to budget stakeholders.

In some MTEFs, budgetary allocations have begun to enable resources to be moved towards higher priority sectors and to achieve a better balance between economic items (e.g., reform of pay structures and greater allocations for operations and maintenance). In more advanced MTEFs, expenditure analyses at program level have focussed on general strategic directions for a relatively small number of basic programmes, in some cases accompanied by the introduction of a programme-level classification. Some MTEF experiences have had success in generating a more active dialogue with the MOF over funding levels and justification for activities to be funded.

In general, the broad lessons from the MTEFs can be linked to issues of timing and sequencing and institutional capacities. In particular, more established MTEFs have tended to develop more slowly and organically, illustrating sequencing and timing issues as discussed in the section below. n other words, these processes have not attempted to go for a “big-bang” approach to MTEF introduction but instead are working through the different stages of development (see separate section on sequencing MTEFs).

In terms of specific lessons, a number of common themes have emerged amongst the countries in the region which have introduced MTEF processes.

Begin with the macro/fiscal framework (top-down process)

In terms of the appropriate timetable and specification for MTEF development, experience indicates that relatively greater progress has tended to be made in establishing and strengthening the macro/fiscal framework, followed by the rest of the top-down process. Slower progress has been made in incorporating sector strategy

98 exercises into the process and getting concrete changes in budget practices leading to resources being spent (not just allocated) according to strategic priorities. This reflects two aspects of transitional countries compared to other regions: (i)the weaker institutional role of MOF; and (ii)the relatively stronger capacities of staff and staff with more relevant experience in MOF than in sector ministries.

First, a strong, directional role for the MOF in the budget process, common in other regions, is less prevalent in transitional countries. This reflects the often unchanged management role and capacities of MOFs. In most cases, the structures and responsibilities of both the MOF and finance departments in the sector ministries have not changed significantly since independence, although their roles have grown in importance as constraints on resources have increased. In particular, the role of the MOF in facilitating policy co-ordination though the budgetary process is not yet widely understood. This can undermine the necessary fiscal management role of the MOF and its overall co-ordination and management of the budget process as a whole (in particular, articulating and defending budgetary trade-offs), which is central to a well-functioning MTEF process. Thus, the first step in developing an MTEF in the region is to build up the macro/fiscal framework in order to set the overall resource constraints for sectoral expenditure decisions.

Second, progress in developing sector expenditure strategies and sub-sectoral policy priorities within constrained resources in the region has been hampered by relatively weaker analytical capacities found in sector ministries. Many budgedfinance departments in sector ministries continue to be small, and budgetary analytical skills tend to be quite limited; these departments tend to be staffed mainly by accountants, reflecting their previous role under central planning. In many MOFs, by contrast, there is evidence of the employment of new staff with relatively greater analytical skills.

Importance of ensuring that the basics are in place jirst

An MTEF process is not a panacea for PFM weaknesses. Lessons from experience indicate that a number of PFM basics need to be in place to undertake successfully the development of an MTEF. First, achieving and sustaining aggregate fiscal discipline is critical. As experience in the region suggests, the MTEF can assist both with the initial achievement of aggregate fiscal discipline and with sustaining it. However, it is easier to begin such a process if the Government has already become accustomed to the discipline imposed by ensuring a realistic macro-fiscal framework.

Secondly, the introduction of MTEFs can be undermined by the lack of a credible budget process at the outset. Above all, a basic budget system needs to be in place. If there is not a credible annual budget process, including basic budget discipline (e.g., in immediate post-conflict countries), then there will be little gained from trying to introduce a strategic phase to such a weak or nonexistent budget process, and it could well prove to be a distraction from building such a process.

Other PFM reforms which assist in the implementation of the MTEF include: (i) having in place a modern budget classification system, such as the GFS, so that expenditure data can be analysed by functiodsector and by economic item; (ii)the promulgation of a revised Organic Budget Law and extended budget calendar to give sufficient time for the analysis, preparation and consideration of MTEF and budgetary

99 parameters; and (iii)strengthening expenditure control systems so that budgets may be more easily executed as planned.

Implementing an MTEF can provide the impetus to the MOF to implement a wider package of PFM reforms, including the necessary institutional capacity building, in order to make the MTEF more successful.

Extent of Political Engagement

It is often assumed, particularly with respect to other regions, that sufficient political engagement is a necessary pre-condition for introducing a successful MTEF process since the MTEF is about ensuring that resource allocations reflect a government’s priorities. In the region, there have been positive MTEF experiences where explicit political commitment has been strong at the beginning of the process, as well as good examples where the process has helped to build political support.

On the one hand, what is often lacking in the process of introducing a more strategic budget process is strong political guidance and leadership on relative policy priorities. In the absence of such guidance, it is difficult for the MOF and other central ministries to make inter-sectoral resource allocation decisions. A t the same time, without strong Government engagement in the process, it is not clear that there is commitment to the medium-term budgetary allocations, thus potentially undermining the credibility of the MTEF process (e.g., Cabinet decisions on budget allocations may not be in line with MTEF ceilings).

On the other hand, in the CIS, governments’ motivation has often been generated subsequently by demand-pull factors by senior technicians, whereby senior budget personnel, usually in the MOF, find that the MTEF facilitates their ability to determine the overall resource framework (in terms of setting a hard budget constraint) to guide sector ministries in budget preparation, evaluate sector ministry budget proposals against this framework, and provide analyses and justification to Government officials on budget parameters, all of which in turn can create the demand for more analytical budgetary information. In addition, at least in theory, the MTEF should enable government officials and MOF staff to have a more informed dialogue with the International Financial Institutions (IFIs), although it is not clear how much this has improved in practice in the region.

Some MTEFs in the region have achieved considerable success in involving governments in the key decision-making points, such as approval of the initial macro- fiscal framework and the overall framework and Budget Framework Paper/MTEF document. This review is often facilitated by the establishment of MTEF Steering Committees to oversee the process. One also sees evidence of a growing interest by Parliaments in some countries to review the MTEF; strengthened Government and Parliamentary scrutiny of the overall MTEF framework and resource movements (ceilings) in relation to governmentPRSP policies will be important. Encouraging wider public interest in the strategic budget process is a longer-term priority.

External pressure has led to the introduction of some MTEFs in the region. The ability of donors to facilitate internal demand for an MTEF depends on the Government’s initial willingness to reform, which in turn often depends on there being prior experience of a productive Government-donor dialogue.

100 Extent of Analytical and Other Institutional Capacities

As indicated above, institutional weaknesses faced by countries introducing more strategic elements in the budget process include weak analytical and policy-making capacities, exacerbated by existing management structures which do not reflect the greater need for these capacities. In many cases, the structures and responsibilities of both the MOF and finance departments in the sector ministries have not changed significantly since the advent of strategic budgeting, although these departments’ roles have grown in importance. In particular, the greater policy analysis role of MOFs is often not evident in the departmental structure, staffing, and job specifications.

Similarly, the analysis of investment and recurrent expenditures often continues to be the responsibility of separate departments within a ministry (or even separate ministries); sometimes due to institutional weaknesses, the MTEF does not always succeed in integrating the processes for allocating recurrent and development expenditures. Effective strategic budgets which aim to link policies to an integrated budget require some co-ordination arrangements (e.g., an Inter-Ministerial Budget Guideline Committee). The lack, or poor specification, of such arrangements can be problematic. Similar issues arise in line agencies with regard to policy development/analysis, planning and strategic budgeting.

Significant difficulties have occurred in processes which have been overly demanding of analyses before sufficient analytical capacities are in place to support these requirements, e.g., for policy development or expenditure analysis. This can be wasteful of staff time and can distract time and attention away from more feasible strategic prioritisation processes which would be informed by simpler analyses.

Finally, there have been difficulties in aligning programmes with budget classifications due to existing classifications being based on organizational (institutional) lines. This has prevented resources from being allocated by programmes during implementation and monitoring information on expenditures programmes to be collected.

Existence of Supporting PFM Reforms

The lack of accompanying and supporting reforms in the wider budget cycle, particularly expenditure execution, accounting, reporting and control processes, can undermine improvements in budget planning. Thus, planned budget allocations which reflect sectoral priorities may not actually be spent according to those priorities.

Without accompanying changes in budget systems and procedures, particularly the use of normative techniques in sectoral resource allocations, it is very difficult to restructure sectoral budgets. The use of norms tends to increase inefficiencies in sectoral allocations through: (i) removing from programme managers the responsibility for introducing innovations that allow for improvements in productivity and in the quality of public service provision; (ii) providing incentives to managers to increase input use as a means of maximising their budget allocations; and (iii) reducing budget analysis in sector ministries to a process of checking whether norms have been applied correctly rather than challenging the assumptions on which budgets are based and considering alternatives.

101 LegaVOrganizational Framework for MTEF

In CIS countries, greater emphasis than in other areas is placed on the legal framework for budgeting. Revising the Organic Budget Law to incorporate the explicit requirement to prepare an MTEF and incorporating it in the Budget calendar has been a critical factor in supporting the MTEF in countries in the region. However, this is not strictly necessary, as the modem budget systems laws in many CIS countries are sufficient to cover the introduction of an MTEF.

Timing for MTEF Preparation The lessons from the MTEFs which have been carried out in the region underline the importance of beginning the process early and devoting sufficient time to it. In practice, in CIS countries, this is likely to mean beginning in January or February, for submission of the BFP/MTEF document to Government in June, prior to the beginning of the Annual Budget process, with the issuing of budget guidelines. Thus, the first half of the year would be devoted by the MOF and sector ministries, if involved, to preparing a Budget Framework Paper/MTEF document. This would give sufficient time for an initial macro-fiscal framework to be submitted to Government, or a governmental Budget Committee, for initial approval before the framework is finalised and included in the Annual Budget Guidelines, which are often issued in May or June. This would facilitate a logical process from strategic budget phase to detailed budget allocations, including involvement of Governments at key decision- making points (Le,, the macro/fiscal framework in MarcWApril and the detailed resource framework including sector/ministry ceilings).

Design of an MTEF

In the design of an MTEF, it is important that expectations be realistic and achievable. A key indicator of initial success is if the Government begins to request explicitly the analyses included in the MTEF. Generating this interest can be facilitated by the requirement for approval by the Government of MTEF parameters. At the same time, in order for the MTEFs to be sustainable, the issue of institutional capacities must be addressed, and in the meantime, governments and MOFs should adopt an appropriate sequence for MTEF preparation.

It is also important that the ceilings are determined at an appropriate level of detail. In MTEFs in other regions, budget and forward estimates are shown by programme. With some exceptions, most MTEFs in the region show their estimates by functiodsector and economic classification or by functiodsector, ministry/agency and economic classification. In the sector expenditure strategies undertaken by some countries in the region, an attempt has been made to analyse the sectors’ expenditures by major programme. Changes in the classification system will be required to introduce programmatic categories, although in some cases these may follow the sub- functional classification.

The program budgets introduced in some countries in the region provide a framework for bottom-up analysis of sector expenditures. These exercises ask sector ministries to set out their mission statement or “vision”, and request narrative justification for expenditure requests, which tend to be presented by sub-function. Sector expenditure analyses, prepared as part of the MTEF, link this type of bottom-up analysis with the

102 top-down policy and strategic framework. The main conceptual difference between programme budgets and sector expenditure strategies is in the level of analytical detail required. In general, analysis of expenditures by sub-functions (as in the program budgets) can be very detailed, making policy and expenditure trade-offs between programmes difficult.

Whilst the MTEF is part of the budget process, the initial focus of many MTEFs has been on producing an MTEF document, setting out the medium-term macro/fiscal strategy and the resource framework, including the sectoral and/or ministry ceilings. This document can help establish the MTEF as a key part of the budget process and provides transparency of future budgetary parameters and thereby can promote fiscal discipline. As the MTEF process evolves, experience shows that the importance of reviewing these documents to ensure that the document does not become overly complicated and the preparation of the document become an end in itself, rather than part of the underlying budget planning process.

103 Example of Annual Budget Calendar in Country with a January-December Fiscal Year

Date Task JanuaryIFebruary Preparation of MTEF methodological guidelines Februaryhlarch Preparation of initial macro/fiscal framework for MTEF period to Government April Presentation of initial macro/fiscal framework for MTEF period to Government Preparation of MTEF budget paper by MOF and presentationto Government Determination of ministerial budget ceilings by Government 1 June Issuance of budget preparation instructions, budget forms and indicative budget ceilings June/July Ministries determine spending priorities and prepare budget requests. Submission to MOF by mid-July. August MOF to scrutinise submissions and to resolve minor discrepancies with spending ministries August/September Meetings with ministries at MOF. Early September Preparation of economic report and revision of MTEF resource envelope. Collation and summarisation of budget data. 15 September Submission of draft Budget to Government 15 October Submission of draft Budget to Budgetary Committee of Parliament October Preparation of Budget Address and update of Economic Report in line with the country’s Economic Development Policy document 1 November Submission of Budget to Parliament 1 December Approval of Annual Budget Law by Parliament 1 January Issuance of authorities to spend 1 January Distribution of Budget March, June, Quarterly reports by line ministries to MoF on budget September, execution December March/April Annual report by Government to Parliament on execution of previous year’s State Budget

104 5. PUBLIC EXPENDITURE AND FINANCIAL ACCOUNTABILITY (PEFA) ASSESSMENT

A. INTRODUCTION

286. This chapter provides a summary of the PEFA Assessment of Tajikistan’s Performance Report undertaken between October 2006 and April 2007.48The purpose of the PEFA assessment is to provide the GOT with an objective assessment of the country’s PFM systems. It aims to support a better understanding of the overall fiduciary environment of the budget and assist in identifying those parts of the PFM systems most in need of reform. As such it should contribute to a common understanding among the GOT and the donors wishing to support further PFM reforms.

287. The PEFA study team was composed of three national and two international consultant^.^^ The main GOTpartner for the PEFA Assessment was the MOF. The MOF appointed a Working Group which coordinated the Government’s participation and input into the PFM Assessment. Meetings were held with a majority of government institutions involved in the PEFA assessment (MOF, MOET, SFCC, MOR, the parliamentary budget committee, Aid Coordination Unit), and with key PEFA partners (IMF, DFID, EU). Subsequently, the MOF prepared a self-assessment report, which was a key component of the assessment process.

Methodology

288. The PEFA methodology is set out in the Public Financial Management Performance Measurement Framework (available at www.~efa.orq).~~It is based on 28 indicators covering a country’s PFM system, and 3 indicators addressing the interaction of donors with a country’s budget process and PFM system. The PEFA methodology has been in use since 2005 and has been applied to more than 60 c~untries.~’

289. PEFA Assessments provide cross-country comparable indications of the effectiveness of PFM systems, and of their improvements over time. They do not provide, however, an analysis of the causes of existing weaknesses, nor an indication of the PFM system’s ability to deliver development objectives, e.g. poverty reduction outcomes.

48 The Main Report will be published separately. 49 The Swiss State Secretariat for Economic Affairs, SECO, was the main sponsoring partner of the assessment. 50 The PEFA assessment framework is also available in Russian through this website. 51 Further information, including the methodology in Russian, can be found at: www.pefa.org.

105 Each indicator is scored on a scale from A to D. The basis for these ratings are the minimum requirements set out in the methodology.

290. The main sources of information used for this PEFA assessment were: (a) official GOT reports and data; (b) external evaluations and reports (WB, IMF, and others); and (c) numerous semi-structured interviews with key users and providers of PFM information and other stakeholders (Government representatives; donor representatives; members of parliament; representatives of selected non-governmental organizations). The self-assessment report played an important role in terms of taking stock of current institutional features, practices, and laws concerning public financial management in Tajikistan; and provided a key source on which the Performance Report draws.

B. INTEGRATED ASSESSMENT OF PFMPERFORMANCE

29 1. The Public Expenditure and Financial Accountability Performance Assessment Results are summarized in Table 1,

Credibility of the Budget (PI-1 to PI-4)

292. In the period 2003 to 2006, the GOTbased its budget planning on conservative revenue forecasts. A s reflected in indicator PI-3, this is generally positive, but it has led to a situation in which the budget outturn has differed considerably from budget plans; i.e., actual revenues have exceeded budgeted revenues and this has enabled actual expenditures to exceed budgeted expenditures. Although this discrepancy has declined, it is still significant. Furthermore, the actual composition of expenditure differs from the composition in the budget plan. Given the tendency for actual revenues to exceed budgeted revenues it is important to increase the transparency of the processes for allocating additional expenditures; including a more consistent involvement of the national parliament. This may also require more precise legal rules. Expenditures should be planned and discussed during the budget formulation process, restricting the use of supplementary budgets within-year to unexpected emergency requirements. Payment arrears have been declining and are currently at low levels;, but the information that is recorded does not provide a clear distinction between payments due and arrears.

Comprehensiveness and Transparency (PI-5 to PI-10)

293. The Government has undertaken initial steps to improve the comprehensiveness and transparency of the budget; but further improvements are needed. The GOThas introduced functional and economic budget classifications that are generally in line with the 1986 GFS standard. However, an administrative classification is still missing; and until 2006, PIP resources were not included in the functional breakdown. The Social Protection Fund was integrated more closely with the budget in late 2006. Sub-national governments are covered by the treasury system and central government monitors their fiscal situation. Reporting, monitoring and oversight of the fiscal risks from State Owned Enterprises (SOEs) is weak; this is a major concern with regard to large enterprises and in the context of still sizable quasi- fiscal activities. The set of information included in the annual budget submitted for parliamentary scrutiny is relatively comprehensive, but lacks analytical material on

106 Table 1: Overview of Assessment Results Indicator I Issue covered I Rating I A. PFM Out-Turns: Budget Credibility I I

I project and program aid D3 I Proportion of aid that is managed by use of national procedures ]D

107 new policy initiatives and their budgetary implications. Parliament is also not informed about tax exemptions which reduce the overall amount of budgetary funds. There is some public access to fiscal information, but the information provided tends to be partial and incomplete. Accessibility of information should be improved further.

Policy-based Budgeting (PI-11 and -12)

294. Some steps towards establishing more policy-based budgeting have been made and a renewed effort at establishing an MTEF was begun in 2006. The annual budget process is orderly, in terms of following key dates set out in the budget calendar. However, the institutional structures and systems for linking policies and budgets are still weak. Cabinet-level discussions of the budget and of key priorities and trade-offs to be reflected in fiscal planning are not well developed. A cabinet-level Budget Commission was established in 2004, but its impact has remained marginal. An initial set of sector strategies has been elaborated; but thus far, only the education sector strategy is at a more advanced stage. Linking sector policy goals to realistic costing estimates will require further work for a number of years.

295. Capital budgeting is fragmented, and needs to be further integrated with recurrent budgets. The high share of capital spending in total expenditures means that discretionary resources are a substantial share of the budget, which underlines the need for policy-based budgeting. Since a large part of capital spending is financed by development partners, it also underlines the need for a more effective approach to aid management on the side of the GOT, and for the harmonization and alignment with national priorities on the side of donors. Some steps in this direction have been undertaken in 2007 in terms of budget presentation. However, institutionally, capital budget planning remains separate and sub-divided into domestically financed and externally financed ‘streams’, and aid policy and aid management structures continue to require greater attention.

Predictability and Control in Budget Execution (PI-13 to PI-21)

296. A basic degree of predictability and control in aggregate budget execution has been achieved, but there are still systemic weaknesses. A core Treasury system has been established; and the Government plans to extend autoimmunization to local levels. Cash management has worked in a situation of continuously higher than budgeted revenues. Debt data recording and reporting has been improved; but debt sustainability analyses still need to include domestic debt.

297. Important weaknesses in budget execution are the following: (a) effectiveness in tax collection is a concern, with significant levels of tax arrears especially from large enterprises, (b) public payroll controls are currently weak, (c) procurement reform has begun, but while legal reform has been undertaken, the institutionalization of the new system is still missing, and (d) the internal audit function is very weak. Furthermore, as indicated above, the process of allocating additional resources during budget execution is not sufficiently clear and transparent. This means that that there is a situation of frequent ad hoc bidding for additional funds from various sources (presidential fund, excess revenue) and of generating own resources by budget organizations which have not been sufficiently taken into account at the budget planning stage. Expenditures budgeted and organized at short notice during the

108 process of budget execution are unlikely to be employed most effectively. Furthermore, further improvements in cash-flow forecasting and management will be needed as the discrepancy between planned and actual revenue is reduced.

298. According to the Public Administration Reform Strategy endorsed by the Government and the President in 2006, payroll controls and internal audit are to be addressed in the short to medium term. It is currently not clear what further reforms are envisaged with regard to the revenue system and the cash-flow forecasting and management function.

Accounting, Recording, and Reporting (PI-22 to PI-25)

299. Basic systems for accounting, recording, and reporting are in place. Accounts reconciliation is carried out in a timely fashion. In-year reports are prepared on a monthly and quarterly basis by the Treasury. The information appears to be reliable. However, a modern reliable “double entry” Financial Management Information System is a vital priority for the future if quality accounting and reporting is to be achieved. A key weakness is the fact that the budget execution report is not being submitted for external audit on an annual basis. Furthermore, accounting standards do not comply with international standards, and the standards used are not systematically disclosed in fiscal documents.

External Scrutiny and Audit (PI-26 to PI-28)

300. External scrutiny and audit are weak. Legislative scrutiny of the annual budget law is satisfactory; but the absence of an administrative classification limits the ability of parliament to hold members of the executive to account. Tajikistan’s External Audit Institution continues to be accountable to the president and currently lacks independence as required by international (INTOSAI) standards.52 The audit is largely focused on compliance with rules (regularity audits) which should be regarded as an adequate interim stage, given that the State Financial Control Committee (SFCC) was only established in 2002. Some training on other forms of audit (performance/value-for-money audits) has been received. Parliament is currently not reviewing audit findings. As noted above, the SFCC does not receive the Government’s financial statement for annual review (contrary to provisions stipulated by law), but rather has undertaken bi-annual audits of the budget (by conducting on- site audits of the MOF - in addition to audits of KBOs). Audit findings are primarily submitted to the president, and to parliament only in summary format. In late 2006, the SFCC was transformed into an Agency of State Financial Control and Combating Corruption. This is a problematic change; as the roles of external audit and of combating corruption are distinct and should be kept separate according to established international practice.

Donor Assistance (D1 to D3)

30 1, Donor assistance to Tajikistan still largely follows a traditional, project-based approach. Use of country systems has been very limited thus far as these systems are

52. Tajikistan’s external audit institution currently is not a member of INTOSAI. It holds membership in ECOSAI, the regional association for the countries of the Economic Cooperation Organization. This regional group is not a member of INTOSAI.

109 perceived to be weak. An intensified dialogue on aid effectiveness and the actions that the GOT and donors can take to improve aid practices would be desirable. Donor coordination should be a higher priority for the Government, given Tajikistan’s substantial aid receipts (around 10% of GDP) and the importance of these relative to domestically generated revenues. Donors need to make greater progress towards harmonization and alignment of their aid. Increasing the share of budget support may be one way of encouraging a more comprehensive approach to Tajikistan’s development needs; but it will require further progress with regard to the quality of the public financial management systems (and in particular internal and external audit functions). c. ASSESSMENTOF THE IMPACT OF PFM WEAKNESSES

302. PFM systems have three objectives: macro-fiscal discipline, strategic allocation of resources, and operational or technical efficiency. These three objectives are linked. Fiscal discipline is the basis without which neither a strategic allocation of resources nor operational efficiency is possible.

303. The findings of the strengths and weaknesses of the current PFM system have the following implications in terms of these three objectives. Tajikistan has made important steps towards establishing macro-fiscal discipline. However, the situation remains fragile due to systemic risks and weaknesses, as shown by the substantially increased fiscal deficit projected in 2007 (which will be funded by external borrowing). It will be essential not to lose sight of the fundamental issue of macro- fiscal discipline and the need to further strengthen relevant structures - including better revenue collection, sustainable debt management, and better oversight of fiscal risks arising from SOEs and quasi-fiscal activities. These issues should be addressed as a priority.

304. Furthermore, significant progress still needs to be made with regard to the strategic allocation of resources and the technical efficiency in the utilization of public funds. The current situation reflects initial reform efforts and priorities since the re-building of the state and of the PFM system begun in 1997 (with an emphasis on re-establishing fiscal discipline). In the early 2000s, efforts were undertaken to improve the strategic allocation of resources (introduction of an MTBF), but they failed to have a systemic impact due to a combination of over-ambitious goals, the absence of sufficient political will in the context of wider political challenges, and capacity constraint^.^^

305. A new round of ‘second-generation’ reforms is currently under preparation. To the degree that these reforms will be executed, they are likely to bring improvements with respect to the strategic allocation of resources (especially the implementation of an MTEF) and to technical efficiency. The earlier experience implies that comprehensiveness and sequencing of these further reform efforts will require careful attention, especially with regard to: (i)avoiding overburdening the MOF and other key institutions during the reform process; (ii) ensuring that refdrms cover implementation and do not stop at the adoption of new legislation; (iii)ensuring that initiatives are complementary and well linked (especially with regard to financial

53 See World Bank, Implementation Completion Report on IBTA2 (2006).

110 and sectoral management information systems; and complimentarity between information systems and underlying processes); (iv) ensuring that reforms are well adapted to the context of Tajikistan.

Macro-fiscal Discipline

306. The use of conservative budget estimates and the ability to limit demands on the budget have been important assets in the context of a very resource-constrained and uncertain economic environment in Tajikistan. However, there is some concern that the commitment to fiscal discipline is weakening with the contraction of large external loans to fund infrastructure projects. A core Treasury system has been established and equipped with the requisite tools for managing budget execution and reporting. While the situation has been stable in recent years, important risks remain. Tax revenue is still rather low, in part due to tax exemptions as well as to weaknesses in tax administration (and resulting high tax arrears). Also, there are substantial quasi-fiscal activities by state owned enterprises (notably in the electricity industry), and the practice of issuing state guarantees for debts incurred by SOEs may be resumed (see PI-17). Monitoring of fiscal risks arising from SOEs is weak. Furthermore, after a period of debt-write offs that helped to reduce the debt burden, Tajikistan is currently acquiring substantial new external debt. The fiscal deficit and debt situation will require careful monitoring in order to ensure that the debt burden does not again become unsustainable.

307. Finally, the current lack of a multi-year expenditure framework poses risks; particularly regarding the insufficient estimation of recurrent costs associated with large investments that are being undertaken or planned.

Strategic Allocation of Resources

308. Open and policy-based competition for resources is not yet well developed in Tajikistan’s PFM system. The fragmentation of the Government at the central level and the evolving system of intergovernmental relations pose challenges for the policy based allocation of resources. Transparency and public discussion of spending priorities in the legislature, civil society, and the media are still in their infancy. The GOT needs to decide whether to enhance the role of the cabinet-level Budget Commission or to consider a different mechanism for the discussion of the budget at this level. Given the high share of investments, the discretionary part of the budget is substantial, and hence requires strategic decision-making. In this regard, there is also an urgent need for greater government leadership and policy debate on the management and allocation of aid resource^.'^ Furthermore, weaknesses in revenue collection that result in an overall reduction of resources should be part of the policy debate. Renewed efforts at establishing an MTEF have been initiated. The full implementation of an MTEF will be a medium-long term effort, in particular with regard to developing information on the cost of programs and realistically costed sector strategies.

309. In the last three years, spending increased on average by more than 8% during budget implementation. The uneven allocation of additional resources has resulted in

54 PIP funding has accounted for between 20 and 30 per cent% of total budgetary resources per year.

111 substantial shifts in the sectoral allocation of funds, thus making allocative decisions taken during budget planning less binding. The main assets of the current system are the relative orderliness of the budget planning and execution processes, which are an important prerequisite for the capacity to translate policy choices into actual fiscal allocations.

Operational efficiency

310. Operational efficiency is important given that public services have to be delivered with very limited resources. Some measure of predictability in resource flows to core activities has been achieved (while there still may be greater difficulties at local levels), and the internal control and reporting environment is in line with current technical capacities (including still limited automating of the Treasury system).

3 11. Key areas of concern with regard to operational efficiency are the following: first, procurement has yet to move from the stage of legal reform to full implementation in order to reap benefits in terms of efficiency. Secondly, institutions for internal and external audit are weak, do not cooperate, and are almost exclusively focusing on regularity in individual transactions, rather than on systemic issues and on the effectiveness with which public funds are used. A strategy for improving internal control and audit is under discussion, and discussions have recently intensified on how to improve the external audit function.

3 12. Furthermore, the expectation of additional funds becoming available during budget execution, distorts some of the incentive for budget organizations and their management to plan within a clear financial envelope. Currently, there is little transparency to citizens (as service users) regarding public funding allocations to front-line service providers such as primary schools or health facilities; and thus the potential for direct accountability between service users and service providers is not used.

D. PROSPECTS FOR REFORMPLANNING AND IMPLEMENTATION

3 13. A new round of PFM reforms and capacity building covering a range of areas is under discussion between the GOT and donors. This follows previous reform and capacity building efforts which have been of mixed success (see World Bank, April 2006); including the successful establishment of a core Treasury and budget management system since the late 1990s, and a range of legal reforms in the 2000s. Donor-supported efforts to reform the system of intergovernmental finances and to introduce an MTBF were not successful in the early 2000s.

3 14. The overall environment is evolving (greater availability of resources, gradually improving capacity, step-by-step familiarization with reform ideas and options) and therefore conditions for further reform are improving.

315. The key reforms and capacity building efforts currently under discussion include the following:

112 budget planning: establishment of an MTEF (including a strengthening of financial planning and management capacities in line ministries); revenue forecasting (already under way); introduction of an administrative budget classification;

budget execution: further development of the Treasury system, reform and expansion of internal control and audit;

reform of intergovernmental fiscal relations): per capita allocation of funds in social sectors; introduction of greater control of sector ministries over sector- related spending at local levels;

revenue administration;

reform of government structures and organization and of the public service;

further donor support for the development of an independent external audit function

Further donor support to develop the role of the parliament in scrutinizing all aspects of the budget process.

As this list reflects, the reform agenda is broad and ambitious and will require good prioritization and sequencing. As a signatory to the Paris Declaration, Tajikistan has committed itself to undertake improvements of its PFM system. Exercising government ownership and active government management of this reform agenda will be crucial in maximizing the benefits from the financial and technical support that is being offered by donors.

3 17. It will be important also to pay attention to links and interactions between the various components of the reform agenda that is under discussion. For example, the further development of the treasury system and changes in the role and capacities for public financial management in line ministries should be well coordinated; similarly, there are important linkages between reforms in sector planning and management and changes to the system of intergovernmental (fiscal) relations.

3 18. Furthermore, while reforms are rightly focused on achieving improvements with regard to strategic allocation of resources, there are concerns about the apparent ambitiousness of these reforms, given the starting position. How exactly the introduction of the MTEF will be sequenced is a vital consideration. For establishing a meaningful medium-term planning system, basic building blocks of the fiscal system need to be strengthened further, including improvements in revenue administration, budget classification, payroll controls, procurement, and internal and external audit.

319. The present tendency of substantial budget deviations at the sector level implies considerable levels of unpredictability even over a 12 month period. Until this unpredictability is reduced, attempts to program resources over the medium term will be futile. Furthermore, an important objective will be to find ways of integrating

113 capital and recurrent budgeting (with implications also for aid policy and management).

320. Even if not the entire reform agenda as outlined can be put on track in the short to medium term (2-3 years), it is likely that PFM systems in Tajikistan will see further improvements over the coming period. This should lead to improvements in repeat rounds of the PEFA assessment at least in those areas where substantive reform and capacity building efforts can be set in motion.

114 REFERENCES

Andrews, Matthew and Alister Moon (2003), PRSP and PEM Capacities in ECA PRSP Countries (mimeo), ECSP World Bank.

Aschauer, DA. (1 989), “Is public expenditure productive?”, Journal of Monetary Economics 23, 177-200.

Barro, R. J. (1990), “Government spending in a simple model of endogenous growth,” Journal ofPolitical Economy 98, S103-Sl25.

Barro, R. J. (1991), “Economic growth in a cross section of countries”, Quarterly Journal of Economics 106,407-43.

Barro, R.J. and Lee, J-W. (200 l),“International data on educational attainment: updates and implications”, Oxford Economic Papers 3, 541 -563.

Bearing Point. Information about the execution of the project “Reform of Economic Policy of Central Asian Republics (Tajikistan) since January 2003, no date.

Bose, N., Haque, M.E. and Osborn, D. (2003), Public expenditure and economic growth: A disaggregated analysis for developing countries’, Working Paper, University of Manchester.

De Broeck, M. and Koen, V. (2000), “The Great Contractions in Russia, the Baltics and the other countries of the Former Soviet Union: A View from the Supply Side”, Working Paper WPlOOl32, IMF.

Devarajan, S., Swaroop, V. and Zou, H. (1996), “The composition of public expenditure and economic growth,” Journal of Monetary Economics 37, 3 13-344.

Easterly, W. and Fischer, S. (1994), “The Soviet Economic Decline: Historical and Republican Data”, Policy Research Working Paper 1284, World Bank.

Easterly, W. and Rebelo, S. (1993), “Fiscal policy and economic growth: an empirical investigation,” Journal of Monetary Economics 32, 4 17-458.

Ebinger, Jane (2006), “Measuring Financial Performance in Infrastructure: An Application to Europe and Central Asia”, WPS3992, World Bank.

European UniordGovernrnent of Tajikistan (2006), Strategy for Developing Public Financial Control in the Republic of Tajikistan.

Government of Tajikistan (2005), Public Administration Reform Strategy.

Government of Tajikistan (2006), Self-assessment on the effectiveness of public expenditure management.

115 Green, C.J. and Murinde, V. (1998), “Modelling the macroeconomic policy framework for an emerging market economy”, The Manchester School, 66, 3, 302- 330.

Harrison, Graham, Esil Martinsen and Sundar Somodi (2005), “A Program for Tax Administration Reform”, Technical Report Fiscal Affairs Department, IMF.

Herrera, Santiago and Gaobo Pang (2005), “Efficiency of Public Spending in Developing Countries: An Efficiency Frontier Approach”, mimeo, World Bank.

Le Houerou, Philippe and Robert Taliercio (2002), “Medium Term Expenditure Frameworks: From Concept to Practise. Preliminary Lessons from Africa”, Africa Region Working Paper Series No 28, World Bank.

Ilolov, Mamdasho, and Khudoiyev Mirodasen (200 l),“Local Government in Tajikistan,” in: Muteanu, Igor and Victor Popa (eds.): Local Governments in Eastern Europe, in the Caucasus and Central Asia, Budapest: Open Society Institute.

International Monetary Fund (2006A), Republic of Tajikistan: Sixth Review Under the Poverty Reduction and Growth Facility. Country Report no. 06/62.

IMF (2006B), “Republic of Tajikistan - Ex Post Assessment of Longer Term Program Engagement”, IMF Country Report 06/63.

IMF and World Bank (2007), “Tajikistan: Joint Fund-World Bank External Debt Sustainability Analysis”, mimeo.

International Monetary Fund (2007). Republic of Tajikistan: Report on The Observance of Standards and Codes (ROSC), Fiscal Transparency Module.

Kronenberg, Tobias (2004), “The curse of natural resources in the transition economies”, Eco.nomics of Transition, Vol 12 (3), 399-426

Landau, D. (1986), “Government and economic growth in the less developed countries: and empirical study for 1960- 1980”, Economic Development and Cultural Change, 35(1), 35-75.

Lensink, R., Hermes, N. and Murinde, V. (2000), “Capital Flight and Political Risk”, Journal of International Money and Finance, Vol. 19, pp. 73-92. Levine, R. and Renelt, D. (1992), A sensitivity analysis of cross-country growth regressions, American Economic Review, 82, 942-63.

Lorie, Henri (2003) ‘Priorities for Further Fiscal Reforms in the Commonwealth of Independent States’ Working Paper WP/03/209, IMF.

Loukoianova E. and Unigovskaya, A. (2004), “Analysis of Recent Growth in Low- Income CIS Countries”, Working Paper WP/04/15 1, IMF.

Matovu, J. (2005), “Sources of Economic Growth” in Tajikistan: Selected Issues and Statistical Appendix, Washington, DC, IMF.

116 Neel, Gerard (2006), Technical Report (EU: TAJIKISTAN: Streamlining Policy Planning Processes: PRSP and PIP).

Pro-gramma (2006) Horizontal Functional Analysis of the State Executive Bodies of the Republic of Tajikistan, Preliminary Results (presentation).

Radev, Dimitar, John Zohrab, John Matovu and Sailendra Pattanayak (2005), “Public Financial Management and Treasury Reforms” Technical Report, Fiscal Affairs Department, IMF.

Segura-Ubiergo, Alex, Alejandro Simone and Sanjeev Gupta (2006), “New Evidence on Fiscal Adjustment and Growth in Transition Economies” Working Paper WP/06/244, International Monetary Fund.

Summers, Victoria and Katherine Baer (2003), “Revenue Policy and Administration in the CIS-7: Recent Trends and Future Challenges”, paper prepared for the Lucerne Conference of the CIS-7 Initiative.

Umarov, K. and Repkine, A. (2004), “Tajikistan’s growth performance: the first decade of transition” in G. Ofer and R. Pomfret (eds), The Economic Prospects of the C.Z.S.: Sources of Long Term Growth, Edward Elgar, Cheltenham.

Varoudakis, Aristomene (2006), “Public Finance, Growth and Governance in Europe and Central Asia”, Preliminary Findings for the 2007 ECA Regional Fiscal Study, World Bank.

Wood, A. and Mayer, J. (2001), “African Development in a Comparative Perspective”, Cambridge Journal of Economics 25, 369-394.

World Bank (2003), Country Procurement Assessment Report.

World Bank (2004A), Country Financial Accountability Assessment.

World Bank (2004B), Trade and Transport Facilitation in Central Asia.

World Bank (2004C), Central Asia: Regional Electricity Export Potential Study, Europe and Central Asia Region, World Bank.

World Bank (2005A), Republic of Tajikistan Poverty Assessment Update, Report NO:30853-TJ.

World Bank (2005B), Tajikistan: Trade Diagnostic Study, Report No 32603-TJ.

World Bank (2005C), Republic of Tajikistan: Health Sector Note, Report No 29858- TJ.

World Bank (2005D), Tajikistan: Public Expenditure and Institutional Review.

World Bank (2005E), Tajikistan Energy Utility Reform Review.

117 World Bank (2006A), “Energy Sector Development of Kyrgyz Republic”, presentation, to the Government and multilateral and bilateral financial institutions, Biskek, June.

World Bank, (2006B), Social Protection in the Electricity and Gas Sectors in Tajikistan: Policy Note.

World Bank (2006C), Europe and Central Asia Inpastructure Study.

World Bank (2006D), Implementation Completion Report: IBTAZ.

World Bank (2006E) Aid Effectiveness Review: Tajikistan.

118 Annex I. The Implications of Public Expenditure for Economic Growth in Tajikistan55

Victor Murinde

1. Introduction

The motivation of this study is to examine the implications of various forms of public expenditure for supporting the priority of the Tajik government to sustain high rates of economic growth. Average growth rates of 8-10% per annum recorded during 1998-2004 were largely driven by recovery of the economy from civil war (1992- 1997), with unused capacity being brought back into production once peace and security had been restored. Although output of the economy is still around 30% below its level in 1990, much of the capital stock of the productive sectors dates from the Soviet era and is ill suited for the demands of a market economy. There are already signs that the recovery is beginning to falter, with growth slowing to 7% in 2005. Consequently, sustaining high rates of GDP growth in future will require new investments or a combination of new investment and more productive use of existing factors of production.

It is anticipated that future growth will be influenced by the nature of public expenditure, through its impact on the factors that drive long term sustainable economic growth, such as human capital formation through education and health services, the macroeconomic environment, as well as construction and maintenance of public infrastructure. The Government is undertaking budgetary reforms, one of the aims of which is to improve the allocation of public expenditure so that it can support sustainable economic growth. Hence, the objective of the study is to model the growth effects of government expenditure, to identify the main constraints to economic growth and to assess their implications for public expenditure.

In spite of the growing research attention to the relationship between government spending and economic growth, a consensus has not emerged regarding what type of government expenditures are growth promoting for developing and transition economies. For example, some studies have found a significant and positive relationship between government size, as measured either by the level of total public expenditure or by the level of public consumption expenditure, and economic growth. However, some other studies have found a significant but negative relationship (see, in particular, Landau 1986; Grier and Tullock 1989; Barro 1990, 1991). Between these conflicting results, some other studies have found the relationship to be insignificant (for example, Levine and Renelt 1992).

In general, the conflicting results have been attributed to differences in empirical models, data sets, econometric techniques, and sample economies. The choice of sample economies is particularly important because most of the existing findings on the growth effects of public spending are drawn either from the experiences of developed countries or large samples pooled from developed, developing and transition economies, with the exception of the work by Landau (1986) and Devarajan et al. (1 996)). What would be more helpful, in the context of this study, is to obtain

55 This background study for the PPER was prepared with the financial support of DFID.

119 more findings that are based on serious empirical work on the institutional structure and the process by which public expenditure variables influence economic growth in transition economies only, in contrast to broad cross-country samples dominated by developed countries. In addition, it is argued that the differences arise from the use of different sets of conditioning variables and initial conditions which are used by various studies on the impact of fiscal variables on endogenous growth (see, Levine and Renelt 1992). Further, it is argued that the conflicting results partly arise from failure by researchers to acknowledge the role of the Government budget constraint in growth models that include fiscal variables. The argument is that in order to undertake a meaningful evaluation of the effects of taxes or expenditures on economic growth, it is important to recognise the flow-of-fund concept about the sources as well as the uses of funds in the growth process. For example, while the studies by Landau (1986) and Devarajan et al. (1996) are important to this study because they focus on developing countries, they suffer from the possibility of omission bias, in the sense that they only focus on the expenditure side of the budget constraint and ignore the revenue side.

Hence, the main objective of this study is to empirically identify the specific components of government expenditure which significantly impact on economic growth, and which could be used as part of a public expenditure review in order to target sustainable economic growth. We invoke endogenous growth theory and apply it in two steps. The first step is to use the model to study the growth effects of government expenditure in eight CIS countries (Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyz Republic, Moldova, Mongolia and Tajikistan) for the period 1996-2004 and identify the patterns of government expenditure that are associated with economic growth regionally. The second step is to use Seemingly Unrelated Regression Equations (SURE) techniques to isolate the coefficients specifically and hence identify the types of public expenditure which are most likely to support sustained economic growth in Tajikistan.

Clearly, the emphasis of the study is on the government expenditure side; the government revenue side and the financing variables, such 'as government budget surplus/deficit and tax revenue, are included in the model in order to incorporate the government budget constraint and to avoid the coefficient biases that would result from their omission. We test the important of each government expenditure variable on economic growth; we also test for the variables as a group.

We explicitly recognize that Tajikistan is a resource constrained transition economy, hence it is important to investigate the impact of each sector-specific government expenditure, for example, government expenditure on education or health. The information from sector-specific government expenditure is useful for informing policy, especially where the allocation of limited public resources between the sectors is a critical issue Le., shall the Government spend more on health or on education, given that resource constraints mean that it cannot spend more on both.

The remainder of this paper is structured into four sections. Section 2 highlights the institutional issues, including geography and historical legacy, which may constrain the possible impact of Government expenditure on economic growth in Tajikistan. The body of existing knowledge on the growth effects of public expenditure is surveyed in Section 3, to provide a point of departure for our specific study on

120 Tajikistan and to provide contrasting experiences in other countries. Section 4 presents the model and reports the estimation and testing results. The main findings, conclusions and policy recommendations are presented in Section 5.

2. Institutional Issues: Economic Geography and Historical Base

Tajikistan shares a common Soviet economic heritage with other CIS countries, and this institutional background influences the sample selection criteria for the empirical work in this report. The argument is that the economic geography and historical developments are important institutional factors in constructing a regional sample for studying the drivers of economic growth in the area. A sample that recognizes both factors includes the CIS-7, namely Tajikistan, Azerbaijan, Uzbekistan, the Kyrgyz Republic, Kazakhstan, Armenia, Georgia and Moldova, as well as Mongolia, which although not a CIS member, is a landlocked country which had a state controlled economy. We exclude some regional neighbors of Tajikistan which have different institutional features, including Afghanistan, Pakistan, India and China. Afghanistan is a failed state which has suffered civil conflict for more than two decades . India and Pakistan are much more established as market-driven countries rather than transitional economies. China is a transition economy but its size makes it very different from Tajikistan.

An important comparative experience of these neighboring countries relates to the changes in economic growth rates following independence from the Soviet Union and formation of CIS in the early 1990s. Like most other CIS countries, Tajikistan suffered a sharp decline in economic growth in the six years following independence in 1991. The economy recovery began in the late 1990s.

Broadly, there are at least two interesting lines of enquiry which are important for empirical work and policy design. The first issue is to disentangle the factors which explain growth, especially in the period since 1996. In view of the existing literature, surveyed briefly in this study, this line of enquiry suggests an endogenous growth model. The second line of enquiry, which is strongly related to the first one, is to identify the most promising public expenditure interventions which will accelerate economic growth. Overall, on the basis of the institutional and historical factors of Tajikistan, and given the existing knowledge of other transition economies, it is plausible to suggest the following key candidate avenues and critical factors for economic growth in Tajikistan during the next 10 years. . Human Capital and Education: Including education at the primary school and secondary school levels. . Life Expectancy and Health: These include health services and expansion of the health and sanitation sectors. . Capital expenditure and private investment: The main type of investment required is mainly private, such as in agro-exports. Labor intensive exports increase wages and value added, and hence growth. . Miwant workers and remittances: Remittances form Tajiks working abroad may be expanded to supplant Foreign Direct Investment (FDI) and exogenous Dinflows. The type of intervention required includes new investment in health, education, and other social service sectors.

121 Physical and human capital in Taiikistan: Current indications are that physical and human capital remains well above the level that is needed to produce the present GDP per capita of US$310; it is also far above the level that can be sustained with the present GDP per capita. It is not clear, though, whether physical and human capital is used efficiently. Indications are that it is not, but we need to invoke a reliable measure of the efficiency of physical and human capital and then apply it to obtain serial indicators over time.

3. Public Expenditure and Economic Growth: What Do We Know?

From the start, it is useful to note that the strand of literature on public expenditure and economic growth is slightly different from the strand of literature on isolating the factors that explain economic growth in a neo-classical growth model. In the case of the latter, researchers tend to use a growth accounting framework to analyse the sources of economic growth; for example, De Broek and Koen (2000) and Loukoianova and Unigovskaya (2004) for CIS countries and Matovu (2005) for Tajikistan. In general, the growth accounting framework disaggregates GDP growth into three components: the growth in the capital stock, the growth in the labour force (and labor productivity) and the growth in total factor productivity (TFP). TFP is a residual which picks up all of the influences on the productivity of factors of production, as well as measurement errors. The fundamental assumptions underpinning the growth accounting exercises arise from a constant return to scale production function with elasticities of output with respect to labor and capital.

The strand of literature on public expenditure which is relevant to this report mainly derives from endogenous growth models. As noted by Le and Suruga (2006), there have been quite a number of empirical studies analysing the impact of public expenditures on economic growth. However, the findings are not unanimous. On the one hand, some findings support the idea that public investment has a positive impact while public consumption has a negative impact on economic growth. For example, the findings by Aschauer (1989), Barro (1991) and Easterly and Rebelo (1993) support the argument that government expenditures spent on main physical infrastructure have a significant positive impact on private-sector productivity. In addition, using a sample of 39 developing countries during the period 1990-2000, Gupta et al. (2002) find a positive relationship between growth and fiscal policy (government budgetary balance, the financing of budgetary deficits, and expenditure composition), and conclude that in order to achieve a sustained fiscal adjustment, governments should reallocate expenditures on wages and salaries to more productivity uses such as capital expenditures.

On the other hand, some researchers have challenged the above findings by specifically arguing that the relationship between government expenditure and economic growth is not positive per se, but tends to be sector-specific and ultimately depends on the type of Government expenditure. For example, Landau (1986) categorises government expenditure into five groups (consumption, education, defence, transfers and capital expenditure) in order to conduct a comprehensive study of the impact of government expenditure on economic ,growth, using cross section data of 96 countries over various time periods between 1961 and 1976. It is found that each type of government expenditure has either significant negative or insignificant positive effect on economic growth. In addition, Grier and Tullock

122 (1989) use a panel data analysis on 24 OECD countries in the period 1951-1980 and 89 other countries in the period 196 1-1980 and find similar inconsistent relationships between public expenditure and economic growth as those found by Landau (1986). Moreover, in a sample of Asian countries over the period 1970-1990, Devarajan et al. (1996) find that an increase in the share of current expenditure has positive and statistically significant growth effects, while Government capital expenditures (transport and communication) have either a negative or insignificant positive impact on economic growth, in sharp contrast to findings by previous researchers, such as Aschauer (1989) and Barro (1990) who regard these expenditures as productive.

A main problem with the existing literature stems from the variations in the models used, estimation and testing techniques and the heterogeneity of the countries included in the sample countries. Another problem arises from the differences in the motivations of the studies, ranging from attempts to model growth and convergence to motivation to tease the main variables that drive growth which could be used as drivers for public expenditure review. The latter is consistent with the objectives of this study. A relevant study along these lines is the recent work by Bose, Haque and Osborn (2003), who examine the growth effects of government expenditure for a panel of thirty developing countries over the decades of the 1970s and 1980~~with a particular focus on sectoral expenditures. It is found that the share of government capital expenditure in GDP is positively and significantly correlated with economic growth, but current expenditure is insignificant. It is also found that, at the sectoral level, government investment and total expenditures in education are the only outlays that are significantly associated with growth once the budget constraint and omitted variables are taken into consideration.

Hence, on the basis of the foregoing, what we know is that a substantial volume of empirical research has been directed towards identifying the elements of public expenditure which influence economic growth. We also note that the literature varies in terms of data sets, econometric techniques, and often tends to yield conflicting results and hence policy recommendations. Following Bose, Haque and Osborn (2003), we aim to address the shortcomings in the existing literature and also draw from our knowledge of the institutional structure of the Tajik economy. For example, we explicitly recognise the implications of the government budget constraint in our empirical work, specifically by considering both the sources and the uses of funds simultaneously for a meaningful evaluation of the effects of taxes or expenditures on economic growth.

4. Econometric Analysis and Results

4.1 An Endogenous Growth Model with Government Expenditures

We model the growth effects of government expenditure for a panel of eight countries (seven CIS countries plus Mongolia) during the period 1996-2004, when these countries were experiencing economic recovery. We invoke a methodology which differs from previous research in two ways. First, we explicitly recognize the role of the government budget constraint. Second, we control for possible biases arising from omitted variables. Specifically, we aim to isolate the specific components of government expenditure which significantly impact on economic growth.

123 We invoke an endogenous growth model following Bose, Haque and Osborn (2003). We specify a baseline model which captures endogenous growth variables as well as the government expenditure variables, as follows:

J K N GROWTH,, = a,,, + xa,CONDS,,, -k xPkGOVTSPENk,,+ x ynOTHERn,+E,, (1) /=I k=l n=l where the superscripts denote country i at time t for each vector of variables; GROWTH represents growth of real GDP per capita; the vector CONDS captures the main explanatory variables that commonly appear as conditioning variables in endogenous growth models; the vector GOVTSPEN represents the explanatory variables which capture government expenditure, hence these are the variables that are of particular interest for the study, namely government expenditures and their major components at aggregate and sectoral levels, expressed as percentages of GDP; the vector OTHER includes all other explanatory variables which are added to the model in order to test the sensitivity of the model, thus we include variables that often have been included in previous studies as indicators for monetary polices, trade polices, and market distortion; and E is a stochastic term.

We follow Bose, Haque and Osborn (2003) to incorporate the government budget constraint in the model. It is argued that empirical investigation of the impact of government spending on growth, which does not incorporate a government budget constraint, is vulnerable to bias that arises from omitted variables. We consider a simple government budget constraint specified in Green and Murinde (1 998), as follows:

GS - TX = AMON + BOND + AFM(2)

Where GS = total government spending, TX = total revenue (including tax as well as well as non-tax revenue), MON = narrow money; BOND = government domestic borrowing, and FRN = foreign borrowing. The Government budget identity specifies that Government expenditure must equal Government revenue or otherwise the balance would be funded by the issue of new money, borrowing domestically or borrowing abroad. For this matter, we cannot include all elements of government spending as well as elements of government revenue; we need to drop at least one element. We therefore include in the model at least one government expenditure variable together with the tax revenue variable. Altogether, we include only the following variables: Government capital expenditure, Government recurrent expenditure, Government expenditure on education, Government expenditure on health, Government expenditure on education and health, domestic revenue and Government budget deficit or surplus; all the variables are given as percentage of GDP.

4.2 Data and Measurement

The transition economies we take as the sample are members of the CIS who share not only geographical proximity with Tajikistan but also trong regional economic ties: these are Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan and Moldova. We also include Mongolia in the sample. Hence, we focus on these 8 transition

124 economies. Our dataset is retrieved from World Bank online data supplemented by World Bank Country Economic Reports and Public Expenditure Reviews, for the period 1996-2004. Hence, overall, we constructed a panel of 72 observations.

The measurement of the variables in the model is as follows. The endogenous variable of the model, GROWTH, is measured as growth rate of real GDP per capita.

The explanatory variables for including in the vector COND are consistent with those used in recent work by, for example, Levine and Renelt (1992), Barro (1991) and Bose, Haque and Osborn (2003). Thus, the vector CONDS includes the main variables that commonly appear as conditioning variables in endogenous growth models: LFEX=Log of life expectancy; PRIM = Primary school enrolment ratio; SECS = Secondary school enrolment ratio; HCGEN = initial human capital (PS) variable; NINV = Non government investment (a proxy for private investment); ILFEX = Log of life expectancy at initial period as at 1996; IGDPL = Log of initial GDP per capita as at 1996; we include this variable in order to control for possible effects of convergence on output growth. The variable HCGEN is constructed in the context of Landau (1986) as the weighted sum of the initial enrolment ratios in primary and secondary schools. The weights are I for primary school enrolment ratio and 2 for secondary school enrollment ratio, hence, these weights are approximations to the relative values of two types of education. We measure PRIM and SECS using school enrolment ratios rather than the average schooling years because the latter were not available; moreover, previous studies, such as Easterly and Rebelo (1993) and Bose, Haque and Osborn (2003) use school enrolment ratios in the sense that the ratios are probably better available measures of investment in education than the average schooling years. In the analysis, we use the HCGEN variable rather than its subcomponents because of the high milticollinearity between the separate enrolment rates of PRIM and SECS. Overall, therefore, the variables included in vector COND capture the key argument of the endogenous growth literature, namely that human capital and institutional factors are important determinants of economic growth in Tajikistan.

The vector GOVTSPEN represents the variables which capture government expenditures and their major components at the aggregate and sectoral levels, expressed as percentages of GDP. We include the following: CEXP = Government capital expenditure; REXP = Government recurrent expenditure; EEXP = Government expenditure on education; HEEP = Government expenditure on health. We also tried combinations of EEXP and HEEP to capture government expenditure on services; we also constructed total government expenditure as the sum of CEXP and REXP.

The vector OTHER includes all other variables which are added to the model in order to test the sensitivity of the model. We follow Easterly and Rebelo (1993) and define the explanatory variables in the vector OTHER to include variables that often have been included in previous studies as indicators for monetary polices, trade polices, and macroeconomic stability; we included the following: BMON = Broad money (M2) (% of GDP); TTDE = Trade ratio (export plus import as % of GDP); ALIT = Adult literacy (as % of total population); INFL = Inflation rate, computed as the rate of change of the consumer price index. The variable BMON is included in order to control for the effects of monetary policy. The variable TTDE is included in order to

125 control for the degree of openness. It has been shown by previous studies, notably Levine and Renelt (1992) that the two variables are significantly correlated with economic growth.

In addition, as discussed above, we include variables to capture the government budget constraint, namely: DREV = Domestic revenue; FDEF = Government surplus / deficit. Because the government budget constraint is an identity (see equation 2), we need to exclude an element of the identity and thus avoid perfect collinearity. Another context for the inclusion of DREV is in order to make a proper assessment about the growth effects of public spending, because while the provision of public goods is growth-enhancing, the distortionary taxes that need to be raised to fund the provision of the public goods may have growth-diminishing effects.

As a summary of the above, Appendix Table A1 presents the variables used in the empirical work, the measurement of the variables and the sources of data.

Table 1: Descriptive statistics of the main variables in a panel of commonwealth

HCGEN 129.46762 109.93558 -0.10328 -1.76059 CEXP 0.059748 0.050000 0.94008 0.047908 REXP 0.20706 0.079493 0.46280 2.22195 ALIT 9.61250 29.49781 2.77749 5.87745 PRIM 49.54023 43.65270 -0.24769 -1.94013 SECS 3 9.963 69 38.54898 -0.03 9259 -1.95863 *

L EEXP 0.035046 0.0241 83 0.6 1222 0.096532 HEEP 0.01 8339 0.0 14560 1.06593 1.0047 1 FDEF -0.043 103 0.039203 -0.27623 -0.30656 1 BMON 10.14413 10.077471 I 1.94156 14.95123 I TTDE 0.8 1803 0.40459 -0.4 1098 0-503 06 INFL 0.12638 0.15527 1.935 15 7.18600

Note: GROW = Average growth rate in GDP per capita; LFEX = Log of life expectancy; ILFEX= Life expectancy at initial period; IGDPL = initial GDP per capita, as at 1996; NINV = Non government (Private investment) (% of GDP); DREV = Tax revenue (% of GDP); HCGEN = initial human capital (PS) variable; CEXP = Government capital expenditure (% of GDP); REXP = Government recurrent expenditure (% of GDP); ALIT = Adult literacy (as % of total population); PRIM = Primary school enrolment ratio; SECS = Secondary school enrolment ratio; EEXP = Government expenditure on education (% of GDP); HEEP = Government expenditure on health (% of GDP); FDEF = Government surplus / deficit (% of GDP); BMON = Broad money (M2) (% of GDP); TTDE = Trade ratio (export plus import as % of GDP); INFL = Inflation rate, computed as the rate of change of the consumer price index.

126 HCGEN, the initial human capital variable is constructed in the context of Landau (1983) as the weighted sum of the initial enrolment ratios (%) in primary and secondary schools. The weights are 1 for primary school enrolment ratio and 2 for secondary school enrolment ratio, hence, these weights are approximations to the relative values of two types of education. We use the HCGEN variable because of the high multicollinearity between the separate enrolment rates of PRIM and SECS.

127

4.3 Patterns of Government Expenditure and Economic Growth Variables in CIS, including Tajikistan

We first check the data for normal distribution and other univariate properties, using the descriptive statistics reported in Table 1. We also check the bivariate relationships of the endogenous variable and the explanatory variables, using a correlation matrix reported in Table 2. Table 1 shows that the endogenous variable has a reasonable mean economic growth rate of 5% but the high standard deviation of4.5% suggests dramatic swings in the growth rates of the sample economies during 1996-2004. In general, however, the main government expenditure variables show a normal distribution in the sense that they are generally symmetric about the mean values. Overall, there are very limited problems of skewness and kurtosis in the data for most variables, according to the statistics reported in Table 1, except perhaps for the inflation rate variable. \ The correlation matrix in Table 2 reports the bivariate relationships of all variables in the model, and thus provides the first insight into how each explanatory variable is correlated with the endogenous variable as well as each of the other explanatory variables in the model. These correlations should give us a first glimpse into patterns of relationships in trying to pin down the impact of each element of government expenditure on economic growth. Hence, Table 2 should be read against Appendix A. T he latter presents 11 charts of scatter diagrams which show the patterns of government expenditure variables and other variables of the model against the economic growth variable for the eight sample countries for the period 1996 - 2004.

Some interesting patterns emerge from these charts, regarding the relationship between government expenditure variables and economic growth in the transition economies. Three main patterns have important policy implications, as highlighted below.

First, EEXP and GROWTH show a pattern that supports the argument of higher economic growth even at low levels of government expenditure on education, hence consistent with the human capital argument. In addition, we find a pattern of high primary school enrolment rate (PRIM) and high economic growth, which is consistent with human capital arguments. Also, there is a pattern of high secondary school enrolment rate (SEC) and high economic growth, which is consistent with human capital arguments. Hence, primary school education and secondary school education are important for economic growth. Again, this is the human capital argument and suggests government intervention to build human capital through education. The correlation coefficients reported in Table 2 support this pattern. The correlation between economic growth rate per capita and the composite human capital variable (HCGEN) is 0.4699; even the components of human capital are highly correlated with growth, primary school enrolment rate at 0.3361 and secondary school enrolment rate at 0.4798. In addition, government expenditure is also correlated with economic growth by 0.293 1. The important policy implication, subject to further investigation in the empirical part of this study, is that government expenditure in education as a proportion of GDP (EEXP) is important for economic growth.

Second, CEXP and GROWTH have a pattern showing positive associated between economic growth and government capital expenditure (CEXP), even at lower levels of capital expenditure, there is positive growth. Table 2 shows a correlation coefficient of 0.15 13 between CEXP and GROWTH. A higher coefficient, though, is reported for non government investment (NINV) and economic growth at 0.2630, which suggests that capital expenditure is positively correlated with economic growth, but that it is private

129 investment that has a higher impact.

Third, there is a high correlation coefficient of 0.604 between the budget surplus and economic growth, which emphasizes not only budgetary discipline but also the importance of an ‘enabling macroeconomic environment’.

Overall, the patterns in the charts in Appendix A are generally consistent with the results from preliminary data analysis presented in Tables 1 and 2. The patterns will be more rigorously investigated using econometric methods in the remainder of this section.

4.4 Initial Estimation and Testing Results

As the starting point in the analysis, we are interested in investigating whether the government expenditure variables in vector GOVTSPEN are significantly correlated with economic growth after controlling for the variables in the COND vector. Hence, we run a restricted version of the model in equation (I),where the gamma coefficients are constrained to be zero. The regressions are run in such a way that we add one GOVTSPEN variable at a time. These initial regressions are intended to capture a central idea of the endogenous growth literature namely that human capital is an important determinant of economic growth. In addition by including initial GDP in these initial regressions, we are trying to explore the effects of convergence on growth.

Table 3: Estimation and Testing Results from an Endogenous Growth Model for a Panel of Commonwealth Independent States (CIS) including Tajikistan: The Impact of Initial

Variable Model Model Model Variant Model Model Variant 1 Variant 2 3 Variant 4 Variant 5 C 0.038 -61.977 -68.624 -52.416 -123.171 (2.8 3 0) (-2.8 17) (-3.194) (-2.492) (-2.3 15) [0.006] [0.007] [0.002] [0.15] [0.024] CEXP 0.509 0.470 (4.639) (3.4 19) [O.OOO] [0.001] EEXP 0.510 0.583 (1.437) (1.653) I [0.156] I [0.103] ILFEX I 14.601 1 16.167 I 12.354 I 28.983 I (2.820) l(3.199) I (2.496) I (2.316) I HCGEN I 0.628 I 0.676 I (1.274) I I (1.407) 1 1 I [0.208] I [0.164] FEDEF (7.230) (3.871) (6.207) (3.742) (6.201) [O.OOO] [O.OOO] [O.OOO] [O.OOO] [O.OOO] AZ 0.028 -0.03 1 0.106 (1.983) (-1.996) (1.967) [0.052] [0.294] [0.054] INFL 0.025 (0.888) [0.3 781 GE 0.055 -0.039 -0.04 1 -0.04 1

130 I (3.583) I (-1.996) (-2.100) (-2.1 15) I [0.001] I [0.050] [0.0401 [0.038] KA I 0.035 I 1.287 1.427 1.087 2.660 I (2.310) I (2.780) (3.157) (2.456) (2.306) I [0.024] I [0.007] [0.002] [0.017] [0.024] KY I 0.026 I 0.748 0.830 0.642 1.58 1 I (1.992) I (2.754) (3.134) (2.451) (2.3 02) I [0.051] I [0.008] [0.003] [O.O 171 [0.025] ML I 0.326 I 0.678 0.759 0.587 1.530 I (0.228) I (2.572) (2.950) (2.290) (2.254) [0.004] [0.025] [0.028] MO I 1.141 1.272 0.980 2.389 I I (2.716) (3.106) (2.414) (2.2 84) J0.0031 [0.019] [0.026] TJ I 1.449 1.609 1.225 2.997 (3.121) (2.425) (2.290) ro.0031 ro.0 I81 r0.0251 SSR I 0.060 1 0.059 0.060 0.06 1 I 0.057 SER I 0.031 I 0.031 0.03 1 0.03 1 I 0.030 ARS I 0.539 I 0.516 0.5 10 I 0.511 ,I 0.549 Note: Below the estimated coefficients are reported tl t-values in (.) brackets and the p-value in [.] brackets. In the above regressions, the endogenous variable is GROW = Average growth rate in GDP per capita. The rest of the variables are defined as follows, as in Appendix Table A 1 : CEXP = Government capital expenditure (% of GDP); REXP = Government recurrent expenditure (% of GDP); EEXP = Government expenditure on education (% of GDP); HEEP = Government expenditure on health (% of GDP); DREV = Tax revenue (% of GDP); FDEF = Government surplus / deficit (% of GDP); NINV = Non Government (% of GDP); BMON = Broad money (M2) (% of GDP); PRIM = Primary school enrolment ratio; SECS = Secondary school enrolment ratio; HCGEN = initial human capital (PS) variable; LFEX = Log of life expectancy; TTDE = Trade ratio (export plus import as % of GDP); ALIT = Adult literacy (as % of total population); INFL = Inflation rate, computed as the rate of change of the consumer price index; IGDPL = Log of GDP per capita at initial period.

Table 3 presents the estimation and testing results from the endogenous growth model with government spending variables Le., with only the explanatory variables in vectors COND and GOVTSPEN. Five model variants are presented to allow for inclusion of all government spending variables. It is shown that of the variables in the COND vector, only initial life expectancy variable shows a statistically significant relationship with economic growth in model variants 2-4. This result is consistent with the endogenous growth literature in which life expectancy is a crucial determinant of economic growth; the sign in each of the five model variants is positive, suggesting a positive relationship between life expectancy and economic growth. The variable for initial human capital is found to have a positive effect on growth in model variants 2 and 4; although the coefficients are not statistically significant.

However, we do not find any evidence to show that there is convergence among this group of transition economies. This is not surprising, perhaps, given that these economies are still in the process of transition to a market economy.

In terms of the explanatory variables in the GOVTSPEN vector, the most significant results are obtained for Government capital expenditure (CEXP) in model variants 1 and 5, and Government expenditure in the education sector (EEXP) in model variants 2 and 3. The results for Government expenditure on health (HEEP) and recurrent expenditure (REXP) were not statistically significant and were dropped from the model variants 1-5.

131 The significant association between the share of Government capital expenditure in GDP and economic growth is not entirely surprising in the light of the conclusions drawn by previous studies (see, for example, Easterly and Rebelo 1993; Cashin 1995; Fuente 1997). Our results are also consistent with those of Bose, Haque and Osborn (2003) in a panel study that included capital expenditure in the regression for developing countries.

Our results which show a positive relationship between Government expenditure on education and economic growth differ from conclusions drawn by previous studies, irrespective of whether these are based on data for a large pool of countries or developing countries (e.g., Landau 1986; Devarajan 1996). These earlier results indicate that the association of this variable with growth is either insignificant or non-robust. We find expenditure on education to be significant (albeit weakly); it is shown that one percentage point increase in government expenditure on education in relation to GDP is associated with an increase in the average growth rate of real GDP per capita by 0.5 percentage points. This effect may be explained in terms of the fact that increased expenditure on education amounts to an investment in human capital, which raises the potential for economic growth, according to the endogenous growth literature.

The model reported, whose estimation and testing results are reported in Table 3, is augmented with variables from vector OTHER and from the government budget constraint. The inclusion of these additional variables also amounts to a check for the robustness ofthe explanatory variables in the GOVTSPEN vector.

4.5 A Check for the Robustness of the Results

We now conduct checks for the robustness of our initial findings by augmenting the model with explanatory variables from the vector OTHER in the model (equation 1). The variables included in this vector are suggested by the previous literature. For example, we include the ratio of broad money (M2) to GDP (BMON) and the trade share of GDP (TTDE), following Easterly and Rebelo (1993) and Levine and Renelt (1992). In our study, the ratio of broad money to GDP is included in order to control for the effects of monetary policy, especially as these countries are in transition to a market economy. The trade share of GDP is included in order to control for the degree of openness, given that economic growth in Tajikistan seems to largely depend on its trading links with the neighbouring countries. We also include the growth rate of the terms of trade (TTDE2) as an additional variable. This variable captures the adverse effect of trade shocks that Tajikistan and a number of transition economies in the region experienced during the sample period. Previous research has shown that the growth rate of the terms of trade is significantly and positively related to economic growth in developing countries (see, for example, Deverajan et al. 1996).

132 Table 4: Robustness Check on the Results from an Endogenous Growth Model for a Panel of Commonwealth Independent States (CIS) including Tajikistan

Variable Model I Model I Model I Model Variant 1 I Variant 2 I Variant 3 I Variant 4 CONSTANT 0.1 14 I 0.103 I 0.094 I 0.864 (3.572) I (3.194) I (2.800) 10.154 [0.001] I [0.002] I [0.007] I [0.878] LFEX 0.8 10 I 0.234 I 0.328 I 0.847 (0.043) (0.121) (0.0 17) (0.441) [0.966] [0.904] [0.9 861 [0.661] NINV -0.896 0.01 1 -0.025 0.029 (-0.102) (0.125) (-0.287) (0.3 3 7) [0.919] I [0.901] I [0.775] I [0.738] DREV -0.205 I -0.083 I -0.089 I -0.035 (-1.076) 1 (-0.437) I (1.267) I (-1.666) [0.286] I [0.664] I [0.211] I [0.102] HCGEN 0.678 I 0.776 I 0.616 I 0.371 (1.237) (1.390) (1.1 00) (0.65 4) [0.221] [O. 1701 [0.276] [0.516] CEXP 0.740 (2.401) I [0.020] HEEP I 0.233 (0.338) [0.737] EEXP 0.772 0.90 1 (1.559) (1.774) [ 0.1251 [0.82] FDEF 0.642 0.561 0.615 0.546 (0.201) (2.789) (3.098) (2.885) [3.196] [0.007] [0.003] [0.006] BMON 0.0 19 0.054 0.156 0.213 (0,139) (0.385) (0.977) (1.391) [ 0.8901 [0.702] [0.333] [O. 1711 BMON(-1) -0.146 -0.087 (-1.396) (-0.804) [O. 1691 [0.425] TTDE 0.969 -0.199 0.026 0.02 1 (0.244) (-0.050) (0.642) (0.5 3 7) [0.808] [0.960] [OS231 [0.594] INFL 0.025 0.0 17 0.675 -0.010 (0.805) (0.566) (0.201) (-0.3 10) [0.424] [0.574] [0.842] [0.758] INFL (-1) 0.045 0.027 (1.463) (0 -030) [O. 1501 [0.372] AZ -0.029 -0.026 -0.030 0.055 (-1.142) (0.942) ( 1.1 68) (1.345) 10.2581 [0.3 501 [0.24 81 [0.185] GE -0.926 -0.020 0.0 12 0.061 -0.471 (-1.027) (0.563) (1.544) [0.639] [0.3 091 [0.5 761 [O. 1291 KA -0.032 -0.034 -0.028 0.055

133 (-1.747) (-1.790) (-1.464) (1.702) [0.086] [0.079] [O. 1491 [0.095] KY -0.044 -0.029 -0.043 0.0322 (-2.00 1) (-1.407) (-1.895) (1.303) [ 0 .O 5 01 [O. 1651 [0.064] [O. 1991 ML -0.065 0.067 -0.045 0.069 (-1.356) (-1.354) (-0.908) (1.2 15) [0.181] [O. 1811 [0.3 681 [0.230] MO -0.060 -0.048 -0.066 (-1.980) (-1.604) (-2.055) [0.053] [0.114] [0.045] TJ -0.068 -0.05 18 -0.070 0.017 (-2.004) (-1.569) (-1.988) (0.455) [0 -05 01 [O. 1221 [0.052] [0.65 11 SSR 0.060 0.062 0.052 0.049 SER 0.033 0.034 0.320 0.03 1 ARS 0.463 0.441 0.05 10 0.521

The test for robustness, which is applied in this study, follows Levine and Renelt (1 992) and Lensink, Hermes and Murinde (2000), among others. According to this test, the variables in the GOVTSPEN vector are robust if they can neither lose their statistical significance nor experience a sign reversal when other variables have been added to the model i.e. the results of the variables in the vector are not sensitive to the addition of new explanatory variables. In this context, the results reported in Table 4, specifically the results for CEXP in model variant 4 and EEXP in model variants I and 3, confirm that there is a robust relationship between the explanatory variables in the GOVTSPEN vector and economic growth and the rest of the transition economies in the sample. Moreover, we find that the statistical significance of EEXP increases, further emphasizing the role of government expenditure on education in economic growth. The initial results regarding the role of government capital expenditure in economic growth are also confirmed. Specifically these results seem to hold for model variant 3 in which the country dummy variable is statistically significant. The evidence therefore further underscores the role of government expenditure on education in economic growth, with specific reference to Tajikistan.

We extend the test for robustness by including explanatory variables from the government budget constraint. As we earlier argued, previous studies on the relationship between government spending and economic growth, which omit the government budget constraint, are potential biased. The results incorporating the variables from the Government budget constraint, are reported in Table 5.

It is shown in Table 5 that the effects of including the budget constraint, and also jointly considering significant expenditure components, improves dramatically the performance of the initial variables in vector COND and the explanatory variables in vector

134 GOVTSPEN. In particular, the new results identify two GOVTSPEN variables, which are important for economic growth in the sample transition economies, namely, government expenditure on education (model variant 1) and government capital expenditure (model variant 2). These results imply that the earlier findings reported in Tables 4 and 5 for the variables in the vector GOVTSPEN are not distorted by the addition or omission of the growth of the terms of trade (TTDE), the ratio of broad money (M2) to GDP (BMON) and the variables for the government budget constraint, specifically, budget deficit or surplus (FDEF) and domestic revenue (DREV). Of particular interest to this study, is that the results identify Government expenditure on education as the key sectoral allocation for public expenditure support in order to promote economic growth, as can be noted from the significance of the country dummy in model variant 1.

135 Table 5: Further Robustness Check on the Evidence from an Endogenous Growth Model for a Panel of Commonwealth Independent States (CIS) including Tajikistan Variable Model Variant 1 1 Model Variant 2 1 Model Variant 3 C 0.103 1 0.016 1 0.073 (3.649) 1 (0.407) 1 (0.015) J0.0011 [0.658] [O.OOO] NINV(-1) 0.036 0.065 (0.657) (1.43 1) Josi3j io. 1571 DREV -0.236 -0.214 (-1.520) (-1.705) io.1341’ I [0.093] HCGEN 0.866 (1.746) J0.0861 EEXP 1.033 0.321 (2.435) (1.386)

FDEF 0.644 I 0.699 (3.946) (6.546) JO.OOO] [O.OOO] INFL (-1) 0.495 0.022 (1.786)

CEXP

FDEF (5.481)

BMON (1.332) [O. 1881 BMON(-I) -0.103 (- 1.714) [0.091] A2 -0.026 0.077 (-1.682) (2.235) 10.0981 [0.029] GE 0.088 (2.771) I io.oo7j KA -0.021 I 0.068 (-1.400) (2.348) LO. 1671 [0.022] KY -0.037 0.038 (-2.037) (2.031) 10.0461 [0.047] AR 0.039 (3.046) [0.003] ML -0.715 0.055 -0.036 (-2.771) (1.771) (-2.672) 10.0071 [0.082] [O. 0 lo] MO -0.054 (-2.23 3) [0.029] TJ -0.058 0.033 (-3.644) (1.204) [0.001] [0.233] SSR 0.056 0.065 0.065 SER 0.03 1 0.032 0.032 ARS 0.529 0.494 0.494

136 Note: In the above regressions, the endogenous variable is GROW = Average growth rate in GDP per capita. The rest of the variables are defined as follows, as in Appendix Table Al: CEXP = Government capital expenditure (% of GDP); REXP = Government recurrent expenditure (% of GDP); EEXP = Government expenditure on education (% of GDP); HEEP = Government expenditure on health (% of GDP); DREV = Tax revenue (% of GDP); FDEF = Government surplus / deficit (% of GDP); NMV = Non Government investment (% of GDP); BMON = Broad money (M2) (% of GDP); PRIM = Primary school enrolment ratio; SECS = Secondary school enrolment ratio; HCGEN = initial human capital (PS) variable; LFEX = Log of life expectancy; TTDE = Trade ratio (export plus import as % of GDP); ALIT = Adult literacy (as % of total population); MFL = Inflation rate, computed as the rate of change of the consumer price index; IGDPL = Log of GDP per capita at initial period.

It is interesting to note from Table 5 that our results suggest that the government budget deficit or surplus as a proportion of GDP has a significant and positive impact on economic growth (model variant 2). This result is consistent with the earlier results reported in Table 4 (model variants 1-5). The evidence in both Tables 4 and 5 suggests that, on average, a one percentage point increase in the budget surplus is associated with an increase in real per capita growth rate of about 0.6 percentage points. Interestingly, although the finding is at variance with some earlier research (e.g., Bose, Haque and Osborn, 2003), it is generally consistent with other earlier findings which suggest that for developing countries, a one percentage point increase in the Government surplus (as a percentage of GDP) is associated with an increase in growth rate of real GDP per capita by an average of 0.15 percentage points. Our interpretation of this variable is that budget outlays are critical in transition economies, and much of their impact depends on how they are financed and how they are utilized.

Furthermore, we test for endogeneity. It is useful to bear in mind the fact that government spending variables and the other economic variables in the model evolve jointly, in the sense that government expenditure variables affect economic growth, but it is also possible to experience reverse causality. To ascertain whether reverse causality is indeed the case in our model, we estimated the growth regression in equation (1) using three stage least squares (3SLS). We chose instruments by using some of the original variables and lags of the other variables. The initial variables in the vector COND enter as their own instruments because these variables are exogenous to the sample. Our results show that even after controlling for possible endogeneity or reverse causality, the signs and levels of significance of the coefficients for the significant expenditure variables (Le., total capital expenditure, expenditure in the education sector) still hold.

4.6 Country-spec$c Evidence for Tajikistan @om Seemingly Unrelated Regression Equations (SURE)

So far, the econometric results we reported above were based on panel data, because we could not estimate the model on Tajikistan alone using annual data for 1996-2004; degrees of freedom problems would have violated the asymptotic accuracy of the estimation results. We now propose to tease out the coefficient estimates for each country so that we can isolate the coefficients for Tajikistan. We estimate and test the model on a regional sample using the Zellner procedure of seemingly unrelated regression equations (SURE) which allow for individual countries to feature in the panel estimates.

The results from SURE estimation and testing are reported for all the eight sample countries, including Tajikistan, in Table 6 for 7 countries, namely Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyz Republic, Mongolia and Tajikistan. We report further results in Table 7, this time for all the 8 countries in the sample.

137 Table 6: SURE (ZELLNER) Estimation Results from an Endogenous Growth Model for Tajikistan (TJ): Country-Specific Evidence

AR AZ GE KA KY MN TJ Constant 0.678 -0.118 0.315 0.246 0.164 -0.116 -0.104 (0.039) (0.464) (0.014) (-0.105) (-0.1 11) 0.932 10.91 13 0.106 -0.228 0.458 0.880 0.880 (9.591) (-11.258) (4.317) (1.060) (1.060) [0.0101~ [o,i ioj [o.oooj [0.0001 [0.0001 [om] 10.2891 NINV 0.794 0.063 1.170 0.145 0.319 -0.050 0.039 (7.896) (1.793) (7.369) (3.509) (-2.540) (-0.312) (0.383) [O.OOO] [0.073] [O.OOO] [O.OOO] [0.011] E0.7551 [0.702] DREV 0.924 -0.132 -2.509 1.137 -0.986 0.067 -0.278 (5.015) (-0.763) (-7.393) (12.255) (-3.595) (0.875) (-2.292) [0 .0001 [0.445] JO.OOO] [O.OOO] [O.OOO] L0.3821 10.0221 CEXP -1.022 1.269 11.267 -1.379 -1.171 0.146 1.716 (-7.909) (4.157) (10.952) (-8.232) (-3.255) (0.794) (7.5 52) [O.OOO] [o.oooj _[O.OOO] [O.OOO] [0.001] [0.427] ~0.000] SSR 0.282 0.466 0.294 0.128 0.588 0.559 0.397 SER I 0.626 I 0.805 0.639 0.421 0.904 0.881 0.743 0.940 0.978 0.680 0.568 0.872

AR AZ GE KA KY MN TJ LFEX -0,158 -0.870 0.784 0.250 (-1.742) (-4.726) (6.227) (2.4 82) [0.082] [O.OOO] [O.OOO] [O.O 131 NINV 0.602 -0.064 0.175 -0.272 0.534 -0.493 (1 1.423) (2.534) (3.976) (-2.385) (5.568) (-5.693) [O.OOO] [0.011] [O.OOO] [0.017] [O.OOO] _[O.OOO] DREV 0.269 0.063 0.093 -1.718 0.657 (3.690) (8.395) (4.146) (-3.968) (8.441) [O.OOO] [O.OOO] [O.OOO] [O.OOO] [O.OOO] HCGEN 0.168 -0.239 0.190 0.708 0.185 (5.95 8) (-6.078) (6.724) (3.442) (4.699) rn nnni rn nnni rn Ann1 rn Ani i [O.OOO] 2.073 (-2.250) (9.05 5) (-2.03 0) (4.524) (-8.437) (4.4 17) [0.024] [O.OOO] [0.042] [O.OOO] [O.OOO] [O.OOO]

SSR 1 0.423 0.270 0.578 0.221 0.551 0.284 0.569 SER 0.767 0.613 0.896 0.554 0.875 0.628 0.889 RS 0.941 0.945 0.882 0.962 0.701 0.781 0.816 Adult literacy (as % of total population); INFL = Inflation rate, computed as the rate of change of the consumer price index; IGDPL = Log of GDP per capita at initial period. The sample countries are: AR=Armenia; AZ=Azerbaijan; GE=Georgia; KA=Kazakhstan; KY=Kyrgyz Republic; ML=Moldova; MN=Mongolia; and TJ=Taj ikistan.

Table 7: Further SURE (ZELLNER) Estimation Results from an Endogenous Growth MI le1 for Tr i$;tan (TJ)~:Country-! iecific Evil I AR AZ KA KY TJ C -0.230 0.1 116 -0.01 8 -0.943 0.523 0.263 (-4 -641) (5.317) (39.196) (12.603) [O.OOO] [O.OOO] [O.OOO] [O.OOO] ILFEX 1.949 0.171 -0.23 1 -0.841 (9.5 14) (10.128) (-6.987) (-22.781) (-99.486) [0 .0001 [O.OOO] [0 I0001 [0 * 0001 NINV 1.460 -3.220 0.155 9.601 (-26.564) (-2 1.768) (7.245) [O.OOO] [O.OOO] [O.OOO] DREV 0.385 -0.237 -0.147 (9.926) (-4.93 1) (-40.05 0) (-27.3 19) [O.OOO] [O.OOO] [O.OOO [O.OOO] HCGEN -4.405 -2.189 -0.320 -2.961 2.824 (-1 8.524) (-4.3 75) (-8.842) (-1 5.728) (1 8.443) [O.OOO] [O.OOO] [O.OOO] [O.OOO] [O.OOO] EEXP 1.251 1.460 -0.459 1.152 0.1332 4.1 13 (23.482) 9.601 (-3.554) (40.659) (6.890) (45.495) FDEF 1 i0.0001 [O.OOO] y.go] 1 [O.OOO] [O.OOO] [O.OOO] -0.908 2.413 L-0.274 (-4.920) (7 1.863) (24.067) (21.891) (-5.643) ~0.000] [0.000] 0.985 -0.479 0.163 0.136 (7.163) (163.688) (-11.717) (13.2 10) (29.862) [O.OOO] [O.OOO] 0.512 0.227 (13.242) (15.3 10)

I [O.OOO] I [0.000] ROW = Average growth rate in GDP per capita. The rest of the variables aredefined as follows, as in Appendix Table AI: CEXP = Government capital expenditure (% of GDP); REXP = Government recurrent expenditure (% of GDP); EEXP = Government expenditure on education (% of GDP); HEEP = Government expenditure on health (% of GDP); DREV = Tax revenue (% of GDP); FDEF = Government surplus / deficit (% of GDP); NINV = Private investment (% of GDP); BMON = Broad money (M2) (% of GDP); PRIM = Primary school enrolment ratio; SECS = Secondary school enrolment ratio; HCGEN = initial human capital (PS) variable; LFEX = Log of life expectancy; TTDE = Trade ratio (export plus import as % of GDP); ALIT = Adult literacy (as % of total population); INFL = Inflation rate, computed as the rate of change of the consumer price index; IGDPL = Log of GDP per capita at initial period.

Table 6 reports the ZELLNER results for explanatory variables from vectors COND and GOVTSPEN. In general, the main results for Tajikistan are clear. First, the evidence suggests that government expenditure in education is positively and significantly correlated with economic growth in Tajikistan (Table 6, model variant B, column 8). The

139 evidence which supports the positive role of education in economic growth is also found for the human capital composite variable in Tajikistan, which by construction encompasses primary and secondary schooling (Table 6, model variant B, column 8). Second, it is shown that the share of government capital expenditure in GDP is positively and significantly correlated with economic growth (column 8 of Table 6, model variant A).

The ZELLNER estimation results in Tables 6 are augmented with explanatory variables from the vector OTHER and from the government budget constraint and the results are reported in Table 7. In general, the results for Tajikistan reported in Table 7 show that the two main government expenditure variables in the GOVTSPEN vector, Government spending on education and Government capital expenditure, are able to meet the robustness test Le., they are robust even if other explanatory variables from the vector OTHER are included in the model.

It is shown in Table 7 that all of the explanatory variables in the COND vector are statistically significant for Tajikistan. Importantly, the non government investment variable shows a significant relationship with economic growth in Tajikistan. This result is consistent with a number of previous empirical studies (e.g., Levine and Renelt 1992). The results are also consistent with the basic prediction of the neoclassical growth theory. The evidence also shows that the trade share of GDP has a significant positive relationship with economic growth. Further, it is found that the ratio of broad money to GDP is positively and significantly related to economic growth in Tajikistan. The domestic revenue and fiscal deficit variables have mixed results, but in most cases conform to the role of the budget constraint.

5. Conclusion and Recommendations for Tajikistan Public Expenditure Review

5. I Summary and Conclusions

This study has focused on identifying the elements of public expenditure that have a significant relationship with economic growth, with respect to the CIS-7 transition economies, and with special reference to Tajikistan.

Our study is innovative in building on the theoretical literature on government expenditure and endogenous growth by proposing and implementing a novel way of estimating an endogenous growth model on Tajikistan and 7 other transition economies. We depart from most of previous research in developing countries by considering the implications of the government budget constraint. We also depart from previous work in this area by focusing on the transition economies.

In terms of econometric methodology, first, we estimated the model on a panel dataset of transition economies, namely Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyz Republic, Moldova, Mongolia and Tajikistan for the period 1996 - 2004, hence 72 observations. We run the base regressions in order to establish whether the government spending variables have significant impact on growth after controlling for the CONDS variables. Basically, we run a restricted version of the model in equation (l), where the gamma coefficients are constrained to be zero. The regressions are run in such a way that we add one GOVTSPEN variable at a time. These initial regressions are intended to capture a central idea of the endogenous growth literature namely that human capital and institutional factors are important determinants of economic growth. In addition by including initial GDP in these initial regressions, we are trying to explore the argument of

140 possible effects of convergence on output growth in the sample economies. We then conducted checks for the robustness of our initial findings by adding to the model vector OTHER, to obtain the augmented model. Further, since we could not estimate the model on Tajikistan alone using annual data for 1996-2005 because degrees of freedom problems would violate the asymptotic accuracy of the estimation results, we tried to tease out the coefficients for each country so that we can isolate the coefficients for Tajikistan. Hence, we estimated and tested the model on a regional sample using the Zellner procedure (SURE, or seemingly unrelated regression equations) which allow for individual countries to feature in the panel estimates. We also checked for robustness within the Zellner procedure.

The evidence from the study is unequivocal in supporting the prevalent view in modem growth theory that education is an important key to economic growth. Hence, the most important finding of this study, from an empirical as well as a policy perspective, is that education is the key sector to which public expenditure should be directed in order to promote economic growth in Tajikistan. From an econometric perspective, this finding is important and reinforces earlier findings by Bose, Haque and Osborn (2003) whose study focused on developing economies. In addition, the study is important in providing evidence which counters the earlier results by Landau (1986) and Devarajan et al. (1996), who find negative or insignificant positive effects of education expenditure on growth for a sample of developing economies. We argue that the results of our study and those obtained by Bose, Haque and Osborn (2003) are more plausible because they incorporate the government budget constraint. Moreover, our study is more plausible because it reflects the economic structure of Tajikistan and the neighboring transition economies which form our sample.

We also find that the share of Government capital expenditure in GDP is positively and significantly correlated with economic growth. This result provides the basis for Ggovernment intervention in supporting public investments in key sectors such as transport and communication in order to achieve economic growth. However, this variable seems to be conditional on the non-government investment variable (which is a proxy for private investment). In fact, the non-Government investment share of GDP is associated with economic growth in a significant and positive manner, especially in the Zellner estimates for Tajikistan.

Also, we find strong evidence to support human capital variables in the 8 countries and in Tajikistan. In particular, the initial human capital variable, expressed as enrolment in primary and secondary schools was consistently significant. This suggests the need for donors and gvernments to support the funding of primary and secondary school education. Also, life expectancy is consistently significant in explaining economic growth in all the 8 countries, and in Tajikisatn, in particular. This finding lends support to increased expenditures to support health and sanitation in Tajikistan.

Finally, we must acknowledge the limitations of our small sample, although the data we used are the most comprehensive database available at the moment for Tajikistan and her regional neighbors.

5.2 Recommendations for Programmatic Public Expenditure Review for Tajikistan

In terms of recommendations for the PPER, the evidence from this study gives rise to information that is particularly useful for the Tajik government, which is resource

141 constrained and where the allocation of limited public resources between the sectors is an issue of paramount importance. First, the empirical finding which unequivocally supports the association between government expenditure on education and economic growth is important for Tajikistan in the sense that the finding underscores the argument that the priority should be to allocate scarce government resources towards the education sector. The results are also important in that they provide evidence to support the argument that expenditure on education should be increased.

Second, the evidence supports the human capital argument in the endogenous growth story, especially in terms of the combination of primary school enrolment rates and secondary school enrollment rates in the 8 transition economies and in Tajikistan. This evidence lends support to the case made above for public expenditure allocation in favor of education and services in the education sector. But, the evidence on human capital variables goes a bit further than that. Specifically, the evidence which supports strong association between life expectancy and economic growth in the sample economics in general and in Tajikistan in particular, lends support to increased expenditures to support health and sanitation in Tajikistan.

Third, the finding that the Government budget surplus as a proportion of GDP has a significant and positive impact on economic growth is important, given the magnitude of the impact: on average, a one percentage point increase in the budget surplus is associated with an increase in real per capita growth rate of about 0.6 percentage points. Our recommendation is based on our interpretation of this variable, which is that the budget outlays are critical in transition economies, and much of their impact depends on how they are financed and how they are utilised. In this context, development assistance can be an important part of budget support, specifically targeted to expenditures in the social services sector such as education and health.

Fourth, we recommend the use of private capital, in addition to the current capital expenditures, perhaps in a public-private partnership. Broadly, our evidence that the share of government capital expenditure in GDP is positively and significantly correlated with economic growth, and the rest of the sample transition economies, provides the basis for Government intervention in supporting public investments and expenditures in other key sectors such as transport and communication in order to achieve economic growth. However, this variable seems to be conditional on the non government investment variable in Tajikistan. In fact, the non government investment share of GDP is associated with economic growth in a significant and positive manner, especially for Tajikistan. Hence, this finding may be interpreted to suggest that while government capital expenditure is associated with economic growth, the financing the construction of roads, bridges and other capital infrastructure projects should at least involve the private sector, perhaps in a public-private partnership, rather than being exclusively the reserve of government spending, especially in: strategically targeted new investment in infrastructure projects and rehabilitation of existing infrastructure network; Fifth, we recommend budget support for the Tajikistan government by international donors in order to deliver an enabling macroeconomic environment that fosters long term sustainable growth. This recommendation is based on the evidence from this study which underlines the role of macroeconomic factors, including broad money, domestic revenue and budget discipline.

142 Finally, the targeting of government expenditure towards capital expenditure in say infrastructure and transport, education and health should bear in mind the institutional, regional and historical factors of Tajikistan.

143 ources, 1996-2004 I data source

World Bank and IMF Country Reports World Bank website CEXP Government capital expenditure (% of GDP) World Bank and IMF Country Reports WorldBank website EEXP Government expenditure on education (% of GDP) World Bank and IMF Country Reports World,Bank website HEEP Government expenditure on health (% of GDP) World Bank and IMF Country Reports World Bank website DREV Tax revenue (% of GDP) World Bank and IMF Country Reports World Bank website FDEF Government surplus / deficit (% of GDP) World Bank and IMF Country Reports World Bank website NmV Private investment (% of GDP) World Bank and IMF Country Reports World Bank website BMON Broad money (M2) (% of GDP) World Bank and IMF Country Reports World Bank website PRIM Primary school enrolment ratio World Bank and IMF Country Reports World Bank website SECS Secondary school enrolment ratio World Bank and IMF Country Reports World Bank website HCGEN Initial human capital (PS) variable Generated from PRIM and SECS- see notes below LFEX Log of life expectancy World Bank and IMF Country Reports World Bank website TTDE Trade ratio (export plus import as % of GDP) World Bank and IMF Country Reports World Bank website ALIT I Adult literacy (as % of total population) World Bank and IMF Country Reports

Notes: GROW = Average growth rate in GDP per capita. The rest o the variables are defined as follows, as in Appendix Table Al: CEXP = Government capital expenditure (% of GDP); REXP = Government recurrent expenditure (% of GDP); EEXP = Government expenditure on education (% of GDP); HEEP = Government expenditure on health (% of GDP); DREV = Tax revenue (% of GDP); FDEF = Government surplus / deficit (% of GDP); NINV = Private investment (% of GDP); BMON = Broad money (M2) (% of GDP); PRIM = Primary school enrolment ratio; SECS = Secondary school enrolment ratio; HCGEN = initial human capital (PS) variable; LFEX = Log of life expectancy; TTDE = Trade ratio (export plus import as % of GDP); ALIT = Adult literacy (as % of total population); INFL = Inflation rate, computed as the rate of change of the consumer price index; IGDPL = Log of GDP per capita at initial period. HCGEN, the initial human capital variable is constructed in the context of Landau (1983) as the weighted sum of the initial enrolment ratios (%) in primary and secondary schools and in higher education. The weights are 1 for primary school enrolment ratio and 2 for secondary school enrolment ratio, hence, these weights are approximations to the relative values of two types of education. We use the HCGEN variable because of the high milticollinearity between the separate enrolment rates of PRIM and SECS.

144 APPENDIX A: Patterns of Government Expenditure and Economic Growth Variables in CIS, including Taiikistan

Figure 1:Patterns of Government Recurrent Expenditure (REXP) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vx REXP

0.500

0.350 0.300 0.250 0.204 +a \.I50 + 0.100 0.050 - ** I -- I I -0.100 0.000 0.100 0.200 Growth

145 Figure 2: Patterns of Government Capital Expenditure (CEXP) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vs CEXP

0.450 - 0.400 - 0.350 - 0.300 - 0.250 - IccExp]

-0.10 -0.05 0.000 0.050 0.100 0.150 0.200 00 Growth

146 Figure 3: Patterns of Domestic Revenue @REV) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vs DRE V

0.500 0.450 + + + + $+ + $+ +

0.050 - n nnn I I 1 -0.100 0.000 0.100 0.200 Growth

147 Figure 4:Patterns of Net Investment (NINV) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vs NINV

0.500 4 4 0.450 0.400 4 0.350 4

0.300 4 0.250 i 4% 'p 4

nn ** 4 r - I I -0.100 0.000 0.100 0.200 Growth

148 Figure 5:Patterns of Broad Money (BMON) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

0.500 1 + 0.400

0.300 -/ + 0.250 + [*BMON]

+ ++ +++

0.050 nnnn I I I I -0.100 0.000 0.100 0.200 Growth

149 Figure 6:Patterns of Terms of Trade (TTDE) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vs TTDE

5.000 4.500 4.000 3.500 3.000 2.500 2.000

Y A.Y I-- I I -0.100 0.000 0.100 0.200 Growth

150 Figure 7:Patterns of Inflation (INFL) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

5.000

3.000

2.000 I + INFL I

1.000

-0.100 0.c00 0.100 0.200 -1.000 - Growth

151 Figure 8:Patterns of Government Expenditure on Health (HEEP) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vs HEEP

0.5 - 0.45 - 0.4 - 0.35 - 0.3 - 8 0.25 - El 0.2 -

-0.100 0.000 0.100 0.200 Growth

152 Figure 9:Patterns of Life Government Expenditure on Education (EEXP) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996- 2005

Growth Vs EEXP

0.3 - B 0.25 - w 0.2 - 0.15 - + n1- +

-0.100 0.000 0.100 0.200 Growth

153 Figure 10:Patterns of Primary School Education (PRIM) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

120

100 sd E 60 PRIM %; 40

I -0.100 0.000 0.100 0.200 GrOWth

154 Figure 1l:Patterns of Secondary School Education (LFEX) and Economic Growth (GROWTH) in Commonwealth Independent States (CIS), 1996-2005

Growth Vs SECS

100I2O 1

-0.100 0.000 0.100 0.200 Growth

155 Annex 11. Expenditures of the State budget of the Republic of Tajikistan 2005-2010

Budget Items

Total Revenues and Granb 1,256.0 1,414.5 1,510.1 1,823.5 1,847.7 2goo.o 2,730.0 3,319.2 3,992.0

in percent to GDP 17.7 19.6 18.0 19.7 19.3 21.2 21.3 22. I 22.5

Total General Government ExDenditures 1.729.9 1.712.7 1.936.7 2.126.8 3.290.9 3.562.9 4.095.5 4,352.0 4.633.1 in percent to GDP 24.4 23.8 23. I 22.9 34.5 32.8 32.0 29.0 26.2 including: -state budget: 1,290.7 1,3 13.4 1,541.7 1,643.4 1,943.2 2,215 1 2,857.9 3,469.6 4,169 1 - wages and salaries 274.6 269. I 3 78.0 381.3 4 74.0 483. I 693.0 0.0 0.0 Centralized State Investment - 137.2 238.4 192.2 2611.8 245.3 349.7 298.0 342.7 394. I Programme (CSIP) - Public Investment Program (PIP) 350.3 310.2 329.5 348.0 489.7 489 7 547.0 344.0 304.0 - State Loan (Chinese) 0.0 0.0 0.0 0.0 778.5 778.5 570.4 398.4 - extra-budgetary (special) funds 88.9 89.1 65.6 135.4 79.5 79.5 120.2 140.0 160.0

1. Governance and Public 198.8 243.2 208.6 292.9 260.0 293.3 319.4 353.0 406.3 Administration in percent to GDP 2.8 3.4 2.5 3.2 2.7 2.7 2.5 2.3 2.3 including: - state budget: 160.0 204.4 180.7 225.7 223.4 256.7 275.0 330.1 396.1 - wages and salaries 33.9 32.6 50.2 46.2 64.3 64.3 91.2 Centralized State Investment - 32.2 75.6 40. I 66.9 38.5 99.0 38.5 42.0 48.5 Programme (CSIP) - Public Investment Program (PIP) 1.6 1.7 2.0 0.7 24.3 24.3 23.2 22.9 10.2 - extra-budgetary (special) funds 37.2 37.2 25.9 66.5 12.3 12.3 21.2

2. Defense 102.4 96.3 124.9 105.0 149.2 157.4 194.0 232.9 279.4 in percent to GDP 1.4 1.3 1.5 1.I 1.6 1.4 1.5 1.5 1.6 including - state budget 102 4 96 3 124.9 103.8 149.2 157.4 194.0 232.9 279.4 - wages and salaries 18 I 16.2 26.9 27.6 33.0 33.0 41.0 Centralized State Investment - 40 15.8 4.5 11.9 4.9 9.5 5.9 7.5 8.7 Programme (CSIP) - Public Investment Program (PIP) 00 00 0.0 0.0 0.0 0.0 0.0 0.0 0.0 - extra-budgetary (special) funds 00 0.0 0.0 I .2 0.0 0.0 0.0 - 3. Law-EnforcementAgencies 111.0 115.1 131.0 138.3 161.0 168.6 216.7 216.1 259.2 in percent to GDP 1.6 1.6 1.6 1.5 1.7 1.6 1.7 1.4 1.5 including - state budget 91 9 96 1 113.8 117.9 138 2 145.9 180.0 216.1 259.2 - wages and salaries 32 0 32.2 49.9 53.6 58.6 58.6 73.0 Centralized State Investment - 4.7 7.9 7.3 6.4 75 13.0 9.2 12.5 15.0 Programme (CSIP) -Public Investment Program (PIP) 00 00 00 0.0 00 0.0 0.0 0.0 0.0 - extra-budgetary (special) funds 19 I 19 1 17.2 20.4 22 8 22.8 36.7 SOCIAL BLOC 665.4 676.1 853.9 907.2 1,103.0 1,219.1 1,477.6 -1,588.9 1,777.8 in percent to GDP 94 9.4 10.2 9.8 11.6 11.2 11.6 10.6 10.0

156 including: - state budget: 600.2 609.8 764.1 801.8 961.4 1,077.5 1,297.0 1,501.6 1,756.4 - wages and salaries 179.5 174.8 237.1 231.2 299. I 308. I 448.3 0.0 0.0 Centralized State Investment - 20.2 31.0 52.6 39.9 69.3 97.7 97.6 120.5 137.1 Programme (CSIP) - Public Investment Program (PIP) 41.5 42.6 71.4 73.4 108.9 108.9 132.4 87.3 21.4 - extra-budgetary (special) funds 23.6 23.7 18.4 32.0 32.8 32.8 48.1 0.0 00

4. Education 281.6 290.2 381.0 372.7 506.4 552.9 689.7 705.7 820.7 inpercent to GDP 4.0 4.0 4.5 4.0 5.4 4.7 4.6 including: - state budget: 244.8 253.1 332.5 317.7 428.5 475.1 562.0 674.6 809.4 - wages andsalaries 139.2 137.9 189.3 183.7 237.1 246. I 308.8 Centralized State Investmenr - 10.9 20.0 26.4 16.9 39.3 47.7 55.0 60.0 60.0 Programme (CSIP) - Public Investment Program (PIP) 18.3 18.5 33.1 30.9 49.3 49.3 85.7 31.1 11.3 - extra-budgetary (special) funds 18.5 18.6 15.4 24.1 28.6 28.6 42.0

5. Health 97.2 90.4 117.9 125.3 178.3 189.7 254.8 311.9 316.9 in percent to GDP 1.4 1.3 1.4 1.4 2.0 2. I 1.8 including: - state budget: 89.3 82.4 105.7 105.4 131.0 142.4 213.0 255.7 306.8 - wages and salaries 29.9 29.0 39.5 38.9 51.2 51.2 122.6 Cennalized State Investment - 6.8 8.2 11.8 9. I 13.0 18.8 16.0 23.0 30.5 Programme (CSIP) - Public Investment Program (PP) 5.7 5.8 11.1 17.7 47.2 47.2 41.4 56.2 10.1 - extra-budgetary (special) funds 2.2 2.2 1.1 2.2 0.027 0.0 0.4

6. Social Insurance and Social Protection 245.8 250.8 295.7 342.1 344.1 382.1 435.8 461.5 508.4 in percent to GDP 3.5 3.5 3.5 3.7 3.4 3. I 2.9 including - state budget 228.2 232.4 268.5 3170 331 7 369.7 430.5 461.5 508.4 - wages and salaries 5.3 2.8 1.9 1.8 3.0 3.0 4.8 Centralized State Investment - 0.7 0.7 0.9 1.9 2.9 8.6 9.7 16.0 21.0 Programme (CSIP) - Public Investment Program (PIP) . 17.5 18.3 27.2 24.8 12.4 12.4 5.3 - extra-budgetary (special) funds 0.1 0.1 0.0 0.3 0.01 0.0

8. Culture and Sport 40.7 44.7 59.3 67.0 97.2 109.8 131.8 in percent IO GDP 0.6 0.6 0.7 0.7 0.8 0.7 0.7 including: - state budget: 37.9 41.9 57.4 61.7 70.2 90.3 91.5 109.8 131.8 - wages and salaries 5. I 5. I 6.4 6.8 7.8 7.8 12.2 Centralized State Investment - 1.8 2. I 13.5 12.0 14.1 22.6 16.9 21.5 25.6 Programme (CSlP) - Public Investment Program (PIP) 0.0 0.0 0.0 0.0 0.0 0.0 - extra-budgetary (special) funds 2.9 2.9 I .9 5.3 4. I 4. I 5.7

ECONOMIC BLOC 522.9 517.8 493.2 605.0 1,452.1 3,196.4 1,363.8 -1,098.0 836.2 in percent to GDP 7.4 7.2 5.9 6.5 10.7 7.3 4.7 including - state budget 210.8 246.7 233.7 319.8 391 5 469.9 563.8 - wages and salaries 7.2 13.4 13.9 22.7 39.5 Centralized State Investment - 75.2 106.9 84.7 133.2 125. I 130 6 138.3 141.3 161.I Programme (CSIP) - Public Investment Program (PIP) 306.1 265.0 255.4 273.4 350 0 350 0 387 6 229.7 272.3

157 - State Loan (Chinese) 0.0 0.0 0.0 778.5 778.5 570.4 398.4 0.0 - extra-budgetary (special) funds 6.0 6.0 :::1 11.8 11.7 11.7 14.2 0.0 0.0 7. Housing and Communal Services 97.6 120.9 125.2 139.7 196.3 148.7 141.2 155.0 in percent to GDP 1.4 I.7 1.5 1.8 1.2 0.9 0.9 including: -state budget: 55.9 80.2 65.0 101.6 82.8 139.4 IO7 6 129.2 155.0 - wages and salaries 1.6 4.2 8.6 9.4 11.2 11.2

~ Centralized State Investment 16.6 29.8 19.2 40.6 23.7 24.9 28.5 34.0 40.0 Programme (CSIP) - Public Investment Program (PIP) 39.1 38.1 19.7 19.6 53.7 53.7 41 1 12 I - extra-budgetary (special) funds 2.6 2.6 0.0 4.0 3.2 3.2

9. Fuel and Energy Complex 97.5 103.2 164.2 I 175.6 571.8 571.8 350.7 408.6 115.7 inpercent to GDP 1.4 1.4 2.0 1.9 6.0 5.3 2.7 2.7 0.7 including: - state budget: 15.1 17.7 25.2 38.2 45.2 45.2 60.0 72.0 86.4 - wages andsalaries 0.0 0.0 0.0 0.0 0.4 0.4 Centralized State Investment - 14.2 16.7 24.2 20.8 43.2 39. I 55.3 42.0 45.0 Programme (CSIP) - Public Investment Program (PIP) 82.4 85.5 139.0 137.4 104 5 104.5 83.9 27.2 29.3 - State Loan (Chinese) 0.0 0.0 0.0 0.0 422.1 422.1 206.9 309.4 - extra-budgetary (special) funds 0.0 0.0 0.0 0.0 00 0.0

IO. Agriculture, Fishery, and Hunting 89.3 84.3 111.0 I 102.6 176.3 184.3 232.9 206.9 254.3 inpercent to GDP 1.3 1.2 1.3 1.1 1.8 1.7 1.8 1.4 1.4 including: - state budget: 41.4 37.9 49.7 44.4 57.4 65.4 73.5 88.2 105.9 - wages andsalaries 4.3 4.3 3.2 6.6 4.7 4.7 Centralized State Investment - 8. I 11.1 8.8 10.7 10.5 18.1 12.6 15.4 17.8 Programme (CSIP) -Public Investment Program (PIP) 46.4 44.8 60.8 56.4 116.4 116.4 151.6 118.6 148.4 - extra-budgetary (special) funds 1.5 1.5 0.5 1.8 2.5 2.5 7.7

11. lndustrv and Construction 18.2 23.7 15.7 I 45.2 20.8 21.5 27.8 31.7 38.0 in percent to GDP 0.3 0.3 02 05 0.2 0.2 0.2 0.2 0.2 including - state budget 17.7 23 2 15 3 44 6 20.3 21 0 26.4 31 7 38 0 - wages and salaries 0.3 0.3 05 05 0.7 0.7 Centralized State Investment - 6.3 19.7 84 41.1 15.4 17.7 18.7 22.6 26.0 Programme (CSIP) - Public Investment Program (PIP) 0.0 0.0 00 00 0.0 0.0 0.0 0.0 0.0 - extra-budgetary (special) funds 0.5 0.5 04 06 0.5 0.5 1.4 - 12. Transoort 215.1 176.1 112.9 I 151.9 536.1 539.2 594.7 298.9 260.2 inpercent IO GDP 3.0 2.4 1.3 1.6 5.6 5.0 4.6 2.0 1.5 including: - state budget: 76.1 78.8 74.8 87.2 99.8 102.9 115.0 138 0 165.6 - wages and salaries 0.3 4.0 0.9 5.3 1.2 1.2

~ Centralized State Investment 30.0 29.6 24.2 19.9 31.3 30.2 22.0 25.3 29.5 programme (CSIP) - Public Investment Program (PIP) 138.3 96.6 35.9 60.0 75.4 75.4 111.0 71 9 94.6 - State Loan (Chinese) 0.0 00 00 0.0 356.4 356.4 363.6 89 0 - extra-budgetary (special) funds 0.6 0.6 2.2 4.6 4.6 4.6 5.1

13. Other Economic Services 5.4 9.7 4.7 I 4.5 7.4 8.0 9.0 10.8 13.0

158 in Dercent to GDP 0.1 I 0. I 0. I 0.0 0. I 0. I 0. I 0. I 0. I including: - state budget: 3.7 3.8 6.4 7.0 9.0 10.8 13.0 - wages and salaries 0.8 0.8 0.9 0.9 - Centralized State Investment ~ 0.8 0.0 0.0 1.0 0.6 1.2 2.0 2.8 Programme (CSIP) - Public Investment Program (PIP) 0.0 0.0 0.0 0.0 - extra-budgetary (special) funds 1 .o 0.7 1 .o 1 .o

14. Other Expenditures 129.5 125.1 78.4 165.6 203.3 524.2 723.1 914.2 in percent to GDP 1.5 0.8 1.7 1.9 4. I 4.8 5.2 including - state budget 124.5 74.4 159. I 196.8 520.4 719.0 914.2 - wages and salaries 0.0 0.4 0.0 0.0 Centralized Slate Investment - 2.9 2.5 0.0 0.0 8.5 18.9 23.7 Programme (CSIP) - Public Investment Program (PIP) 0.6 0.5 6.5 6.5 3.8 4.1 - extra-budgetary (special) funds 30 31 0.0 3.5 0.0 0.0 - Budget DeficiUSurplus -34.7 101.1 -31.6 180.0 -95.5 84.9 -127.9 --150.3 -177.1 inpercent to GDP -0.4 1.9 -1.0 0.8 -1.0 -1.0 -1.0

15.030. GDP 8,400.0 9,272.1 9,550.0 10,865.0 12,790.0 17,710.0 0 1 percent of GDP 71.0 I 72.0 84.0 92.7 95.5 108.7 127.9 150.3 177.1

159