Georgia - Public expenditure review

Report No. 15779-GE Georgia Public Expenditure Review June 27, 1996 Country' Operations Division II Country Department IV Europe and Central Asia Region 0 X - * V , ,, - -, Docun*nt of lW.&ju  CURRENCY EQUIVALENTS (as of January 1996) Currency Unit = Lari US$1 1.23 WEIGHTS AND MEASURES Metric System ABBREVIATIONS AND ACRONYMS cIS - Commonwealth of Independent States CPI - Consumer Price Index EU - European Union EF - Employment Fund FSU - Former Soviet Union GDP - Gross Domestic Product GFS - Government Finance Statistics IMF - International Monetary Fund IMR - Infant Mortality Rate NGOs - Non Governmental Organizations OECD - Organization for Economic Cooperation and Development O&M - Operations and Maintenance PIP - Public Investment Program UFSS - United Fund for Social Security WDR - World Development Report GEORGIA - FISCAL YEAR January I - December 31  Preface The preparation of this report was managed by Ana Revenga and Michelle Riboud. The study draws on background papers prepared by Arup Banerji (Government Program & Employment), Gary Burtless (Social Protection), Stephen Lister (PIP), Franqois Orivel (Education, Science), Chandra Pant (Infrastructure), and H. Barnum and Laura Rose (Health). The report also benefited from contributions by Afsaneh Farzin, Hafez Ghanem, Pedro Taborga, Marijn Verhoevens and Vlado Vucetic. Kyle Peters and Jeff Hammer were peer reviewers. Una Raymond assisted with document preparation. The report was prepared under the general direction of Wafik Grais (Division Chief) and Basil Kavalsky (Director). The mission team wishes to thank the Georgian authorities for their excellent support and cooperation and their useful comments and disc,ussion on an earlier draft.   REPUBLIC OF GEORGIA PUBLIC EXPENDITURE REVIEW TABLE OF CONTENTS EXECUTIVE SUMMARY ................................................. i-v CHAPTER 1: THE MACROECONOMIC CONSTRAINT ................................................. 1 A. The Collapse of Government Finance ..........................................................1 B. Medium-Term Fiscal Framework ..........................................................4 CHAPTER 2: THE NEW ROLE OF GOVERNMENT ................................................. 10 A. Rebuilding Government ......................................................... 10 B. Expenditure Priorities by Sector .......................................................... 13 C. Policies for Expenditure Reform .14 CHAPTER 3: SELECTED SECTORAL EXPENDITURE ISSUES ............................... 24 A. Health ......................................................... 24 B. Education ......................................................... 27 C. Science ......................................................... 32 D. Social Protection ........................................................ 34 E. Energy ......................................................... 36 F. Transport ......................................................... 39 Annex 1: Distribution of Consolidated Government Expenditure, 1995-1996 Annex 2: Means of Government Expenditures in Developing Countries by Regions, (Functional Classification, Average of 1985-89) Annex 3: Fiscal Expenditures under Base Case Scenario Annex 4: Brief Summary of Georgia's System of Social Protection   GEORGIA: PUBLIC EXPENDITURE REVIEW Executive Summary The Macroeconomic Constraint The breakdown of government finances in Georgia is unlike anything seen elsewhere in the Former Soviet Union. Following independence, in a context of economic collapse, civil conflict and political chaos, tax revenue virtually disappeared. By 1993, total revenues had fallen to 2.3 percent of GDP, covering barely 6 percent of total expenditure. The collapse of government revenues has implied drastic cuts in the financing of key public services. On a per capita basis, spending on education and health is now well below the average for the the rest of the region, and below levels seen in much of the developing world. Spending on maintenance of essential infrastructure has also basically disappeared. Table 1: Per Capita Spending on Education and Health for Selected Countries Country GDP per capita Population Education Health (US$) (millions) $ p. cap % GDP $ p. cap % GDP Armenia 570 3.7 6 1.0 7 1.2 Sri Lanka 600 17.2 17 2.8 8 1.4 Indonesia 740 187.2 12 1.6 3 0.4 Bolivia 760 7.1 22 2.9 14 1.8 Moldova 1060 4.4 63 6.0 41 3.9 Lithuania 1320 3.7 51 3.9 41 3.1 Greece 7390 10.4 228 3.7 358 3.2 Georgia 440 5.4 4 0.9 3.3 0.8 Source: World Development Report 1995. The challenge for Georgia to deliver key public services is enormous. With revenues projected to reach only 9-13 percent of GDP during the 1996-98 period, the pressure to maintain a tight expenditure program will continue. On the other hand, the government clearly needs to make some essential investments in the economy (in human capital and infrastructure) to ensure sustained growth and the continued welfare of its population. Georgia will not be able to meet this double challenge without a thorough redefinition of the role of the state in its economy. Public expenditure policy has an important role to play in fostering this change. A sustainablefiscalframework The main constraint Georgia faces in designing its public expenditure policy is the availability of resources. With little ability to borrow domestically and limited foreign exchange reserves, the government has only two sources of financing for its budget deficit: central bank credit and foreign borrowing. However, the extent to which it can rely on either one is greatly constrained by the need to consolidate the progress made against inflation, on the one hand, and by creditworthiness considerations on the other. To consolidate the stabilization, Georgia will need to maintain a tight fiscal program, which aims to reduce the budget deficit to 3 percent of GDP by 1998, and to under I percent by 2004. This degree of fiscal discipline is necessary to reduce public dissaving and restore positive domestic savings in the economy by 1997. Higher domestic savings are critical to the recovery of domestic investment that is needed to sustain growth. Reducing public dissaving is also crucial to improving Georgia's external accounts. Lower fiscal deficits would lead directly to significant improvements in the current account-from a deficit of 15.3  ii percent of GDP in 1995 to 1.7 percent by the year 2004. This, in tum, would contribute to lessening Georgia's dependence on foreign inflows, and would keep Georgia's external obligations within the limits of its ability to pay. Even with increased domestic savings, however, extemal financing needs will remain substantial: gross financing requirements for 1996-2004 are projected at about US$420 million per year. The bulk of these resources would have to come from official sources, initially on concessional terms. Table 2: A Sustainable Fiscal Framework % of GDP 1995 1996 1997 1998 2000 2004 GDP growth rate 2.8 8.0 10.0 10.0 8.0 5.0 Current Account (Excld Grants) -15.4 -7.2 -5.4 4.2 -3.1 -1.7 Domestic Savings -8.8 -0.6 3.6 7.8 12.4 17.5 Investment 3.7 5.1 8.0 11.2 14.7 18.4 o/w: Public 1.0 1.2 2.2 2.9 4.1 4.8 Fiscal Deficit (Excid Grants) -7.5 -4.2 -3.8 -3.5 -3.0 -1.0 Debt Service after rescheduling/Exports 28.6 10.7 28.0 11.8 20.4 27.9 Total Debt/Fxports 241.9 248.7 245.5 242.2 228.9 186.6 Can Georgia sustain largerfiscal deficits? Whether a higher level of spending could be sustained over the medium termn would depend on how the incremental deficit was to be financed, and on the impact of additional spending on growth. If higher spending were to lead to a higher growth rate-for example by allowing for more public investment in high-return infrastructure-a higher fiscal deficit may indeed be sustainable. However, this argument can be easily oversold: even if an investment has a high return (and not all public investments do) the government must be able to capture part of it if the increase is to be self-sustaining. In the case of Georgia, this would require a significant improvement in the government's ability to collect revenues from the pnvate sector through cost-recovery and taxes. Table 3: Key Creditworthiness Indicators under a Larger Fiscal Deficit Scenario as % of GDP 1996 1997 1998 1999 2001 2004 Financed via external borrowing: Fiscal deficit (excl. grants) 6.0 6.6 6.2 6.4 5.4 4.8 Current account (excl. grants) -9.8 -7.5 -6.5 -7.2 -6.9 -6.1 Total DOD (US$M) 1462 1783 2147 2585 3587 5240 Debt Service afler rescheduling/Exports 10.7 29.3 14.7 17.5 41.0 60.5 Financed through credit expansion: Fiscal deficit (excl. grants) 6.1 7.0 6.8 6.8 4.4 3.0 Current account (excl. grants) -9.8 -6.6 -5.7 -5.3 -4.7 -4.3 Total DOD (US$M) 1462 1734 2081 2463 3522 5992 Debt Service after rescheduling/Exports 10.7 29.3 14.5 16.6 34.7 56.4 A larger fiscal deficit path could be financed either through an expansion of credit from the central bank, through increased external borrowing, or through some combination of both. Financing a larger deficit path through foreign borrowing on commercial terms would add to Georgia's already large debt burden, and make it difficult for the country to meet its debt service obligations. Even a relatively modest increase m the fiscal deficit of 2 to 3 percentage points per year, would more than double key debt and creditworthiness within a ten year period. The current account deficit would double relative to the sustainable scenario. To finance a larger external imbalance, Georgia would have to accumulate twice as  iii much debt; debt service to exports would increase to 50 to 60 percent within ten years. Given this scenario, it is highly unlikely that commercial external financing would be available. Foreign borrowing on commercial terms is not a viable financing option for Georgia. A larger deficit path could also be financed by an expansion of credit from the central bank to the government. Given Georgia's recent hypenrnflation history, however, faster credit growth could quickly lead to more inflation. Higher total spending and higher inflation would in the short term work to appreciate the real exchange rate, and lead to a worsening of the current account. In the medium term, slower export growth would also feed back to lower total GDP growth. The total stock of debt would reach US$ 6 billion and debt service to exports would increase to 56 percent. r Given its already high debt burden and the legacy of high inflation, the government has very little scope to relax its fiscal stance: it can increase its aggregate spending envelope only to the extent that it can pay for it out of increased revenues. Running a fiscal deficit higher than 2-3 percent of GDP is financially unsustainable over the medium term. Additional external financing on concessional terms could play a critical role in relaxing this severe resource constraint. Such financing would allow Georgia to address its key investment needs and provide an additional spurt to growth. By underpinning necessary public investment, concessional financing would also help draw in private flows, and lay the basis for improving Georgia's commercial creditworthiness in the medium term. The amounts needed are not huge: an additional US$ 80 tc US$ 110 million of concessional financing per year over the 1996-1998 period would be sufficient to allow an increase m public investment of about 2 percent of GDP per year. The New Role of Government Rebuilding Government Until now, the government has reacted to the drop in revenues on a "day-to-day" basis: addressing the most pressing needs with available resources, and delaying all other spending until it becomes unavoidable. Energy payments, and the necessity of paying wages and social entitlements have dominated the spending agenda, leaving little room for maintenance or more medium-term investments. A more careful expenditure strategy is needed, to concentrate resources only on those minimum public functions that absolutely must be met, while putting in place a framework that will allow other priorities to be addressed by the private sector. =. Priority expenditures are those in which there is both a strong case for public intervention, and in which the costs of delaying spending are highest. As a first priority, these include spending on pure public goods-public safety, the legalframework-and on mixed private/public goods for which expenditure needs are urgent-emergency maintenance in the power sector, critical maintenance of main roads and water distribution/sewage systems, and expenditures on basic education and essential public health. Comparing this set of priorities to the current allocation of spending suggests that some major expenditure shifts are needed: * An increase in the resources devoted to priority public good activities: public order and safety, the court system, and further development of key economic institutions and public administration. These expenditures presently account for about 15 percent of the total spending budget or 2 percent of  iv GDP. This is approximately one-half the average level for developing countries. A reasonable goal is to aim for devoting somewhere between 3.5 and 4 percent of GDP to these tasks. * A withdrawal from directly productive activities in areas such as agriculture, forestry, industry, media and filmn Government involvement in these areas should be primarily indirect, and focus on provision of an appropriate institutional and regulatory framework. In agriculture, which will remain a priority sector, expenditures should focus on: (a) institutional support to land reform; and (b) quality and animal health control. The total share of spending going to economic services or "national economy" should probably not change much. * An increase in the resources devoted to maintenance and rehabilitation of existing infrastructure. The top priority in this regard is addressing the backlog of maintenance needs in critical infrastructure sectors- power, roads and railways. This will demand a significant up-front investment of resources: about US$40 million per year for maintenance and rehabilitation of the main road system, and another US$ 50 to 60 million per year for the power sector. To accomodate this, the share of resources devoted to capital expenditure will have to increase significantly to reach 18 to 20 percent of total spending by 1998. * A reorientation of the social-cultural budget to focus primarily on education, health and social protection, reducing expenditures on cultural activities, sports, and subsidies to the mass media. While there may be a case for public spending on some of those activities, these should not be given priority over health and education. Curtailing these secondary social expenditures would allow spending on critical health and basic education services to increase significantly, without a need to increase the total social-cultural spending envelope. * Within education, health and social protection, a re-focusing of expenditures on the priority sub- sectors: public health and minimum essential clinical services, basic education, and a minimum social assistance program (which would include the current minimal, universal pension). This would ensure that public resources go to those activities where social returns are highest and would concentrate resources on those services which the poor tend to use more. * To make room for the increase in the share of spending going to pure public goods, the share of "other expenditures" will have to decline. This should be facilitated by a gradual decline in interest payments after 1998, but will also require a reduction in unclassified, discretionary expenditures. Policiesfor expenditure reform To achieve a better provision of public services within the overall spending constraint, reforms are needed in several critical areas: * Improving government pay and employment practices. The primary objective of civil service reform should be to reduce employment to an efficient minimum, while paying remaining employees reasonable wages. This will involve, at a minimum, a reduction in government employment of 5 to 6 percent per year over the 1996-1999 period. Assuming a moderate increase in the share of spending going to wages and salaries, such a reduction would allow the average public sector wage to converge to 75 percent of per capita GDP by 2004. To sustain a larger wage increase, however, more drastic cuts in employment would be needed.  v * Widening private sector participation. There is a whole range of goods that are supplied publicly but that can also be privately provided. Given its limited resources, the government should try to encourage private sector participation all of these areas. Putting in place an adequate regulatory framnework for these activities and facilitating entry is of utmost priority in this regard. The areas in which the private sector should play a significant role over the medium term include: transport services, basic utilities, communications, higher education, clinical health services, and social insurance (through private pension schemes). * Increasing cost recovery. Where the state is to retain a role in provision, it should seek at least partial cost recovery. For most public services, cost-recovery levels are currently very low. This is due both to low collection levels (energy) and because user charges are too low (transport, education). Household tariffs on electricity are one-half below the supply cost; more importantly, collections are abysmally low, in the range of 10 to 15 percent. Charges to enterprises are higher, but on the whole the electricity entities are recovering only 30 to 40 percent of costs. Basic municipal services, such as trash collection or urban transport, are also heavily subsidized, with charges recovering between 2 and 40 percent of costs, depending on the municipality. Cost-recovery should be increased in all of these areas. Table 4: Priority Maintenance and Rehabilitation Needs in Selected Sectors, 1996-98 (in millions of USS) Emergency Maintenance (1996) 55.7 Health Facilities 0.5 Main roads 6.6 Power - thermal 12.2 Power - hydro 36.4 Full Rehabilitation (1996-98) 266.6 Health Facilities 9.2 Roads 86.4 Power - hydro 63.0 Power - thermal 60.0 Power - transmission & dispatch 33.0 Tbilisi - roads 15.0 Increasing maintenance and rehabilitation expenditures. To prevent existing infrastructure from deteriorating beyond a point of recovery, the govermnent needs to allocate significantly more resources to meeting its maintenance needs. The short term goal for key infrastructure sectors should be that they operate on a commercial basis, and generate enough revenues through cost-recovery to cover at a minimum their O&M costs. However, it is unlikely that in the short-term, cost recovery will generate sufficient resources up-front to address the backlog of maintenance and rehabilitation needs. Some budgetary resources will have to be allocated to these tasks. The bulk of these resources should be allocated to the backlog of delayed maintenance and urgent repair work, followed by rehabilitation of existing assets.   Chapter 1: The Macroeconomic Constraint A. The Collapse of Government Finance 1. The breakdown of government finances in Georgia is unlike anything seen elsewhere in the Former Soviet Union (FSU). Following independence, in a context of economic collapse, civil conflict and political chaos, tax revenue virtually disappeared. By 1993, total revenues had fallen to 2.3 percent of GDP, covering barely 6 percent of total expenditure (on an accrual basis). This dismal revenue performance continued into the first half of 1994: during January - June 1994, revenues represented only 2.8 percent of GDP, or 5.4 percent of current expenditures. At the same time, the expenditure program was heavily burdened with subsidies (most notably for energy and bread, which in the first half of 1994 amounted to over 30 percent of GDP). The result was a soaring budget deficit, which reached 26 percent of GDP in 1993. The overall deficit increased further during the first half of 1994, to reach 44 percent of GDP, before the implementation of the Government's adjustmnent program restored some degree of fiscal balance (Table 1.1). Table 1.1: Key Fiscal Indicators, 1993-95 (% of GDP) 1993 Jan-Jun 1994 Jul-Dec 1994 1995 Total revenue 2.3 2.8 4.5 5.2 Grants 7.4 5.8 3.1 2.0 Expenditure and 35.9 52.3 19.5 12.9 net lending Accrual deficit 26.2 43.7 11.9 5.6 Source: IMF/World Bank on the basis of data provided by national authorities. 2. To finance the deficit the Government relied mainly on the accumulation of large domestic and external arrears, and on central bank credit. The unsustainable fiscal situation was largely accommodated by monetary policy, fueling a hyperinflation of 60-70 percent per month; this, in tum, led to a depletion of reserves and a sharp depreciation of the domestic currency. The extemal debt mounted to unsustainable levels. By the end of 1994, three years after entering independence with modest extemal obligations, the country had accumulated a stock of extemal debt of almost US$1 billion-equal to almost 80 percent of Georgia's GDP. The huge imbalance between govemment revenues and expenditure needs, moreover, forced the Govemment to increase its reliance on intemational humanitarian assistance, which at the end of 1994 still represented nearly 45 percent of govemment revenues. 3. Faced-with limited budgetary resources, the Govemment ceased to perform many of its basic functions-most importantly those of maintaining law and order, providing basic safety net protection for the poor, and maintaining critical infrastructure. Expenditures on health and education declined from 12 percent of GDP in 1991 to less than 2 percent in 1994-a decline of over 90 percent in real terms. Spending on maintenance of essential infrastructure basically disappeared. The result was a marked deterioration in many measures of social welfare: infant mortality rates (IMRs) rose by 13 percent to reach  2 21.4 in 1993;' life expectancy declined for both men and women, and the age-adjusted mortality rate rose by 18 percent; net school enrollments fell by 11 percent; and critical transport assets such as the main road system (the main east-west road from Azerbaijan to the Black Sea and north-south road from the Armenian to the Russian borders) fell into a state of disrepair. 4. Politically and economically, the situation has improved markedly since late-1994. Following the establishment of a cease-fire in Abkhazia, the Government was able to focus once again on the task of reforming and rebuilding the Georgian economy. With support from the IMF and the World Bank, the Government designed a successful stabilization and reform program. The cornerstone of the program was a drastic fiscal adjustment, accompanied by tight monetary policy. The impact has been impressive. The overall budget deficit, on an accrual basis, has been reduced from 26 percent of GDP in 1993 to about 6 percent in 1995, mainly through sharp cuts in expenditure. The ratio of revenue (excluding grants) to current expenditure has increased from 5.4 percent in the first half of 1994, to 27.2 percent in the second half, and to 56.1 percent in 1995. The measures supporting this adjustment included the reduction of subsidies through price increases for bread, gas and electn'city;2 improved monetization of wheat grants; a wide range of measures to improve tax administration and tax policy; and a close to 40 percent reduction in the number of government positions during Figure 1.1: Revenues as % of GDP, 1994 1995. Georgia , : .... Georgia . .. .... .. 5. Despite this initial success, China.... ....... Tx Rvnu Georgia's fiscal situation remains precarious. Notwithstanding government actions to Peru U Total Current increase tax collections, revenues are still Ghana Revenue------extraordinarily low: estimates for 1995 show Turkey . . ... tax revenues reaching only 3.7 percent of Cidonesi .. GDP; the lowest tax to GDP ratio anong the Chile former FSU econormies, and below the level Vietnam observed in some of the poorest countries in Albania _ L , . . . the world (Figure 1. 1). As a result, Georgia .:n ........... continues to depend on external, mainly in- kind, grants for a quarter of its total revenues Estonia . ................(nearly 2 percent of GDP). Latvia Buga i .-.- ..,,.'.:..: :'' :" :''... ............ Bulgare _ _ _ _ , 6. On the expenditure side, fiscal 0 10 20 30 40 adjustment has translated into an almost unparalleled compression of public spending. Total expenditure has fallen from 52.3 percent of GDP in the first half of 1994 (24 .2 percent for the year as a whole) to 12.9 percent in 1995 (Table 1.1). While essential to restore fiscal balance, such a compression of spending has implied drastic cuts in the financing of key public services. On a per capita basis, spending on education and health is now well below the average for the rest of the region, and below levels seen in much of the developing world (Table 1.2). Given that one of Georgia's sources of comparative advantage is its relatively high level of human capital, this is a source of concern. Spending on maintenance of essential infrastructure has also basically disappeared. In 1995, for example, outlays for routine road maintenance represented less than Compare to LIMRs of 13-16 for the Baltics, Belarus and Ukraine; of 12-14 for much of Central and Eastern Europe, and of 21 for Russia. 2 Subsidies were reduced from close to 39 percent of GDP in the first half of 1994 to less than I percent in 1995.  3 one percent of the estimated minimum needs. Similar pattems were seen in other key infrastructure sectors, such power, where improper operational practices and insufficient maintenance has led to a serious deterioration of facilities.3 Inadequate and unreliable electricity supply has become one of the major bottlenecks in the economy; yet because of the financial deterioration of the electricity company, few if any resources are being devoted to maintenance mad rehabilitation. There is a real concem that current levels of spending on these essential public functions may prove insufficient to ensure economic recovery and may hamper Georgia's prospects for long-term growth. Table 1.2: Per Capita Spending on Education and Health for Selected Countries Country GDP per capita Population Education Health (US$) (mnillions) $ p. cap % GDP $ p. cap % GDP Armenia 570 3.7 6 1.0 7 1.2 Sri Lanka 600 17.2 17 2.8 8 1.4 Indonesia 740 187.2 12 1.6 3 0.4 Bolivia 760 7.1 22 2.9 14 1.8 Philippines 850 64.8 24 2.9 5 0.6 Moldova 870 4.4 52 6.0 34 3.9 Lithuania 1320 3.7 51 3.9 41 3.1 Greece 7390 10.4 228 3.7 358 3.2 Georgia 440 5.4 4 0.9 3.3 0.8 Source: World Development Report 1995. Data for Georgia refer to 1995. 7. There are many distortions in the allocation of public expenditure which contribute to lowering the efficiency with which scarce public resources are used. Because of Georgia's energy deficit, a disproportionate amount of public resources must be devoted to utility payments: over a third of the educational budget in 1995, for example, was allocated to this task. This made it necessary to squeeze all other expenditures, and most notably wages. In 1994, wages and salaries of public employees represented only 3 percent of total expenditure, or a miniscule 0.8 percent of GDP. Since then, the share of wages and salaries has increased to about 13 percent of total expenditure (or 1.7 percent of GDP); but remains well- below the norms observed in other countries.4 Combined with considerable overemployment in the public sector, this has translated into very low average wages (about US$12 per month in December 1995). Such meager levels of pay cause government workers to allocate much of their time to alternative jobs, to the detriment of public services. Another source of inefficiency is that many public services (especially at the municipal level) are still being provided in a highly subsidized manner, when scarcity of resources would warrant greater cost-recovery. Even when services are provided at a nominal charge, non-payment means de facto free provision: as of February 1996, collections on electricity payments from households were just 15 percent of supply. In the energy sector in particular, lack of payment discipline by users has led to the financial deterioration of the utility companies, and to the consequent deterioration of services. The result is a vicious circle in which lack of resources reduces the amount and quality of services, which in turn reduces people's willingness to pay and lowers collections. With resources as limited as Georgia's, the failure to allocate spending efficiently is particularly costly. 3 Sakenergo's 1994 budget, for example, allocated less than 0.3 percent of total expenditure to repair and maintenance (about 3 million lari). 4 The average for East Asia, Latin America and Africa is 33 percent of total expenditures (or 6-8 percent of GDP), whereas in the OECD public wages account for about 25 percent of total expenditure (8-10 percent of GDP). In the Baltic countries, the corresponding figure ranges from 15 percent of total expenditure in Estonia to 21 percent in Latvia.  4 8. The challenge for Georgia to deliver key public services within its very constrained resource envelope is enormous. With revenues projected to reach only 9-13 percent of GDP during the 1996-98 period, the pressure to maintain a tight expenditure program will continue. This is especially so if Georgia is to become creditworthy, and hence eligible for receiving significant foreign financing. On the other hand, the government clearly needs to make some essential investments in the economy (particularly those in human and physical capital) to ensure sustained growth and the continued welfare of its population. Georgia will not be able to meet this double challenge without a thorough redefinition of the role of the state in its economy. 9. In many ways, the decline in the size of government in Georgia may prove to have a silver lining. Unlike other countries in the FSU, the Georgian economy is not burdened by a huge and omnipresent government which absorbs and wastes resources. And given the collapse of government finance, there should be fewer vested interests in maintaining the old state regime. Moreover, the political economy of many key reforms, such as the reform of the pension system, could be facilitated to some degree by the earlier collapse. All of this should make the transition to a market economy easier. However, none of it will prove ultimately beneficial unless the Government is able to fulfill its key mandate of delivering sustained growth and improvements in the living standards of the population. Public expenditure policy has an important role to play in achieving this goal. B. Medium-Term Fiscal Framework 10. The main constraint Georgia faces in designing its public expenditure policy is the availability of resources. With little ability to borrow domestically and limited foreign exchange reserves, the government has only two sources of financing for its budget deficit: central bank credit and foreign borrowing. However, the extent to which it can rely on either one is greatly constrained by the need to consolidate the progress made against inflation, on the one hand, and by creditworthiness considerations on the other. In this sense, the government has little scope to relax its fiscal stance: how much it is able to increase expenditures over the next few years will depend primarily on its ability to pay for them through increased revenues, and to a lesser extent on the availability of external financing on purely concessional terms. 11. This section first presents a medium-term fiscal framework that is consistent with Georgia's projected stream of revenues and expected foreign inflows. It then considers whether a higher spending path could be sustained, and whether there are viable sources of financing for that incremental spending. A sustainable fiscalframework 12. To maintain macroeconomic stability, Georgia needs to ensure that its medium-term fiscal position is compatible with its balance of payments, monetary and inflation targets. Table 1.3 shows the evolution of key macroeconomic variables under a sustainable fiscal policy scenario. The key characteristic of this scenario is that it assumes a tight fiscal program, with deficits declimnig to 3 percent of GDP by 1998, and down to under 1 percent by 2004. Such fiscal discipline is a prerequisite to reducing public dissaving, and to restoring positive domestic savings in the economy by 1997. This, in turn, is critical to the recovery of domestic investment that is needed to sustain growth. Total domestic investment needs to triple over the next few years if Georgia is to sustain growth rates of 8-10 percent per year. This cannot happen without a significant increase in government savings. 13. Reducing public dissaving is also crucial to improving Georgia's external accounts. With a totai stock of debt of about US$1.0 billion (nearly 80 percent of GDP), and a current account deficit of more than 15 percent of GDP, Georgia is at present very far from being creditworthy. As a result, its access to  5 external financing from anything but official sources (and on concessional terns) is very limited. However, concessional flows will not be available indefinitely. In anticipation of such flows drying up, Georgia needs to move toward restoring some degree of commercial creditworthiness. As illustrated by Table 1.3, lower fiscal deficits would lead directly to significant improvements in the current account-from a deficit of 15.3 percent of GDP in 1995 (excluding grants) to 1.7 percent by the year 2004. This, in tum, would contribute to lessening Georgia's dependence on foreign inflows. As a result, total debt to GDP would decline gradually over the period, from 39.5 percent in 1995 to just under 20 percent in 2004. Total debt to exports would fall in parallel from 250 percent in 1996 to 187 percent in 2004, and debt service to exports would remain below 30 percent. Table 1.3: A Sustainable Fiscal Framework 1995 1996 1997 1998 2000 2004 Real Growth Rate GDP 2.8 8.0 10.0 10.0 8.0 5.0 Inflation (end of period) 64.8 23.1 8.5 6.6 6.2 6.2 As % GDP Current Account (Excld Grants) -15.4 -7.2 -5.4 -4.2 -3.1 -1.7 Domestic Savings -8.8 -0.6 3.6 7.8 12.4 17.5 Investment 3.7 5.1 8.0 11.2 14.7 18.4 o/w: Public 1.0 1.2 2.2 2.9 4.1 4.8 Government Revenue 6.9 9.4 11.3 12.7 15.5 17.6 o/w: Tax Revenue 3.6 6.7 8.1 9.7 13.2 14.9 Government Expenditure 12.5 12.8 14.5 15.7 18.2 18.5 Fiscal Deficit (Incld Grants) -5.7 -3.4 -3.1 -3.0 -2.7 -0.9 Fiscal Deficit (Excld Grants) -7.5 -4.2 -3.8 -3.5 -3.0 -1.0 Total Debt/GDP 37.7 27.0 24.7 24.4 23.4 19.5 Debt Service after rescheduling/Exports 28.6 10.7 28.0 11.8 20.4 27.9 Total Debt/Exports 241.9 248.7 245.5 242.2 228.9 186.6 14. Achieving these fiscal targets will not be easy as they require both significant improvements in revenue mobilization, and continued control over expenditure. On the revenue side, the projections assume that efforts to expand the tax base via the removal of exemptions and to improve tax administration are successful. This would allow tax revenues to increase to 6.7 percent of GDP in 1996, and reach nearly 15 percent by the year 2004. Total revenues would increase in parallel from 9.4 percent of GDP in 1996 to 17.6 percent by 2004. Although modest by most countries' standards, achieving even this level of government revenue would allow for a gradual recovery of expenditure, to reach 17 percent of GDP in 1998, and stabilize at 18-19 percent thereafter. 15. A tight fiscal program alone, moreover, will not guarantee the recovery of growth. Fiscal efforts need to be complemented by progress on the structural measures-privatization, trade liberalization, institution building-that are key to the development of a dynamic private sector. Equally important is that Georgia address the problems of its energy sector, which lie at that core of the economic collapse. A first priority in this regard is to enforce payment discipline, so that the energy companies can operate efficiently and deliver critical energy services, without which the economy cannot function. A second priority is to reduce the burden of energy imports-through a combination of increasing energy use  6 efficiency and developing domestic energy resources. The composition of public spending needs to change to reflect these needs: devoting more resources to institution building and indirect support of private production, while shifting out of directly productive activities (Chapter 2). 16. Even with improved revenue performance and increased domestic savings, Georgia's domestic resources will remain insufficient to finance its full expenditure needs. Substantial additional financing from external sources will continue to be needed. Average gross financing requirements for 1996-2004 are projected at about US$420 million per year. The bulk of these resources would have to come from official sources, and, at least initially, on concessional terms. Private foreign financing is exected to play an important role, but only in the medium term. Can Georgia sustain larger fiscal deficits? 17. The deficit path projected in Table 1.3 leaves room for only a modest recovery of public expenditure. But is there room for a bigger recovery of spending? In the absence of parallel improvement in revenue performance, higher spending would imply running a larger fiscal deficit. Whether this could be sustained over the medium term would depend on how the incremental deficit was to be financed, and on the impact of additional spending on growth. If higher spending were to lead to a higher growth rate-for example by allowing for more public investment in high-return infrastructure-a higher fiscal deficit may indeed be sustainable. However, this argument can be easily oversold: public investment can also be inefficient and generate low returns; more importantly, even if public investment has a high return, the government must be able to capture the additional retums if the increase is to be self-sustaining. In the case of Georgia, this would require a significant improvement in the government's ability to collect revenues from the private sector through cost-recovery and taxes. 18. In principle, a larger fiscal deficit path could be financed either through an expansion of credit from the central bank, through increased external borrowing, or through some combination of both. Financing a larger deficit path through foreign borrowing would tend to appreciate the exchange rate, encouraging imports and worsening the current account deficit. Direct foreign borrowing on commercial terms would also add to Georgia's already large debt burden, and make it difficult for the country to meet its debt service obligations. As illustrated by Table 1.4, even a relatively modest increase in the fiscal deficit of 2 to 3 percentage points per year (to average 5.5 percent per year over the 199 7-2004 period), would more than double key debt and creditworthiness within a ten year period. The current account deficit would double relative to the sustainable scenario (remaining above 6 percent of GDP for the whole period). To finance a larger external imbalance, Georgia would have to accumulate twice as much debt; as a result, the stock of debt would climb to over 5 billion by the year 2004. Total debt to GDP ratios would reach 40 percent, and debt service to exports would increase to 50 to 60 percent within ten years. Given this scenario, it is highly unlikely that commercial external financing would be available. Foreign borrowing on commercial terms is simply not a viable financing option for Georgia at this time. 19. A larger deficit path could also be financed by an expansion of credit from the central bank to the government (or by a combination of increased credit and foreign borrowing). Faster credit growth, however, would quickly lead to more inflation. Under the best of scenarios-assuming fairly high money holdings relative to GDP and fast GDP growth- the government could finance a deficit increase of 1 to 2 percentage points of GDP through seignorage. Anything beyond that, however, would be inflationary. In the case of Georgia, moreover, and given its recent hyperinflation history, the inflationary impact of a credit expansion would probably be felt sooner. Higher total spending and higher inflation would in the short term work to appreciate the real exchange rate, damaging export growth and favoring imports. This would lead to a worsening of the current account. In the medium term, slower export growth would also  7 feed back to lower total GDP growth. A continued credit expansion, moreover, would ultimately require a devaluation of the nominal exchange rate, which would further fuel inflation. The illustrative projections suggest that under this scenario, the current account deficit would remain above 4 percent of GDP for the full period. The total stock of debt would reach US$6 billion and debt service to exports ratios would increase steadily to reach 56 percent. Inflation would remain stuck at above 20 percent. Table 1.4: Key Creditworthiness Indicators under a Larger Fiscal Deficit Scenario as % of GDP 1996 1997 1998 1999 2001 2004 Financed via external borrowing: Fiscal deficit (excl. grants) 6.0 6.6 6.2 6.4 5.4 4.8 Current account (excl. grants) -9.8 -7.5 -6.5 -7.2 -6.9 -6.1 Total DOD (US$M) 1462 1783 2147 2585 3587 5240 Total Debt/GDP 30.9 29.8 31.0 33.6 37.6 42.6 Debt Service after rescheduling/Exports 10.7 29.3 14.7 17.5 41.0 60.5 Growth rate of GDP 8.0 10.0 10.0 8.0 8.0 5.0 Financed through credit expansion: Fiscal deficit (excl. grants) 6.1 7.0 6.8 6.8 4.4 3.0 Current account (excl. grants) -9.8 -6.6 -5.7 -5.3 -4.7 -4.3 Total DOD (US$M) 1462 1734 2081 2463 3522 5992 Total Debt/GDP 30.9 28.7 27.5 26.4 24.3 25.6 Debt Service after rescheduling/Exports 10.7 29.3 14.5 16.6 34.7 56.4 -Growth rate of GDP 8.0 10.0 10.0 8.0 6.0 2.5 20. These two alternative scenarios highlight the severe constraints facing Georgia's expenditure policy. The legacy of hyperinflation seriously limits the government's scope for financing larger deficits through central bank credit, while the possibility of borrowing domestically is still some years away. At the sarme time, Georgia's large stock of existing debt undermines its creditworthiness and precludes foreign borrowing on commercial terms. In the absence of additional concessional financing or debt reduction, running a fiscal deficit higher than 2-3 percent of GDP is unsustainable over the medium term. Composition of expenditure. 21. While the sustainable scenario envisages a moderate increase in expenditures, total spending has to be kept within the available resource envelope, so as to allow for a reduction in the fiscal deficit to under 2 percent of GDP by the year 2001. The key to fulfilling essential public functions within this limited spending envelope will be to maximize the efficiency with which scarce resources are used. The composition of expenditures hence becomes particularly important. Because of this, the sustainable fiscal scenario envisages several key expenditure policy measures: * An increase in public investment, which would rise from 1 percent of GDP in 1996 to 4.8 percent of GDP by 2004. This represents a minimal increase in investment, without which it would be difficult to sustain the projected high growth path. To accommodate such an increase in capital spending, it will be necessary to contain current expenditures. * An increase in the share of expenditures going to wages and salaries, from 13 percent of total spending in 1996 to 20 percent by the year 2004 (Table 1.5). In combination with a rationalization of  8 government employment, this would allow for a substantial increase in the real public sector wage-a prerequisite to improving efficiency in the civil service. * Successful reforms in the energy sector that ease supply constraints and allow for a gradual reduction of energy-related payments in the state budget. This would allow expenditures on other goods and services (which are primarily utility payments) to decline from their current level of 24 percent of total spending to under 20 percent, releasing resources that can be devoted to other critical activities. * A reduction in extra-budgetary social security expenditures, as a result of pension reform. Given Georgia's age structure, pension liabilities will inevitably rise over the next three decades; financing these liabilities under the existing pay-as-you-go system will require either raising the already-high contribution rate or allocating more budgetary resources to financing the deficit of Pension Fund. Either option is undesirable: the first because it will distort the labor market further; the second because it would strain the total spending envelope. Table 1.5: Projected Government Expenditure by Economic Classification, 1995-2004 (as percent of total expenditure, excluding net lending) 1995 1996 1997 1998 2000 2004 Current Expenditure 0.92 0.91 0.85 0.82 0.78 0.74 Wages and salaries 0.14 0.13 0.15 0.17 0.18 0.20 Other goods and services 0.24 0.28 0.25 0.21 0.23 0.19 Transfers 0.09 0.09 0.10 0.11 0.11 0.11 Extrabudgetarysocial 0.12 0.14 0.12 0.10 0.07 0.06 Interest payments 0.13 0.07 0.08 0.07 0.06 0.06 Other expenditure 0.20 0.19 0.14 0.14 0.14 0.14 CapitalExpenditure 0.08 0.09 0.15 0.18 0.22 0.26 Source: Annex 3. 22. What are the risks associated with the failure to implement these expenditure shifts? Less public investment would be reflected in lower quality infrastructure services to support the private sector, which would work against private investment, and risk lowering the overall GDP growth rate. A lower growth rate would, in tum, undermine the sustainability of the projected level of fiscal deficits, which would have to be adjusted accordingly. Similarly, the failure to successfully reform the energy sector would risk perpetuating the current situation, with a sizable fraction of govemment resources going to finance utility payments, to the detriment of public wages and other key elements of public expenditure. Continued energy problems, moreover, would constrain the development and growth of a private sector, and act as a disincentive to foreign investment. Failure to control mounting pension liabilities could risk blowing the spending ceiling; or require compressing other-more productive-social expenditure. All of these would lower the expected growth path. A role for additional concessionalfinancing 23. Given Georgia's extremely tight aggregate spending envelope, additional public spending could, at the margin and if directed at critical bottlenecks, be very productive. Incremental spending on key public investments in infrastructure, for example, could contribute to increasing the productivity of many private activities and to raising the growth path. Similarly, additional expenditure on building key market institutions (an appropriate legal framework and court system, for example), could have a big payoff in  9 terms of attracting private (both domestic and foreign) investment. But because resources are so constrained, and financing possibilities so mirted, many of these key investments may not get made; or may be delayed for several years. 24. In this context, additional external financing on concessional terms could have an important role to play over the next few years. By relaxing the resource constraint, such financing could allow Georgia to address its key investment needs and provide an additional spurt to growth. By underpinning necessary public investment, concessional financing would also help draw in private flows, and lay the basis for improving Georgia's commercial creditworthiness in the medium term. In absolute terms, the amounts needed are not huge: an additional US$ 80 to US$ 110 million of concessional financing per year over the 1996-1998 period would be sufficient to allow the government to increase public investment by approximately 2 percentage points of GDP per year. Assumning that these resources were directed to high- return, high-priority activities, and assuming the govermment improves its capacity to capture part of the return through taxes, the payoffs to these investments amply should exceed the government's borrowing costs. 25. However, the need for additional concessional financing comes on top of already-identified needs of about US$420 per year, a substantial part of which would already have to be concessionary in nature. Georgia's ability to mobilize these magnitudes of concessional financing will require demonstrating a particularly strong commitment to reform, and putting together a prioritized and focused public investment strategy.  10 Chapter 2: The New Role of Government A. Rebuilding Government 26. Unlike many other economies of the FSU, Georgia is faced with the task of rebuilding its government rather than reducing it. And it must do so while facing an incredibly severe revenue constraint which is unlikely to ease significantly in the near term. In such a context it has little choice but to limit its public interventions to a bare minimumn-to those core functions which are essential to ensure the future growth and creditworthiness of the economy and an adequate level of welfare for the population. The task of defining these functions is not easy: the choice must combine economic criteria of efficiency with distributional considerations and social equity objectives. And given Georgia's limited revenues, it is bound to involve difficult tradeoffs between competing uses, all of which may have legitimate claims on public resources. 27. The definition of what constitutes a core function can be guided by two criteria . First, by the relative role of the private versus public sectors in a market economy (see Box 2. 1). Govermnent resources should not be wasted on services that the private sector can provide on its own. Instead, public intervention should concentrate on those activities that it does best, and where market and socially desirable outcomes diverge significantly. Second, spending allocations should be based on the relative urgency of the expenditure. Certain areas of public spending may be neglected for a period of time without irreversible danage; other less so. Taking this intertemporal dimension into consideration can help establish priorities. Expenditures on agricultural extension services, for example, may be part of a set of functions that the state ultimately wants to keep; yet if the state does not spend any money on these activities for a few years, the damage to agriculture will not be all that expensive to recoup. In contrast, two or three more years of neglect of basic maintenance could lead to the irreversible deterioration of Georgia's main roads, and a significant larger cost of reconstruction would have to be incurred to recoup the damage. Box 2.1: The role of the private versus public sectors In a market economy most economic decisions are made by the private sector on the basis of market forces. However, governments often do intervene in markets; usually to address what is perceived as a failure of the market outcome, or to pursue broader social equity and fairness objectives. The extent to which governments should or should not intervene has long been a subject of much debate. While there is no standard recipe for the precise mix of private versus public activities in a market economy, there is nevertheless broad consensus on the reasons for government intervention. Most economists would agree that a role for government is warranted in the following areas: . Ensuring the provision of goods and services subject to market failure. This includes goods and services which private markets will not provide (public goods such as defense or law and order), or which they will provide in insufficient quantity (services such as immunization, for which private and social returns to the activity may differ). . Establishing a legal, regulatory and policy environment to facilitate private sector activity and the effective operation of competitive markets. . Pursuing poverty alleviation and other distributional objectives through income transfers and other measures to protect the poor. Although these three arguments provide clear theoretical rationales for government intervention, in practice choices are much more difficult. Few goods are pure public goods; many more are mixed-goods that are supplied by private markets although in insufficient quantities, or goods that are supplied publicly but could be privately provided. Moreover, govermnents need to decide not only when and where to intervene, but also how they will do so. Should government provide a public good directly? Should it encourage its production indirectly? Or should it subcontract its provision to private firms? All these decisions have implications for the efficiency with which services are delivered, and for fiscal expenditure. For Georgia, limited public resources dictate the need to maximize private sector participation wherever possible.  11 The role of the private versus public sectors 28. The respective roles of the private versus public sector, and the distinction between public financing and public provision, can be used to derive a taxonomy of goods and services in terms of the importance of public sector involvement (Table 2.1). This taxonomy then provides a basis to prioritize items for Government spending. A first priority is spending on pure public goods-law and order, the legal framework, municipal lighting, the basic safety net; these the government will have to continue financing and providing, as the private sector largely cannot and will not do so. Second, are "mixed" private/public goods in which the public sector will retain a primordial role even in the medium term: the provision of basic health and education, the maintenance of the main road system, and water supply and sanitation services. Third are "transitionally public" goods, activities that are part of the public domain now but that are to be shifted to private ownership as soon as an appropriate regulatory and institutional framework is in place. This applies, most importantly, to energy services, but also to a number of educational and health services, telecommunications services, urban transport, and railways. Finally, there is a whole set of goods that can be adequately provided by the private sector, and on which the govermnent should not waste its resources. Table 2.1: A Taxonomy of Activities by the Degree of Public Sector Involvement Public Sector Involvement Activity Pure public Defense, Law and Order, Regulatory and Legal Framework, Social Assistance Municipal Streets and Lighting, Environmental Protection. Public/Private Education, Health ,Water Supply and Sewage, Highways, Social Insurance, Agricultural Extension. Predominantly private w/ role for government Electricity, Heating and Gas, Municipal Transport, Telecommunications, Railways, Airports and Ports, Research and Development. Pure Private Agriculture, Mining, Manufacturing, Commerce and Trade, Tnicking and Shipping, Construction, Housing, Finance, Tourism, Sports, Media and Communications. The time path of expenditures. 29. Georgia is very much an economy in transition: not only in the midst of a difficult passage from administered to market-based economy, but also in the process of rebuilding its state after its post-war collapse, and within a context of very limited resources. In setting priorities for public expenditures, this "transitional' aspect needs to be taken into account. Georgia 's long-range preferences as regards the role of the state in the economy may be quite different from what it can afford today. Decisions on expenditure allocations should distinguish explicitly between medium-term objectives and immediate needs; allocating expenditure to where it is more urgently needed, while putting in place a strategy that will allow for remaining public functions to be adequately fulfilled as revenues increase. 30. How can the urgency of different expenditures be adequately compared? A simple rule is in terms of the "irreversibility" of the damage done by neglecting spending in a particular area. This has to weighed against the cost of maintaining that spending. An example of this is presented in Box 2.2. Using  12 this analysis, it is possible to draw a preliminary assessment of the urgency of different expenditures (Table 2.2). Table 2.2: The Timing of Different Public Expenditures Time ath Activi Most urgent - neglect over the next 3 years highly * Public safety costly; little room for private activity and/or cost- * Courts and legal system recovery; no alternative to public resources. * Emergency maintenance in power and main transport routes. * Emergency maintenance of water distribution and sewage systems. Urgent - neglect over the next 3 years costly; . Basic public health (prevention) and essential however, private resources and/or informal clinical services. mechanisms partially addressing these needs; costs of temporarily delaying spending are lower than for * Pnmari educatlon: ensunng access and first group. maintaining minumal standard of quality. * Minimal safety net. Important, but less urgent - private sector and/or * Secondary education: maintaining access for cost-recovery can generate sufficient resources to the poor and maintaining quality. meet the bulk of priority needs. . Full rehabilitation of main roads, and backlog of maintenance on secondary roads * Backlog of maintenance and rehabilitation in other key infrastructure sectors: irrigation, power/gas, railways and ports. Box 2.2: Assessing the costs of delaying spending: an 31. The areas where public expenditures example from the transport sector are needed most urgently include: police Routine maintenance if the main road network is estimated to protection and maintenance of public safety; require outlays of approximately US$6 million per year. This emergency maintenance in the power sector is equal to the costs of resurfacing I km of road * length of the (which remains a bottleneck to growth in the main road network (US$6000* 1000kin of main highway). If economy); emergency maintenance of main this expenditure were to be postponed for 3 years, the savings transport routes (to ensure they are not lost generated would amount to US$16.4 million (discounted). beyond a point of repair); addressing critical However, 3 years of additional neglect would almost certainly maintenance needs of water distribution and lead to a situation where the roads could no longer be sewage systems; and all expenditures related to resurfaced at the cost of US$6000 per km, but a substantially setting up a proper regulatory and legal larger cost of reconstruction would have to be incurred. At framework (which will be essential to US$132,000 per km, reconstruction of the main 1000km would require an investment of US$90 million (discounted). developing a pnrvate sector, and attracting And this would not take into account the economnic costs of foreign investment, and hence will be crucial to longer transport times. reducing medium-term demands on government Hence, for every $1 spent on current maintenance, $6 would resources). Second in order of urgency are have to be incurred to recoup the damage. To assess the expenditures on basic education and on relative urgency of this expenditure, this ratio needs to be essential public health activities; as well as compared to similar ratios for alternative uses of resources. expenditures on maintenance of a minimal level of social protection for the poor. While these  13 social expenditures are definitely of top priority, the costs of temporarily delaying spending in these areas are likely to be smaller than for the first set of activities. Many of these activities have a private retum, and informal mechanisms for delivering these services in the face of declining govemrnment revenues have sprung up already (see Box 2.3). Third in order of importance are expenditures on rehabilitation of existing infrastructure-full rehabilitation of existing energy facilities, main and secondary roads etc. Since these expenditures can probably be delayed for a year or two, the bulk of resources should be raised through cost-recovery and/or by drawing in private (domestic and foreign) investment. B. Expenditure Priorities by Sector 32. The two criteria described above provide useful guidelines for deciding on the appropriate cross- sectoral mix of expenditure. Priority sectors are those in which there is both a strong case for public intervention, and in which the costs of delaying spending are highest. Comparing this set of priorities to the current allocation of spending (Figure 2.1) suggests that some major expenditure shifts are needed. These include: * An increase in the resources devoted to priority public good activities: public order and safety, the court system, and further development of key economic institutions and public administration. At present , these expenditures account for about 15 percent of the total spending budget or 2 percent of GDP. This is approximately one-half the average level for developing countries.5 A reasonable goal is to aim for devoting somewhere between 3.5 and 4 percent of GDP to these tasks. * A withdrawal from directly productive activities in areas such as agriculture, forestry, industry, media and film. Government involvement in these areas should be primarily indirect, and focus on provision of an appropriate institutional and Figure 2.1: Functional Distribution of Expenditures regulatory framework. Overall, the total 1996 share of spending going to economic Other incl. Nabonal services or "national economy" will interest Econony probably not decline much in the near 22% 16% Heakh & term, as resources released from directly lucation productive activities will have to be Adrrinl & _ 15% employed in setting up the regulatory Rib Safety - 7--77 ^^ 15% n - and legal framework of a market 9%fn Social economy. As resources expand in the 23% future, however, the share allocated to these 'economic services" should indeeed decline. An increase in the resources devoted to maintenance and rehabilitation of existing infrastructure. The top priority in this regard is addressing the backlog of maintenance needs in critical infrastructure sectors- power and transport (mainly roads and railways). This will demand a significant up-front investment of resources: about US$40 million per year for maintenance and rehabilitation of the main road system, and another US$50 to 60 million per year for the power sector. To accommodate this, the share of resources devoted to capital expenditure will have to increase significantly. Pradhan (1995) reports a mean for developing countries in 1990 of about 5.5% of GDP (see Annex 2).  14 * A reorientation of the social-cultural budget to focus primarily on education, health and social protection, reducing expenditures on cultural activities, sports, and subsidies to the mass media. While there may be a case for public spending on some of those activities, these are second priority and may have to wait until the resource envelope expands sufficiently. Curtailing these secondary social expenditures would allow spending on critical health and basic education services to increase significantly, without a need to increase the total social expenditure envelope. Over the medium-term, new financing mecanisms for activities such as culture will have to be plored, to bring in financing from the private sector and to increase cost-recovery. * Within education, health and social protection, a re-focusing of expenditures on the priority sub- sectors: public health and minimum essential clinical services, basic education, and a minimum social assistance program (which would include the current minimal, universal pension). This would ensure that public resources go to those activities where social returns are highest and would concentrate resources on those services which the poor tend to use more. * To make room for the increase in the share of spending going to pure public goods, the share of "other expenditures" will have to decline. At present, "other current expenditure" and "unclassified" represent almost 17 percent of total current spending. Lower energy-related payments, and an appreciating real exchange rate, should greatly contribute to reducing the share of "other expenditures". This should be further facilitated by a gradual decline in interest payments after 1998. * In parallel, some consideration should be given to what is a desirable level of defense expenditures. Military expenditures currently represent 9 percent of spending or 1 percent of GDP. However, the trend is one of slow increase which, if continued, may imply reducing expenditures on high-priority areas-such as basic health or public administration-where resources are already inadequate. C. Policies for Expenditure Reform 33. To achieve a better provision of public services within the overall spending constraint, reforms are be needed in three critical areas: (a) improving government pay and employment practices; (b) widening private sector participation and cost recovery; and (c) increasing maintenance and capital expenditures. Government Pay and Employment Practices6 34. An essential part of rebuilding government is strengthening the capacity of government employees to do their jobs. In the case of Georgia, this implies a profound reform of public administration. Fundamental to this task is the overhaul of government pay and employment practices, which have become completely dysfunctional and incompatible with a motivated and effective civil service. 35. Government employment in Georgia today is marked by two distinct traits: considerable overemployment and extremely low levels of pay. The meager remuneration causes government workers to spend much of their time attempting to supplement their income through alternative work. In effect, much of the civil service functions as a transfer mechanism akin to a public works program by which employees receive a small income supplement from the government for putting in a few hours of work per week. This 6 This discussion draws extensively on the background paper prepared by Arup Banerji, on "Georgia: Issues in Government Pay and Employment Reform", World Bank, February 1996.  15 situation has not just lowered efficiency in the government, but has in some instances jeopardized its the basic civic functions. 36. Overemployment. In November 1995, the budgetary sector had positions for an estimated 366,900 employees (about 13 percent of the working-age population), down from 621,200 people in January 1995 (Table 2.3). While comparable to the situation in many FSU countries, the number of civil servants is extremely high when compared to most developing countries. Prior to reforms, Ghana's civil service employed 1.7 percent of its population; in Zambia the comparable figure was 2.1 percent. In 1990, the Argentine central government employed about 2.1 percent of the population, with local governments employing another 3.4 percent. Even in pre-reform Peru, total gove,nment employment accounted for about 3.4 percent of the population or 8 percent of the total labor force. In all of these countries, moreover, retrenchment of the civil service has been an important component of recent economic adjustment programs. Table 2.3: Employment in Budgetary Organizations, 1995 (thousands of employees) Sector January 1995 November 1995 Rep. Local Total Rep. Local Total Education 90.0 174.3 264.3 30.0 133.0 163.0 Health 23.5 122.2 145.7 0 0 0 Science 30.7 0.3 31.0 24.3 0.2 24.5 Administration 12.0 12.0 24.0 10.0 10.5 20.5 Culture 9.6 10.0 19.6 8.7 8.6 17.3 Tourism 2.7 0.8 3.5 2.2 0.5 2.7 Sports 2.0 0.5 2.5 1.5 0.4 1.9 Defense/Law Enforcement 67.7 2.0 69.7 67.7 2.0 69.7 Other* 34.6 15.3 49.9 48.4 18.9 67.3 Notes: * includes forestry, agriculture, environmental protection and railways. 37. In the case of Georgia, extensive overemployment is combined with a scarcity of skilled workers in some key areas and with a substantial misallocation of personnel resources. There is significant redundant employment in institutions that are responsible for activities irrelevant to a market economy (such as the planning of production targets or coordination of foreign trade). At the same time, there is a shortage of skilled people trained in market economics and finance. Many ministries and institutions have overlapping responsibilities, leading to a duplication of jobs. And some sectors (health and education, for example) are clearly overstaffed by market economy standards. 38. Extremely low pay. By August 1994, real wages of civil servants had declined to I percent of their 1992 level in real terms. The purchasing power of wages has since been partially restored by successive wage increases and by the appreciation of the currency. However, wages remain far below their pre- independence levels in real terms. After the latest increase in February 1996, the monthly salary in the budgetary sector ranges from 7.5 laris (about $6) for the lowest-paid unskilled worker to 26 laris (about $20) for the top grade.7 These wages are a fraction of what the same workers could earn elsewhere. A government accountant, for example, could earn 10 to 16 times her government wage by working in a privatized firm. Such wage differentials will make it difficult to retain skilled workers within the Personnel in defense and law enforcement receive higher salaries. As of Novemebr 1995, their average wage was 23.1 laris per month.  16 government as soon as private employment demand starts to pick up. Moreover, government wage scales are incredibly compressed: a senior specialist in the public administration earns only 1.4 times the wage of an unskilled laborer; a department head only 2 times the wage of an entry-level clerk. This degree of compression exceeds even those seen in highly-compressed pay structures in countries such as Tanzania, where in the mid-1980s a top public servant earned only 6 times an unskilled entry-level wage, or Mozambique, where a senior department head earned 5.4 times the wage of an unskilled clerk. 39. A large scale reform of the civil service is now underway. The major thrust of the reforn has been the removal of thousands of workers from the budgetary payrolls, mostly in a few sectors such as health, education and tourism. At the same time, there have been small increases in the pay received by remaining government workers. However, because of the budgetary constraints, these increases have been modest. They have done little to relieve workers of the impossibility of making ends meet on a single government salary. A more fundamental reform is still needed. The focus on reducing employment and increasing pay should be maintained, as ensuring appropriate employment and pay practices are a necessary (if not sufficient) element in building government administrative capacity. Efforts to increase pay, however, need to be combined with a substantial decompression of salaries and with attempts to link remuneration to job performance and productivity. 40. Recommendations. For Georgia, the prinary objective of civil service reform is to reduce employment to an efficient minimum, while paying remaining employees reasonable wages. Doing this will almost certainly involve increasing the present wage bill. This objective has been worked into the sustainable scenario discussed above, which projects an increase in the share of total spending going to wages and salaries from its present level of 13 percent of total spending ( 1.7 percent of GDP) to 20 percent of total spending ( 3.7 percent of GDP) by the year 2004. Combined with a steady decline in government employment of 5 to 6 percent per year over the 1996-1999 period (a cumulative 16 percent decline), this will allow for a significant increase in the average real wage of public sector workers. The bulk of the employment decline should come from the overstaffed education sector, which accounts for over 40 percent of total public employment. Additional cuts in employment will have to come from non-priority sectors such as culture, sports, science, and in particular from the "other" employment category, which has actually grown over the last year. Under this scenario, the average public sector wage would converge to 75 percent of per capita GDP (a rough proxy for the average economy-wide wage) by 2004. To sustain a larger wage increase, however, more drastic cuts in employment would be needed 41. In the short term (next twelve to eighteen months), reforms will have to focus on implementing agreed-upon redundancies, especially in the education sector. Meanwhile, the processes of finding and eliminating "ghost" or non-existent workers from various organizations' payrolls, elimination of all unfilled vacancies and enforcement of the retirement age should be continued and accelerated. At the same time, there needs to be a continued hiring freeze except in specialized, high-need areas. In parallel, the groundwork for later reforms needs to be laid-by deciding on desirable government structure and size, and using this information to identify the targets of further redeployment. To do this, the govemment has to agree on what its core functions should be. This in tum will affect the plans to rationalize the structure of ministries and administration, eliminating some functions and consolidating others. Once the core functions of ministries are understood, functional reviews will have to be carried out, beginning with a few key government agencies (such as the Ministries of Finance, Economy, Education, Health, etc.). These will attempt to determine whether existing staffing and organizational structures are appropriate for carrying out each agency's purpose. 42. A process of reducing and simplifying the salary grade structure should be initiated and followed up during the medium term. The current multi-layered pay structure is unnecessarily complex. The  17 existing 22-level structure should be reduced to fewer levels-explicitly differentiated between top management (including ministers), upper administrative staff, professionals and workers. The US Federal Pay Structure, for example, has 15 levels, plus a separate "Senior Executive Service" for the highest ranking federal officials. Instead of having a single salary rate for each grade, a small range can be set within grades; within-grade salary increases can be given as rewards for discernible achievements on the job. Moreover, the pay structure should be decompressed much further, concentrating the decompression on the intermediate professional grades and at the very top-where the salary differentials with the private sector appear to be largest. Within this framework, pay should be increased to levels which provide a greater degree of incentive. Initially, the level of government pay need not aim to be the same as in the private sector, as government work has some compensatory factors (usually more employment security) which can allow the salary to be below he alternative private sector wage. Private Sector Participation and Cost Recovery 43. There is a whole range of goods that are supplied publicly but that can also be privately provided. These include services such as education, where the existence of social externalities and of equity objectives may warrant public intervention even though privately-provided substitutes exist; and may include goods such as electricity or telecommunications, which can be fully provided by private markets but require that an appropriate institutional and regulatory framework first be put in place. Given Georgia's scarcity of public resources, the government should try to maximize private section participation whenever possible, and encourage private sector entry into all of these services. 44. Realistically, it will take some time to mobilize private financing for sectors such as education, housing or municipal services; the public sector will have to continue to play a key role in their direct provision. However, for sectors such as electricity generation (especially hydropower), ports or urban transport, it should be possible to attract private involvement fairly quickly, provided an adequate market and regulatory framework exists (it will be hard to attract resources to electricity as long as collections remain abysmally low, for exxample). Putting in place an adequate regulatory framework for such "privatizeable" activities and facilitating entry is hence of utmost priority. In parallel, where the state is to retain a role in provision of these mixed goods, it makes sense for it to seek at least partial cost recovery. 45. Cost recovery policies involve imposing part or all of their costs on the actual users and beneficiaries through user charges, rather than on the general public through the tax system. 9 The increased application of cost recovery policies has several benefits. First, it reduces the dependence on tax revenues and increases the resources available to the government for provision of key services. Second, it increases efficiency by moving the price of the good in line with the cost of provision, hence encouraging consumers to rationalize their demand. And third, it encourages the "commercialization" of public sector activities, making them more efficient, increasing their quality, and paving the way for private sector competition. 46. Cost recovery policies make sense only when users andlor beneficiaries can be identified and non- payers excluded from using the service. In such a case, a key aspect of setting cost-recovery targets is assessing users' willingness to pay for the good. Much can be learned in this regard from Georgia's recent experience, as the collapse of government revenues meant that de facto many services the state had 8 For example, studies of U.S. federal and state government employees indicate that they earn on average 20 to 40 percent less than comparable workers in the private sector. 9 Given low tax compliance in Georgia, cost recovery has the side benefit of eliminating the implict subsidization of non tax payers by those who do pay.  18 traditionally performed went unfulfilled or were provided infonnally at a charge. For example, in the case of education and health, households' responses indicate a high willingness to pay for these services (see Box 2.3). The existence of privately provided equivalents or near substitutes can also assist in identifying the appropriate price. Box 2.3: Assessing households' willingness to pay for social services The collapse of government finances in Georgia meant that many services which the state had traditionally performed went unfulfilled. To what extent did private agents take over provision and/or financing of these services? Can anything be learnt about private and social valuations of these previously publicly-provided goods? * In education, the private sector responded to the collapse of government resources by taking over a significant fraction of sectoral expenditures. Households presently pay fees for textbooks, extra tutoring, and many also provide in-kind supplies to schools for heating. Moreover, there has been significant growth in the number of students attending private schools, especially at the secondary and university level. Rough estimates suggest that at present households account for over one- half of total spending in the sector. Within the public system, households are contributing about US$29 per pupil (or 52 percent of unit cost), up from close to zero in 1990. Private fees are estimated to be within the US$150 to US$200 range. In response to the increase in the price of public schooling, there has been a decline in public enrollments of 23 percent.'0 This has been partially offset by an increase in private enrollments equal to abuat 12 percent of the public sector figure. Assuming no double counting, this translates into a net decline in enrollments of 11 percent. Given the magnitude of the price increase, the decline in enrollments is relatively moderate, indicating a fairly inelastic demand for education. Optimal tax theory would suggest that in this case there is a high willingness to pay for educational services on the part of households, and hence scope for (some) cost-recovery. * A similar argument can be made for health. With the decrease in government resources, the private sector became the major source of revenue for the health sector (totaling for possibly as much as 75 percent of total sector spending). While most services continued to be provided in public facilities, patients paid for all aspects of treatment, including drugs, supplies and provider fees. The increase in price had an apparent effect on demand, with annual outpatient visits per person falling from 10 to 5 and hospitalization rates falling from 0.11 admissions per capita to 0.06. This decline should reflect, in part, a "rationalization" of use of health services following the increase in its price; it may also reflect the effects of declining quality, supply constraints and declining incomes. Although previous utilization levels were probably inefficiently high, the worsening health status of the population also indicates that there could be extenal costs associated with the new equilibrium. For example, if fewer visits means people enter the system later in the course of their illness, when treatment costs are higher. While inter-sectoral comparisons are tricky, the evidence on changes in demand suggests a higher price elasticity for health than for education. 47. For most public services, cost-recovery levels are currently very low. In education, households are paying for textbooks, as well as other informal fees; but official charges are still negligible. Cost-recovery from households on electricity is abysmally low, in the range of 10 to 15 percent; this is in part subsidized by higher charges to enterprises, but on the whole the electricity entities are recovering less than 40 percent of costs. Basic municipal services, such as trash collection, are also heavily subsidized, with charges recovering between 2 and 20 percent of costs, depending on the municipality. In urban transport, tariffs recover an average of 30 to 40 percent of O&M costs. There is scope for increasing cost-recovery in all of these areas. 48. Users' willingness to pay is not the only relevant criteria for determining the level of the user charge. The social benefits of providing a particular service also have to be taken into account. In the case of education and health, for example, the existence of extemal effects implies that full cost-recovery will not be necessarily optimal. In contrast, publicly-provided transport and communications services (roads, railways, telephones) should aim at full cost-recovery. The same applies to basic public utilities '