Transition economies: how appropriate is the size and scope of government?

Gupta, Sanjeev ; Leruth, Luc ; De Mello, Luiz 等

INTRODUCTION

As markets expand in transition economies, the size and scope of government should also be reformed. Rather than engaging in a wide range of economic activities as under central planning, the government should be shifting its focus towards providing public goods and services, achieving society's distributive goals, and ensuring macroeconomic stability--the so-called classical functions of government (Tanzi and Schuknecht, 2000). Not only should the structure of government be changing in terms of its expenditure and revenue mobilisation roles at different levels, but new institutions (eg, in the budget area) should be emerging to shape the interface with the test of the economy.

This paper assesses the changes observed so far in the size and scope of government in transition economies and the extent to which these changes are appropriate and adequate. It has often been argued that governments in transition economies are trying to do too much relative to their limited capacity for raising revenue. While there are several indicators suggesting a decline in the size of government in transition economies, this paper finds that the size continues to be large and the scope, inappropriate, in many cases.

An analysis of the size and scope of government in transition economies is complicated by at least four factors. First, the coverage of government accounts as reported in the budget is often incomplete and may vary across time. (1) Second, a full record of commitments entered into by all spending units is usually not available. Third, the use of noncash transactions by the government (eg, tax offsets or 'netting operations', and in-kind payments) to settle accounts with the private sector further complicates the analysis. Finally, governments in some cases influence private sector activities in ways that are difficult to capture in reported statistics. For example, many governments in transition economies engage in quasifiscal activities and impose regulations to support public programmes and activities in order to cope with falling revenues (Tanzi, 1998). The rest of this paper is organised as follows. The following section surveys the literature. The next section describes recent trends and indicators of government size. Finally, policy lessons are drawn in the last section.

WHAT DO ECONOMIC THEORY AND EMPIRICAL STUDIES TELL US?

The economic literature offers limited guidance on the appropriate size and scope of government in transition economies. There is no consensus on what is 'appropriate' from a practical point of view. The literature has sought to identify variables affecting the size and scope of government to make comparisons across groups of countries and to estimate the 'optimal' size and scope on the basis of theoretical considerations. Important determinants of the size and scope of government include the following:

* Trade openness and the degree of integration in the world economy: Alesina and Wacziarg (1998) argue that the size of government correlates negatively with country size and positively with trade openness. Wei (2000) decomposes openness into 'natural openness" (exports and imports, geographical location, language and ethnic diversity, and size of population) and 'residual openness' (captured by other socioeconomic factors) and notes that governments in countries with 'natural openness' tend to spend more on payroll. In a similar context, Barro (1998) and Rodrik (1998) argue that bigger governments are needed to absorb external shocks in open economies.

* Business and political cycles: Gali (1994) shows that taxes and public spending on certain programmes act as automatic stabilisers and are correlated with output variability. Soh (1986) and Schuknecht (1994) distinguish between political and business cycles and observe an increase in government expenditure during and before an election year in many countries.

* Demographics: A high dependency ratio and ethnolinguistic fragmentation have been found to increase demand for public spending on education, health care, social security including pensions, and defence, and on programmes to cater to the needs and demands of regional and ethnic interests (Alesina et al., 1997; Annett, 2000).

* Budget institutions: The absence of strong budgetary institutions can result in a mismatch between the costs of raising revenue and the benefits of expenditure programmes. This might result in an overestimation of the marginal benefits of public spending. Public spending may, therefore, increase beyond desirable levels (von Hagen and Harden, 1996).

* Preferences and the heterogeneity of taxpayers and voters: Governments spend more in societies with a relatively unequal income distribution because the median voter is poorer than the mean voter. The benefits to the median voter of redistributive spending outweigh the costs borne by such a voter of increased taxation to finance spending (Persson and Tabellini, 1999). La Porta et al. (1998) provide empirical evidence that countries with a large share of transfers and subsidies in spending have larger governments.

* The structure of government: If taxpayers are mobile, the devolution of tax bases to subnational governments encourages competition for tax bases and may help to reduce the size of government. Reliance on grants and transfers from higher levels of government to finance subnational governments is associated with larger governments and fiscal imbalances at the subnational level (de Mello, 2000).

While all the variables noted above can influence the size and scope of government, it is almost impossible to aggregate their net effect. There are also variables whose influence is difficult to ascertain empirically (eg, the heterogeneity of taxpayers and voters" preferences, or inertia on the part of voters when the government is too large).

Another strand of literature compares the size and scope of governments across countries and over time in order to assess the impact of selected spending categories on economic growth and social indicators. Measures of government size have varied (eg, the number of employees in the public sector, the ratio of public spending to GDP, or the share of government consumption in total consumption). Tanzi and Schuknecht (1997) observe a shift in the composition of public expenditure away from defence, law and order, and property rights to social programmes and health care, education, and environmental protection for industrial countries during the period 1870-1990. They find, however, that higher spending on social programmes has not commensurately improved critical social indicators such as life expectancy, infant mortality, or school enrolment, suggesting that increase in public spending is not necessarily productive beyond a certain level. (2)

Finally, some studies have advocated the use of an allocative efficiency rule to establish the optimal size and scope of government. Although intellectually appealing, the rule that the size and scope of government are optimal when the social marginal cost of public resources is equal to their social marginal benefit is difficult to operationalise (Dahlby, 1998). (3) In particular, it is difficult to take into account all relevant country-specific determinants of social costs and benefits. However, the allocative efficiency principle has a number of important policy implications:

* Changes in the structure of public spending and taxes affect the balance between social costs and benefits. For example, a change in the scope of government would affect the social marginal benefit of some programmes and, hence, the overall marginal benefit of public spending. Similarly, a less distortionary tax system would decrease the marginal cost of raising funds. This would consequently affect the balance between social costs and benefits.

* All sources of finance should be taken into account. At the 'optimum', the social costs of raising resources must be equated across all sources. Hence, the efficiency of the tax system, the cost of debt financing, and the buildup of expenditure arrears, as well as quasifiscal operations, must be taken into account. The same principle applies to the benefits accruing from public spending that must be equated across programmes. For example, debt financing generates a flow of future interest payments, thereby complicating the computation of marginal costs and benefits. There is no unique optimum applicable to all countries. The social cost of raising revenues as well as their social benefits can be expected to vary among countries because of political economy factors such as differences in voters' preferences and in the effectiveness of budgetary institutions.

Unfortunately, empirical studies based on allocative efficiency rules yield unrealistic results. This is partly because these studies are unable to integrate voter preferences into the cost-benefit framework--an important political economy determinant of government size. Using a variation of the allocative efficiency rule, Karras (1996) estimates the optimal size of government (measured as government consumption) to be in the range of 14-33 per cent of GDP (with an average of 23 per cent). (4) Scully (1995) estimates that the average tax rate of 22 per cent for federal, state, and local taxes combined would maximise output growth in the US--a ratio that has been exceeded since 1949. (5) The typical size of government, even in countries considered to be market oriented, exceeds these estimates. Richards et al. (1994) show a gradual convergence in the size of government in OECD countries during the period 1979-93 to the range of 40-50 per cent of GDP. In these countries, upward pressure on government spending has been exerted primarily by the welfare state. One more reason why results have been unrealistic is that, with the exception of Sachs et al. (2000), empirical studies have traditionally focused on outcomes and the analysis has not been cast in terms of what determines a transition path, that is, the starting conditions and the policies adopted.

GENERAL TRENDS IN INDICATORS OF GOVERNMENT SIZE

Conventional measures of government size--aggregate public spending measured on a cash basis in relation to GDP and the share of public employment--show that there has been widespread downsizing along with a steady expansion of the private sector in transition economies. At the same time, there are indications that the government's role in transition economies has not diminished enough: public sector indebtedness has grown, expenditure arrears have persisted, the noncash system for settling accounts is often used, and the regulatory burden remains high. There is also evidence that the scope of government activity may be misdirected. For example, governments are unable to provide critical public services to their populations and most transition economies now have a higher incidence of poverty than before.

Indicators suggesting a decline in government size

Public expenditure--measured on a cash basis--has declined substantially, contracting more than output in most transition economies. However, there is considerable variation in expenditure shares across countries. Data for transition economies show that public outlays declined by 6 per cent of GDP during the period 1993-99 (Figure 1). (6) While government spending measured on a cash basis remained steady at 41-46 per cent of GDP in the European Union (EU) border states, it fell from ah average of 43 per cent in 1993 to 36 per cent in 1999 in the remaining transition economies. The contraction in cash expenditures has been particularly severe in central Asia and the Caucasus where total outlays nearly halved as a share of GDP over the period 1993-99. In contrast, government expenditure relative to GDP rose from 35 per cent in 1993 to 44 per cent in 1999 in the Baltics.

Caution needs to be exercised in interpreting these spending trends. Most budgets were narrowly defined at the beginning of the transition period. Social funds, parastatal and State-owned enterprises, and local governments were typically not consolidated in the fiscal accounts. In some countries, consolidation was carried out on a net basis including only central government transfers to these funds and enterprises. More importantly, the broadening of the coverage of fiscal accounts has occurred over time along with the declining share of government spending in GDP. The fall in the spending-to-GDP ratio is likely to have been higher if measured only for the narrowly defined government at the beginning of the transition period. Moreover, to the extent that GDP was overestimated at the beginning of the transition period, the fall in expenditure-to-GDP ratio over time is underestimated, thereby indicating a less drastic downsizing of the government in the course of transition.

The expenditure decline has been accompanied by changes in the scope of government. In many transition economies, governments have been called upon to provide services such as kindergartens, health-care centres, and housing that were previously the responsibility of the state-owned enterprises. These and other social mandates have not been fully funded and exert pressure on overall public spending particularly at the subnational level. (7) This lack of funding has contributed to the buildup of expenditure arrears in the social sectors. Explicit subsidies for food items have generally been eliminated, but subsidies for heating and for other communal services have taken forms that are not always transparent. Privileges to different population groups for the use of these services complicate the analysis of incidence of subsidies) Public outlays on transfers to households have fallen as a share of total expenditures in most transition economies, particularly the Baltics and central Asia, to a level far below that of the OECD countries (Figure 2). The introduction of market-based instruments in central Asia to finance government deficits and the rapid accumulation of debt have increased interest payments from the budget, thereby squeezing other public spending. In these countries, public outlays on wages and salaries have risen significantly as a share of total government spending. The share of capital outlays in total public spending has fallen, particularly in the Baltics.

Although government spending levels have gone down on average, the fall in expenditure has partly been driven by the decline in the ability of governments to mobilise domestic revenue. On average, revenues have fallen as a share of GDP, but by less than the decline in spending. In recent years, revenues have recovered in some countries; however, in others, persistent budgetary deficits have led to substantial debt accumulation. Revenues declined as a share of GDP by between 3 and 5 per cent in the Baltics, the Balkans, and the EU border states during the transition years. The decline in revenues (by around 6 per cent of GDP) was particularly sharp in the countries of Central Asia and the Caucasus where revenue-to-GDP ratios were already low at the beginning of the transition period.

Public sector employment--another conventional measure of government Size--has also fallen, although it remains high by international standards. Consistent time-series data on employment are not readily available for transition economies. The data available show that government employment (measured per 1,000 population) has fallen in most transition economies particularly in the EU Border States and the other CIS countries including Russia (Table 1). On average, the government employed 60 in 1,000 people during the period 1995-99 as opposed to nearly 80 in 1,000 at the beginning of the transition. In the course of transition, over 7.5 million government jobs were retrenched in economies. Inspite of this decline, government employment in transition economies remains hi8her than in all other country groupings for which data are available (except the OECD) as shown in Table 1. Average civil service retrenchment in transition economies masks sizeable regional disparities. Government employment has increased substantially in Azerbaijan, the Czech Republic, Moldova, and Uzbekistan. In these countries, government employment in health care and education fell during this period suggesting that the social sectors were affected more by employment downsizing. More importantly, falling government expenditures in transition economies have not been accompanied by a commensurate reduction in government employment. Public outlays on wages and salaries constitute on average nearly 6 1/2 per cent of GDP and have risen during the transition years, particularly in the Baltics and the other CIS countries including Russia. Information comparing public and private sector wages and salaries is scarce, but labour compensation in the government has typically not kept pace with the private sector in some countries including Armenia, Georgia, and Hungary as seen from their low compression ratios (Table 1). (9) Government pay scales also remain compressed in most transition economies particularly in Central Asia. These governments are unable to raise wages in part because government employment continues to be high. The government's wage bill has increased in most countries and accounts for as much as 10 per cent of GDP in Lithuania and 16 per cent of GDP in Uzbekistan over the period 1995-99.

The share of the private sector has expanded in virtually all transition economies and now accounts for a large share of GDP in most economies (Table 2). There has been a gradual transfer of ownership of previously government-owned assets to the private sector and some reduction in the regulatory burden through the liberalisation of prices and foreign exchange markets. The private sector has also increased participation in the production of public goods, ah area where the government had hitherto been the sole producer as well as provider. However, the overall regulatory burden is still perceived to be high and continues to distort economic incentives and foster rent-seeking behaviour and corruption in most transition economies (Table 2). The EU border states and the Baltic countries score better than other transition economies (eg, Central Asia, Caucasus, other CIS) in terms of effectiveness of government and the degree of regulatory burden. This is also confirmed by the indicator of openness constructed by Sachs et al. (2000), which seeks to capture the ease with which economic activity can take advantage of the foreign sector for markets, know-how, competition, financing, investment, source of inputs, and other components linking its markets and firms to the global economy. Countries such as Armenia, Azerbaijan, the Kyrgyz Republic, and Uzbekistan have relatively low openness scores.

Indicators suggesting that government size is still too large and scope Inappropriate

The persistence of expenditure arrears suggests that the actual size of government measured on a commitment basis may be larger than that suggested by the cash expenditure-to-GDP ratios. The wedge between cash and commitment expenditures further indicates that the scope of government, as gauged by the entitlements granted to the population, is broader than that suggested by the magnitude of total cash spending. Reliable information on expenditure arrears is not available for most transition economies and those that track this information tend to be the ones where the problem is less severe. In addition, there are methodological difficulties in measuring expenditure arrears. (10) The limited data suggest that the bulk of expenditure arrears comprises wages and salaries, social security payments, and arrears to suppliers including energy companies. In 1998, the stock of arrears relative to GDP was 3.5 per cent in Georgia, 11 per cent in Moldova, and 6 per cent in Ukraine. Energy arrears accounted for nearly half of the flows of total arrears in Ukraine. The inability of most governments to curb public consumption, particularly in the energy sector (such as heating subsidies), has increased indebtedness and led to the accumulation of expenditure arrears and contingent liabilities in the budget. Failure to revise expenditure commitments in line with revenue capacity has compromised the actuarial sustainability of pension funds in most transition economies. Difficulties faced in reforming expenditure commitments, particularly in the social area, as well as in stemming the accumulation of arrears, are due to entrenched interests in these countries.

Rising national indebtedness also implies that in transition economies, spending is high compared to the government's ability to raise domestic resources. Most transition economies had virtually no national, and relatively little foreign debt in the early 1990s. However, between 1993 and 1999, external debt relative to GDP rose from 8.5 to 24 per cent in the Baltics, from 22 to 42 per cent in the Caucasus, from 35 to 70 per cent in Central Asia and from 38 to 49 per cent in the other CIS countries (including Russia). The rapid increase in indebtedness during economic transition is traceable to several factors. (11) These include the practice of extending public guarantees for private debt in many countries, the liberalisation of financial markets that increased the cost of servicing debt, the dependence of some energy-deficient transition economies on foreign energy suppliers, the weak public expenditure management systems that have prevented adjustment of expenditures over time in line with budget allocations, and inefficient public spending.

The changing composition of government spending raises questions about the appropriateness of the scope of government. The composition of government spending seems to have shifted in favour of areas with a relatively high social return. For example, public spending on defence fell in relation to GDP, the decline being particularly large in the CIS countries (including Russia) and the Caucasus. (12) Central Asian countries have experienced significant reductions in capital spending and outlays on social security and welfare. However, some authors (Milanovic, 1997) have suggested that the worsening income distribution in many transition economies reflects the failure of governments to target social programmes to the poor and other social groups adversely affected by reforms. The decline in the share of outlays on social programmes has been more pronounced in the Caucasus and the other CIS countries (including Russia) than in the other country groups.

Public spending on education and health care declined in transition economies as a per cent of GDP but remained stable in relation to total government outlays (Table 3). Although the composition of education spending favours primary and secondary education, a large proportion of these outlays is directed at providing child care to children between ages 1 and 5 years. (13) Teaching loads are typically low. For example, the normal workload for a secondary education teacher in Moldova in 1998 was 18 teaching hours per week. Budget allocations for books and teaching material have been squeezed, and per capita spending for tertiary education is a multiple of spending on primary and secondary education. Some transition economies are spending between 30 and 50 per cent of their total education budget on energy. In Ukraine, outlays on maintenance (including energy and heating) accounted for 17-20 per cent of total public spending on health care on average between 1995 and 1998. Further, curative health care accounts for a large share of health spending. Informal charges to secure admission to secondary schools and universities are common in many transition economies (World Bank, 2000). The prevalence of informal user charges has limited the access of the poor to basic health services and most social indicators have worsened. (14) There is also a marked disparity in the amounts spent by different countries on health care. While Azerbaijan and Tajikistan are spending less than 1 per cent of GDP on health, the Czech Republic, Croatia, and the Slovak Republic are spending between 6 and 7 per cent of GDP. The coverage and quality of basic health care is deteriorating as compared to that in the pretransition period. Most transition governments are unable to do much to combat the emergence of diseases such as tuberculosis and AIDS. The health-care system in most countries therefore, needs to be rationalised. They are plagued by several problems such as an oversupply of poorly maintained health-care facilities, an excessive number of hospital beds and long average stays at hospitals, (15) inadequately paid health-care personnel, and an overstaffing of administrative personnel.

POSSIBLE LESSONS

Although some indicators in transition economies point to a contracting size of government, others suggest that it continues to play too large a role. Growing national indebtedness, the accumulation of arrears, the proliferation of noncash transactions, and a heavy regulatory burden are reflections of governments seeking to do more than they should. The reliance on sources other than domestic revenue to finance public spending has increased the social marginal cost of raising resources. Further, the failure to align expenditure commitments to available revenues can be attributed, at least in part, to the absence of well-functioning institutions that strengthen governance and promote transparency and accountability in the government. The stronger the budgetary institutions, the greater the control of expenditure. The apparent downsizing of the government has thus been prompted more by falling revenues, rather than a conscious strategy to constrain the government's role and to enhance allocative efficiency.

EU border states appear to be more efficient than other country groups, in terms of both raising resources and spending them. This conclusion stems from the aggregation of selected elements of the costs of raising resources and benefits from spending programmes (Figure 3). (16,17) For example, the Central Asian republics have higher costs of raising resources and lower benefits from government spending than EU border states. Caucasus and the other CIS countries are in a similar situation. Although the benefits are slightly lower in the Baltics than in EU border states, costs are similar and the overall level of expenditure is lower. These results are consistent with the earlier discussion that there is no unique optimum for the size of government.

More attention needs to be paid to rationalising the scope of government in transition economies than to expenditure cuts per se. Social programmes need to be better targeted and allocations shifted within these sectors in favour of activities that yield higher social rates of return. This could mean reducing allocations for kindergartens (day-care centres) and tertiary education as well as for specialised hospital services, which mainly benefit the well-off. Government programmes for improving quality of primary and secondary education and preventive health care would, therefore, be more appropriate. It could also involve reforming existing pension and other social programmes to make them more efficient and equitable. Such reforms pose a challenge for the governments of transition economies in light of pressures exerted by the rapid aging of the population. The observed fall in aggregate expenditures in government remains inappropriate, inconsistent with the required functions of government. Moreover, public provision of public goods and services does not imply that a government should also be involved in their production.

The political economy determinants of the size and scope of government also need to be addressed. Failure to reform the role of government in transition economies may be due, at least in part, to strong vested interests. The political process may reflect the preferences of regional interest groups rather than those of the electorate at large. Poor governance has hampered the resumption of sustained growth, expansion of private sector activity, and the reduction of poverty. Strengthening budgetary institutions, particularly in the areas of tax administration and expenditure management and control, will also help limit the size and improve the scope of government. (18) The stronger the budgetary institutions, the greater the control on expenditures and the more successful the implementation of structural reforms in the fiscal area.

Table 1: Government employment, 1990-99 (in per cent of population
except where noted)

 Government employment (a)

 Total

 1990 1995 Change
 -95 -99 (in thou-
 sands)

Transition economies 7.9 6.0 -7569.0

Balkans 4.5 4.0 -165.6
 Bulgaria 7.7 4.2 -314.8
 Macedonia 3.9 4.8 21.0
 Romania 3.3 3.9 128.0

Boltics 8.2 6.3 -164.0
 Estonia (c) 7.8 1.8 -93.6
 Latvia (c) 8.2 4.9 -95.0
 Lithuania 8.3 9.0 24.4

Caucasus 10.5 10.1 -5.5
 Armenia (c) 13.2 7.9 -186.0
 Azerbaijan (c) 9.5 13.5 346.0
 Georgia (c) 10.0 7.0 -165.0

Central Asia 7.7 12.0 2550.0
 Kazakhstan (c) 6.9 5.6 -240.0
 Kyrgyz Republic (c) 12.1 7.2 -210.0
 Tajikistan (c) 6.5 0.9 -307.0
 Turkmenistan (c) 8.4 4.4 -135.7
 Uzbekistan (c) 7.6 21.6 3442.0

EU border states 5.8 5.0 -671.0
 Croatia 7.2 6.4 -50.0
 Czech Republic (c) 4.3 8.7 448.0
 Hungary 8.7 8.0 -83.0
 Poland 4.9 3.4 -568.0
 Slovakia (c) 8.6 1.6 -369.0
 Slovenia (c) 6.9 4.5 -49.0

Other CIS (including Russia) 9.0 4.7 -9112.0
 Belarus (c) 5.5 4.2 -132.0
 Moldova (c) 7.4 9.9 106.0
 Russia (c) 9.3 4.3 -7411.0
 Ukraine (c) 8.8 5.7 -1676.0

Memorandum items
 OECD 6.5 7.8 12289.2
 Sub-Saharan Africa 1.2 0.7 -2517.0
 Middle East 3.2 5.1 6238.7
 Asia 2.3 1.6 -18442.0

 Education

 1990 1995 Change
 -95 -99 (in thou-
 sands)

Transition economies 3.8 1.5 -8905.0

Balkans 1.8 1.1 -254.0
 Bulgaria 3.0 1.0 -172.0
 Macedonia 0.9 1.0 2.7
 Romania 1.4 1.1 -84.7

Boltics 3.5 2.4 -93.0
 Estonia (c) 2.8 -- --
 Latvia (c) 3.5 1.5 -55.2
 Lithuania 3.8 3.9 5.2

Caucasus 4.7 1.1 -590.0
 Armenia (c) 4.1 2.2 -65.5
 Azerbaijan (c) 5.2 0.5 -348.4
 Georgia (c) 4.5 1.3 -1755.8

Central Asia 3.6 3.0 -330.8
 Kazakhstan (c) 1.8 3.1 191.0
 Kyrgyz Republic (c) 3.6 1.6 -86.6
 Tajikistan (c) 4.0 0.2 -209.0
 Turkmenistan (c) 4.3 1.8 -89.3
 Uzbekistan (c) 4.8 3.8 -137.0

EU border states 2.0 1.2 -517.0
 Croatia 1.4 1.2 -10.1
 Czech Republic (c) 1.7 0.7 -99.3
 Hungary 2.9 2.3 -62.6
 Poland 1.6 1.2 -149.3
 Slovakia (c) 3.2 -- --
 Slovenia (c) 2.5 1.3 -24.0

Other CIS (including Russia) 4.7 1.4 -7121.0
 Belarus (c) 1.1 1.6 57.2
 Moldova (c) 4.2 1.1 -138.8
 Russia (c) 5.1 1.5 -5331.3
 Ukraine (c) 4.4 1.1 -1707.9

Memorandum items
 OECD 1.7 1.7 659.8
 Sub-Saharan Africa 0.3 0.2 -648.1
 Middle East 0.8 1.0 853.5
 Asia 0.7 0.6 -383.4

 Healthcare

 1990 1995 Change
 -95 -99 (in thou-
 sands)

Transition economies 2.4 1.4 -4163.0

Balkans 1.2 0.8 -142.0
 Bulgaria 2.3 1.1 -107.9
 Macedonia 1.4 0.8 -9.3
 Romania 0.7 0.6 -24.9

Boltics 2.6 1.5 -93.0
 Estonia (c) 1.9 -- --
 Latvia (c) 2.8 0.9 -52.9
 Lithuania 2.8 2.5 -11.4

Caucasus 2.2 0.8 -219.0
 Armenia (c) 2.0 0.1 -69.7
 Azerbaijan (c) 2.3 1.5 -54.3
 Georgia (c) 2.1 0.4 -94.8

Central Asia 2.5 1.6 -428.0
 Kazakhstan (c) 3.1 1.7 -236.8
 Kyrgyz Republic (c) 2.4 0.8 -69.2
 Tajikistan (c) 2.0 0.6 -73.2
 Turkmenistan (c) 2.0 1.7 -4.7
 Uzbekistan (c) 2.3 1.9 -44.0

EU border states 2.0 0.9 -751.0
 Croatia 1.5 0.7 -37.4
 Czech Republic (c) 1.2 0.3 -90.6
 Hungary 2.4 2.2 -21.1
 Poland 2.0 0.9 -439.6
 Slovakia (c) 2.4 -- --
 Slovenia (c) 2.7 0.9 -35.1

Other CIS (including Russia) 2.8 1.6 -2531.0
 Belarus (c) 2.8 1.2 -163.2
 Moldova (c) 2.5 1.1 -58.5
 Russia (c) 2.8 1.3 -2234.3
 Ukraine (c) 2.9 2.8 -74.7

Memorandum items
 OECD 0.6 0.7 825.6
 Sub-Saharan Africa 0.1 0.0 -424.5
 Middle East 0.3 0.3 108.9
 Asia 0.2 0.1 -3111.3

 Government wage bill (in per cent of GDP)

 1990 1995 Change
 -95 -99

Transition economies 3.7 6.3 2.6

Balkans 4.8 6.4 1.6
 Bulgaria 2.7 5.2 2.5
 Macedonia 6.0 8.5 2.5
 Romania 5.8 5.6 -0.2

Boltics 5.2 9.3 4.2
 Estonia (c) 5.6 8.3 2.7
 Latvia (c) 6.3 9.4 3.1
 Lithuania 3.6 10.3 6.7

Caucasus 1.5 2.5 1.1
 Armenia (c) 1.8 2.9 1.1
 Azerbaijan (c) 1.8 2.4 0.6
 Georgia (c) 0.8 2.3 1.5

Central Asia 3.9 7.3 3.4
 Kazakhstan (c) 2.8 4.9 2.1
 Kyrgyz Republic (c) 4.9 4.6 -0.3
 Tajikistan (c) -- 3.5 --
 Turkmenistan (c) -- -- --
 Uzbekistan (c) - 16.0 --

EU border states 3.3 5.2 1.9
 Croatia 3.0 - --
 Czech Republic (c) -2.7 2.7 --
 Hungary 3.3 7.3 4.0
 Poland 6.1 3.9 -2.2
 Slovakia (c) 3.7 3.7 3.0
 Slovenia (c) -5.8 5.8 --

Other CIS (including Russia) 3.7 7.3 3.6
 Belarus (c) 3.9 7.0 3.1
 Moldova (c) 5.5 6.2 0.7
 Russia (c) 1.7 -- --
 Ukraine (c) -- 8.7 --

Memorandum items
 OECD 4.8 3.3 -1.5
 Sub-Saharan Africa 6.7 6.8 0.1
 Middle East 10.3 11 0.7
 Asia 7.7 7.5 -0.2

 Compression
 Ratio
 (1995-99) (b)

Transition economies --

Balkans --
 Bulgaria --
 Macedonia --
 Romania 2.3

Boltics --
 Estonia (c) 6.0
 Latvia (c) --
 Lithuania 6.8

Caucasus 8.0
 Armenia (c) --
 Azerbaijan (c) 10.0
 Georgia (c) --

Central Asia 4.2
 Kazakhstan (c) 3.4
 Kyrgyz Republic (c) --
 Tajikistan (c) --
 Turkmenistan (c) --
 Uzbekistan (c) --

EU border states --
 Croatia --
 Czech Republic (c)
 Hungary 3.6
 Poland --
 Slovakia (c) --
 Slovenia (c) --

Other CIS (including Russia) --
 Belarus (c) 9.7
 Moldova (c) --
 Russia (c) --
 Ukraine (c) --

Memorandum items
 OECD --
 Sub-Saharan Africa --
 Middle East --
 Asia --

(a) Total government employment includes employment in the central
government, and the subnational governments. Education employment
covers primary, secondary and university education. Health employment
covers employees of government hospitals and health institutions at all
levels of government.

(b) Compression ratio is defined as the ratio of the highest to the
lowest grade wages and an update of the same work by Guilio de Tomasso
and Amitabha Mukherjee, forthcoming.

(c) The trend in total employment should be interpreted with caution
since data for all components are not available for initial and
latest years for these countries.

Sources: Schiavo-Campo et al. (1997); and an update of the same work
by Guilio de Tomasso and Amitabha Mukherjee, forthcoming.

Table 2: Indicators: governance, regulation, and private sector share

 Government Regulatory
 effectiveness (a) burden (a)

Balkans -0.6 0.2
 Romania -0.6 0.2

Baltics 0.2 0.6
 Estonia 0.3 0.7
 Latvia 0.1 0.5

Caucasus -0.8 -1.0
 Azerbaijan -0.8 -1.0

Central Asia -0.8 -0.4
 Kazakhstan -0.8 -0.4

EU border states 0.5 0.6
 Croatia 0.1 0.2
 Czech Republic 0.6 0.6
 Hungary 0.6 0.9
 Poland 0.7 0.6
 Slovenia 0.6 0.5

Other CIS (including Russia) -0.7 -0.7
 Belarus -0.7 -1.5
 Moldova -0.5 -0.3
 Russia -0.6 -0.3
 Ukraine -0.9 -0.7

Memorandum items

 Advanced economies 1.5 1.0
 Asia -0.2 -0.2
 Middle Eastern -0.2 -0.5
 Sub-Saharan Africa -0.4 0.1
 Transition economies -0.2 0.0
 Western hemisphere 0.0 0.6

 Voice and Rule of
 accountability (b) law (c)

Balkans 0.4 -0.1
 Romania 0.4 -0.1

Baltics 0.7 0.3
 Estonia 0.8 0.5
 Latvia 0.6 0.2

Caucasus -0.9 -0.6
 Azerbaijan -0.9 -0.6

Central Asia -- --
 Kazakhstan -- --

EU border states 0.8 0.5
 Croatia -0.3 0.1
 Czech Republic 1.2 0.5
 Hungary 1.2 0.7
 Poland 1.1 0.5
 Slovenia -- --

Other CIS (including Russia) -0.4 -0.8
 Belarus -0.5 -0.9
 Moldova -- --
 Russia -0.3 -0.7
 Ukraine -- --

Memorandum items

 Advanced economies -- --
 Asia -- --
 Middle Eastern -- --
 Sub-Saharan Africa -- --
 Transition economies -- --
 Western hemisphere -- --

 Corruption (d)

Balkans 3.0
 Romania 3.0

Baltics 4.0
 Estonia 5.0
 Latvia 3.0

Caucasus 2.0
 Azerbaijan 2.0

Central Asia --
 Kazakhstan --

EU border states 3.8
 Croatia 2.0
 Czech Republic 4.1
 Hungary 4.6
 Poland 4.6
 Slovenia --

Other CIS (including Russia) 3.4
 Belarus 4.0
 Moldova --
 Russia 2.8
 Ukraine --

Memorandum items

 Advanced economies --
 Asia --
 Middle Eastern --
 Sub-Saharan Africa --
 Transition economies --
 Western hemisphere --

 Private sector share
 (in per cent of GDP)

 Mid-1993 Mid-2000

Balkans 32.0 60.0
 Romania 32.0 60.0

Baltics 40.0 70.0
 Estonia 39.8 75.0
 Latvia 39.2 65.0

Caucasus 13.3 45.0
 Azerbaijan 13.3 45.0

Central Asia 10.6 60.0
 Kazakhstan 10.6 60.0

EU border states 38.0 69.0
 Croatia 31.0 60.0
 Czech Republic 33.1 80.0
 Hungary 54.0 80.0
 Poland 47.0 70.0
 Slovenia 25.0 55.0

Other CIS (including Russia) 20.5 50.0
 Belarus 13.3 20.0
 Moldova 17.0 50.0
 Russia 33.0 70.0
 Ukraine 18.7 60.0

Memorandum items

 Advanced economies -- --
 Asia -- --
 Middle Eastern -- --
 Sub-Saharan Africa -- --
 Transition economies -- --
 Western hemisphere -- --

(a) Measured on a scale of about -2.5 to 2.5 with higher values
corresponding to better outcomes. The index of government effectiveness
combines perception of the quality of public service provision,
the quality of the bureaucracy, the competence of civil servants,
the independence of the civil service from political
pressures, and the credibility of the government's commitment to
policies. The index of regulatory burden includes measures of the
incidence of market-unfriendly policies, such as price controls or
inadequate bank supervision, as well as perception of the burden imposed
by excessive regulation in several areas such as foreign trade and
business development, among others. Scores refer to 1996 or 1997 for
most countries.

(b) The index of voice and accountability comprises several
measures relating to the political process, civil liberties,
and political rights, and is based on information on the extent to
which citizens of a country are able to participate in the
selection of governments, and on measures of the independence of the
media. Scores refer to 1970-95 averages.

(c) The index of rule of law includes several indicators measuring
the extent to which agents have confidence in, and abide by,
the rules of society, and is based on information on the perceived
incidence of both violent and nonviolent crime, the effectiveness
and predictability of the judiciary, and the enforceability of
contracts. Scores refer to 1970-95 averages.

(d) The ICRG index measures a country's corruption as perceived
by foreign investors. It varies from 0 (most corrupt) to
6 (least corrupt). Corruption is defined as the likelihood of a
government official to demand special payments, and whether
illegal payments are expected throughout the lower levels of
government in the form of bribes connected with import
and export licenses, exchange controls, tax assessment,
police protection, or loans. Scores refer to 1970-95 averages.

Sources: Kaufmann et al. (1999); ICRG, and EBRD Transition Reports; and
IMF staff calculations.

Table 3: Education and health spending, comparing transition
economies and other developing countries

 In per cent
 of GDP

 1993 2000
Education

All transition economies 5.4 4.6
of which
 Centra-Asia 6.0 4.6
 Balkans 4.3 3.5
 Baltics 6.9 6.8
 Other CIS (including Russia) 5.8 4.2
 Caucasus 4.5 3.8
 EU border states 4.7 4.8

Sub-Saharan Africa 4.1 4.6
Asia 3.8 4.2
Middle East 4.9 4.6
Western hemisphere 3.5 4.3
OECD 5.5 5.7

Health

All transition economies 3.9 3.4
of which
 Central Asia 3.2 2.3
 Balkans 3.7 3.7
 Baltics 3.9 4.8
 Other CIS (including Russia) 4.3 3.1
 Caucasus 2.1 0.9
 EU border states 6.1 5.6

Sub-Saharan Africa 1.7 2.3
Asia 1.7 1.9
Middle East 1.8 1.6
Western hemisphere 2.4 2.7
OECD 8.4 8.7

 In per cent of Sample size
 total expenditure

 1993 2000
Education

All transition economies 14.5 14.3 18
of which
 Centra-Asia 16.0 19.1 5
 Balkans 10.1 9.3 2
 Baltics 17.3 16.7 1
 Other CIS (including Russia) 13.6 11.4 4
 Caucasus 20.1 18.4 1
 EU border states 9.8 10.7 5

Sub-Saharan Africa 14.0 15.3 32
Asia 14.0 16.0 18
Middle East 13.3 14.4 15
Western hemisphere 15.8 17.1 24
OECD 15.2 19.0 23

Health

All transition economics 9.9 9.3 18
of which
 Central Asia 8.7 9.5 5
 Balkans 8.6 9.8 2
 Baltics 9.9 11.8 1
 Other CIS (including Russia) 9.9 8.3 4
 Caucasus 9.4 4.2 1
 EU border states 12.7 12.4 5

Sub-Saharan Africa 5.7 7.4 32
Asia 5.5 6.9 18
Middle East 4.8 5.1 15
Western hemisphere 10.3 10.3 25
OECD 23.6 31.0 23

Source: National authorities; IMF staff estimates; and World
Development Indicators 2002.

Acknowledgements

We thank Benedict Clements, Stefano Fassina, Robert Gillingham, Maria-Teresa Guin-Siu, Oleh Havrylyshyn, Ali Mansoor, Thomas Richardson, Gerard Roland, Thomas Wolf, and three anonymous referees for their helpful comments, and Rama Chandra Reddy for his assistance in preparing the literature review.

(1) Government can be defined narrowly to cover the operations of the central government or more broadly as the central government plus subnational governments, and in some cases, even the State-owned enterprises. Typically, the data in most transition economies cover the operations of the central government and subnational governments.

(2) The authors conclude that government spending needs to be no higher than 30 per cent of GDP to achieve socially desirable goals. They further conclude that 'big' governments (those with expenditures exceeding 50 per cent of GDP) do not fare better than 'small' governments (those with spending in the range of 30-40 per cent of GDP).

(3) This concept was later developed in the context of optimal commodity taxation by Atkinson et al. (1984) and Sandmo (1998).

(4) Karras (1996) uses Barro's rule (ie, the size of government is optimal when the marginal product of capital equals unity) for a sample of 118 countries during 1960-85. There are other similar studies. For example, Grossman (1988) estimates the optimal tax rate for the US government at 19 per cent of GDP. In a similar study covering the period 1889-1986, Peden (1991) estimates the optimal size of the US government at 20 per cent of GDP. He also notes, as did Feldstein (1997), that increased expenditure hurts output growth because of dead-weight losses arising from higher taxation.

(5) Economic growth can be fostered through government spending only if the productivity of public outlays exceeds the dead-weight loss associated with distortionary taxation. Government expenditure on health care and education may have ah indirect impact on economic growth through human capital formation (Landau, 1983; Kneller et al., 1999; Commander et al., 2000). Using data for 1970-90 for 43 developing countries, Devarajan et al. (1996) show that seemingly productive expenditure, when financed through distortionary taxes, may be counterproductive. They find that even though expenditure on goods and services in their sample grew by 8 per cent over the last two decades so reach 26 per cent of GDP, public outlays on health care, education, and transport and communication have a negative or statistically insignificant impact on economic growth.

(6) This paper covers 24 transition economies. These are classified in terms of the criteria suggested by Sachs et al. (2000)--five central Asian countries (Kazakhstan, the Kyrgyz Republic, Tajikistan, Turkmenistan, and Uzbekistan); three Balkan countries (Bulgaria, the former Yugoslav Republic of Macedonia, and Romania); three Baltic countries (Estonia, Latvia, and Lithuania); four other Commonwealth of independent States (CIS) countries including the Russian Federation (Belarus, Moldova, and Ukraine); three countries in the Caucasus (Armenia, Azerbaijan, and Georgia); and six countries bordering the EU (Croatia, the Czech Republic, Hungary, Poland, the Slovak Republic, and Slovenia). Expenditure analysis covers the post-1993 period to exclude relatively low-quality data available for the early years of transition. Deficiencies in measuring GDP, particularly in light of rising inflation, distort the relevant indicators of government size at the beginning of the transition period.

(7) For example, low cost recovery in the provision of energy and other communal services has imposed additional costs on the budgets of local governments in many countries, In Belarus, cost recovery in the provision of communal services was estimated at approximately 40 per cent in February 1999, at a cost to the budget of 1.4 per cent of GDP. The total (budgetary and nonbudgetary) cost of subsidies for housing and communal services was estimated at 3.3 per cent of GDP in the Russian Federation in 1999. Effective cost recovery rates vary between 20 and 80 per cent across regions and averaged between 40 and 45 per cent in early 2000.

(8) In the Russian Federation, there were more than 220 population categories (covering about 100 million people) entitled to approximately 150 privileges in 2000. The cost of privileges and exemptions (free transportation and concessions for housing and communal services) for veterans of labour was estimated at roughly 1.1 per cent of GDP in 1999.

(9) International experience suggests that, it the compression ratio, defined as the ratio of the highest-to-lowest-grade wage, is less than 10, the wage scale does not provide sufficient incentives for higher levels of productivity. The compression ratio is on average low in transition economies. In Moldova, in 1995, the highest public sector wage was 6.6 times the lowest. In Azerbaijan, the compression ratio was 5.5 in the executive bodies in 1996. In Belarus, it was 6.5 in 1999.

(10) Information on tax rather than expenditure arrears is more readily available. In the Russian Federation, noncash transactions through the issuance of promissory notes, tax offsets, and the accumulation of arrears make up a significant proportion of total transactions. It is estimated that the total overdue debt of enterprises accounted for nearly 40 per cent of GDP in 1998. Total payables to large- and medium-size enterprises rose from 20 per cent to 70 per cent of GDP between 1994 and 1997, while total receivables rose from 20 per cent to 45 per cent of GDP over the same period. (See Commander et al. (2000) for more information.) In Azerbaijan, pension arrears were nearly eliminated in 1997, but increased to more than 1/2 per cent of GDP by the end-of 1998 and rose further in 1999. Arrears to utility companies accounted for almost 2 1/2 per cent of GDP by the end of 1998.

(11) There is considerable variation in debt ratios in transition economies. Maturity structures and the share of concessional debt in total foreign liabilities also differ widely. (See Hamann and Mourmouras (2000) for more information.)

(12) Data from Stockholm International Peace Research Institute (SIPRI) show that military spending for the 24 transition economies fell from around 4.7 per cent of GDP in 1993 to 3.1 per cent in 1999. In Russia, for example, the decline was from 5.8 per cent of GDP to 3.8 per cent of GDP over this period.

(13) In Ukraine, between 1995 and 1998, spending per student at the preschool level was on average nearly 70 per cent higher than for general education, and was equivalent to spending per student in vocational training and higher education. In Lithuania, outlays on preschool education, at approximately 14.5 per cent of total public spending on education, were nearly as high as those on tertiary education in 1993-98.

(14) It is estimated that in 1999, the share of patients who had to make informal payments to get health care was 91 per cent in Armenia, 70 per cent in Moldova, 66 per cent in Tajikistan, and 60 per cent in the Slovak Republic (World Bank, 2000).

(15) In 1996-97, the average length of stay in acute care hospitals was 17.6 days in Moldova. 16.3 days in the Russian Federation, 14 days in Tajikistan, 10 days in Romania, and 9 days in the Czech Republic and Hungary, as compared to about 5-6 days in OECD countries such as France, Denmark, Sweden and the UK (World Bank, 2000).

(16) The costs are measured by the regulatory burden, an indicator of quality of government including the burden of taxation, and the level of public or publicly guaranteed debt. The benefits are proxied by life expectancy at birth, secondary school enrolment, and ah index of government effectiveness.

(17) The area covered by the lower portion of the diamond can be interpreted as the cost of raising resources from three major sources of funds. Everything else being equal, a higher point on the axis indicates a higher cost of raising funds. The same applies to benefits. A graphical comparison is made between the EU border states, the Caucasus, Central Asia, and the Baltics.

(18) Hallerberg and von Hagen (1997) provide empirical evidence of an association between budget deficits and budgetary institutions in EU countries. Alesina and Perotti (1995, 1996) show that coalition governments, tax smoothing and intergenerational distribution concerns, as well as electoral systems and the strength of budgetary institutions are important determinants of budget deficits.

REFERENCES

Alesina, A and Perotti, R. 1995: Political economy of budget deficits. IMF Staff Papers, 42: 1-31.

Alesina, A and Perotti, R. 1996: Budget deficits and budget institutions. NBER Working paper no. 5556, National Bureau of Economic Research: Cambridge, MA.

Alesina, A, Baqir, R and Easterly, W. 1997: Public goods and ethnic divisions. NBER Working paper no. 6009, National Bureau of Economic Research: Cambridge, MA.

Alesina, A and Wacziarg, R. 1998: Openness, country size and government. Journal of Public Economics 69: 305-321.

Annett, A. 2000: Social fractionalization, political instability and the size of government. IMF Working paper 00/82, International Monetary Fund: Washington.

Atkinson, AB, King, M and Stern, NH. 1984: Evidence to social security reviews. Economic and Social Research Council: UK.

Barro, RJ. 1998: Notes on growth accounting. NBER Working paper no. 6654, National Bureau of Economic Research: Cambridge, MA.

Commander, S, Dolinskaya, I and Mumssen, C. 2000: Determinants of barter in Russia: An empirical analysis. IMF Working paper 00/155. International Monetary Fund: Washington.

Dahlby, B. 1998: Progressive taxation and the social marginal cost of public funds, Journal of Public Economics 67: 105-122.

de Mello, LR. 2000: Fiscal decentralization and intergovernmental fiscal relations: A cross-country analysis. World Development 28: 365-380.

Devarajan, S, Swaroop, V and Zou, H-F. 1996: The composition of public expenditure and economic growth. Journal of Monetary Economics 37: 313-344.

Feldstein, M. 1997: How big should government be? National Tax Journal 50: 197-213.

Gali J. 1994: Monopolistic competition, business cycles, and the composition of aggregate demand. Journal of Economic Theory 63: 73-96.

Grossman, PJ. 1988: Government and economic growth: A non-linear relationship. Public Choice 56: 193-200.

von Hagen, J and Harden, I. 1996: Budget processes and commitment to fiscal discipline. IMF Working paper 96/78, International Monetary Fund: Washington.

Hallerberg, M and von Hagen, J. 1997: Electoral institutions, cabinet negotiations, and budget deficits within the European Union. In: Poterba JM and von Hagen J. (eds). Fiscal Institutions and Fiscal Performance, University of Chicago Press: Chicago.

Hamann, J and Mourmouras, A. 2000: External debt in the Baltics, Russia, and other countries of the former Soviet Union. Unpublished. International Monetary Fund: Washington.

Karras, G. 1996: The optimal government size: further international evidence on the productivity of government services. Economic Inquiry 34: 193-103.

Kaufmann, D, Kraay, A and Zoido-Lobaton, P. 1999: Governance matters. World Bank Policy Research paper no. 2196, World Bank: Washington.

Kneller, R, Bleaney, MF and Gemmell, N. 1999: Fiscal policy and growth: evidence flora OECD countries. Journal of Public Economics 74:171-190.

La Porta, R, Lopez-de-Silanes, F, Shleifer, A and Vishny, R. 1998: The Quality of Government. Journal of Law, Economics, and Organization 15: 222-279.

Landau, D. 1983: Government expenditure and economic growth: A cross-country study. Southern Economic Journal 49: 783-792.

Milanovic, B. 1997: Income, inequality, and poverty during the transition from planned to market economy. The World Bank: Washington.

Peden, EA. 1991: Productivity in the United States and its relationship to government activity: An analysis of 57 years, 1929-1986. Public Choice 69: 153-173.

Persson, T and Tabellini, G. 1999: Political economics and public finance. NBER Working paper no. 7097, National Bureau of Economic Research: Cambridge, MA.

Richards, J, Watson, WG and Brown, DM. 1994: The case for change: Reinventing the welfare state C.D. Howe Institute: Toronto.

Rodrik, D. 1998: Why do more open economies have bigger governments? Journal of Political Economy 106: 997-1032.

Sachs, J, Zinnes, C and Eilat, Y. 2000: Patterns and determinants of economic reform in transition economies: 1990-1998. CAER II Discussion paper no. 61, Consulting Assistance on Economic Reform II Project: Cambridge, MA.

Sandmo, A. 1998: Redistribution and the marginal cost of public funds. Journal of Public Economics 70: 365-382.

Schiavo-Campo, S, de Tommaso, G and Mukherjee, A. 1997: An international statistical survey of government employment and wages. Policy Research Working paper no. 1806. World Bank: Washington.

Schuknecht, L 1994: Political business cycles and expenditure policies in developing countries. IMF Working Paper 94/121, International Monetary Fund: Washington.

Scully, GW. 1995: The 'Growth Tax' in the United States. Public Choice 85: 71-80.

Soh, BH. 1986: Political business cycles in industrialized democratic countries. Kyklos 39: 31-46.

Tanzi, V. 1998: Fundamental determinants of inequality and the role of government. IMF Working paper 98/178, International Monetary Fund: Washington.

Tanzi, V and Schuknecht, L. 1997: Reconsidering the fiscal role of government: The international perspective. American Economic Review, Papers and Proceedings 87:164-168.

Tanzi, V and Schuknecht, L. 2000: Public spending in the 20th century: A global perspective, Cambridge University Press: New York.

Wei, S-Jin. 2000: Natural openness and good government. Discussion paper no. 305, School of Economics and Finance, University of Hong Kong: Hong Kong.

World Bank. 2000: Making transition work for everyone: Poverty and inequality in Europe and Central Asia. World Bank: Washington.

SANJEEV GUPTA, LUC LERUTH, LUIZ DE MELLO & SHAMIT CHAKRAVARTI

Fiscal Affairs Department, International Monetary Fund, 700 19th Street, N.W., Washington, DC 20431, USA. E-mails: sgupta@imf.org; lleruth@imf.org; ldemello@imf.org; schakravarti@imf.org