IV Key Feature 3: Budgets That Are More Pro-Poor and Pro-Growth

Benedict Clements, Gabriela Inchauste, Nita Thacker, Thomas Dorsey, Shamsuddin Tareq, Emanuele Baldacci, Sanjeev Gupta, and Mark Plant
Published Date:
September 2002
  • ShareShare
Show Summary Details

This key feature sets out the expectations that PRGF-supported programs increase poverty-reducing expenditures, improve the efficiency and targeting of these and other expenditures, and include tax reforms that simultaneously improve equity and efficiency.

Reorient Government Spending Toward Activities That Benefit the Poor

The evidence available thus far suggests that the composition of budgeted and actual public spending is shifting towards poverty-reducing activities under PRGF-supported programs, as sought under the PRSP approach. Countries are allocating more to education and health care, both as a percentage of GDP and as a share of total government spending (Figures 1 and 2). These changes represent broad-brush shifts in expenditure composition towards poverty-reducing activities (although imperfectly, because not all public education and health spending is poverty-reducing). While not available for all countries, data on budget outturns suggest that such increases are being realized. Programs were initiated in 25 countries in the sample in 2000; for the 13 countries with available data for that year, spending on education and health care rose by an average of 0.4 and 0.2 percentage points of GDP, respectively. The emphasis in PRGF-supported programs on increasing these outlays is in concert with the view that government has a critical role to play in the provision of social services to support economic growth and poverty reduction. To improve social outcomes, though, this spending will need to be targeted to the poor (see below).

Figure 1.Education and Health Spending in PRGF-Supported Programs, by GDP: Pre-PRGFYear to 2001–02 1

(In percent of GDP; number of countries in parentheses)

Sources: National authorities; and IMF staff estimates.

1 The figures in parentheses indicate the number of countries for which data are available.

2 1999 in most cases.

3 2000 refers to actual expenditure level for that year.

4 Refers to the average projected spending level for 2001–02.

Figure 2.Education and Health Spending in PRGF-Supported Programs, by Total Government Spending: Pre-PRGF Year to 2001–02 1

(In percent of total government spending; number of countries in parentheses)

Sources: National authorities; and IMF staff estimates.

1 The figures in parentheses indicate the number of countries for which data are available.

2 1999 in most cases.

3 2000 refers to actual expenditure level for that year.

4 Refers to the average projected spending level for 2001–02.

Real public spending on education and health care (including spending on HIV/AIDS) is expected to rise sharply on a per capita basis (Figure 3). These spending increases build further on the gains realized during ESAF-supported programs, where real per capita public outlays on education and health care each rose by an average of more than 3 percent per year between 1985 and 1999.13

Figure 3.Annual Change in Real Per Capita Education and Health Spending in PRGF-Supported Programs 1

(Number of countries in parentheses)

Sources: National authorities; and IMF staff estimates.

1 Excludes Guinea-Bissau; extreme changes in spending distort the mean. Including Guinea-Bissau, the average annual increase in real per capita spending in education and health is 11.7 and 11.8 percent per year, respectively, over the period 1999–2001.

The figures in parentheses indicate the number of countries for which data are available.

2 Refers to the annual change in real per capita spending between 1996 and 1999. For 1999 alone, the average increase in all PRGF-supported programs for education and health is 8.9 percent and 11.0 percent, respectively.

3 Refers to the annual change in real per capita spending over the program period up to 2001 or 2002.

Projected increases in education and health care spending vary across countries. In the transition economies,14 for example, smaller-than-average increases are envisaged. This reflects the slightly higher levels of initial spending and the substantial scope for further rationalizing education and health care systems inherited from the pre-transition era. This contrasts with nontransition PRGF-supported program countries, where the need for expanding public education and health—as well as improving the efficiency of such spending—is great. Higher spending increases are also envisaged in nontransition PRGF-supported program countries because of debt relief under the enhanced HIPC initiative.

Substantial increases in spending identified as poverty-reducing in PRSPs are also envisaged (Figure 4). PRSPs have defined a range of programs as poverty-reducing, including spending on primary education, primary or basic health, roads, rural development, agriculture, judicial systems, and anticorruption measures.15

Figure 4.Poverty-Reducing Spending in PRGF-Supported Programs: Pre-PRGF Year to 2001–02 1

(In units as indicated; number of countries in parentheses)

Sources: National authorities; and IMF staff estimates.

1 The figures in parentheses indicate the number of countries for which data are available.

2 1999 in most cases.

3 Refers to the average projected spending level for 2001–02.

Over time, as PRSPs are updated and revised in light of the impact of policies on social outcomes, the definition of poverty-reducing activities is expected to be refined. Based on budgetary data in 19 countries that most closely approximate the PRSP definition of poverty-reducing spending, these outlays will rise, on average, by about 2 percent of GDP from the pre-PRGF year; for new PRGF-supported programs, the increase is slightly lower. The share of total government spending absorbed by these outlays will rise slightly more in new PRGF-supported programs than in the sample as a whole.

For most countries, existing budget classification systems do not allow for a precise matching of expenditure allocations and the programs identified as poverty-reducing in the PRSPs. Only 8 of the 19 countries (Azerbaijan, Bolivia, Guinea-Bissau, Madagascar, Mozambique, Niger, Tanzania, and Uganda) compile spending data on the specific poverty-reducing activities identified in PRSPs. There are also significant lags in the dissemination of data on poverty-reducing spending, including spending under the approximate definition described above; for 2000, figures are available for only five countries.16 As countries move forward in their efforts to improve public expenditure management systems and track poverty-reducing spending (see Section VII)—including efforts with respect to budget classification—it is expected that more countries will be able to provide data consistent with the PRSP definition of poverty-reducing outlays.17 Further progress could also be expected as countries move from interim to full PRSPs and Board documents report more systematically on these data.18

The ultimate objective of reorienting public spending in favor of poverty-reducing programs is to achieve better social outcomes. Data on indicators that gauge social progress, however, are available only with a substantial lag, and they may not be available for every year—making it difficult to iso-late the effect of recent changes in policies. For instance, the only social indicator for which recent data are available for a large number of countries is access to clean water, for which 20 countries report data for 2000. For five countries (Burkina Faso, Lao PDR, Mozambique, Tanzania, and Uganda), data on an additional two to nine indicators are available for the first year in which a PRGF-supported program was in place (2000).19 With due caution in light of the small sample size, preliminary evidence indicates that progress on these indicators since the pre-program year was somewhat better than before the start of the PRGF-supported program.

Higher government spending on education and health, and the shift in the composition of public spending from current to capital outlays, will help facilitate poverty reduction and higher economic growth. In developing countries, public expenditure is a powerful tool for shaping equity or reducing poverty.20 In particular, well-targeted and efficient public programs in education and health are essential to ensure that the poor have the skills needed to contribute to—and benefit from—economic development. Higher outlays for education and health in PRGF-supported programs, in combination with efforts to improve the efficiency and targeting of this spending (see below), can help facilitate higher economic growth. In a similar vein, programs also envisage higher outlays for capital expenditures, which include the provision of critical infrastructure. These outlays will rise, on average, by three-fourths of a percentage point of GDP in PRGF-supported programs, and will also climb as a share of total government outlays (Appendix Table A.1). In combination with the containment of current expenditures (see Section V), the improved composition of public expenditure envisaged in PRGF-supported programs constitutes an important ingredient in countries’ poverty-reduction strategies. At the same time, it is important to ensure that public investment is productive, which may require complementary efforts to strengthen governance.21

In sum, while the available evidence indicates a good beginning in reorienting spending toward pro-poor activities, weaknesses in budgetary classification and reporting suggest due caution in this assessment. Given existing weaknesses in the data on poverty-reducing spending, a more precise assessment of how the composition of spending has been altered under PRGF-supported programs will have to await the strengthening of public expenditure management systems. While these improvements will only be fully realized over the medium term, there are several steps that could be taken in the short term to shore up the reporting of poverty-reducing spending (see Section VII).

Improve the Efficiency and Targeting of Spending in Key Sectors

Higher spending on poverty-reducing activities must be accompanied by improvements in efficiency and targeting to significantly improve social outcomes. Programs are seeking sizable increases in poverty-reducing spending. But this is not sufficient for reducing poverty, because the efficiency and targeting of these outlays must also be improved. The agenda of reform in this area is large, given that the poor reap a disproportionately small share of the benefits from education and health outlays in low-income countries, and a majority of studies show a weak relationship between government social spending and social outcomes.2223 The limited scope for reallocating a large portion of public spending to poverty-reducing activities in the short run further underscores the need to realize improvements in this area.

Enhancing the efficiency and equity of spending in key sectors, such as education and health care, is a critical component of countries’ poverty-reduction strategies. More than three quarters of all PRGF-supported programs incorporate such measures (Figure 5). In many cases, these policy actions are drawn from the PRSPs and advice from the World Bank.

Figure 5.Measures for Improving the Efficiency and Targeting of Public Spending in PRGF-Supported Programs

(Percent of countries)

Source: IMF staff estimates.

PRGF-supported programs are placing more emphasis on improving the efficiency and equity of spending than did ESAF-supported programs. Only one-third of ESAF-supported programs focused on this facet of expenditure policy. Moreover, PRGF-supported programs deal with these issues with more breadth and depth.

Two complementary but distinct aspects of efficiency have been addressed in PRGF-supported programs.

  • The productive efficiency of spending can be improved by using a more productive mix of spending inputs—for example, shifting spending to critically needed nonwage inputs in health and education, such as medicines and textbooks; or making efficiency improvements that allow the government to provide the same level of public services with lower spending—for example, by reducing waste.
  • The allocative efficiency of spending can be strengthened by reallocating public spending within sectors to programs that are most useful for meeting the government’s policy goals. For example, a reallocation of spending from tertiary to primary education and to rural areas can potentially enhance the efficiency of spending if the principal goal of education policy is to increase primary enrollment and completion rates.

The measures to improve productive efficiency vary, depending on country circumstances. A number of countries are improving the input mix by re-straining the growth of the wage bill, thus providing room for increased or higher quality social services (The Gambia, Kenya, Lao PDR, Niger, Senegal, and Zambia).24 On the other hand, some countries (Azerbaijan, Cameroon, Mali, and Uganda) are granting targeted wage increases, or decompressing the wage scale, so as to maintain or attract skilled workers, including workers in the social sectors. In some circumstances, further streamlining of public employment is also envisaged, as is a rehabilitation of infrastructure to enhance the returns from public investment. In Uganda, the efficiency of spending in a number of ministries is also expected to improve with the use of output-based budgeting.

To improve allocative efficiency, the share of spending in primary education and health will rise. A reallocation of spending towards primary education and health is envisaged in a number of countries (Benin, Ghana, Guinea-Bissau, Kenya, Rwanda, and Uganda). Spending in these areas is seen as important in meeting targets for immunization and/or lowering infant mortality; providing health care to pregnant women and/or the elderly; and improving access to kindergarten and/or primary school. Outside the social sectors, some countries also target a shift of spending from urban transportation projects to rural roads.

Policy measures are also envisaged to better target spending toward the poor. In transition economies, a common aim is to make social assistance a more effective instrument of poverty alleviation by targeting this spending to the poor. In a similar vein, some countries plan to replace subsidies that are enjoyed by all consumers—poor and nonpoor alike—with those that benefit only low-income groups. In other countries, the objective is to increase equity by improving the access of the poor to health care and education, facilitated by the elimination of primary school fees (e.g., Tanzania, Uganda). In general, measures to improve targeting are not very specific, and programs provide only general statements of principle.

There is substantial scope to improve the quality and specificity of expenditure advice in PRGF-supported programs. As more countries move to full PRSPs, there will be further opportunity to fully articulate specific measures, and for PRGF-supported program documents to cross-reference these measures and assess progress on earlier initiatives. Full PRSPs will also provide occasion to integrate more fully the policy advice available from development partners, including that found in the World Bank’s Public Expenditure Reviews.

Implement Tax Policies That Simultaneously Improve Efficiency and Equity

Programs attempt to foster more efficient tax systems—which can facilitate growth—and improve the administration of taxes. To facilitate the higher growth necessary for poverty reduction, tax systems should be efficient—that is, they should minimize distortions to the best use of resources across the economy. Efficient taxation also requires that the tax system be impartially but effectively administered based on the rule of law. Building on the substantial progress to reform tax systems in the past, about two-thirds of tax-related measures under the PRGF-supported programs focus on improving the administration of the tax system, rather than on the design of the tax system perse. The average number of new tax measures per program request or review is 8, compared with 10 under the last annual ESAF-supported program.

Tax policies in IMF programs have also sought more equitable taxation. Policy measures have included those that promote greater “horizontal equity”—that is, taxpayers with similar incomes are treated equally. There has been less emphasis on “vertical equity”—that is, making sure that high-income taxpayers pay higher taxes. This is because there is limited scope to implement progressive in-come taxes in low-income countries, given administrative constraints and the high share of agriculture and the informal sector in economic activity.

Tax revenues are expected, on average, to rise by about one percentage point of GDP over the three-year program period under the PRGF, based on higher receipts from indirect taxes (Appendix Table A.1). The likely effect of rising indirect taxes on the distribution of income is difficult to ascertain. While some studies indicate that the poor pay a higher share of their income for indirect taxes than other income groups do (i.e., these taxes are regressive), other studies suggest that indirect taxation is progressive.25

Most measures to change the tax system focus on consumption and trade taxes, rather than income and property taxation (Figure 6). In some countries (Ethiopia, Lao PDR, Lesotho, and Rwanda), this includes efforts to introduce a modern and broad-based value-added tax (VAT). Reform of trade taxes mainly encompasses measures to simplify import taxation and reduce distortions to resource allocation caused by high tariff rates (see below), at times in the context of regional integration movements.

Figure 6.Tax Policy Measures in PRGF-Supported Programs by Type of Tax

(Number of measures per country) 1

Source: IMF staff estimates.

1 Number of countries implementing measures in each category indicated in parentheses.

2 Measures that cover more than one category or the tax system as a whole.

Distributive considerations are important in some programs. In Mauritania, for example, a lower VAT rate was allowed for basic goods consumed by the poor, even though it increased the burden of administering the tax (see Section VIII). In Georgia, greater progressivity in the income tax was envisaged to improve income equality, and in Kenya, further adjustments of tax brackets and the personal tax allowance were undertaken to offset the effects on low-income groups of increases in indirect taxes.

Programs are also supporting a more equitable and efficient tax system by reducing exemptions and broadening the tax base. Almost three-fourths of programs incorporate measures to broaden the tax base by removing exemptions (Cambodia, The Gambia, Guinea, Kenya, Lao PDR, Macedonia, Mauritania, Mozambique, Rwanda, Tajikistan, Tanzania, and Uganda), abolishing tax holidays, or paring special tax regimes that benefit foreign investors (Cambodia, Georgia, Lao PDR, Mauritania, Mozambique, and Rwanda).

Efficiency gains are also expected from the simplification of the tax system or reductions in high marginal tax rates. About three-fourths of all PRGF-supported programs seek to simplify the tax system and reduce distortions (Figure 7). This is to be accomplished, for example, by lowering tariff rates or reducing the number of import tariff rates (Albania, The Gambia, Ghana, Malawi, Mauritania, Mongolia, Niger, Sierra Leone, Tanzania, Vietnam, and Zambia); tightening tariff exemptions; reducing excessive excise tax rates; and simplifying the rate structure of the personal income tax.26

Figure 7.Tax Policy Measures in PRGF-Supported Programs by Type of Reform

(Number of measures per country) 1

Source: IMF staff estimates.

1 Number of countries implementing each measure indicated in parentheses.

2 Includes measures to introduce VAT, eliminate surtaxes, and make general reforms of the tax system.

3 Includes removing tax exemptions and abolishing tax holidays.

4 Includes reductions in the number of different tax rates and lowering of marginal tax rates.

5 Includes temporary duties.

A wide array of measures are expected to strengthen tax administration. About half of these measures are devoted to improving taxpayer compliance by widening the revenue collection base and strengthening enforcement (Figure 8). These efforts include the introduction of taxpayer identification numbers (Cameroon, Chad, Mali, Senegal, and Tajikistan) and strengthened tax audits or inspections (Albania, Armenia, Benin, Cambodia, Georgia, Ghana, Macedonia, Mauritania, Moldova, Mozambique, Rwanda, and Uganda). Other measures seek to improve the technical capacity and organization of the revenue authorities through computerization and improved operating procedures, including improvements for large taxpayer units (Bolivia). The majority of tax administration measures focus on the domestic tax system, and a larger share of customs measures are subject to conditionality (i.e., the measures are prior actions, performance criteria, or structural benchmarks under the program).

Figure 8.Tax Administration Measures in PRGF-Supported Programs

(Number of measures per country) 1

Source: IMF staff estimates.

1 Number of countries implementing each measure indicated in parentheses.

PRGF-supported programs place slightly less conditionality on tax measures than the ESAF-supported programs did. PRGF-supported programs average between two and three tax measures subject to conditionality per annual program request or review; about half take the form of structural benchmarks. This average compares with about three such measures per country under the last ESAF-supported program. Furthermore, the percentage of all tax measures subject to conditionality has declined. These developments reflect efforts to streamline conditionality (see Section VI).

Distributive aspects of taxation are not always discussed in PRGF-supported program documents. As poverty and social impact analysis (PSIA) is strengthened (see Section VIII), the implications of taxation on income distribution and poverty reduction can be expected to be treated more systematically and in greater depth.

A vigorous program of technical assistance will be required to achieve continued progress in modernizing tax administration. Despite the impressive strides made in this area, there is a long road ahead to building modern and efficient tax administrations in PRGF-supported countries. A well-coordinated effort among development partners is needed to help accelerate the pace of improvements in domestic capacity.

    Other Resources Citing This Publication