The most welcome way of increasing revenue would be for the prince to abolish superfluous expenditure, to disband redundant offices, to avoid wars and foreign tours, … to check the acquisitiveness of officialdom, and to pay more attention to the just administration of his territory than to its expansion.
—Desiderius Erasmus (1469–1536)
Fiscal adjustment should ideally be carried out through high-quality structural measures early in the process. These include sound revenue and expenditure policies for the medium term, as well as short-run options that are best for growth. Social safety nets can mitigate any short-term costs imposed by fiscal consolidation on the poor.
Improving the Tax System and Mobilizing Revenue
Structural problems in the tax system may well be a major contributor to fiscal deficits, and also to poor growth and unemployment. The ability to generate revenue by raising tax rates may be limited, particularly when an economy is undergoing substantial structural change, tax bases are narrow, hard-to-tax sectors (e.g., agriculture and the self-employed) are large, existing rates are high, or tax administration is weak—conditions that often reinforce one another. Thus, fiscal consolidation is often accompanied by measures to improve the tax system, while taking into consideration the equity and efficiency impact of reforms.
Tax systems need to balance macroeconomic and microeconomic objectives (Tanzi and Zee, 2000). In particular, a tax system should ideally be:
Productive. A central goal of the tax system is to raise revenue to finance government spending. This requires a system that can generate revenue increases at least in line with the growth in income without frequent changes in tax rates or introducing new taxes. Charging low rates on a broad base typically best meets this objective. Exemptions and incentives—such as (and arguably worst of all) tax holidays—can severely undermine revenue-raising capacity, often with little if any offsetting benefits (Box 6).
Efficient. Taxes distort relative prices and thus affect the pattern of production, consumption, investment, and income. Unless there is strong reason to suppose that market prices are themselves sending the wrong signals—which may be the case, for example, when some activity causes environmental damage—an efficient tax system imposes low and reasonably uniform tax rates on as broad a tax base as possible. It also avoids exemptions or special tax rates that artificially encourage investment in projects with below-market returns. However, when environmental or other externalities are present—where parties to some transaction do not bear all the social costs or enjoy all the social benefits—then particular taxes (e.g., on pollution) or, subsidies (e.g., to basic education) although less likely (given their revenue cost and potential for abuse) are in principle appropriate.
Fair. Each country must decide for itself exactly what constitutes fairness. Distinguishing between vertical and horizontal equity can be useful. Vertical equity refers to differentiation of the tax burden according to ability to pay, and horizontal equity, to equal treatment of those in similar economic circumstances. Certain types of taxation may affect income distribution—for example, a progressive income tax or a reduced rate of value-added tax (VAT) on basic foods. But experience has shown that taxation is relatively ineffective in influencing income distribution in general, and in helping the poor in particular. And this is becoming increasingly true as tax bases (especially, but not only, capital income) become more mobile internationally. Expenditure policy is generally better suited to influencing income distribution, and equity effects should be designed and assessed comprehensively, embracing not only all taxes but also the spending they finance.
Simple and transparent. Taxes with a single rate or a low number of rates and minimal exemptions are easier to administer and easier for the taxpayer to understand, comply with, and—ultimately—express views about. Tax rates should be stable and predictable. Once tax laws that generate buoyant revenue growth are in place, frequent amendments should be avoided. If changes are planned, taxpayers should ideally know the implications in advance (unless this creates unacceptable scope for tax avoidance).
Dangers of Tax Holidays
Tax holidays—exemptions from tax, sometimes for many years—are a particularly ill-designed form of investment incentive for the following reasons:
They attract the most footloose forms of business, since they can easily move elsewhere at the end of the holiday. These are the firms least likely to offer spillovers to the wider economy in terms of training or deep linkages with the domestic economy.
They are open to abuse, undermining tax revenue not only directly but indirectly, by providing entrepreneurs with an incentive to use transfer pricing and financial arrangements that shift taxable income to holiday companies from companies that would otherwise be taxed. For example, taxpaying companies (able to deduct the interest payments) can arrange to borrow from holiday companies (not taxable on interest received).
In many cases, the appeal of tax holidays has been that they protect companies from exploitation by corrupt tax administrations. By the same token, offering a holiday can be seen as signaling a low-quality civil service.
They are relatively inefficient at encouraging employment (often the claimed objective) since, like other investment-based incentives, they encourage the use of capital, not labor.
Developing countries have felt compelled to offer such incentives because of tax competition with each other, and this appears to be one of the reasons for the reduction in corporate tax revenues in many countries. But tax holidays do nothing to address the underlying problems that may deter foreign investment (e.g., instability of the tax system, an ineffective judicial system, arbitrariness in administration and red tape, and foreign exchange restrictions). The evidence suggests, indeed, that tax factors, while they do matter, are far from the main concern of foreign investors in deciding where to place their funds.
There are better options for stimulating investment, such as accelerated depreciation and capital allowances. Maintaining a reasonably broad base for the corporate income tax makes it easier to set a reasonably low corporate tax rate. This in itself is likely to protect the revenue base and provide a supportive environment for investment. Realizing this, a number of countries have scaled back tax holidays and other tax incentives for investment. Sometimes, however, countries feel pressure to retain or expand such incentives to compete with those available elsewhere. This leads to a mutually damaging form of tax competition. In such cases, regional agreements can be helpful—for example, on a code of conduct on business taxation, as adopted by the European Union.
The design of the major taxes should draw on the following criteria:
Sales tax or VAT. This should be a broad-based tax on final domestic consumption that does not tax intermediate consumption or exports, and does not differentiate by source of production (foreign or domestic; see Ebrill and others, 2001). Because of its efficiency (in not affecting business use of inputs) and revenue security (in collecting revenue at all stages in the production chain, not just at final sale), the ideal instrument to achieve this goal is usually a VAT levied at a single positive rate—key elements being crediting provisions to remove businesses’ input purchases from tax and zero rating22 of exports. Exemptions should be avoided as far as possible, since they reduce the efficiency of the VAT by causing it to “cascade” through the chain of production (Harrison and Krelove, 2005). Zero-rating other than for exports should also be avoided, since it creates a need to pay refunds that can cause great administrative difficulty. Nevertheless, exemptions are often given either on technical grounds (for financial services) or—albeit misguidedly, since most of the benefit often goes to the better off, who buy more of these items than the poor—for equity reasons, as for selected basic foodstuffs and basic education and health services.
Excises. Imposing excises (which, unlike the VAT, are not creditable) on a limited number of commodities can be a good tax handle when demand for the commodities is inelastic and the commodities can be easily monitored. They are appropriate for dealing with negative externalities (e.g., a gasoline tax to address problems of local air pollution and global warming) and have, for some, a particular appeal in limiting consumption of “harmful” items (e.g., alcohol and tobacco). Excises should be levied equally on domestic production and imports and, with a few exceptions—notably for petroleum products (to avoid worsening volatility)—on an ad valorem basis.23
Customs duties. A low, across-the-board tariff is useful for revenue reasons in countries where other, preferable, taxes may be hard to administer. Schemes are needed to relieve exporters of the anti-export bias caused by customs duties on inputs, either by a system of drawback (refunding duties paid on imports used to produce exports) or by one of suspension (excluding such imports from duty altogether). Exemptions from customs duties should be limited and clearly defined to avoid abuse. As trade liberalization proceeds, revenue from this source is likely to fall. For many developing countries, a key challenge is to strengthen their domestic revenue systems to replace lost trade tax revenues.
Export taxes. Export taxes should generally be avoided as they tend to draw resources from the export sector toward less efficient uses, compromising growth. But they can be useful on a limited basis for hard-to-tax activities (common in the agriculture sector), as a temporary substitute for income taxation, and to absorb one-time windfall gains (e.g., from devaluation or from exceptional movements in world commodity prices).
Business income taxes. A tax on profits should ideally be levied at a single rate, broadly comparable with the top marginal rate of the personal income tax. This minimizes the likelihood of tax-induced shifts between personal income and enterprises. Deductions, allowances, and credits are best applied neutrally across sectors and assets to foster efficiency. If tax incentives are considered, they should consist of accelerated depreciation—that is, allowing assets to be written off for tax purposes more rapidly than they actually depreciate. A minimum profits tax based on turnover or gross assets may be used in some circumstances to promote compliance and equity, and to secure revenue. For small businesses and the informal sector, a simple presumptive tax based on gross assets or other indicators can be used, given the difficulty of assessing their actual incomes.
Natural resource taxation. This generally requires specialized tax instruments to ensure that risk and revenues are shared appropriately between the government and investors, and that the resource itself is responsibly exploited.
Personal income taxes. A basic personal income exemption should be set high enough to exclude the poor. Sufficient progressivity can be achieved with only a few income tax brackets. Tax bands should be adjusted periodically for inflation to avoid “bracket creep,” while incentive and compliance considerations argue for keeping rates as low as possible. It is not prudent, however, to expect a cut in high marginal rates—as seen recently in a number of countries that introduced a “flat tax”—to produce an actual increase in revenue. One textbook view is that income taxes should be levied on a “global” base—charging tax on the sum of income of all kinds. However, it is often administratively convenient to establish schedular taxes24 on different sources of income (e.g., through final withholding taxes on interest income). A number of countries have recently experimented successfully with a “dual” income tax of this kind, combining a low, flat rate of tax on capital income and a progressive tax on labor income.
Major reforms in tax design and administration take time to implement and to achieve their expected revenue impact. New legislation is often required and basic systems and procedures frequently need modification. Short-term measures may thus be needed. They should be assessed in terms of their effect on resource allocation, fairness, administrative feasibility, and their consistency with the desired direction of longer-term tax reform. The most promising measures for producing short- (and indeed long-) term revenue increases typically involve expanding the tax base by eliminating tax exemptions (especially tax privileges). Not only can this raise revenue, it can also simplify administration, freeing up resources for more productive activities, and it fosters taxpayer compliance as well.
Rationalizing Public Expenditure and Protecting the Poor
Public expenditure reform is typically undertaken to reduce government spending. But even when public spending need not shrink, expenditure reform can still improve the productivity of existing spending, free resources to help meet new needs, and improve governance and transparency (Gupta and others, 2005a). Reducing expenditures while improving their composition need not undermine growth or social indicators.
Spending cuts should be pragmatic—adequate to achieve the intended goals yet economically sound and socially feasible. Many measures can be taken quickly, but durable, good-quality expenditure reform typically demands a review of underlying government policies, the composition of spending, the coverage of activities by the public sector, and the modes of delivery of public services. Quite often, a thorough structural reform of government spending policies can be done only over a number of years (Box 7).
Fundamental Public Expenditure Reform
Fundamental structural reform requires asking basic questions about whether government activities are needed, should be provided by the public sector, or could be made more market based. For activities to remain in the public sector, specific objectives need to be set, desired outputs quantified (where possible), inputs determined, and managerial freedom given to pursue the most efficient delivery of services.
A comprehensive review might generate the following types of reforms:
Eliminate unproductive or low-priority services.
Privatize activities that can, and should, be carried out in the private sector.
Introduce a more commercial approach to public activities, including competitive tendering and contracting out of some services to the private sector.
Simulate market discipline, including in government purchasing and provision of services, for example, in health care.
Enhance management of those services that are to remain in the public sector. This would entail, for example, more devolved managerial authority and linking of managers’ salaries to performance.
In countries where public expenditure management is advanced, introduce “performance-based” budgets that strengthen links between the results delivered by agencies and the funding they receive.
Longer-Term Expenditure Reform
Public spending should be judged on its impact on growth and investment, as well as on poverty and equity. Apart from core government functions, market failures (e.g., positive externalities, public goods, and imperfect credit markets) are a main justification for public sector activity. Public expenditures also play a redistributive role, especially when targeted at the poor.
The private sector may refrain from certain activities that have a large social return because the private return is too low. This will be the case when there are positive spillover effects (“externalities”) and for public goods. In the case of positive externalities, the market will produce too little of a good relative to the socially optimal amount—for example, too little education relative to the benefits accruing to the community at large. In the extreme case of public goods, where it is impossible to charge people for the benefits they receive, these goods will likely not be produced by the private sector at all. Alternatively, the private return of an investment may be sufficiently high, but financing, owing to market failure, may be unavailable.
Productive expenditures are those that have a high social rate of return. Such rates of return are difficult to measure, but at the functional level, they are often highest in infrastructure, primary education, preventive and primary health care, and basic public services. In terms of their impact on poverty, programs in primary education, basic health care, water and sanitation, roads, rural development, agriculture, judicial systems, and anti-corruption appear to have the largest impact. Expenditure composition in many countries suggests that there is great scope for improvement:
About one-fifth of education spending is allocated to tertiary education, which has a lower rate of social return than education spending at the primary and secondary levels.
Similarly, in the health care sector, spending on basic preventive health care—such as immunizations—have a high social return and also a relatively larger impact on the poor. Still, almost two-thirds of public health care outlays tend to be absorbed by curative care rather than basic and preventive health care.
Some governments provide generalized subsidies. These are usually distortionary. For example, subsidies that encourage wasteful power, water, and fertilizer use generally damage the environment and typically benefit the better-off in society as they consume more than the poor. In some instances, subsidies can be targeted at the poor, while in others, they can be eliminated and the proceeds used to benefit the poor (for example, by developing rural transportation networks).
Spending on goods and services other than wages can be highly productive, but it often bears the brunt of adjustment. Highly productive nonwage inputs, such as medicines and textbooks, are often crowded out by wages and salaries. As a result, health care workers and teachers are left without the necessary complementary inputs. Similarly, public investment programs often ignore the recurrent costs of a project, resulting in rapid deterioration of infrastructure.
Other than by targeting specific sectors, the productivity of government spending can be increased by addressing governance and management issues. In education, the quality of teachers, and whether they turn up and teach, is more important than class size. This underscores the relative importance of reasonable wages over the number of teachers. Investment expenditure is notorious for containing “white elephants,” such as underused airports and bridges. The productivity of investment can be increased by following some basic principles. These include subjecting large projects to tighter scrutiny and rigorous cost-benefit analysis and monitoring all central and local governments proposed and ongoing projects.
Reducing corruption will tend to improve the quality of spending and the amount of revenue to finance it (Tanzi, 1998). In particular, corruption tends to be linked with less spending on health and nonwage operations and maintenance and with more allocations for less productive investment projects and military spending.
Countries with aging populations will require two very difficult adjustments, entailing not only changes in spending patterns but also in the promises that governments make.
Pension reform. Pension systems face two challenges. First, most public pension systems promise benefits that cannot be financed with existing contributions. Countries have several options for addressing this medium- to long-term problem (Box 8). At one extreme, they can restructure existing pay-as-you-go systems to bring benefits into actuarial balance with contributions. At the other extreme, they can shift to private, funded systems, gradually reducing current implicit pension liabilities. Even in this latter case, adjustments in the legacy pay-as-you-go systems will be necessary to make the transition affordable. Second, the coverage of many pension systems is often narrow and provides preferential treatment for certain groups, such as public sector employees. It is important to widen coverage; in developing countries, this will require labor markets to become more formal. However, coverage should not be widened until existing systems have been reformed to ensure their long-run sustainability.
Health care. Adjustments in how health care is rationed and financed may be even more daunting than pension reform. The elderly are the most intensive consumers of health care, and the aggregate demand for health care will grow rapidly as populations age. Rapid technological change in health care has been, and will likely continue, expanding available treatments, often at substantial cost. Other demands on the health care system, such as HIV/AIDS, may compound the growth in the demand for health care. Allocating health care spending in a manner that is both fiscally responsible and socially equitable is one of the biggest challenges currently facing economic policymakers.
Key Issues in Pension Reform
The fiscal unsustainability of many public pay-as-you-go pension schemes has become increasingly apparent in recent decades. Many countries—especially developed countries with large unfunded liabilities—are attempting to remedy this problem. In so doing, several choices must be made:
Public versus private. A threshold question is what role the government should play in the pension system. It may want to take an active role in the system if workers are myopic about their future income needs or might refrain from saving on their own in the belief that the government will provide assistance in any event. The government may also want the system to have a redistributive component.
Defined-benefit versus defined-contribution. Pension systems are typically either defined-benefit, in which benefits are determined by formula based on prior income, or defined-contribution, where benefits are determined by the amount of accumulated contributions. These two choices differ in the manner in which risk is allocated between the plan sponsor—for instance, the government—and the participant.
Funded versus pay-as-you-go. A system can be either funded, in which contributions are invested and accrue market returns, or pay-as-you-go, in which the contributions of current workers are used to pay the benefits of current retirees.
Choices along these dimensions must be carefully considered and will—at least for some developing countries—be dictated by the capacity of the private sector to administer the system and provide investment opportunities. Moreover, shifting from a pay-as-you-go to a funded system entails significant transition costs, and a country’s ability to finance these costs will determine the feasibility and speed of such a shift. A transition to a funded pension system may be desirable, but there is no substitute for reforming existing pay-as-you-go systems to reduce unfunded liabilities.
Short-Term Expenditure Reform
There are no hard and fast rules about how public expenditure should be cut in the short run, when needed. This will depend partly on the factors driving the growth in spending (e.g., wages and salaries or the capital program), as well as on the social and political constraints facing policymakers. However, experience suggests some guidelines.
Protect core programs. While fiscal consolidation in the short run may require lower overall spending, it is still possible to safeguard most, if not all, core productive expenditures and to protect the poor with well-targeted social safety nets.
Identify specific program reductions. Many programs can and should be dropped, pruned, or consolidated as economies develop and priorities change. For instance, free milk distribution may become unnecessary when income levels rise above a certain level. Program elimination usually leads to effective savings—because it requires governments to redefine their priorities and is a first step toward a more fundamental expenditure review—and it preserves the efficiency of operations elsewhere in the public sector.
Cut the public sector wage bill. Wage restraint and hiring freezes can be a major source of savings in the short run, but they are politically difficult to sustain and not necessarily desirable from an efficiency point of view. Thus, they should be seen as interim substitutes for structural reform that entails a deeper review of employment and pay policies and program staffing needs.
Target social programs narrowly. General subsidies could be eliminated and transfers made more efficient by targeting eligibility and by reducing income replacement rates (Gupta and others, 2000a). Where possible, multiple programs for social protection should be consolidated into more global schemes of income transfers. This avoids significant overlap in entitlements provided by uncoordinated agencies.
Review the capital program. The capital program is often the prime target for short-term retrenchment. Postponement of projects not yet begun can save resources with relatively little disruption of day-to-day government operations. The cost, however, may be lower growth and development. Capital programs are best cut in the context of an overall public investment review—often possible only as part of a medium-term strategy.
Raise fees and charges. Governments are often reluctant to reduce volumes or standards of delivery in high-priority areas like education and health. Savings in these areas may best be achieved by cost recovery through an increase in the fees and charges for services, while protecting the poor.
Change public enterprise tariffs and subsidies. If public enterprises are in deficit, their pricing structures may need to be adjusted and subsidies eliminated. The scope of their activities may also need redefining; their employment policy, adjustment; and their capital program, rationalization.
Social Safety Nets and Poverty and Social Impact Analysis
Social safety nets consist of a combination of measures aimed at protecting the poor from the adverse consequences of economic shocks and structural reforms, and helping them escape poverty. Because social safety nets need to work quickly and reliably, they must be tailored to the specific circumstances of each country, including its administrative capabilities, the strength of its informal and formal social support systems, and the characteristics of the poor. Typically, the major components of social safety nets include:
Targeted cash compensation, commodity subsidies, and fee waivers. With effective targeting, these are the preferred schemes to protect consumption by the poor in the face of falling incomes or rising prices. When fees were introduced for budgetary reasons, fee waivers have been used to help maintain access for the poor to education and health services. Providing basic minimum services (such as electricity and water lifelines) at low cost is another option.
Enhanced unemployment benefits, severance pay, and public works schemes. Because reforms may lead to a temporary rise in unemployment, assistance can be channeled through schemes designed to mitigate the fall in employment. Governments should maintain the distinction between social insurance, typically financed through contributions, and social assistance, which should be financed through the budget. A well-defined reemployment strategy (e.g., retraining schemes) can also play a role.
Targeting and incentives are the key issues in the design of social safety nets. Sophisticated means testing is generally not possible, owing to the lack of administrative capacity. Many countries rely instead on categorical targeting, such as limiting benefits to children or pensioners, or to households in certain especially poor regions. Another form of targeting that requires little administrative capacity is to limit subsidies to goods consumed disproportionately by the poor, or to limit the quantity that each household can consume, for example, via coupons. The fiscal cost of the social safety net is reduced the more sharply benefits are phased out with rising household income. This, however, increases the implicit marginal tax rate facing beneficiaries, and thus the potential adverse impact on work incentives.
PSIA can be used to determine the impact of expenditure (and other policies) on the poor. It is the analysis of the positive and negative consequences of changes in policy (e.g., taxation, trade reforms, public enterprise retrenchment, and social expenditures) on the welfare of different groups in society—with an explicit focus on the poor and vulnerable. Depending on country circumstances, PSIA may employ a variety of tools to gauge the impact of policies. These include incidence analysis, social impact surveys, micro-simulation models using household surveys, and computable general equilibrium models. These tools rely on inputs from governments, which, in turn, derive from an open consultative process, including with civil society. A PSIA may suggest modifying proposed policies or strengthening social safety nets to accompany new policies.