Vito Tanzi1

Vito Tanzi1

In recent years many countries have adopted economic programs aimed at “adjusting” their economies. By and large, the adjustment efforts have focused on reducing the rate of inflation, improving the balance of payments, and promoting economic growth. All these adjustment programs have required that substantial attention be paid to the fiscal situation since a reduction of the fiscal deficit is now recognized to be a necessary condition for improving the macroeconomic situation. Yet the experience of many countries indicates that fiscal reform is very difficult.

Many of the issues that arise in pursuing fiscal reform concern (i) the determination of the correct size of the fiscal adjustment needed, (ii) the difficulties in measuring the existing fiscal disequilibrium, and (iii) the type and timing of the required fiscal reforms. These issues are discussed in the following sections.

Fiscal Adjustment Needed

In many developing countries a major cause of the external imbalance of a country is excessive monetary expansion. By increasing nominal incomes, the expansion leads to increases in nominal spending and in domestic prices. When exchange rates are pegged, it encourages imports/discourages exports, and, by creating expectations about devaluation, induces capital flight and the dollarization of the economy. In developing countries, excessive monetary expansion is often the result of fiscal deficits that cannot be financed through noninflationary channels.2 Thus, a kind of “fiscal approach to the balance of payments” arises that establishes a close connection, through liquidity expansion, between the financing of the fiscal deficit and the outcome of the balance of payments.

Limiting the fiscal deficit to a level that can be financed internally in a noninflationary manner (say, by selling bonds to the public) or externally through available, ordinary credit, a country may be able to eliminate the short-run problems associated with the fiscal imbalance. Thus, an adjustment program that focuses on the short run requires that the fiscal deficit be reduced enough to eliminate the need for inflationary financing. The program, therefore, would not focus on the part of the deficit that can be financed, in the short run, in noninflationary ways. This does not mean, however, that the part of the deficit not financed by monetary expansion is of no concern. If this part leads to an increase in the share of public debt in GDP, crowds out the private sector, or skews portfolio distributions in the financial sector, good adjustment programs must also attempt to set limits on the total fiscal deficit. In some countries the continuous and prolonged reliance on bond financing has increased the share of domestic debt in GDP to worrisome levels. Eventually this domestic debt leads to serious difficulties and to the need to introduce major and painful fiscal restructuring.

It is perhaps necessary to emphasize that the part of the fiscal deficit that can be financed in noninflationary ways varies from country to country. In relatively closed economies, it depends on the saving rate of the private sector and on the after-tax real rate of return on government bonds. The higher the private sector saving rate, the higher the after-tax interest rate that the government pays on the public bonds, and the lower the private sector borrowing, then the higher is the fiscal deficit that a country can finance in a noninflationary way. Countries such as India and Italy were able to finance large fiscal deficits in relatively noninflationary ways for long periods of time. However, as domestic (or foreign) debt accumulates, this strategy becomes progressively more costly unless the public sector borrowing is channelled to very productive investments that lead to a higher rate of economic growth.

Measuring the Fiscal Disequilibrium

Correctly measuring fiscal disequilibrium is formidably difficult.3 The commonly used measures of the fiscal deficit can be highly inadequate. They may not be comparable over time or for a given country (or across countries) in a given period of time. And they are often imprecise gauges for determining the size of the needed fiscal correction. Several other difficulties also complicate matters.

The first difficulty is connected with the quality of the data. With few exceptions, governments have not given a high priority to gathering and providing reliable and up-to-date statistics. The second issue concerns the comprehensiveness of the available fiscal data. The central government, which is often the main actor in adjustment programs, represents in many countries only a limited part of the public sector. The whole public sector comprises the central government, local governments (provinces and municipalities), public enterprises, the central bank, stabilization funds, and various forms of extrabudgetary accounts, including social security. These subsectors are often interconnected, and implicit transfer prices determine the scope of financial or resource flows among them. These transfer prices often do not reflect market prices. For example, the central bank may finance other parts of the government at subsidized rates or assume some of their budgetary functions, as has happened in several Latin American countries where governments have run very large deficits. The public enterprises may sell services to the central government at below market prices or may borrow to reduce transfers from the central government. Social security institutions as well as public enterprises may be forced to buy government bonds at interest rates below what the government would have paid in the free market.

To avoid this problem, the data on which the fiscal adjustment will be based must be comprehensive enough to encompass the whole or at least much of the public sector. But achieving this is extremely difficult, even impossible, because the information required is often not available and when it is available it is not up to date. For example, data from the provinces, the municipalities, and the public enterprises can be hard to obtain and are often unreliable. This limitation may create serious difficulties in the conduct of fiscal policy.

In the presence of domestic debt, inflation can bring about dramatic changes in the size of the conventionally measured fiscal deficit by increasing nominal interest rates and thus nominal interest payments. Inflation disguises as interest payments what in reality are amortization payments,4 and while amortization payments are conventionally not counted as part of the deficit, interest payments are. As a result, the conventional deficit becomes a direct function of the rate of inflation. In such situations the deficit ceases to provide a useful measure of the size of the fiscal correction needed by the adjustment program.

A country with a large foreign debt can influence the measurement of the fiscal deficit (as a share of GDP) by changing its exchange rate. An overvalued exchange rate will convert the foreign exchange interest payment on the debt into fewer domestic monetary units, and thus give the impression of a lower fiscal deficit.5 This lower deficit may mislead policymakers into believing that the fiscal problem is less serious than it is or may raise their resistance to a necessary devaluation.

Timing issues also create difficulties. A deficit can be measured on the basis of cash flows (that is, actual cash receipts and payments) or on a commitment basis for expenditure and on an accrual basis for revenue. A government can increase its spending but delay the actual payment for the additional goods and services that it has bought from suppliers. When arrears are increasing, an adjustment program that employs the cash concept may misjudge the pressure on resources and on demand that is associated with expenditures made but not yet paid for. In other words, the cash concept will underestimate the real size of the deficit. The impact on the real economy is likely to be felt mostly when the resources are transferred to the government rather than when the government’s check is actually cashed.

Essential Fiscal Reforms

A given reduction in the fiscal deficit may be genuine and of good quality or largely cosmetic and of low quality. The economic effect of these two types of “adjustments” especially over the medium run are likely to diverge widely. Unfortunately, cosmetic changes are frequently easier and politically less costly to make.

It is often not appreciated that a specific reduction in the fiscal deficit is usually not the result of a single policy decision but the summation of many specific policy decisions, both on the revenue side and on the expenditure side.6 A high-quality fiscal adjustment must be associated with measures that are efficient, durable, and equitable. In other words, these measures must not introduce avoidable distortions; they must not self-destruct in the near future; and they must not eliminate expenditures that are important for economic or social reasons when alternatives are available. The public spending to be reduced must be that spending which contributes least to the efficiency and fairness of the economic system.

Revenue Measures

On the revenue side, and broadly in order of preference though not in order of facility of introduction, the following measures could be taken.

First, a general consumption tax, possibly one with the characteristics of a value-added tax (VAT), should be introduced or, when it exists, broadened.7

Second, the government should fully explore the revenue possibilities of excise taxes. These excises should be imposed on commodities with inelastic demand or on those whose consumption generates substantial negative externalities.

Third, important changes should be introduced to the personal income tax. Personal income taxes still contribute little to total revenue in developing countries. If basic changes are made, the threshold of the tax can remain generous and high marginal tax rates can be cut without much revenue loss and with potential gains in work effort and taxpayer compliance. Given the income distributions prevailing in developing countries, personal income taxes do not need to cover the entire population to generate substantial revenues. Focusing on the top 10 percent of the population would be sufficient.

For taxes on enterprises, similar considerations apply. Enterprise income taxes are often eroded by excessive incentives and complex laws. These incentives rarely achieve the intended objectives but create major problems for tax administration. The elimination of these incentives often achieves a great simplification of these taxes and makes them more equitable.

The next measure, which in the short run can be especially important from a revenue standpoint, consists of raising public utility prices and introducing user charges for particular services provided by the public sector, such as higher education and health. The real prices at which electricity, water, telecommunications, transportation, and other public sector services are sold normally fall quite drastically during inflationary periods. This fall increases the demand for these services and, given common capacity constraints, leads to crowding and overuse. Because of the losses experienced, the public enterprises cannot muster the resources to expand capacity. Often these enterprises do not even have the funds necessary for operation and maintenance, especially in light of the more intensive utilization of their plants. Thus, capacity declines and the quality of the service deteriorates. The greater the fall in the real tariffs charged by the public enterprises and the greater the demand response to that fall, the greater will appear the need to expand investment in these activities. Thus, an artificial justification for capacity expansion will be created by the fall in real prices especially when the resources necessary to satisfy that expansion are sharply reduced.

A correction of these prices will thus (i) generate more revenue, (ii) reduce the need to expand capacity by lowering demand, and (iii) lower operational and maintenance costs by reducing overuse. Generally, this revenue increase can also be defended on the basis of equity considerations since the very poor often have no access to these services. In fact, by keeping prices too low, the resources needed to provide the poor with access to these services are reduced. Broadly similar considerations are relevant to the implementation of user charges, especially for higher education.

Finally, imports should be made to generate larger revenues either by dismantling quotas and other quantitative restrictions and replacing them with import duties or by removing the excessive erosion of the import tax base created by incentives and special exemptions and introducing a minimum tax on all imports.8 In both cases the additional revenue would be accompanied by improved efficiency.

Expenditure Measures

On the expenditure side, a variety of steps should also be taken.

First, and most important, unproductive investment projects must be eliminated. The popular argument that investment must be protected during adjustment is simply misguided. While productive investment is an important source of growth and must be protected, unproductive investment, especially if associated with imported machinery and capital equipment, is a major burden on the economy. In many developing countries the investment budget is padded with politically motivated and unproductive investments which can and should be eliminated. Unlike consumption expenditure, which always benefits someone within the country, investment, especially if associated with imported capital equipment, may contribute little to the welfare of the country’s citizens. Furthermore, if it is obtained with foreign credit, it becomes a long-term drag on the economy. Hence, unproductive investment must be the first area where cuts are made.9 Some of the savings from this source should be allocated to operations and maintenance expenditure, which would increase the efficiency of the existing capital structure and would permit that structure to support a higher level of income.10

A second area where reductions should be made is in the wage bill of the public sector. During adjustment, many countries have attempted to reduce the wage bill of the public sector but have often preferred to reduce real wages than the actual level of public employment. Yet in many developing countries the public sector is overstaffed. Therefore, fiscal adjustment that aims to reduce the wage bill permanently must cut (in some cases quite considerably) the number of public sector employees. This may require privatizing some activities. Since a public employee receives not just a wage but also workspace, tools, family allowances, and other benefits, the reduction in public employment would also reduce the necessity for some public investment (buildings) and for some nonwage expenses (paper, electricity, and so forth). These potential savings disappear when real wages, not public employment, are reduced.11

The third important area for reduction, although a politically difficult one, is military expenditure. An argument can be made that military expenditure remains excessive in all countries. In developing countries it absorbs as much as 5 percent of GDP and about 30 percent of total tax revenue, thus greatly reducing the government resources available for education, health, and other important sectors. It does not take much imagination to see how a country could benefit from the wise alternative use of these resources. For example, by cutting military spending a country could raise its growth rate by at least 1 percentage point a year; or it could double the real income of the lowest 20 percent of the income distribution; or it could raise the level of human capital or the quality of life. And tragically, military spending often increases “national security” only in the abstract, since it commonly leads to equivalent increases by potential enemies. International pressures to downsize military establishments, particularly in rival countries, could generate very high returns, although it may lead to complaints about interference in the domestic affairs of countries.

Fourth, many countries engage in various forms of “unproductive” expenditure, from the building of monuments to the subsidization of unnecessary activities. In many of these countries the enforcement of useless regulations also requires substantial public sector resources, and compliance with those regulations can require even higher private sector resources. Reducing regulation lowers public spending while often enhancing the productivity of private sector activities. In most countries the budgets of many of these activities could stand pruning.

Subsidies must be closely scrutinized. Those which are essential, because of well-defined social objectives or because of significant positive externalities, should be protected, provided that they reach their objective in the most efficient manner. But generalized subsidies, provided through artificial reduction in the prices of general consumer goods, should be eliminated. These subsidies often result in large transfers to families in the upper percentiles of the income distribution and little help, perhaps even hurt, those most in need. Worthwhile social objectives can often be achieved far more cheaply and efficiently than is generally the case.

Cosmetic Versus Genuine Fiscal Reforms

A conflict arises in adjustment programs between the need to achieve quick results and the time necessary to develop, legislate, and implement sound policies. The need for quick results is required by (i) the usual precariousness of the economic situation (rising inflation, exhausted foreign exchange reserves, building up of arrears), (ii) the fear that if changes are not made immediately, they will not be made, and (iii) the arrangements with international institutions that are often time constrained.

While changes in interest rates, exchange rates, and many other elements of economic policy can be made relatively quickly and seldom require legislative approval, good fiscal reforms, including tax reform, public sector reorganizations, privatization, and reform of public expenditure programs, require time, much effort, and in many countries legislative approval. As a consequence countries often opt for “quick fixes”—fiscal reforms that reduce the fiscal deficit in the period immediately ahead but are neither durable nor efficient.

Common elements of these quick-fix solutions have been (i) sharp reductions in public sector real wages to levels below their likely long-run equilibrium, (ii) sharp and indiscriminate cuts in investment expenditure without much assurance that the projects eliminated are the least productive,12 (iii) sharp cuts in expenditures on operation and maintenance, leading to a faster deterioration of the existing capital infrastructure and to a much reduced capacity utilization,13 (iv) emergency tax legislation, including the temporary introduction of distortionary taxes, such as those on exports and financial instruments, and of temporary surtaxes on import duties, income taxes, and others, (v) excessive increases in some excises (such as those on petroleum products, beer, and several others), (vi) anticipation of tax payments, sometimes by providing discounts for anticipated payments to taxpayers, thus reducing future tax collection, (vii) tax amnesties, (viii) quick sales of some assets, and (ix) delayed payments to suppliers and other creditors, thus building up arrears.

Most of these measures are self-destructing or of questionable quality, or both. They are not the kind of measures that characterize a good program or that result in durable adjustment. They increase uncertainty and send negative signals to investors, thus discouraging capital repatriation and encouraging capital flight. Therefore, a deficit reduction achieved through these means should not be expected to improve economic conditions.14 Such a reduction can only be justified if it is clearly announced and believed to be a transitory step toward a more durable and higher-quality package. Unfortunately, these measures often exhaust the political will of the government to make more basic reforms or they are seen as the only way to reduce the fiscal deficit. They also exhaust the time of economic and finance ministers and other relevant policymakers.

Sustainable fiscal policy requires, almost by definition, measures that will survive the test of time and that will not impede the efficient allocation of resources. It requires a coordinated and closely monitored effort carried out by competent and determined individuals. It must involve good macroeconomists working in close cooperation with public finance specialists experienced in both policy and administration. The group in charge of this effort must be given a clear mandate, the human and material resources needed to carry it out, and a realistic but short deadline to produce concrete proposals. The group must coordinate closely with key individuals in the legislature in order to explain to them the need for given actions and to coordinate various positions, and it must report routinely to the finance minister and other relevant authorities.

This group, which could be referred to as the “fiscal reform commission,” should make a careful assessment of the magnitude of the fiscal correction needed, an inventory of the possible actions required (quantifying to the extent possible their effects), and a strategy related to the optimal phasing in of the various changes. The commission would recognize that some measures take more time to introduce than others and could thus contemplate the use of temporary measures, provided that their cost in terms of inefficiency and depletion of political capital is not too high and that they are justifiable, transitory, and necessary steps in an overall strategy. The main consideration must be that the general strategy is credible and consistent and that it is clearly and forcefully articulated early in the process. A schedule for the introduction of the measures must be prepared and adhered to as closely as possible. International financial and technical support for a country that sets out to deal seriously with its fiscal problem should be particularly generous.

Although the details will differ from country to country, and only an in-depth analysis can provide precise guidance for a particular country, in broad terms the sequence of important fiscal reform measures might be as follows.

Increasing Revenue

A revenue measure that can be introduced relatively quickly and that can have several beneficial economic effects is an increase in public utility tariffs when these tariffs have been allowed to fall because of inflation or political considerations. It is important, however, to take steps to prevent the benefits from being dissipated in higher salaries, higher employment, or in unnecessary or postponable investments by the public enterprises. If the central government is unable to control the behavior of the public enterprises, their privatization might provide the only solution to their being a drain on the budget. Assuming that public enterprise spending can be controlled, the tariff increase may be larger in the short run than would be necessary if other measures were in place. For a while, the government can exploit the monopoly power of these enterprises to raise needed revenue. The increase in tariffs must not be neutralized by future inflation, and the government must resist the temptation to fight inflation by freezing tariffs since these attempts never work. After the initial adjustment, these tariffs should be adjusted as needed to maintain their real level.15

The revenue potential of particular excises should also be fully exploited. Increases in taxes on tobacco, alcoholic beverages, gasoline, and a few other commodities can be justified on various efficiency grounds (inelasticities, externalities, or benefit-received criteria); they also give quick revenue, and, in modern economies, their equity cost is usually small. But once again the initial increases must be protected against the eroding influence of future inflation. Indexation of specific excises or the use of ad valorem rates would be desirable to ensure that the price elasticity of these taxes is at least unity. Excises levied with specific rates almost always result in inelastic revenue. Thus, until the fundamental reforms are in place, these excises can be made to generate more revenue than might be desired over the long run. On the basis of what is known or presumed about the demand elasticity of these products, taxes that would maximize their revenue generation and minimize their efficiency cost could be imposed. These taxes could eventually be reduced to levels justified by the traditional reasons for imposing excises. Selected “luxury” products could also be taxed (cars, televisions, cameras, and so forth) to generate revenue and introduce equity into the tax system. The taxation of these products with ad valorem rates also raises the elasticity of the tax system.

Some administrative changes for raising tax revenue in the short run can also be considered. For example, in inflationary situations, collection lags can be shortened. When the inflation rate is high, this reduction in lags can raise revenue by significant amounts. Tax evasion can be reduced by the judicious use of penalties. For example, giving the tax administration the faculty to close, for specified periods (one week, one month), shops or enterprises whose owners have not filed returns or have evaded taxes in other ways can go a long way toward encouraging taxpayer compliance, especially if these closures are well advertised. Postponement of tax payments can be discouraged by steeply raising the interest charges on delayed payments to high positive levels. Tax administrations can also focus on the largest taxpayers since in most countries a small proportion of taxpayers generates the overwhelming share of tax revenue.16

While the above changes are being made, the fiscal reform commission should be preparing the more fundamental reforms. With the properly trained people or with the assistance of foreign experts, the required studies can be done in a relatively short time, certainly in a period of months. The search for excessive precision, however, could significantly delay the conclusion of the studies. The longer the delay, the less likely it is that serious reform will be carried out. Long and detailed studies by tax commissions often become excuses for delaying action. Being approximately right is normally good enough, and being approximately right is much faster than attempting to be precisely right.17

The introduction of a value-added tax with the widest possible tax base and a single rate should be a central element of the fundamental reform. If the country does not already have a general sales tax, it will need at least two years to implement a value-added tax. If it already has a general sales tax, the time for reforming it can be cut considerably, though such a change still requires at least a year. Estimations of the size of the potential VAT base will be necessary in order to determine the rate required to raise the desired revenue. Of course, the larger the share of the potential base that is exempt, the higher the rate must be. The argument that many products will need to be exempt for equity reasons should be resisted as it rarely has much rationale. Equally, the arguments in favor of multiple rates regardless of their theoretical validity cannot be allowed to prevail for purely administrative reasons. Many important administrative actions (the design of forms, the organization of the tax office, the use of computers, the determination of the list of taxpayers, the campaign to inform taxpayers, and the development of auditing procedures) require time. Administrative mistakes can be costly, and time can help reduce the number of mistakes. The introduction, or broadening, of the VAT often has beneficial effects on the revenue raided from other taxes, especially from the income taxes.

The reform of income taxes (personal and corporate) also requires time, though much less than the introduction of the VAT. Most developing countries already have income taxes so that it is mainly a matter of modifying the existing structure and improving the administration. Frequently these taxes are applied to tax bases that have become too narrow because of legal decisions to exempt whole sectors (for example, agriculture), regardless of the income of the taxpayers in those sectors. The tax bases also become too narrow because of a proliferation of tax incentives to industries and regions. These incentives rarely achieve their intended objectives but instead create distortions, inefficiencies, and inequities as well as reduce tax revenue.

For the personal income tax, the main policy changes concern (i) a reduction in the often very high level of personal exemptions, (ii) the elimination of many deductions and special treatments of particular income, (iii) a change in the rate structure, probably by raising the base rate to at least 10 percent and reducing the highest rates to the 30-40 percent range, after careful analysis of the income levels to which these rates should apply, (iv) the introduction of withholding in all cases where it is feasible (wages and salaries, interest payments, dividends, and payments to suppliers or others by the government and public enterprises), and (v) the forceful use of presumptive taxation in connection with hard-to-tax activities (farming, professional services, and shopkeeping, for example).18 Some of these changes require more time than others. A serious effort should also be made to improve the administration of these taxes by relying on all the information available to the government (imports, social security payments, and property ownership, to name a few) to reach potential taxpayers and determine their income.

Changes in corporate income taxes can also be introduced relatively quickly, although, as with personal income taxes, the revenue effects may not be felt for two years. The basic changes involve several measures: the elimination of deductions associated with various forms of tax incentives, which can nearly wipe out the tax base; the lowering of the tax rate after the elimination of the incentives; and better administration. The use of minimum taxes can also be contemplated, but they must be very simple to administer. In some recent reforms, these minimum taxes have been based on the imports of corporations, their turnover, and their gross assets. These minimum taxes now exist in many countries.

In developing countries, land and buildings are potentially important tax handles and can contribute needed revenue, especially to local governments. However, if property taxes exist, the assessed values are probably well out of date. One method of generating quick revenues is through proportionally adjusting those values. This adjustment could reflect the change that has occurred in some relevant price index since the last time property values were determined. Adjustments in relative values must wait for a more appropriate time because these reassessments take many years to carry out. The main obstacles to raising property taxes are political and legal rather than technical. When income taxes on properties (land and buildings) are very low or zero, it is important that this tax base contribute to tax revenue through property taxes.

Changes in import taxation can also be made relatively quickly once the legislation is approved. The most important and quickest change is the imposition of a minimum tax of perhaps 20 percent on all imports. For very open economies, such a change can generate substantial revenue in a very short time. This change also improves the efficiency of the economy by reducing the dispersion in effective protection. Changes involving the tariff structure of imports and changes aimed at dismantling quantitative restrictions are also very important but they require more time to implement.

Decreasing Expenditure

In recent adjustment programs, expenditure reductions have tended to be more significant than revenue increases. However, the approach taken in cutting expenditures has not always been efficient.

If public sector wages are high (when compared with those in the private sector), their reduction could be part of an adjustment package. However, wage reductions must not be so excessive as to lead to inefficiency or to expectations that real wages will be raised in the near future. The cuts must not exceed what is justified by wage levels in the private sector, and wage differentials, reflecting different skill requirements and responsibilities, must not be overly compressed.19 Real wages can obviously be reduced quickly and thus can result in rapid expenditure reductions. Over a longer period, however, it is the size of the public sector work force that is the major determinant of public expenditure on wages and salaries. The reduction of public employment is difficult and may be financially costly in the near term, but it must be pursued vigorously if a durable adjustment is to be achieved. Therefore, a concerted effort must be made to reduce the number of public employees, even if this action involves short-run increases in public spending (for severance pay and other compensation).

The second important category of public expenditure that should receive early attention is the investment budget. A close evaluation of all planned and ongoing investment projects must be carried out, with the help of the World Bank or other institutions where indicated. Projects planned for but not yet initiated should be carefully scrutinized regardless of whether they have foreign financing or not. Often the availability of foreign financing has been used as an excuse for carrying out projects of dubious merit. Only if their (realistically estimated) expected rate of return is very high should they be pursued. For the ongoing projects, considerations of sunk costs play a role as do the costs associated with stopping these projects and restarting them at a later date. In some cases this stop-go approach can be very costly. The criterion for maintaining these projects must be the same as that for going ahead with planned projects: those that cannot be justified on the basis of high productivity must be terminated.

The third obvious expenditure area for close scrutiny and quick expenditure reduction is subsidies. Some subsidies may be very important to the poorest groups and must be retained. Others may be important in subsidizing activities that generate significant positive externalities (for example, subsidies to urban mass transport). However, in many countries, much of the expenditure on subsidies meets neither of these criteria. When subsidies are not targeted for the poor but are general in nature, especially when they encourage the consumption of a traded good or support inefficient enterprises, they must be cut. If adjustment means anything, it must mean that the state cannot continue to subsidize the consumption habits of the middle classes or the production activities of inefficient enterprises.

Obviously, as argued earlier, the military budget must not escape close scrutiny. Unfortunately, expenditure on these programs has proven very resilient during adjustment programs. Nonetheless, the fiscal reform commission should put downward pressure on this spending by requiring a full accounting of the military budget and by analyzing the opportunity cost of using these resources for defense purposes rather than for economic or social ones.

As a longer-run objective, the commission should reassess public sector involvement in the economy. In most countries many activities have been taken over by the government, perhaps for reasons that appeared legitimate at the time. A period of adjustment is a good time to make an inventory of these activities and question anew their justification. Vested interests and pressure groups will argue for the status quo, but a determined government with a clear sense of direction can show the absurdities of many regulations and governmental activities and can make a concerted effort to eliminate many of them.

Concluding Remarks

This paper has surveyed some major issues that arise when trying to control the public finances of a country during an adjustment program. It has emphasized the much greater complexities and difficulties that arise in the fiscal area as compared with other areas of economic policy. Most economists do not fully appreciate that good fiscal policies are difficult to develop, legislate, and implement. For this reason easy-to-introduce, quick-fix policies have been used not as necessary transitory steps toward a more durable and efficient policy package but as the ends in themselves. These policies, however, are at most a palliative. They are unlikely to provide a sustainable solution to the fiscal problem.

In recent years there has been a lively discussion among economists and policymakers on whether countries should opt for shock therapy or for gradualism in their adjustment programs. This is not a meaningful debate, however, if applied to fiscal policy because good fiscal measures always require time. Nonetheless, two observations are important. First, good macroeconomic policies in other areas often improve the fiscal accounts, especially if they reduce inflation and bring about realistic exchange rates.20 Second, gradualism can be interpreted in two ways. One is that gradualism delays decisions about and the implementation of important policies. The other is that the basic policy decisions in a gradualist approach are made early, but the implementation follows a sequencing pattern that allows good policies to be developed, legislated, and implemented. If a clear strategy is developed early, the introduction of the fiscal measures can follow a pattern as close to optimal as possible. Given the importance of expectations for economic performance, even the elaboration and announcement of a sustained long-run policy path can be a first step toward stabilization.

A final comment concerns the legal and constitutional impediments that exist in many countries. These can create formidable difficulties during fiscal reform. For example, in some countries arrangements regarding revenue sharing with local governments, or even with public enterprises and other institutions through earmarking, can impede reform if no major legislative changes are made. Special treaty arrangements with some trading partners can impede reforms in import taxation (an example being the introduction of a minimum tax on imports). Tax incentives provided to some enterprises may impede the reform of corporate income taxes or even other taxes. Constitutional limitations on the use of some taxes can prevent their introduction.21 Various forms of legal entitlements reduce flexibility on the expenditure side as well. Without the removal of these impediments, sound fiscal reform is difficult. The removal of these impediments will ensure that the fiscal adjustment is spread throughout all sectors of the economy and all levels of government rather than concentrated in those areas over which the central government has some control.


Azizali F. Mohammed

This paper by Vito Tanzi is a valuable distillation of countries’ experiences with the complexities of fiscal adjustment. He has correctly identified fiscal reforms as the most difficult of the various policy changes required in adjustment programs. Indeed, the difficulties that he enumerates are so many and the need for “quick fix” solutions to immediate problems so pressing that one is actually at something of a loss to explain successful instances of “high-quality” fiscal adjustment. Tanzi is evidently aware of this dilemma for he consoles us, “Being approximately right is normally good enough, and being approximately right is much faster than attempting to be precisely right.”

While recognizing the limits of quantification, one must confess to disappointment that Tanzi does not venture beyond qualitative statements to define the size of the required fiscal adjustment and the proportion of it to be completed during the period typically covered under an IMF or World Bank program. Tanzi speaks of a fiscal approach to the balance of payments that suggests limiting the fiscal deficit to a level that can be financed internally in a noninflationary manner and externally through “available, ordinary credit.” The internal aspect is linked to liquidity expansion through a change in the domestic credit extended to the government. Tanzi notes, however, that even that part of the deficit that is not being financed by monetary expansion is a cause for concern if it leads to an increase in the ratio of public debt to GDP, if it crowds out the private sector, or if it skews portfolio distributions in the financial sector. Tanzi would perform a valuable service by developing a modeling framework that delimited the boundaries of a sustainable fiscal deficit.

This exercise would entail, as a first step, an assessment of that portion of public expenditures which represents domestic, but not foreign exchange, outlays as well as that portion of public revenues which subtracts from the domestic income stream. A distinction between domestic and external financing is implicit in Tanzi’s thinking since, as noted above, he does refer to that portion of the deficit that is financed externally. However, his reference to “available, ordinary credit” is not clear. How, for instance, would he treat a deficit that is financed by recurring foreign grants? Moreover, the portion that is financed by external credits raises different issues of debt-servicing liability and real resource transfers. True, there are connections between external and internal financing, such as the need to find supporting domestic funds for foreign-aided projects, the treatment of counterpart funds arising from the sale proceeds of nonproject aid, and the preempting implications for domestic resources in “utilizing” foreign aid. However, these connections do not quite justify lumping domestically and externally financed deficits together into a single figure of the permissible deficit and fixing targets for fiscal reduction as if these two forms of financing were economically equivalent in their implications or consequences.

Tanzi has provided a useful catechism of “dos” and “don’ts” in dealing with both revenue and expenditure aspects of correcting fiscal imbalances. He has little to say, however, on issues of tax administration and expenditure control mechanisms, which often lie at the heart of implementation. To design a package of high-quality measures that meet efficiency, durability, and equity criteria would be of little help if they are not implementable in the political, constitutional, or administrative circumstances of a given country.

This last point is particularly relevant to conditions prevailing in Pakistan where tax administrators are compensated at levels far below the higher-priced employees available to the private sector (who are paid to resist the legitimate claims of the tax authorities). One can only agree with Tanzi when he argues that over a longer period it is the size of the public sector work force that is the major determinant of public expenditures on wages and salaries. There is a need to protect real wages in the public sector if corruption is to be constrained and an even greater need to raise real wages for those dealing with tax and expenditure matters in the public sector.

This does not mean, however, that public enterprise expenditures on wages and salaries must necessarily be validated by corresponding increases in the prices of the services these enterprises provide. Tanzi is well aware that increases in public utility tariffs can be dissipated in higher salaries, higher employment, or lower-priority investments; if such spending cannot be controlled, he would rather see these enterprises privatized. But since most public utilities are natural monopolies, this immediately raises the issue of effective regulation. It is a moot point whether a government that is unable to control the wage-setting behavior of its public enterprises will be any more effective in regulating the behavior of privatized monopolies.


The author would like to thank Ehtisham Ahmad and George A. Mackenzie for their comments. The views expressed here are strictly personal. They do not necessarily reflect official positions of the International Monetary Fund.


The basic accounting identity in a financial programming exercise is ΔM = ΔR + ΔD. That is, the change in the money stock (AM) is the sum of the change in the domestic currency value of international reserves (AR) and the change in domestic credit (AD). The change in domestic credit, in turn, can be decomposed into credit to the government and credit to the private sector. Credit to the government is a reflection of the fiscal deficit. The higher the fiscal deficit, the larger is the credit to the government, and thus the higher is the increase in liquidity. By restricting credit to the private sector, overall liquidity creation can be reduced. However, the cost of this action is crowding out of the private sector.


For discussions of fiscal deficit measurement issues, see Tanzi, Blejer, and Teijeiro (1987) and Blejer and Cheasty (1991).


In other words, during inflation nominal interest payments come to contain an element of amortization so that the government is actually repaying its debt at a faster pace. See Tanzi, Blejer, and Teijeiro (1987).


This has led some economists to conclude that devaluation always increases the size of the fiscal deficit. See Tanzi (1989) for a different view.


Even changes in one tax, say the personal income tax, involve many decisions. For example, reform of the personal income tax requires decisions about the level of the personal exemption, the extent of deductions, the treatment of the family, the special treatment of particular incomes, the levels of the various rates, the incomes at which they become effective, the form of payments, the desirability of inflation adjustments, and so forth. The official description of the 1986 income tax reform in the United States ran into the thousands of pages.


The value-added tax is now in existence or in the process of being introduced in almost 60 countries. The latest country to introduce such a tax with a uniform rate of 10 percent is Thailand, where the tax was enacted in October 1991.


The replacement of quotas and other quantitative restrictions with tariffs shifts the power to tax from those who benefited from these restrictions to the government.


Of course, the determination of what is an unproductive investment may be difficult. Nonetheless, a close technical scrutiny of the investment budget is always useful. Decisions should relate to specific investments not to the total investment budget.


The trade-off between a better utilization of the existing capital structure and an addition to that structure has not received the attention that it deserves.


There are other costs associated with a reduction of real wages below some efficiency level. These costs are not discussed in this paper.


In some cases, projects may not be eliminated but their completion may be stretched out. This process can lead to sharp increases in the final discounted cost of projects. In some cases the politically protected investment may survive even though its productivity may be lower than that of a less protected project.


Because of drastically reduced expenditure for maintenance, the capacity utilization of some plants and infrastructure is sometimes reduced to a small proportion of its potential.


If the improvement in the fiscal situation is seen as temporary, investors will not want to sink their financial wealth into fixed assets that may lose value as soon as the situation worsens again. There should be no surprise, then, when growth does not quickly follow this kind of adjustment.


When normalcy returns and fundamental fiscal reforms have been put into place, the tariffs can return to levels justified by purely allocative (rather than revenue-generating) considerations.


A common conflict for tax administrators is whether to concentrate on the largest taxpayers or to administer as large a number of taxpayers as possible. When the revenue objective becomes important, it is more efficient to focus on the largest taxpayers (both enterprises and individuals). Of course, this may conflict with the equity objective of taxation. Thus, the group in the spotlight should progressively become larger as administrative capacity improves.


Very few tax commissions have produced studies that resulted in major tax reforms. The reason is the time that these commissions require to fulfil their mandate.


Ahmad and Stern (1991) have strongly recommended the use of presumptive taxes on agricultural incomes in Pakistan. For a discussion of these taxes, also see Tanzi and Casanegra de Jantscher (1989).


In some countries, this compression has been so extreme that government ministers and their drivers have received comparable cash salaries.


See Tanzi (1989).


For example, in India the central government cannot introduce a VAT that extends beyond the manufacturing sector. In Pakistan, the constitution excludes agricultural income from the federal income tax income.

  • Ahmad, Ehtisham, and Stern, Nicholas H. The Theory and Practice of Tax Reform in Developing Countries (Cambridge: Cambridge University Press, 1991).

    • Search Google Scholar
    • Export Citation
  • Blejer, Mario I., and Cheasty, Adrienne “Analytical and Methodological Issues in the Measurement of Fiscal Deficits,” Journal of Economic Literature, Vol. 2 (December 1991).

    • Search Google Scholar
    • Export Citation
  • Blejer, Mario I., and Cheasty, Adrienne “The Impact of Macroeconomic Policies on the Level of Taxation and the Fiscal Balance in Developing Countries,” Staff Papers, Vol. 36 (September 1989).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, Blejer, Mario I. and Teijeiro, Mario O. “Inflation and the Measurement of Fiscal Deficits,” Staff Papers, Vol. 34 (December 1987).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, and Jantscher, Milka Casanegra de “The Use of Presumptive Income in Modern Tax Systems,” in Changes in Revenue Structure edited by Chiancone Aldo and Messere, Kenneth proceedings of the 42nd Congress of the International Institute of Public Finance (Detroit: Wayne State University Press, 1989).

    • Search Google Scholar
    • Export Citation