Growth-Oriented Adjustment Programs: Fiscal Policy Issues

Mohsin Khan, Morris Goldstein, and Vittorio Corbo
Published Date:
September 1987
  • ShareShare
Show Summary Details
R. J. Chelliah

If the external payments problem facing a country is considered a short-run phenomenon, attributable, say, to a temporary fall in agricultural output or generation of excess demand through the budget, the remedy might be sought in a stabilization program which emphasizes contraction of demand or reduction in “absorption.” In the short run, productive capacity is taken to be fixed and while the degree of utilization of the existing capacity could be increased to an extent, the demand side may be said to offer greater scope for manipulation. Since the 1973 oil shock, however, several events have aggravated the balance of payments problems of developing countries, and it has become more and more clear that the restoration of equilibrium would require longer term measures leading to structural changes. In the context of fundamental disequilibrium owing to structural deficiencies, a large reduction in absorption, while often imposing an unacceptably severe cut in real income or welfare, would leave the fundamental causes of low growth, underutilization of capacity and the tendency for balance of payments deficits unattended. It is now recognized more clearly—by the Bretton Woods Institutions as well as by the member countries—that a more comprehensive set of policy prescriptions are called for, including supply-side measures, and that a longer period of adjustment is often needed during which the inflow of external capital is to be maintained at the required minimum level. If the criterion of growth is explicitly brought into the analysis, there is less danger that whatever demand-contracting measures are recommended would be such as to slow economic growth.

In the simple financial programing framework employed by the Fund, say, during the 1960s, fiscal policy played the role of restricting government demand for credit. In other words, fiscal policy was considered an aspect of monetary policy; by itself, it was supposed to have no independent effects on aggregate demand and balance of payments. Such a view was taken partly as a result of the concentration on the size of the unweighted budget deficit to the neglect of the composition of revenues and expenditures and any favorable supply-side effects that could be expected from fiscal policy changes would require a fairly long gestation period. In the context of medium-term programing such as that under the extended fund facility or the structural adjustment lending program, fiscal policy could be effectively used not merely for demand restraint but also for effecting structural changes. This paper discusses the role that fiscal measures could play in a medium-term adjustment program.

Before the role of fiscal policy is discussed, it is necessary to specify the basic objectives that a country has set for itself and the institutional structure that it has adopted. I shall assume the institutional structure of a mixed economy in which the public sector is to play an important, if not a dominant, role. This role will be played through the operation of public enterprises in certain core areas thereby determining the broad directions of development. At the same time, public ownership would act as a counterforce to concentration of wealth in private hands. Such an institutional structure has been consciously adopted in several developing countries to promote growth along with social justice. Growth with social justice also means that while growth is emphasized, the government should strive to change the composition of output so as to benefit the poorer sections of society and to ensure for all the minimum of essential goods and services. While in a mixed economy a very large part of the economy would be operated by private enterprises guided by market forces (which could, however, be influenced by indirect means of control, such as taxes), a part of the economy would be planned on the basis of needs. This would comprise the core sector as well as those segments that produce essential consumer goods. The government must take steps, insofar as it is possible, to ensure that sufficient resources will go into the production of these essential goods and they are within the reach of the poorer sections of the population.

It is also widely recognized that certain services, such as primary education and rudimentary health facilities, should be provided free of charge or at subsidized rates. It may also be necessary to provide a certain minimum of food grains and cheap cloth at subsidized prices to people below the poverty line. The core objectives of Fund-supported adjustment programs are the achievement of external and internal balance and of adequate economic growth. I would add the qualification “with due attention being paid to the aspect of social justice.”

The Bank works out its program and the magnitudes of the variables needed for it in a medium-term framework. It appears that a variant of the two-gap model is generally used for this purpose. The rate of growth in the medium term is naturally brought in. The theoretical framework employed by the Fund as underpinning for its financial programing exercise, however, remains basically short term in character but can be recast in a medium-term framework. For example, instead of using the expected short-term increase in real output, the maximum feasible increase in output to be targeted in the medium term could be used. The desired rates of growth of government revenues and expenditures could only be worked out against the background of the assumed or targeted rate of growth of the economy.

In the context of the medium term, fiscal policy can come into its own and play an independent role, particularly on the supply side, while in the short run, the impact of fiscal policy is transmitted mainly through the budget deficit. This occurs because significant changes in the composition of revenues and expenditures and in the structures of different taxes cannot be brought about instantaneously or in the short run, and the effects of such changes take time to work themselves out.

The objective of a medium-term adjustment program is to create conditions, through policy measures, which would enable the country to attain an adequate rate of economic growth with social justice, internal price stability (or a fairly low rate of inflation), and a sustainable external payment position. Several measures are called for to achieve these objectives. I confine myself to fiscal policy measures that could be used to help regulate absorption, increase domestic savings and investment, promote the more efficient use of resources, and channel an adequate share of the benefits of growth to the poorer sections of society.

Fiscal Deficit

In many developing countries suffering from balance of payments disequilibrium, the major cause of the problem is often excess demand emanating from the public sector. In these circumstances, a reduction in the budget or fiscal deficit is made part of the adjustment program. Once the size of the budget deficit is brought down to the “correct” level, one needs to plan the subsequent increase in its size in relation to the growth of real output and other relevant factors.

In this context, it is necessary to consider carefully the definition of budget deficit. The Fund and the World Bank make use of the concept of fiscal deficit, which can be taken to correspond to the “overall deficit” used in public finance literature. The overall deficit is defined as the difference between government’s total expenditures1 and current revenues; thus it becomes equal to total net borrowing by the government that finances the deficit.

Limitations to the use of the fiscal deficit as an indicator of stimulus to aggregate demand are well known. The use of the unweighted budget deficit ignores the differential impact of different types of revenues and expenditures; the impact of the deficit depends also on the source of financing it; and insofar as tax revenues are endogenous, the size of the deficit is not fully under the control of the government. In spite of these limitations, the size of the deficit, properly defined, could be taken as a rough indicator of the impact on demand, even though it cannot be wholly attributed to government action. Besides, the budget deficit is an important variable from other points of view, such as the growth of public debt and of interest burden on the budget.

The figure of the government’s net borrowing is of significance because it reflects the absolute increase in public debt. Also, to the extent the government is borrowing from abroad, the external liabilities of the country are increased, and domestic borrowing by the government would have implications for the availability of credit to the private sector. Obviously the government’s borrowing program must be carefully regulated. Hence, the figure of overall fiscal deficit is of significance, particularly if it is taken net of net lending. However, in countries in which the government owns and runs important public enterprises, such as electricity undertakings, railways, and telecommunications, whether government borrowing undertaken on behalf of such enterprises and invested in them is to be excluded from the measurement of overall deficit (because that part of the credit is strictly not for government purposes) must be considered. In fact, one of the reforms suggested for countries in which the government borrows on behalf of public enterprises is that such borrowing be transferred to an independent financial agency that can be enjoined to act on commercial principles.

As pointed out earlier, the impact of the budget deficit on aggregate demand depends on the source of financing. It could be legitimately argued that government expenditure financed by borrowing from the non-bank private sector will not result in any significant net expansion because the diversion of private savings will result in the reduction of purchasing power in private hands. In this view, the amount of overall deficit cannot be used as a correct indicator of the expansionary impulse emanating from the budget. In India, a prominent committee has recently suggested that the total net Central Bank credit to the government sector (which leads to an increase in high-powered money) should be taken as the proper measure of budget deficit for measuring the expansionary impact of the budget.2 Others would argue that total bank credit to the government should be used instead. In any case, the overall deficit generally used by the Fund is not a satisfactory indicator of the expansionary impact of the budget.

In standard Fund programs, the performance criteria include ceilings on total bank credit and often sub-ceilings on net bank credit to the government. One sometimes hears statements about the need to eliminate the fiscal deficit, and the figure of fiscal deficit is presented as a percentage of GDP to show how large it is. Often, this gives an unduly alarmist picture because, for one thing, it includes government borrowing to finance government investment in public enterprises. (If government lends to public enterprises, such borrowing would get excluded if net lending is excluded from total expenditure.) Although, as I said earlier, it is useful to see how public debt is growing, it is necessary to look at the magnitudes of borrowing from different sources for an analysis of the impact of the budget. Just as total bank credit is allocated between the government and the private sector when a sub-ceiling on net bank credit is prescribed, so also a proper allocation of the financial savings of the household sector flowing through non-bank channels should be worked out.

Public borrowing can be looked at from the point of view of demand impact and of growth. For measuring impact on aggregate demand, one should take net central bank credit and net bank credit to the government. (For measuring the budget impact on demand for domestic output, one should consider the excess of domestic budgetary expenditures over all domestic receipts other than borrowing from the banking system.) For considering the impact on growth, one should take the balance on current budget account that reflects government sector’s contribution to domestic savings. Government sector saving can be taken to equal government current revenues minus current expenditures.3 To the extent that current revenues exceed current expenditures, the government is saving a part of the purchasing power it is transferring from private hands through taxes and other means. If there is a current account deficit, it would mean that the government is using a part of the savings of the private sector to finance a portion of its current expenditures.4 Where savings are falling short of needs, as in most developing countries, government’s dissaving affects growth adversely. Hence, I would urge that in any adjustment program a stipulation that the current account deficit of the government would be speedily eliminated.

In terms of Keynesian fiscal policy prescription, it would not matter if the proceeds of government borrowing are spent on capital formation or on public consumption. If there is widespread unemployment and underutilization of capacity owing to a deficiency of effective demand, deficit financed spending by government would, so it is argued, lead to significant growth in output and employment, thereby increasing government revenues. So long as such increases continue, the interest burden in terms of ratios to government revenue and GDP may not rise. But the Keynesian type of underemployment resulting from deficiency of demand is not typical in most developing countries. Deficit-financed public consumption will not lead directly to any significant increase in real output through the operation of the multiplier.

When one argues that current expenditures should be met through current revenues, one is not saying that current expenditures are unimportant. In the longer term, in which the promotion of the growth of the economy is a major goal, the population has to limit its total current consumption—private and public consumption taken together. It is not that more public consumption may not be desirable, but if it is, private consumption should be correspondingly curtailed and for this higher current revenues should be raised.

In brief, in my view, where public finances have gone off track and the government is incurring deficit both on its current account and on the total of current and capital accounts, as much attention should be given to eliminating the current account deficit as to bringing down government borrowing from the banking system.

The question arises as to how the deficits should be brought down. Obviously, circumstances differ from country to country and the same measures cannot be applied everywhere; however, a few generalizations could be made. To begin with, a distinction has to be made between short-run and medium- or long-term measures. Taking expenditures first, in the longer term, it should be possible to regulate the growth of particular items of expenditure in relation to GDP making the ratio of those items to GDP remain constant or fall; in the short run, the actual level of expenditures would have to be brought down. This would be difficult because, in the short run, several items of expenditures are likely to suffer from rigidity, for example, total employment in the government sector. However, there is some room for maneuver even in the short run. First, expenditure increases could be prevented in regard to government sector employment through a short-term freeze on recruitment and on wages for government employees. Second, action could be taken to reduce subsidies other than those on exports and on essential goods for the poorer sections of society. Third, the introduction of zero-base budgeting in all areas other than key developmental areas would enable savings in current expenditure and re-deployment of staff from nondevelopmental to developmental departments. Fourth, there could be some pruning of capital expenditure of lower priority.

As regards revenues, tax rates are high in a number of developing countries (the tax ratios ranging between 25 and 30 percent of GDP). In these cases, it will be difficult to increase revenues further in the short run. There are other cases where tax ratios are much lower. Short-term increases in revenues would be more possible in those countries, although even there the low level of revenues in relation to GDP may not be attributable to low rates but to an inadequate structure and loose administration.

The problem about raising tax rates in the short run is that higher direct tax rates might induce further evasion and might not result in higher revenue, because rates are already high in many countries; on the other hand, increasing the rates of indirect taxes also might not give any significant addition to revenue in real terms, if there is indexation of wages and salaries and if many prices are administered on a cost plus basis. In fact, often the addition to revenue from rate increases is overestimated.

Net Marginal Product of Additional Taxation

In traditional theory, an increase in tax revenue is taken to be deflationary, that is, it is considered to reduce private demand. In this connection, no distinction is made between increase in the yield owing to a rise in tax rates (including imposition of a new tax) and that owing to an increase in the base because of either improvement in administration or growth. An increase in the rate of a tax (or in the rates of taxes) will normally produce a direct effect in the form of an increase in nominal revenue.5 But how much of net increase in real revenue there will be depends on several factors.

In many countries, several prices are administered on a cost plus basis. Also, wages in the organized sector of industry and government are indexed to inflation. Hence, an increase in the rates of indirect taxes, which increases prices of inputs or directly affects consumer goods prices, leads to a further round of increases in prices affecting several goods as well as wages. To maintain government expenditure on goods and wages in real terms, government expenditure in money terms will have to rise. This means that the real value of each dollar of revenue falls. This must be set off against the original increase in tax revenue owing to the hike in tax rates. But at the same time, the secondary rise in prices may bring in some extra revenue. Even allowing for that, taking the original increase in revenue would generally overestimate the net accretion to the budget.

There is another reason why the marginal product of additional taxation may be less than the yield of the additional tax measures. Where several bases are subject to tax, and these bases are interrelated, an extra tax on one base has the effect of reducing the size of other tax bases. Thus, when a number of commodities is subject to tax, an increase in the rate of tax on some commodities subject to inelastic demand would divert purchasing power away from commodities with more elastic demand and would thus cause a fall in the yield of taxes on the latter group of commodities. Similarly, a rise in income tax yield owing to an increase in income tax rates would reduce private consumption and saving. Decrease in private consumption would effect a corresponding reduction in revenue from taxes on consumer goods. And again, taxes that raise prices of inputs would have the effect of reducing the profits of public enterprises. For example, an increase in the tax on diesel, ceteris paribus, reduces the profits of public road transport corporations.

So, one can conclude that the marginal product of additional taxation will generally be less than the nominal yield of the measures introduced and further that the net contribution of such measures to the reduction of budget deficit would be even lower than the marginal product in the case of increases in indirect taxes.

An increase in direct tax rates has a greater chance of bringing about a net reduction in private demand than an equivalent increase in indirect tax rates. Direct taxes, however, generally account for a relatively small proportion of total tax revenue in developing countries. Hence, in the short run, particularly when there are inflationary pressures, reduction in government expenditure would be a more effective way of bringing down the budget deficit and aggregate demand. In the longer run, however, tax revenues can be made to increase faster through a rationalization of the tax structure and improvement in administration. Tax increases that arise from growth and improved administration do not lead to price increases or to any significant disincentives.

If some increase in revenues in the short run becomes imperative in order to reduce the budget deficit, the best way of doing so would perhaps be to levy a uniform surcharge on all taxes other than the corporate tax. That way, the basic structure will not get altered in a hasty manner, and the extent of rate increase will be the minimum needed since all taxes but one will be covered by the increase.

Long-Term Fiscal Policy

As generally agreed, in the longer-term perspective, the fiscal system could be employed for producing the desired supply-side effects. The most important of these is the increase in the efficiency of resource use and in the rate of growth of productive capacity. For achieving the first, there should only be the minimum of undesired distortions in relative prices of factors of production and of producer choices; for achieving the second, savings and investment must be increased, and investment should be redirected to more productive channels. At the same time, the fiscal system must mobilize sufficient resources to be used for development purposes.

Revenue Side

Taxes form an overwhelming proportion of current revenues in most developing countries and are capable of being altered to suit particular objectives. Therefore, only tax revenues will be considered here.

In terms of the basic objective of a medium-term adjustment program, tax revenues must rise substantially to bring the budget deficits, appropriately defined, under control and, at the same time, the tax structure must exert a favorable influence on the economy. This favorable influence leading to higher growth would, in turn, be reflected in higher revenues. Thus, the level of revenues and the structure of the tax system are both important. The design of the structure would be partly influenced also by administrative considerations.

In most developing countries, the pattern of taxation has been evolved mainly on the criterion of ease of securing revenues rather than on the basis of well-accepted economic principles. Moreover, there has been too much imitation of western models without sufficient adaptation to the different conditions and requirements of developing countries. Often the structures are at variance with the capacity to administer, and the effects are contrary to what is desirable. Take, for example, two countries in many respects very different from each other but both members of the Commonwealth, namely, India and Zimbabwe. Both had developed their personal income tax and the corporate profits tax in imitation of models in the advanced industrial nations. The high progressivity of the personal income tax, the high rates of the corporate profits tax, and the complicated legislation needed to plug loopholes for avoidance induced by high rates—all these were unsuited to the requirements of the two economies as well as to their capacity for administration. The high nominal rates did not lead, at least in India, to progressivity because of substantial tax evasion.

The governments of developing countries depend predominantly on indirect taxes for obtaining their revenues. It is important therefore that indirect taxes levied in these countries do not lead to distortions in the relative prices of factors of production or to any significant degree of cascading. For economic reasons as well as for ease of administration, there must be a general tax with only one or two rates. This could be supplemented by a selective consumption tax levied at higher rates so that the indirect tax structure, taken as a whole, would be progressive with respect to consumption expenditure. (The selective tax should also be applied to petroleum products.) In countries that have built up sufficient administrative capacity the general tax could take the form of value added tax, but it would not be prudent to make it very comprehensive on the European pattern. Most services, for example, would have to be excluded. (A separate tax could be levied on entertainment.) In smaller countries, it might be preferable to have a retail sales tax instead of a value-added tax. Where a retail sales tax is found to lead to a great deal of evasion, the sales tax could be introduced at manufacturing and import levels with provision for setoff for tax paid on inputs by manufacturers. The sales tax and the consumption tax would apply to domestically produced goods as well as to imports. Imports may be subject, in addition, to protective tariff.

In suggesting the adoption of a noncascading general sales tax, I am saying nothing new. What I wish to stress, however, is that the available administrative capacity should decide which of the three kinds of sales taxes I have mentioned should be imposed, in place of whatever exists. In many cases, a sales tax with relief for taxation of inputs might be the best proposition.

If it is desired that the indirect tax system should be biased toward investment, it could be stipulated that the tax paid on machinery would also be eligible for setoff. In most developing countries, however, it might be considered desirable not to encourage capital-intensive methods of production. If so, setoff for tax paid on machinery may not be allowed, but capital goods themselves may be taxed at the lower of the two rates.

An attempt can be made to promote savings and investment through direct taxes that fall on corporations and on the richer sections of the community. Ideally, given the limitations of an income tax, which contains a bias against savings and which can only be levied on a realization basis, direct taxation should take the form of a tax on personal expenditure supplemented by some variant of a cash flow tax on corporations. Nevertheless, neither a universal expenditure tax nor a two-tier expenditure tax (as defined by the Meade Committee)6 can be administered satisfactorily in a typical developing country. My own study of the way in which income taxes are administered in many developing countries leads me to conclude that these countries are not equipped to administer even a progressive personal income tax. I would suggest that many of them opt for a single-rate income tax—say, levied at 30 percent—which would be mildly progressive because of the exemption limit. Tax theorists would concede that a single-rate or nominally proportional income tax would enable one to get rid of several problems encountered in the administration of a progressive income tax.

A single-rate income tax would not only enable a broadening of the base and a more equitable treatment of income from different sources but also make it easier to check avoidance and considerably ease the task of administration. With a moderate rate of tax, several exemptions and concessions could be done away with; this would broaden the tax base as well as simplify administration. Second, it could be specified that while the individual would be the unit of assessment, there would be a fixed amount of tax-free allowance for each family consisting of husband, wife, and minor children. This amount would have to be divided among the taxpayers in the family. The same amount of tax-free allowance would be granted to partnerships (one individual would be allowed to claim only one allowance either as a partner or as an individual in his own right). Such provisions would effectively prevent any attempts to split incomes that may be made even after a single rate of tax has been introduced. Third, the taxation of capital gains would be much simplified and could be made more equitable: after adequate indexation, capital gains can be included as part of income and there would be no need to make any distinction between short-term and long-term gains. Fourth, the profits of corporations could be taxed at the same rate as personal income; with the exemption of dividends paid to residential holders, there would be proper integration of taxation of individuals and corporations. Fifth, a progressive income tax can only make sense if it can be ensured that full aggregation of income from all sources is reflected in the returns submitted. Unfortunately, it has not been possible to ensure this to a satisfactory extent in most developing countries, with the result that there is higher taxation of income subject to withholding tax than of other types of income. Tax administration in developing countries is also not in a position to check and verify numerous returns. Under the single-rate income tax, individuals whose incomes are from salaries, interest, or rent (dividends may be exempted from tax in the hands of residents) need not be asked to file returns. This would enable tax administration to concentrate on business and professionals whose proclivity for tax evasion could be expected to be reduced considerably because of the moderate single rate.

I recognize, however, that for political reasons it may not be possible to switch over to the single-rate personal income tax. It could also be argued by some that the mere fact that the rate of income tax is moderated would not ensure that savings would be encouraged. They would prefer higher rates of tax that could be moderated through exemptions for savings or particular forms of financial investments. If a progressive income tax is to be retained, it is highly desirable that the rates should be moderate and that the number of rates should not exceed three or four. In this respect, the present Indian system can be taken as a model. The Indian personal income tax falls on each individual taxpayer separately and contains no allowances other than the personal allowance. There are only four rates with the maximum rate being 50 percent. As in the Indian case, introduction of various types of allowances and abetments should be avoided so as not to complicate the administration and to erode the base. If the progressive income tax is retained, it would be necessary to provide for incentives for savings under the tax. Income tax laws in most countries already contain such incentive provisions. But most of these provisions are not such as to reward net savings and often result in widely differing after-tax rates of return to different forms of investment because of variations in the degree of tax concessions. They need to be modified and rationalized. Under the rationalized scheme, deductions from taxable income should be granted in respect of investments in specified accounts up to a monetary limit, while withdrawals from the accounts should be added to taxable income. Funds in these accounts would get invested in the public and private sectors. The eligible investments could include insurance premia and provident/pension fund contributions.

As far as business taxation is concerned, it would be necessary to provide for accelerated depreciation so as to encourage and facilitiate capital formation. The tax on corporate profits is also in need of radical reform in several developing countries. There is no justification or rationale for a high tax on corporate profits except in the case of multinationals engaged in the exploitation of natural resources, such as minerals. I tend to believe that high rates on corporate profits have been introduced in developing countries because such rates exist in a number of developed countries. Even if a progressive income tax is retained for equity reasons, the corporate income tax rate must be kept low for companies in general. A fairly low rate of tax on corporations would induce greater risk taking and investment and leave enough resources for financing the increased investment. Along with the lowering of the rate of corporation tax, many of the exemptions eroding the tax base should be removed.

It has been reported that “changes in direct taxation have not played as important a role in Fund-supported programs as revisions of indirect taxation.”7 Out of the 94 programs surveyed by the Fiscal Affairs Department of the Fund, direct taxes were proposed to be changed in only 44 and the personal income tax in only 20 programs. It also appears from the report that major revisions of the current tax system were not contemplated. I would urge that in the context of a medium-term adjustment program such as the extended Fund facility or the structural adjustment lending program of the Bank, direct tax reform should be given high priority.

The simplification of the tax structure and the introduction of moderate rates would not only strengthen incentives but also make the tax system more administrable. As things are, the quality of the tax administrators and the logistic and statistical support they receive are far from adequate to administer satisfactorily the complicated systems which have been set up in the developing countries partly on the advice of visiting experts from the developed countries. Improving tax administration will take time but is indispensable for achieving any degree of success through tax reform. Tax structure simplification and reform and improvement in tax administration are mutually reinforcing, and one will not succeed without the other. A Country Policy Department discussion paper of the World Bank points out that work on country adjustment programs has indicated considerable scope for improving tax administration even in such middle-income countries as Argentina, Brazil, Chile, Colombia, and Mexico.8 The need for improving, rather than entirely revamping, tax administration is far greater in several countries of Africa and Asia.

It is heartening to learn that “there is an emerging consensus (including apparently the Fund and the Bank) that lower marginal tax rates are desirable as a way to preserve the incentives for work and savings and as a pre-condition for further changes that will improve equity among taxpayers.”9 It may be useful to point out how the kind of tax system that I have outlined would increase savings and investment and promote the growth of the economy. I have no doubt that the income elasticity of the tax system would go up considerably after the simplification of the tax structure and with improvement in tax administration. At the same time, the low rate of tax on corporate profits (with the exceptions mentioned above) and accelerated depreciation would give a fillip to corporate savings and investment. The single-rate personal income tax would enable much better collection and also increase compliance. With the reduction in the proportion of black income generation, there would be a considerable increase in savings brought into official accounts. Apart from the fact of the low rate itself, the elimination of double taxation of dividends and the removal of the bias in favor of capital gains in the tax system would also serve to promote savings and investment in the right direction. If a single-rate income tax is not acceptable, specific provision for rewarding savings must be introduced into the income tax. Indian experience shows that such provisions can help mobilize savings and draw them into the formal financial sector.

As regards progressivity, I doubt that the overall system recommended here would be any less progressive than the existing systems in their actual operation in most developing countries. If the reforms could succeed in taking far more away from the income taxpayers and corporations than they are paying now, the capacity of the government to use money taken from the richer sections for the benefit of the poorer sections would increase. But it must be admitted that the lack of progressivity in the income tax or the blunting of the degree of progression in it, while enabling greater savings and accumulation by individuals, would accentuate inequalities of wealth. In order to counteract this, an inheritance tax with moderate rates could be introduced, if it is not already being levied. Here again, I would prefer a single-rate tax. Another instrument that may be used to tax the well-to-do and raise revenue, particularly for local authorities, is the property tax. In many developing countries, the structure of this tax has to be rationalized and its administration vastly improved.

In the medium-term adjustment program, revenue increase has to play an important part in bringing down the budget deficit, in particular the current account deficit, and in enabling the government to finance essential services on an increasing scale. If tax ratios are already fairly high, however, it would not be desirable to attempt to raise them any further. In a number of developing countries, the informal and non-monetary sectors constitute a substantial part of the economy. Taxes fall mainly on the formal monetized sector. Hence, a 25 percent tax to GDP ratio would mean a much higher tax ratio with respect to the formal sector that is to spearhead development. The “limit to taxation” has to be decided with reference to the conditions prevailing in each country. Clearly if taxation proceeds beyond a limit in relation to GDP, growth is likely to be retarded. Hence the importance to be attached to the control of expenditure.


In several developing countries suffering from macroeconomic disequilibrium and external payments problems, government expenditures, particularly recurrent expenditures, are seen to have clearly gone out of hand. Strenuous attempts have therefore to be made in the medium term to bring the growth of expenditures in line with that of revenues. In a growth-oriented program, obviously, priority should be given to the objective of eliminating dissaving by the government sector. We therefore concentrate on current expenditures. Four components of current expenditure could be distinguished and considered: interest payments, current development expenditure, current nondevelopment expenditure (on goods and services), and subsidies.

The ratio of interest payments to current revenues (and to GDP) has continued to grow in several countries because government has borrowed money to meet current expenditures, government investment has not been productive enough to raise commensurately GDP and current revenues, and public enterprises in which borrowed money has been invested by the government have been making insufficient profits or are incurring losses. The first and the third causes may be dealt with here.

Although it could be argued that certain types of current expenditures would lead to human capital formation and therefore should be treated on a par with expenditure on material capital formation, as a rule it is safe to proceed on the basis that all current expenditures should be covered by current revenues. This means that the growth of the other three components of current expenditure mentioned above (other than interest payments) should be regulated to bring down the level of total current expenditures to the level of total current revenues within a reasonable period of time. Thereafter the rate of growth of current expenditures must be made to keep pace with that of revenues. As regards loss-making public enterprises, it is necessary to divide them into enterprises in the core sectors such as the infrastructure sectors and those in non-core sectors. In the medium term, enterprises in the latter category that are continuously making losses must be closed down or sold off to the private sector. Public enterprises in the former category must be asked to work out time-bound programs for making their operations profitable; and provided they are working efficiently and wages are not rising faster than labor productivity, they must be allowed price increases justified by increases in the costs of inputs.

In most countries, the size and the composition of capital expenditures during a plan period are decided upon without any reference to the capacity to meet, out of the current budget, increased maintenance or committed expenditure in the subsequent period. Inevitably, in course of time, current expenditures tend to grow faster than current revenues. This then leads governments to put restrictions or squeezes on recurrent expenditures creating the so-called “recurrent cost problems.” A recent examination of the situation in 18 African countries identified all but 3 as being affected by this problem. “The problem arises because, faced with a need to restrain recurrent spending, it is difficult to reduce the size of the civil service or to cut wages and salaries. It is the non-wage recurrent items which tend to be cut: transport allowances, supplies of materials and spare parts, sometimes even the use of electricity and telephones.”10 This kind of restraint obviously leads to wastage and inefficiency and reduces the productivity of previous investments. It should therefore become an important element in public investment strategy that in formulating public investment schemes and in determining the total size of the capital budget, the consequential rise in maintenance and running expenditure in the subsequent periods must be kept in view and the composition and the size of the investment plan must be tailored to the expected availability of revenues at the end of the plan period. In the long run there is no other way of averting a fundamental imbalance arising between revenues and expenditures.

Given that the ratio of tax to GDP in many developing countries is under 25 percent, and in some cases below 20 percent, these countries cannot afford to divert a large proportion of GDP to defense and other nondevelopmental expenditures. Nevertheless, countries have to decide for themselves the fraction of resources that they wish to, and can afford to, spend on general administration, law and order, and defense. In the interest of economic development, the rate of growth of such expenditures must be limited to the rate of growth of GDP, after their relative level has been suitably adjusted.

Expenditure on subsidies is also a significant component of public expenditure in a number of countries. We may consider four kinds of subsidies: subsidies to cover losses of public enterprises, subsidies to encourage the use of a particular input in the interest of the national economy, export subsidies, and subsidies to lower the prices of particular consumer goods for the benefit of all consumers or particular groups of consumers.

One sometimes gets the impression that the Fund and the Bank consider that subsidies, as a rule, are a “bad thing.” I am sure that this impression is wrong, because it can be shown that a positive Or negative tax intervention in the price system can often be justified just as intervention through the public expenditure side, such as the provision of free education or free highways.11 Subsidies can, of course, be objected to if they are growing too fast in relation to revenues and in the context of other more urgent demands on revenue. They could also be objected to if they tend to distort resource allocation or incentives for producers. But if incentives for producers are maintained and if the price to consumers or to a targeted group of consumers is lowered in respect of goods that are not subject to a high price elasticity of demand, there could be no objection on economic or efficiency grounds. Take, for example, a public distribution system such as the one operating in India. The provision of wheat, rice, and kerosene through the fair price shops at subsidized rates does not mean any disincentive to producers. In the case of rice and wheat, the cost of the subsidy is borne by the State, and in the case of kerosene, the production and supply is through a State monopoly. The quantities supplied are limited to fortnightly rations. Although this set of arrangements cannot be criticized on grounds of efficiency, it could be argued that given the scarcity of resources and the demands on them for various developmental purposes, the food subsidy must be targeted toward only the poorer sections of the society.12 Moreover, food subsidies could get out of hand if the degree of subsidization is increased over time. This can happen (it has already happened, to some extent, in India) if the price at which foodgrains are bought from the farmers is adjusted upward in line with the increase in costs of production, but the price at which it is sold to the targeted group of consumers is not raised correspondingly.

It is clear that subsidies that involve disincentives to producers or that lead to significant distortions in the allocation of resources should be eliminated. Also, subsidies on unlimited amounts of consumption, particularly in respect of noninferior consumer items (such as meat in Zimbabwe and general higher education in India), should be avoided because they would lead to artificial inflation of demand.

Equally objectionable are subsidies to cover losses of inefficiently run public enterprises. As already suggested, all those in non-core sectors should be closed down or sold off, and it must be one of the objectives of the adjustment program to improve the efficiency of public enterprises in the core sectors. In the case of even efficiently run enterprises, losses would occur if administered prices are kept below costs and these losses would then have to be covered by budgetary subsidies. Such subsidies also tend to militate against efficient allocation.

This leaves export subsidies, subsidies in favor of the poor, and subsidies on the use of particular inputs that would further public interest. Even assuming that these subsidies are justifiable under particular circumstances, the total of these subsidies in the budget cannot be allowed to grow out of proportion to the growth in national income and government revenues. Depending upon the targeted tax ratio, the permissible ratio of such subsidies to GDP should be determined in the light of other government commitments of higher priority.

We may conclude that government subsidies that tend to distort allocation of resources or worsen income distribution through unjustified lowering of the prices of goods and services consumed by politically powerful groups or other favored sections are harmful or wrong in principle and must be eliminated. Other subsidies, such as those favoring the poor in a limited way and those encouraging particular activities or methods of production, as part of promotional effort, could be justified, but care must always be taken to keep them under control. In a growth-oriented adjustment program, reduction in subsidies should figure as an element but it is important to ensure that the poorer sections are not unduly hurt, or that desirable promotional efforts are not prematurely terminated.

Concluding Remarks

In a growth-oriented adjustment program, fiscal policy has to play an important role both in the short term and in the medium term: in the short term because often increasing budgetary imbalance is one of the major causes of macroeconomic disequilibrium, and in the medium term because the structures of government revenues and expenditures affect savings and investment and the efficiency with which resources are used.

In the short term, the main task of fiscal policy is to reduce budget deficit, thereby bringing down government borrowing from the banking system and government dissaving, if any. In general, however, it is difficult to increase revenues in the short run without fueling inflation or creating other problems. Unless direct tax rates happen to be low, which may be true of some countries, raising direct tax rates significantly may only lead to greater evasion and further disincentives. Since there is generally indexation of wages and salaries and administrative fixing of several prices on a cost plus basis, raising indirect tax rates so as to obtain a significant increase in revenues will lead to a chain cost-push reaction. On the whole, in the short term, a cut in government expenditure other than essential developmental expenditure is preferable and would be more effective in reducing the demand for real resources. In a very difficult budgetary situation, however, it may become necessary to raise tax revenues and at least some public enterprises’ prices. Then one may resort to a general surcharge on all taxes so as not to change the differential rates structure without proper study. Also, in regard to public enterprise prices, preference may be given in the short run to raising, where needed, prices of enterprises not producing vital inputs (their turn can come later with resumed growth).

In the medium or longer term, unless the tax ratio is quite high in a country, it would be possible to increase revenues through the reform of the tax structure and improvement in administration. The four guiding considerations should be: simple structures with only limited exemptions and concessions, broad bases, moderate rates, and efficient and strict administration. Moderate rates themselves would increase incentives; but specific incentives to savings and investment could be incorporated, if they do not already exist under the personal income tax, such as accelerated depreciation, low tax on machinery, and exemptions for net savings, at least to a limited extent.

Over time, the growth of current expenditures should be regulated (and in respect of some items reduction should be effected) so that, given the growth in revenues, the current account deficit of the budget can be eliminated and growth of expenditures kept in line with the expected growth in revenues. In this exercise, attention should be paid to the elimination of subsidies that are unjustifiable or that lead to inefficiency in resource use. On the capital account, government expenditure should be restructured so that public investment increases the productivity of private investment through the strengthening of the infrastructure.

When a country has got into serious balance of payments and inflation problems, the process of adjustment is hard and long. The population will undoubtedly be subjected to hardship in the initial stages. Under such circumstances, it is important that the poorer sections of society are not made to suffer equally with others. Hence, some elements of the program must be directed to safeguarding their interests. Fiscal policy may be fashioned accordingly.

Note: The views expressed here are those of the author and do not reflect the views of the Indian Planning Commission.

For the present purpose, we can take total expenditure (current plus capital) as excluding net lending.

Reserve Bank of India, Report of the Committee to Review the Working of the Monetary System (Bombay: Reserve Bank of India, 1985), p, 152.

Public sector saving can be defined as government sector saving plus the saving of public enterprises.

If the deficit is financed through foreign borrowing, then there would be an addition to deadweight debt and that much of resources obtained from abroad would not be available for investment.

The exceptions are where a rate increase is accompanied by at least a proportionate increase in evasion or a similar decrease in demand.

The institute for Fiscal Studies, The Structure and Reform of Direct Taxation (London: Institute for Financial Studies, 1978).

Fiscal Affairs Department, fund-Supported Programs, Fiscal Policy, and Income Distribution, IMF Occasional Paper No. 46 (Washington: International Monetary Fund, 1986).

Armeane M. Choksi, Adjustment With Growth in the Highly Indebted Middle-Income Countries, World Bank, Country Policy Discussion Paper, Washington, October 1986.

Chad Leechor, Tax Policy and Tax Reform in Semi-industrial Countries (Washington: World Bank, 1986).

Report of the Commission of Inquiry into Taxation (Harare, Zimbabwe: Government Printer, 1986), p. 302.

In this connection, note the examples mentioned in Michael Bruno, “Comments,” in IMF Conditionatity, ed. by John Williamson (Washington: Institute for International Economics, 1983), p. 127.

The Fiscal Affairs Department of the Fund has argued that food subsidies are not of much value because foods even at subsidy prices are likely to be beyond the reach of the “ultra poor.” It is difficult to accept this argument because there is no justification to deny a benefit to the poor just because the ultra poor cannot avail themselves of it. In any case, for the ultra poor, food for work programs can be started, as in India.

    Other Resources Citing This Publication