Charles A. Sisson
The Fund has long recognized the significance of distributional issues in adjustment programs. Adjustment policies are framed in light of the size of the imbalance confronting the economy and how much adjustment is realistically possible and can be maintained beyond the program period, so that distributional issues have always been an inherent, if unspecified, element in programs. However, in the absence of any satisfactory means of addressing this matter, the Fund has generally maintained that distributional issues are primarily an internal political concern. This stance has led some observers to suggest that Fund-supported adjustment programs (hereafter referred to as Fund programs) favor vested interests and are inimical to the needs of the poor.
This paper examines the methodological difficulties of analyzing the distributional effects of Fund programs and then presents, in summary form, the analytical approach and conclusions of a recent Fund staff study. (This study, Fund-Supported Programs, Fiscal Policy, and the Distribution of Income, is to be published in the Fund’s Occasional Papers series. Mr. Sisson was one of the principal authors of this study—Ed.)
The question of the relationship between Fund programs and issues relating to income distribution can be addressed in several ways, each emphasizing slightly different aspects of, and concerns about, the effect of economic policies. Perhaps the most encompassing approach would be to try to determine whether Fund programs support a more desirable distribution of income. Second, one could try to determine whether there typically is an alternative set of policies that would more adequately address social objectives. A third approach would be to determine if, generally speaking, Fund-supported programs have worsened income distribution. Fourth, Fund programs could be analyzed to determine which policies would best address distributional issues without compromising the pursuit of other objectives.
The first two approaches provide a more comprehensive view of the choices available to the authorities, but they also greatly complicate the analysis. Ideally, Fund programs would have to be compared to the alternative program that a country would have pursued without Fund support, and the period for this comparison would have to be sufficiently long to allow the economy to achieve a new equilibrium. Clearly, economic analysis has not progressed to the point where such an analysis can be successfully undertaken. Further, most countries do not identify an adequate adjustment program before approaching the Fund so that an alternative approach is not available for comparison. It is ultimately for the authorities to determine what policies are politically feasible to achieve the stated objectives of a program. Changes in income distribution as such are not a performance criterion in adjustment programs (and, given the difficulties described in this paper, it is hard to see how it could be); and the authorities have discretion in seeking to achieve any politically acceptable distribution of income.
The last two approaches may be more manageable, mainly in that they avoid the trade-off between growth and equity which is central to the other two, and indeed to any attempt to discover an “optimal” set of policies. It has long been recognized that there is some element of trade-off between the rate of growth which an economy can achieve and the degree to which a particular income redistribution lowers the returns that factors can command in the marketplace. In other words, over time the size of the economic “pie” measured by output can become more important than how it is initially divided.
Even using these more limited approaches, however, leaves some difficult, if not insurmountable, methodological issues and data problems. First, there are a number of ways in which the distribution of income can be classified—such as by individuals, households, urban-rural distinctions, factors of production, type of employment, among others—and no one classification is fully adequate to encompass distributional comparisons for all purposes.
Second, the existing income distribution of a country seeking balance of payments support from the Fund is a result of past policies and may be unsustainable. Furthermore, the indices commonly used in making comparisons of income distribution do not normally provide clear results about distributional changes. Moreover, price movements during the course of a Fund program will change the consumption opportunities available to individuals, leading to further difficulties in making comparisons. Therefore any simple comparison of pre-program and post-program income distribution would result in misleading conclusions about the distributional effects of Fund-supported programs. The more appropriate comparison would be a “with-and-without” comparison, i.e., a comparison with the income distribution that would have evolved in the absence of a program or under a different program or an alternative set of policies. Unfortunately, economic analysis has not progressed to the point where such an approach is feasible.
Third, to the extent that growth is an objective of Fund programs—and a primary objective of the Fund is to promote growth within the context of a sustainable balance of payments position—it implies that programs focus on resource allocation rather than income distribution per se, although resource reallocation obviously affects income distribution as incentives are adjusted. However, the effect of corrective policies takes longer than any simple income transfer (just as the adverse effects of price distortions are often evident only after a protracted period). Thus, the period of most Fund-supported programs is too limited to allow a full study of their distributional implications.
A further consideration is the need for extensive information on individual income groups (.including detailed information on consumption and personal income levels) as well as information on government transfers and services received, public sector, nonmonetary and irregular sources of income, and the composition of investment portfolios; and all of this for both before and after the program period. Such information is not necessarily available in countries with highly developed statistical services and is certainly unavailable in most countries that have Fund-supported programs.
In view of the considerable difficulties associated with analyzing the impact of Fund programs on income distribution outlined above, the Fund staff recently conducted a more limited inquiry reviewing the measures contained in 94 Fund-supported adjustment programs in the 64 countries which had undertaken such programs in 1980–84, and assessed their likely distributional effects on various socioeconomic groups.
The programs surveyed contained a variety of approaches to the adjustment problem and consequently encompassed a wide range of policy measures (see accompanying table), while also containing some common elements. Virtually all programs used limits on monetary or credit expansion, while 55 percent included measures to liberalize and reform external trade. Most programs linked the external imbalance to fiscal performance, with 91 percent containing measures to restrain government expenditure, and 96 percent measures to increase revenues. Among specific expenditure control measures, about three fifths of all programs sought to limit the wage bill; a similar number limited purchases of goods and services, and also subsidies. Revenue measures focused on the personal income tax in 46 percent of the programs, corporate taxes in 34 percent, taxes on domestic goods and services in 73 percent, import duties in 57 percent, and export duties in 24 percent.
While the study concerned the distributional effect of fiscal policies in Fund-supported programs, it recognized that these policies are normally framed in the context of monetary and external sector policies that have far-reaching implications for income; hence it attempted to provide a general overview of these effects on income distribution, even though a detailed analysis was beyond its scope. Exchange rate, monetary, and credit policies are commonly acknowledged to be important for adjustment, but their full effect on income distribution is obscured because their impact is indirect, through changes in resource allocation. The initial effects of measures to increase the price of foreign exchange favor factors of production in export industries and those who hold capital abroad, while policies to make capital more expensive reward those who own capital. This is not to suggest, however, that the wealthy will be the primary beneficiaries of even the initial redistributive effects, as they may not be involved in export-oriented industries (indeed, often they are involved in importing activities), or they may have a relatively small share of their wealth held abroad. Moreover, in many countries, small farmers are helped by more profitable exports and less competition from (undervalued) imports; and the wage share of manufactured exports has been shown to be higher than that of (usually protected) import substitution. Over the longer term, the effects of these measures are not clear-cut; the macroeconomic measures should promote growth and employment throughout the economy, leading to a higher standard of living for all, including the poor, so that the ultimate effect of the policies will not necessarily be to increase inequality.
There is a relationship between fiscal policies and exchange rate and monetary developments, and these links are an important aspect of the adjustment process. On the revenue side of the budget, the most significant tax sources surveyed were income taxes, taxes on domestic goods and services, and trade duties. Measures designed to reform individual and corporate taxation were seen to have had little effect on income distribution because of the limited reliance on these taxes and their inherent limitations as a potential source of revenue in developing countries. While the impact of income taxes has been limited, the overall effect has been to improve the distribution of income. Taxes on goods and services have, in keeping with their relative importance in the revenue systems of developing countries, been more important sources of revenue. The evidence indicates that these taxes affect the distribution between different sectors of the economy (e.g., rural versus urban) and between individuals with different consumption habits (e.g., between smokers and non-smokers), rather than redistribute income across income classes.
Taxes on trade are an equally, if not more, important tax source for developing countries, and generally their impact has been supportive of distributional concerns. Taxes on imports, under adjustment programs, have three important characteristics that help improve income distribution. Individuals in higher income classes tend to consume more imports than those with lower incomes, so that they tend to pay more import taxes. Second, goods that are generally identified to represent luxury consumption by the wealthy tend to have higher tax rates. Finally, to the extent that tariffs replace quantitative restrictions on imports, the economic rent that importers gain when imports are administratively limited is reduced or taken away from importers, who tend to be in higher income classes.
|Liberalization and reform of exchange rate arrangement||52||55|
|Limit on credit expansion||92||98|
|Measures to mobilize domestic savings||51||54|
|Measures affecting wages and prices||83||88|
|Structural adjustment measures||70||74|
|Improve or reform of tax administration||52||55|
|Restructuring of personal income tax||43||46|
|Measures affecting corporate tax||32||34|
|Domestic tax on goods and services||69||73|
|Restrictions on central government functional expenditure outlays||4||4|
|Restraint of central government current expenditure||86||91|
|Restraint of wages and salaries||59||63|
|Restraint on capital outlays and net lending||56||60|
|Improved expenditure administration||40||43|
|Capping or reduction in subsidies||39||41|
|Curtailment of current transfers to NPEs2||26||28|
|External debt policies||86||91|
Classifications are not necessarily mutually exclusive.
Nonfinancial public enterprises.
Classifications are not necessarily mutually exclusive.
Nonfinancial public enterprises.
Who do subsidies affect?
Empirical information concerning the direct effect of price and subsidy policy on income distribution is scarce and of questionable quality. In addition, existing studies do not cover a sufficient range of countries to provide an adequate basis for generalization. The conclusions of the few studies available, however, indicate that for large, general subsidy programs, the hardship imposed in the short run by eliminating subsidies to the poor cannot be ignored. They also indicate, however, that such programs are inefficient and ineffective mechanisms for redistributing incomes.
In one member country, for example, spending approximately 2 percent of GDP on consumer subsidies, a Fund study based largely on household survey data found that even a relatively open-ended subsidy scheme resulted in rather perverse distributional effects. As such open-ended food subsidy schemes are typical of those attracting Fund attention, precisely because of the budgetary resources they consume, the major results are worth reviewing here. In urban areas, it was found that less than 10 percent of the benefits of food subsidies accrued to the poorest 20 percent of the population, while 46 percent of the subsidy went to the richest 27 percent of the urban dwellers. Looking more directly at nutritional status, and taking the figure of 1,800 calories a day as representing the limit of serious undernutrition, it emerged that in the rural areas only 15 percent of the budgetary cost of subsidies was transferred to those under this limit, and in the urban areas the corresponding figure was 25 percent. Despite this poor targeting, however, data indicate that food subsidies provided an income transfer equivalent to up to 6 percent of per capita household expenditure for the poorest urban dwellers.
In a second member country for which data were available, food subsidies have been allocated through ration shops that distribute foodgrains at below free market prices. Rural people, who made up some 90 percent of the population, received less than 20 percent of the available food ration, and even this proportion was reported to be reduced in the event of shortages. More than 60 percent of the recipients were government employees, and the privilege of a ration card was reportedly used to hold down government wages and salaries.
A recent study in another member country, which spends approximately 7 percent of GDP on food subsidies, found a substantially less skewed distribution of benefits, no doubt due to the broader coverage (and higher cost) of the system. The amount of income transferred equaled 17 percent of the income of the poorest rural expenditure quartile, and 16 percent of the income of the poorest urban expenditure quartile. The amount of income transferred for the richest rural and urban quartiles was 4 and 3 percent, respectively. The distribution of subsidy benefits in absolute terms appeared to be much less regressive than in the country previously mentioned, with per capita subsidy transfers being relatively equal in each quartile.
Another member country decreased the budgetary cost of food subsidies from 6 percent to 3 percent of GDP, while at the same time increasing the per capita transfer to the poorest classes by more than 30 percent through the introduction of a food coupon system that targeted food subsidies to the lower income classes.
The effects of policies concerning export duties in these programs on income distribution are less important, but also less supportive. The final incidence of export taxes depends on the relative elasticities of supply and demand for a country’s exports, but generally developing countries are price-takers in international markets and their producers must bear any increase in an export tax. Since these producers often tend to be agricultural small-holders or other low-income producers, increased export duties may not have desirable redistributional effects. However, if export duties do not discourage growers from growing their crops, but merely from exporting them, it is consumers who may benefit from lower domestic prices of export crops.
Fund-supported programs have not focused on specific functional expenditures: less than 5 percent of programs referred to a specific functional expenditure, and less than one third of them made any reference at all to functional expenditures. Therefore it may be concluded that the impact in this area has been minor. In contrast, efforts to limit excessively expansionary fiscal policies through policies to control economic categories of expenditure—specifically government wages and salaries, investment outlays, food and petroleum subsidies, and transfers to cover the losses of public enterprises—are perhaps the most important and controversial aspects of Fund-supported programs.
About 63 percent of Fund programs contained measures to restrain the wage bill. By trying to limit the growth of wage bills by such measures as wage freezes, limits on employment, and reduced fringe benefits, wage and salary policies in Fund-supported programs have tended to redistribute income away from the urban middle class. To the extent that the poor are predominantly in rural areas, these efforts tend to equalize the income structure, although they may adversely affect the efficiency of government operations. Three fifths of all Fund-supported programs have also tended to constrain investment to available resources, often by limiting new projects, or by improving the efficiency of such investment programs by focusing on quick-yielding projects. Given the long gestation periods of these investment programs it is difficult to determine how these measures have affected income distribution.
The issue of subsidies and transfers to public enterprises is a controversial one (see box). The reassessment of food subsidies under Fund-supported programs often focuses on increasing the price of basic commodities to levels nearer world market prices. This frequently increases domestic agricultural incomes and, in this way, distribution is improved. Where the increase would have a particularly adverse effect on both the urban and rural poor, some programs have attempted to target subsidies more accurately to those who most need them. In devising policies with respect to subsidies on the consumption of petroleum, for example, there has been an attempt to redistribute resources away from factors employed in inefficient industries and, by removing subsidies to the use of private vehicles, to promote the development of mass-transit systems which are typically used by lower income groups. Thus, ironically, one of the most controversial aspects of Fund-supported programs—the reduction of subsidies—has been the one that has probably had the greatest success in improving income distribution.
Government participation in the market economy through nonfinancial public enterprises is often justified by its presumed favorable impact on income distribution. In practice, however, these enterprises frequently become a financial burden on the economy, so that the favorable effects of producing goods at controlled prices may be offset by the limited amount of goods produced and the inflation caused by government reliance on domestic bank borrowing to finance their losses. In some cases, the only elements in the economy that benefit from these activities are the factors of production used by these industries, and those who can control the sale of these products in parallel markets. When these firms are encouraged to raise prices, the effect is not so much to penalize consumers as it is to validate prices already being charged in unofficial markets. Thus, the primary redistributive effect of pricing measures is to transfer income from the unofficial and untaxed sector to the official and taxed sector.
While the types of policies undertaken in the context of Fund-supported adjustment programs do, in their role to restore the framework for sustained economic growth, have implications for the distribution of income, they have not in general been inimical to the goal of improving distribution of income. In particular, Fund programs have not been directed against the poor. Indeed, in some cases the policies in these programs have been framed to protect the poor as far as possible. It has to be recognized, however, that the existing income distribution and nominal income levels in many countries embarking on adjustment programs with the assistance of the Fund are frequently the result of overly expansionary fiscal and monetary policies, and hence are not sustainable under any set of policies. The resultant demand-management policies enacted as part of prudent financial planning are thus one (and it is hoped a sound) means of instituting a necessary decline in domestic consumption levels. The ensuing income distribution does not necessarily become more regressive in that higher income groups are affected more than lower income groups. The debate over some Fund-supported adjustment programs may be more a reaction to the required adjustment, and reflect the views of vocal and well-organized affected groups, than an indictment of the type of adjustment measures implemented under Fund-supported programs.
A final consideration is the broader issue of what, if any, adjustment policies could be adopted that would lessen further any adverse distributional effects of adjustment policies on the lowest income groups. The Fund staff have made recommendations to improve the targeting of adjustment measures away from low-income groups. These include, in particular, exchange rate action aimed at providing adequate incentives for an agricultural sector dominated by small farmers, improving access to credit markets, movement toward taxation of global income, establishment of a truly broad-based tax on goods and services, replacement of quantitative restrictions on imports by import duties, and focusing on expenditure programs which will yield quick results, such as provision of basic skills and vocational training. However, the absence of more complete socioeconomic and demographic data—factors particularly lacking in developing countries—will pose obstacles to the implementation of the above recommendations and, more generally, to improvements in income distribution within the context of adjustment programs.
Two Studies on the Impact of IMF Programs
Occasional Paper No. 41
Fund-Supported Adjustment Programs and Economic Growth
by Mohsin S. Khan and Malcolm D. Knight
A review of the evidence available in the literature on the impact of Fund-supported adjustment programs on economic growth in developing countries. In the context of why programs are typically needed, and what they generally consist of, the study examines both econometric evidence on the effects of individual measures and cross-country experience with complete policy packages. Illustrative simulations are given to reconcile conflicting evidence and to examine the different effects of demand and supply measures on growth.
Occasional Paper No. 42
The Global Effects of Fund-Supported
by Morris Goldstein
As an increasing number of countries implement Fund programs, questions on their aggregate impact become relevant. Will they give a deflationary bias to the world economy? Are the implications for trade of individual programs globally compatible? Would simultaneous devaluation by many program countries lower their export prices? After a detailed discussion of how the impact of programs can be measured, this study identifies the ways in which program policies can have global effects and reviews the evidence on these.
Price: US$7.50 (US$4.50 to university libraries, faculty members, and students)
Publications Unit • Box A-861
International Monetary Fund
700 19th Street, N.W. • Washington, D.C. 20431, U.S.A.
Telephone: (202) 623-7430