The distribution of income and wealth in developing countries has recently become a matter of great concern to all those interested in development. In part, this new concern reflects increased awareness of the social and political costs of income inequality. As Mr. McNamara, President of the World Bank Group, stated at the Joint Fund-Bank Annual Meeting in 1972: “When the highly privileged are few and the desperately poor are many—and when the gap between them is worsening rather than improving—it is only a question of time before a decisive choice must be made between the political costs of reform and the political risks of rebellion.”
In part also, however, attention has come to be focused on the distributional question for economic reasons. When, as has been estimated for Latin America, the poorest half of the population receives about the same share of income as the top 1 per cent and the lowest 70-75 per cent of the population the same share as the top 5 per cent, then it is clear that the distribution of income and wealth will have substantial implications for the pattern of consumption and production in developing countries. Problems related to an extremely unequal income distribution are by no means confined to Latin America although the degree of income concentration appears to be less in other parts of the world. It has been suggested that an unequal pattern of income distribution may make it difficult to develop broad-based national markets for the products of modern industry. Furthermore, as a result of the high direct and indirect foreign exchange component of some of the goods consumed by the well-to-do classes in many developing countries, income inequality may also tend to accentuate pressures on the balance of payments.
Despite these comments, it must not be assumed that a more equal income distribution will in itself necessarily raise the rate of economic growth. It may, as has been the traditional view, in fact decrease it by, for example, reducing savings. In reality, the relation between the distribution of income and wealth and the level and pattern of economic growth in societies at different stages of development is largely an unresolved question. There are very few careful empirical studies of the interrelationship of growth and distribution over time in particular countries and no very convincing general theories bearing on this question. Nevertheless, it is safe to say that the new concern with problems of distribution has convinced most development economists that distribution must be considered as a separate—though not necessarily secondary—goal to growth. Partly for this reason, there has recently been a marked revival of interest in the extent to which governments can alter the distribution of income by means of fiscal policy.
The potential for fiscal redistribution
Traditionally, the fiscal system has been considered, as a recent Colombian plan put it, “the most effective and least disruptive instrument of the State for bringing about better distribution” (Guidelines for a New Strategy, National Planning Department, 1972). For decades it has been a common, if often merely implicit, assumption of economists that any unwelcome effects on distribution resulting from other government policies or from the general course of economic growth could be readily corrected by means of the fiscal system. Not only has the fiscal system been thought of as the supreme equalizer of public policy, but it is also commonly considered to be used extensively for this purpose, at least in the industrial countries. This fiscal redistribution is assumed to take place both through progressive income taxes (which take relatively larger shares from the rich than from the poor) and through expenditure programs which transfer income to the poor and provide them with proportionately larger benefits from free government services.
Recently, however, increasing doubt has been expressed about the efficacy of the fiscal system as an instrument of redistribution. For example, Mahbub ul Haq of the World Bank referred in a recent speech (“Employment in the 1970’s: A New Perspective,” International Development Review, December 1971) to the prevalence of “misguided faith in the fiscal systems of the developing countries and a fairly naive understanding of the interplay of economic and political institutions.” Until recently this misguided faith was taken to justify the preoccupation of development economists with growth policies to the neglect of income distribution policies. “We know now,” Mr. Haq went on, “that the coverage of these fiscal systems is generally narrow and difficult to extend.” It is thus coming to be the new conventional wisdom that the fiscal systems of developing countries have not done much to alter the distribution of incomes generated by the economic system—and, indeed, that they cannot do much.
Two reasons have been offered in support of this distributional pessimism. The first is that only a relatively small part of the national income in most developing countries is channeled through the fiscal system so that its potential as a redistributor is inherently limited. The second is that, because extensive direct income transfers are impractical in poor countries, the major task of redistribution falls on the tax system, but the tax systems in most poor countries are not, in fact, progressive. On the contrary, they are either regressive (that is, fall relatively more heavily on the poor) or exhibit, at best, a sort of wandering proportionality from income class to income class. Furthermore, it is generally argued, this failure to tax the rich more heavily than the poor is not offset on the side of public expenditures by policies which tend to spend more on those things which benefit the poor rather than the rich. In fact, some studies indicate that, at least in some countries, most of the benefits from public expenditure accrue to the better-off section of society, thus accentuating the general ineffectiveness, or even perversity, of the fiscal system as a means of correcting the distribution of income and wealth.
This picture is indeed a somber one. Fortunately, however, this pessimism is only partly supported by the facts. In the first place, quite a few developing countries, such as Brazil and Tunisia, collect 20 per cent or more of the national income in taxes. These countries would thus appear to have substantial scope for redistributing income through the fiscal system if they chose to do so. On the other hand, it is true that many other countries, ranging from India to Guatemala, collect a considerably smaller proportion of the national income in taxes, and thus, on the face of matters, would find it more difficult to alter substantially the general pattern of income distribution through fiscal means. Even in these cases, however, it would, in theory, be possible to reduce inequality significantly by taxing the rich—although this would in itself do very little to help improve the lot of the poor since taxes can only make the rich poorer, not the poor richer. This brings us back to the second reason for distributional pessimism noted above: the observed regressivity of the fiscal systems in developing countries.
Here again, the facts do not appear quite as gloomy as some seem to believe. A recent intensive review in the Fund of the large number of studies on the distributional impact of fiscal systems in developing countries suggests that the great variation within the developing world makes it very difficult to generalize about the progressivity or regressivity of taxes. About half of the tax incidence studies examined (in total there were 67 studies covering 23 countries) indicated that the tax system was at least mildly progressive, with the remainder exhibiting proportionality or even regressivity over some ranges of the income distribution. Although the incidence of expenditure systems has been less studied (13 studies for 9 countries) the results were again mixed, with half of the studies suggesting a regressive (which means pro-poor in this case) pattern and the remainder more or less proportionality. It should be remembered, however, that even a high degree of relative favoritism for the poor in expenditures may still mean that far less is spent on them than on the rich in absolute per capita terms. It is the latter which is important so far as redressing unequal income distribution is concerned, and one might in fact argue that more in absolute terms must be spent on the poor than on the rich if public expenditure policy is to be truly equalizing in its effects. No country examined came even close to meeting this standard.
Although the conceptual and statistical bases of these studies of tax and expenditure incidence leave a great deal to be desired, it can at least be concluded that the available evidence provides no reason for believing that the fiscal system in developing countries is inevitably regressive and ineffective as a means of distributing income. On the contrary, at least in the more advanced of the developing countries—in which the size of the government budget also tends to be relatively greater—there appears to be considerable scope for affecting distribution through the fiscal system if the countries want to do so.
Taxation and resource allocation
Rather than throwing up one’s hands in despair, then, it seems worthwhile to look a little more closely at how tax and expenditure policy can be used to affect the distribution of income and wealth in a developing country. The first point to note in this respect is that the implicit taxes and subsidies which are introduced into the economy of many countries by fiscal (and other) policies which distort relative prices may well be more important than the explicit taxes and expenditures which appear in the budget. In countries with considerable unemployment, for example, payroll taxes which raise the private cost of employing workers above the social cost may lead to fewer workers being employed in the modern sector than would otherwise be the case. Similarly, exemptions from customs duties for imported capital equipment, especially in countries with overvalued exchange rates, tend to maintain the cost of capital at artificially low levels and to encourage the relative expansion of capital-intensive lines of production, particularly by the larger firms which have easier access to import licenses and foreign suppliers’ credits and which tend anyway to use less labor per unit of capital. Furthermore, tax and other incentive policies which induce more investment through lower capital prices will tend, simultaneously, to reduce the desired level of utilization of existing capital equipment (which depends in part on the relative prices of capital and labor), and hence will further reduce employment.
The misallocation of resources resulting from such capital-favoring policies has recently received much attention in countries concerned, as are most developing countries, with an excessive and growing amount of open urban unemployment. The substantial distributional implications have not been so commonly noted although it is clear that policies which tend to lower the price of capital and raise the price paid to labor will tend, perhaps paradoxically, to increase the inequality of the distribution of income earned through work through their effects on employment and the choice of technology. A reversal of such policies will thus tend to reduce inequality. In short, in countries where fiscal policies have artificially fostered economic growth by lowering the relative price of capital below that which should prevail in the light of the scarcity of capital and the abundance of unskilled labor, a reversal of these policies should have substantial and beneficial allocative and distributive effects. Indeed, this appears to be one of the few instances in development economics in which one can, so it seems, have one’s cake and eat it too.
One way in which tax and expenditure policy may affect the distribution of income and wealth is thus through policies which affect employment and the relative rewards received by labor and capital and which thus alter the pattern of economic activity and the distribution of income and wealth generated by that activity: this may be called primary redistribution. The effect of taxes on the structure of industry and the demand for unskilled labor may well be the single most important distributional aspect of tax policy in developing countries.
A second way in which tax policy can alter this initially-generated distribution is through the taxation of wealth. It is true that taking away wealth from the few, however wealthy, and giving it to the many, however poor, would not in most developing countries have much visible impact on the well-being of most people, given the relative numbers in the two groups. However, this is, in a sense, irrelevant since the primary aim of taxing wealth—in addition to encouraging the more efficient utilization of capital—is not to make the poor richer but to make the rich less powerful. One relevant point of particular interest in many developing countries is that an effective land tax will in itself tend to have an initial primary redistributive impact through reducing land prices. A more important point is that heavier taxes on wealth (or on income from capital) must be imposed to offset the effects on incomes of the higher returns to owners of capital—the rich—which would result from those policies designed to raise the relative price of capital to its true scarcity value in poor countries. If this is not done, the presumably undesirable distributional impact of making wealthy capitalists wealthier will have to be counted as a partial counterbalance to the poverty-relieving effect of increasing employment through more rational factor pricing.
There are, of course, well-known limitations in most countries on the political and administrative feasibility of land taxes, net wealth taxes, and death taxes. There is also some possibility of adversely affecting savings, though the importance of this effect is generally greatly exaggerated: there are few other taxes which can tap the same segments of the population with less adverse impact than wealth taxes. Although the impact of wealth taxation will no doubt usually be minor, on the whole it seems clear that those who wish to use the fiscal system for redistributive purposes should pay much closer attention to the role of wealth taxation, both as a means of redistribution in itself and as a support for such more direct approaches to the problem as land reform and more rational factor prices.
The other approach to fiscal redistribution, and a more traditional focus for fiscal discussion, is to redistribute income after it has been initially distributed by the workings of the economic system: this may be called secondary redistribution. Conventionally, it is argued that the major instrument which should be used for this purpose is a global progressive personal income tax. Unfortunately, in many developing countries this tax is neither global in coverage (since many forms of income escape effective taxation), nor (for that reason) particularly progressive nor, indeed, really on personal income. In fact, the income tax in practice all too often amounts to a tax on certain forms of earned income, with only scattered coverage (largely through presumptive assessments) of many of the other forms of income—rents, dividends, interest, profits—which accrue particularly to the better-off groups in society.
In the political and administrative circumstances of many developing countries, an attempt to levy a truly effective personal income tax may in fact end up by imposing a still heavier burden on income from labor while continuing to fail to reach income from property effectively. The same degree of failure in attempts to tax wealth—if one may speak in this very relative fashion—is likely to lead to much better results, both distributionally and allocatively. While the potential of the income tax to produce significant revenues over the long run should not be neglected, it would thus be a mistake to rely solely upon it to correct all distributional ills. In many countries, progressive taxes on consumption combined with wealth taxes may prove as good or better for this purpose.
The attack on poverty
More generally, as noted above, taxes cannot make the poor richer, which is, after all, the main concern of distributional policy. Even the complete removal of all taxes on the poorest members of society would not make them much better off, simply because of the low absolute amounts of income and tax involved. Furthermore, many of the poorest people, particularly those in rural areas, take part only marginally in the economic life of the country and are thus little affected by taxes. While the regressivity of the tax system, where it exists, ought to be reduced as much as possible in order not to make things worse, it is clear that, if our main concern is with poverty as such, with the waste and misuse of human resources and the stunted opportunities in life afforded those with incomes below some minimum level, any fiscal corrective must be exercised primarily through the expenditure side of the budget.
Two major points may be mentioned on the effects of public spending on distribution in developing countries. First, expenditures on health, education, and similar government activities related directly to the well-being of individuals can clearly be so directed as to benefit most the poorer members of society. The fact that in some countries it appears that education expenditures have in reality helped most those who need it least—for example, by subsidizing university education for the children of the well-to-do—does not affect this conclusion. What it does do is reinforce the importance of being selective and of paying careful and close attention to the precise nature of expenditures if one is concerned with the distributional impact. It is obvious, for example, that money spent on primary education in the rural areas will do more to benefit the poor than the same amount spent on university education in the capital city. The poor, like the rich, are by no means a homogeneous group, and policies designed to help particular subgroups of the poor will need, as a rule, to be selective and often localized in their incidence.
Second, one should pay close attention not only to the precise identity of the beneficiaries of government expenditure, but also to the identity of those whose incomes are indirectly increased as a result of the expenditure. The ultimate gainers from improved roads may, for example, be those rich enough to own an automobile, but the total effect of the road expenditure on distribution depends also on the way in which the road is constructed: how much employment does the construction generate; whose income is increased by the money payments made in the course of construction? These are among the questions that must be explored by those interested in improving income distribution. The choice of technique in the public, as in the private, sector is thus a central focus of distributional concern. While there is much we have still to learn about designing projects in terms of their distributive as well as allocative efficiency, it seems clear that the proper design and placement of investment projects can, in principle, contribute substantially to improving the distribution of income in many developing countries while simultaneously fostering growth.
Both sides of the public finances in developing countries—expenditures and taxes—can thus, it appears, play a substantial role in distribution policy. Indeed, since the fiscal system is, in every country, one of the most pervasive instruments of government policy, its distributional implications must always be carefully considered in any case. It is true that fiscal redistribution is, in the long run, inherently less important as a means of alleviating poverty in the poorer parts of the world than economic growth, and that care must be exerted not to reduce unduly the long-term benefits from growth for short-term gains from redistribution. Nevertheless, there appears to be considerable scope in many countries to pursue fiscal policies which will foster both growth and redistribution. Even if one shrinks, for administrative and other reasons, from an active redistribution policy, it can at least be said that there appears to be little economic justification for such inegalitarian fiscal policies as now exist in some countries. The public finances cannot in themselves cure the ills of the world, but they can, and should, be employed carefully and consciously to improve the distribution of the gains from economic growth as well as to improve the rate of growth itself.