Development and Income Distribution: Economic growth has been accompanied by increasing inequality in many developing countries. But a strategy of “nibbling” at the problems of income distribution through financial and technical assistance seems promising.
Author: John H. Adler

Contributor Notes

It is increasingly recognized that countries can be successfully “developed” while leaving masses of their citizens in poverty and misery. The author tentatively suggests a method whereby economic development can be skewed in favor of the poor.

This paper focuses on the subject of development and income distribution, and suggests a method whereby economic development can be skewed in favor of the poor. The paper underscores that improvements in the distribution of income can be achieved by applying shadow cost significantly below money cost to determine the social cost of employing members of low-income groups and to use the social consolidation strategy in the choice of technology in the physical construction of projects. The application of this method would result in the more extensive use of labor instead of capital equipment.


This paper focuses on the subject of development and income distribution, and suggests a method whereby economic development can be skewed in favor of the poor. The paper underscores that improvements in the distribution of income can be achieved by applying shadow cost significantly below money cost to determine the social cost of employing members of low-income groups and to use the social consolidation strategy in the choice of technology in the physical construction of projects. The application of this method would result in the more extensive use of labor instead of capital equipment.

John H. Adler

In a paper published in 1955 which was, in his words, “5 per cent empirical information and 95 per cent speculation” yet which bears the marks of prophetic insight, Simon Kuznets advanced the proposition that “in the early phases of industrialization in the underdeveloped countries income inequalities will tend to widen before the leveling forces become strong enough first to stabilize and then reduce income inequalities.”

Since 1955 ample evidence has accumulated to bear out Kuznets’ “speculation” that in the developing countries income is distributed even less equally than it was in the industrialized countries before the trend toward even greater inequality was reversed—in some countries before the turn of the century, in others not until after the first World War; the evidence is indeed that in most, though perhaps not in all developing countries, inequality has increased.

Despite such speculations and growing evidence of their correctness, income distribution, when carefully examined, presents two elements that surprise and shock: the extent of the inequality and the low absolute level of per capita income of the poorest quintile of the population. In 21 out of a total of 40 developing countries for which data are available, the average per capita income of the poorest 20 per cent of the population is less than 28 per cent of the national average. There are two deeply disturbing facts. First in the “hard core” least developed countries which have been singled out in various resolutions of the United Nations bodies for special consideration, the per capita income of the majority of the population is abysmally low. The second is that in a number of countries in which the national average per capita income is well above the bench mark of say, $100, such as Brazil, Colombia, El Salvador, Jamaica, Lebanon, Mexico, Panama, Peru, and Tunisia, a substantial proportion of the population still has a per capita income below that international poverty line.

There is another group of countries in which the low per capita income of the poorest quintile—or for that matter of the poorer half—of the population is a matter of special concern on humanitarian grounds and, in policy terms, of urgency. This group includes India and Pakistan in Asia and some of the major countries of Africa such as Kenya, Nigeria, and Tanzania. In these countries the relative share in the national income of the poorest quintile of the population is significantly higher than that in the “least-developed” group; but because of the low national average, the absolute amount is so low that it is clear that minimum requirements of caloric intake and nutritional balance are not met. According to the data, the per capita income of some 110 million Indians is $38 compared with a national average of $100; in pre-partition Pakistan, the poorest 20 per cent of the population—some 22 million—”enjoyed” a per capita income of $36; and so on.

These figures are the more disconcerting since the experience of the last 20 years has shown that it is virtually impossible in developing countries to mitigate the plight of the lower income groups by redistributive fiscal operations. For a variety of reasons, of which the low national per capita income itself is presumably the most important one, efforts aiming at redistribution for the benefit of the poor do not work. Moreover, there is evidence which suggests that in the great majority of developing countries the benefits of economic development accrue chiefly to the upper income groups—the highest 20 per cent or 40 per cent of the population—and that in some countries the poorest 20 per cent or even a larger percentage do not participate in the process of economic advancement at all. This state of affairs is in sharp contrast with the fundamental objectives of economic development—the diminution of poverty and human misery.

Faced with this fearful disparity between objectives and results, the observer is faced by three questions: what are the causes of the inequalities in income in developing countries; why has little or nothing been done so far to mitigate the inequalities; what, if anything, can be done about them?

The causes of inequality

The causes of inequality of income have been dealt with in a refreshing, imaginative, and comprehensive way by Adelman and Morris in their U. S. AID report, An Anatomy of Patterns of Income Distribution in Developing Nations. Because of the complexity of their findings, it would be futile to attempt to summarize them here. Only four remarks that are relevant for a better understanding of the answers given below to the third question seem appropriate.

One is to underline the importance of what Adelman and Morris have called dualism in the structure of the many developing countries, i.e., the existence side by side of a technologically and institutionally backward sector and a technologically advanced and well-organized modern sector. The more pronounced the dualism, the greater the inequality. But since the growth of a modern sector is an essential and unavoidable ingredient of the growth process itself, it becomes immediately clear that economic development itself is one of the prime causes, if not the prime cause, of income inequality—just as Kuznets had suggested. Conversely, low growth and growth potential are associated with less inequality of income.

Other causes of inequality are unemployment and underemployment—the former prevalent in urban and the latter in rural areas—and a rise in the rate of population growth. This must be considered as a significant cause of unemployment and underemployment, in the first place, and of the unevenness of income distribution if it is assumed, presumably with ample justification, that the increase in the birth rate is greater among the lower income groups than among the upper which practice birth control.

The fourth cause of inequality which deserves to be singled out—at least as a consolation to those of us who have been brought up to believe that knowledge and education are the key to everything—is the close relation between education and the level of family income. On general grounds it is of course difficult to establish the direction of the causal relation, because there is logic—and empirical evidence—for either of two assertions: (1) that poor or nonexistent formal education is one of the causes of low income; or (2) that low family income is one of the reasons why children do not go to school, or get only a minimum, and an inadequate minimum at that, of education.

Inequality is ignored

Why has so little attention been paid so far to the problems of income distribution associated with the development process? Several admittedly speculative explanations may be advanced. One is that although most of the intellectual and political proponents of development would assert as a matter of course that the objective of development is to enhance human welfare, the aim of their development operations, in practice, is more output, preferably more industrial output. The measure of their success is not the reduction of poverty, but a high growth rate of the national product. This essentially nineteenth century concept of development—or progress as it was called then—has been reinforced by the advice of the “development economists” bred in the rich countries who have brought their neoclassical professional training to bear on the problem of economic development. They understand the functioning of the price and market mechanism. And they can readily explain why entrepreneurial talent and professional skills should be highly regarded in the developing countries, and why the abundance of unskilled workers and peasants prevents those persons’ income from rising to any significant extent. Development economists generally are much readier with explanations than with remedies for this state of affairs.

What can we do?

In the face of some complacency, then, we must ask, what can be done to mitigate the income inequalities and their deleterious effects on social stability in the developing countries?

Of course the answer to this question is not easy; indeed there is no single straightforward answer. The phenomenon of income inequality is excessive and virtually universal in the Third World, but the constellation of circumstances, the social structures, and especially the locus of poverty, differ so widely from country to country that it is difficult to come up with a generally meaningful and operationally useful answer. For example, it is “generally true” that the per capita income in rural areas is lower than in urban areas; from this “generally true” observation it seems to follow that measures to raise agricultural income would automatically help to mitigate the rural-urban inequality of income. But it is also “generally true” that commercial farmers and rural landlords are among the richest groups in developing countries and that because of the source of their income they pay little or no taxes. Thus it would be a mistake, for instance, to attempt to alleviate rural poverty through tax relief or subsidies on agricultural production—because it may well be that the rich rather than destitute farmers would reap most of the benefits from such measures.

By the same token, it may be agreed that in many countries the redistribution of land would go a long way toward alleviating rural poverty; recently the general case for land reform has been strengthened by evidence that smaller holdings are by and large more intensively cultivated than larger ones, and given the abundant supply of rural labor, total production is likely to be increased by a redistribution of land. However, differences in soil fertility, climatic conditions, production techniques, and a host of other relevant factors make it virtually impossible to determine, in general terms; what the best pattern of land distribution should be and what rules and regulations must be built into a system of land reform to prevent an open, or clandestine, reversion to the system of inequality. Every case is sui generis, and for every case new rules have to be developed.

At this point a simple distinction may be useful. The problem of income inequality can be approached either at the macroeconomic level by policies affecting the economy as a whole, or it can be dealt with at the microeconomic level, on a case-by-case basis, in the selection, evaluation, and execution of individual development projects. This distinction is useful not only as an expositional device; its importance derives from the fact that a frontal attack on the uneven distribution of income by broad measures of economic policy is not likely except in a revolutionary situation. For the structure of income distribution is of course associated with the social structure which in turn determines the structure of political power. Any attempt to bring about changes in the distribution of income is bound to run up against economically and politically powerful groups, even in countries in which the need for reforms to improve social justice is widely recognized. Resistance to reforms is likely to be even stronger, and more effective, if the initiative toward policy changes comes from abroad, from foreign observers or advisers, or from national or international sources of development finance. Therefore a strategy, or at least an initial strategy, of “nibbling” at the problems of income distribution via financial and technical assistance—in which external bodies have considerable influence—may be more successful than advice on general policies giving preference to the economic advancement of the poor.

A first step

As a first step, which involves research and intelligence rather than explicit policy advice and guidance, it should be possible to collect, analyze, and publish data that measure and compare for a number of countries the rate of growth of income of the lower half (or 40 per cent, or the lowest third) of the population. Information of this sort would constitute a move away from the preoccupation with aggregate growth—the international pastime of growthmanship—and substitution for it of the idea of growth with social justice.

I realize that the comparison between national growth rates and growth rates of the income of lower income groups raises a host of issues in welfare economics to which it is hard to find conclusive answers. For example, is it “better,” in some objective way, to have a period of sustained economic growth during which per capita national income grows at 3 per cent, and per capita income of the lower income groups at 4 per cent, or a growth pattern in which the national average rate of growth is 7 per cent, but the average of the lower income groups only 5 per cent? If I confess that I do not know, I am in good (and large) company—that of the great majority of welfare economists. However, I have the feeling that the juxtaposition of the two patterns of a high growth rate with a low social justice performance and of a poor growth rate with a high social justice accomplishment, which implies the existence of a measurable tradeoff, is in reality hardly ever the way in which the problem poses itself.

A more realistic and more practical way to put the question is to ask: could the higher national growth rate have been achieved without a deterioration in the income distribution? And what can be done to maintain the growth rate while at the same time correcting the distribution of income? In other words, it is doubtful whether the idea of a conscious choice among various combinations of income growth and income distribution is useful and whether the notion of a tradeoff, while logically unassailable, has any operational significance. This implies that there is little or no connection between the pattern of income distribution and the rate of growth—a proposition which I suggest can be demonstrated as essentially correct.

But we must return from the rather slippery road of welfare economics and speculations on the ultimate determinants of economic growth to the subject at hand. From the short comments about the causes of income inequality this at least seems clear: that the chances are good that social justice can be successfully pursued by public concern with, and public expenditures for, rural development, education, and the elimination of urban unemployment.

Rural development

The suggestion that the economic growth of the agricultural sector should receive increasing attention is of course based on ample evidence that in virtually all countries for which data are available, the average per capita income in this sector is significantly below the national average and that a very high proportion of the poor are peasants. Moreover, data on income distribution are likely to underestimate the “welfare gap” between the rural poor and the generally less poor urban population, because income distribution data fail to take into account the scarcity of free public services—in education, public health, and medical services—in rural areas compared with the availability of such services in urban centers. The experience of the last 20 years has shown that the stimulation of agricultural development is difficult and costly. Nevertheless, the prospects that the benefits of this “green revolution” can be adapted and extended to meet the requirements of many countries open up the possibility that the lag in rural development, which seems to be virtually universal, can be closed. By this means one of the main causes of the growing inequality of income distribution would be eliminated.

In many countries the support of agriculture will require two kinds of measures: one, a more even distribution of land and, two, rural public works to absorb some of the rural underemployment. But even without tackling the politically and administratively difficult tasks of land reform and rural public works, much can be accomplished simply by directing a larger share of development expenditures (which must include current as well as capital expenditures) into agriculture. A study of one of my colleagues at the World Bank indicates rather conclusively that projects of agricultural development, especially in the form of credit schemes, irrigation works, and mixed schemes—involving land clearance, the provision of improved seeds or nursery stock, storage facilities, and above all, much extension service—will go a long way toward the elimination of income inequalities. Preliminary results of the study show that in the case of 42 out of the 53 projects investigated the income of the beneficiaries was below the national per capita income when the project was undertaken; after completion of the project the number was expected to be reduced to 33. More significantly, the medium income of the beneficiaries of all projects was expected to increase from 40 per cent of the national average to 63 per cent, and, incidentally, from 72 per cent of the national average farm income to 165 per cent. The last two figures indicate that these projects were selected without explicit regard to their income distribution effects. It is obvious that a conscious effort in the selection of projects for the distribution of income would have brought even better results in that respect.


Turning now to education as a means of attacking one of the prime causes of the uneveness of income distribution, it is clear that considerations of income distribution indicate that rural primary education should receive the highest priority, to be followed by the provision of education and training at the secondary level for the staff of agricultural extension services, and perhaps by nonformal adult education aiming at the rural population. This suggestion runs counter to the preference that in recent years development economists and educational experts have given to secondary education and vocational training. This preference is a secondary reflection of the preoccupation of the development economists with output, irrespective of income distribution. There is no need to throw it overboard, but to balance it off, in terms of overall budgetary allocation, with increased expenditures for primary education in rural areas.

Alleviating urban unemployment

The third avenue of a direct attack on the unevenness of income distribution is a policy to alleviate urban unemployment. In most countries this is not only one of the major causes of poverty and misery but also a significant focal point of social tension and political dissatisfaction, especially if members of the educated middle class as well as the urban poor are unemployed. Development economists have argued that the chief means of curbing urban unemployment must be the expansion of industrial production, if necessary through the persistent pursuit of substitutes for imports. This argument overlooks one important drawback: policies of import substitution, if they involve articles of mass consumption, might lead to a deterioration of the distribution of real income. More recently the expansion of industrial production for export is being promoted chiefly for the purpose of improving the balance of payments; but its effects in creating employment are also emphasized. The promotion of export industries is of course preferable, on the grounds that the production of exports is required by international competition to be efficient; the production of import substitutes is not subjected to any such discipline. But in view of the almost insurmountable obstacles in the world markets to a rapid expansion of industrial exports of developing countries and the likelihood that only a few of them will be able to break through the barriers of protection and commercial inexperience, the growth of industrial exports can be relied upon at best as making only a small contribution to solving the problem of urban unemployment. For these reasons one of the major remedial measures of unemployment in developed countries, the provision of public housing, should also have a place, and perhaps an important place, in the economic policies of developing countries—provided architects and construction technologists can provide realistic ideas about minimum standards of housing for the poor in developing countries.

However, irrespective of the specific measures or combination of measures that are taken toward equalizing income distribution, it should be clear that they should not involve outright redistributive measures which would be reflected in the price level. For example, schemes to support the prices of agricultural products for the domestic or the export market for the purpose of increasing farm income—the common method of income equalization in the advanced countries—or the introduction of maximum prices for staple products to keep down the cost of living of the urban population should be avoided because of resulting price distortions. On the other hand, there is a case for intervening in the price mechanism if the purpose of such an intervention is the stimulation of supply or demand, or the reduction of supply or demand for a limited, or perhaps even for a prolonged, period provided the income distribution effects of such measures are realized.

Outlines of a method

If the distribution of income can be influenced by the selection and preparation of specific development projects, is there a method, or a set of methods, by which this can be done?

The simplest and in a sense most radical method is to base the selection of development projects not on the size of the total net social return (or total net benefits), but on that part of total returns that accrues to beneficiaries below a certain level of income. Whether this approach is easy or difficult depends on the characteristics of the project. In an irrigation project, for example, the evaluation of the benefits should not be based on the increase in income of all beneficiaries but only that of the poorer beneficiaries. Since usually much is known about the distribution of land that can be brought under irrigation and presumably there exists a close correlation between the size of holdings and farm income, the distinction between rich and poor farmers should not be too difficult to establish. If the suggested criterion is applied, the selection would be “warped” in favor of the lower income groups, as against the use of the available water resources for other projects with fewer poor beneficiaries. If it turns out, however, that for technical reasons such as topography or soil characteristics it is reasonable to include a large area of farm land owned by rich farmers and landlords, then the project should be undertaken only if provisions for its financing are combined with provisions for the distribution of lands, or, as a minimum, for the regulation of land rent so as to assure the tenants of most of the benefits of the project.

In industrial projects, to follow the suggestions to exclude from the evaluation the benefits of richer beneficiaries would in most cases mean that profits would be disregarded; and two other types of benefits would become more important. One is the increase in wage income associated with the project, assuming that workers would not be bid away from other employment. This type of benefit is itself reflected in the net social return if shadow wages are introduced in the appraisal. The introduction of the increase in workers’ income on account of employment in the proposed project is usually done as a reduction of cost, but it can of course be added on the benefit side with the same result. The second modification would be to take account of the increases in the wage bill in the sectors supplying inputs to the new industry previously not produced, or, more generally to include on the benefit side those additions to the income of the lower income groups associated with the project. For a fruit or vegetable cannery, for example, the increased income of poor farmers supplying the fruit or vegetable should be considered a benefit.

Finally, improvements in the distribution of income can be achieved by applying shadow cost significantly below money cost to determine the social cost of employing members of the low income groups—which for practical purposes would be the urban unemployed and the rural underemployed—and to use the social cost in the choice of technology in the physical construction of projects. The application of this method would of course result in the more extensive use of labor as against capital equipment wherever there is a technological possibility of such a substitution. It must be realized, however, that this would increase the money cost of investment (by the difference between shadow wages and money wages) and that this difference should not be included in the capital cost of the project (and should be passed on to the consumers of public utilities if a policy of full cost pricing is applied) but should be absorbed by the fiscal authorities as, in effect, a welfare expenditure.

Any suggestions about income distribution are bound to run into all kinds of political, conceptual, and practical difficulties; I am sure that those I have offered here will do so. These difficulties should not, however, be insurmountable, because the application—even the rough and ready application—of such suggestions would make a significant contribution to the solution of one of the major problems of development. After all, it can hardly be repeated too frequently that the objective of development is not simply the increase in per capita income but the elimination or at least mitigation of poverty and human misery.

Finance & Development, September 1973
Author: International Monetary Fund. External Relations Dept.