Journal Issue

Poverty and progress—choices for the developing world: Conclusions from recent Bank research, and a pragmatic approach to poverty alleviation

International Monetary Fund. External Relations Dept.
Published Date:
June 1980
  • ShareShare
Show Summary Details

Hollis B. Chenery

Concepts of progress in most developing countries are heavily conditioned by their colonial past. Many of them express their objectives in terms of “catching up” with the advanced industrial societies and pattern their economies on this model. This tendency is reinforced by political objectives in countries that wish to acquire military power and influence.

One drawback to the emphasis on growth is that its benefits have usually been concentrated on the modern sectors of the economy, and increasing inequality of incomes has often led to political tensions. An alternative view of progress focuses more on achieving an equitable society and reducing poverty, with growth regarded as a necessary but by no means sufficient condition.

The postwar experience of relatively rapid growth in developing countries has provided a rich body of data on this set of relations that is only now being analyzed. Since there is relatively little established theory to guide this analysis, the collection of data and the formulation of hypotheses have gone hand in hand. Although substantial progress has been made in understanding the economic forces at work, the results to date are largely speculative and fall considerably short of the needs of policymakers. This article explores some of the implications of the prevalent views of progress in developing countries in the light of the information available on the results.

This article is adapted from a paper to appear in G. Almond, M. Chodorow, and R. Pearce (editors), Progress and Its Discontents, Stanford University Press.

Concepts of progress

Catching up. The material success of the industrialized West has been a powerful incentive to the rest of the world to adopt elements of Western experience that are conducive to accelerated growth. The success of countries with different historical backgrounds and economic and political systems has served to reinforce this objective.

The concept of catching up with the industrial leaders is a product of the industrial revolution and its outward spread from Western Europe. This concept both provides a goal for social action and suggests a means by which this goal can be achieved. The technology and forms of economic organization created by the advanced Western countries have provided the means for accelerated growth for countries in all parts of the world. Nations following this model have differed primarily in their choice of the economic and social elements to be incorporated in their societies.

The prototype of a successful process of catching up is Japan, whose economic structure and income level in 1910 were not significantly different from those of the poor countries of today. Econometric estimates of the sources of Japanese growth suggest that the process of borrowing technology from more advanced countries is now virtually completed and that Japan is likely to attain the income level of the United States by 1990 (Jorgensen and Nishimizu, 1978).

The Japanese example has had a powerful effect on Taiwan, Korea, Singapore, Thailand, and other countries of East Asia. All of these economies are now growing considerably faster than those of the advanced countries, and some may be able to complete the transformation from a state of underdevelopment to one of maturity in less than the 60 years taken by Japan.

Several of these East Asian countries provide modern approximations to the earlier idea of progress as a process in which “good things come in clusters” (Keohane, 1979). Unlike most developing countries, the benefits of growth have been widely distributed in Japan, Taiwan, Singapore, and Korea, and the incomes of the poor have grown almost as fast as those of the rich. Postwar governments have been growth minded and authoritarian but not very repressive, and these countries have ranked high on most indicators of social progress. In more typical cases growth has been achieved at the expense of increasing the concentration of wealth and income, however, and the poor have benefited much less.

Equity. Although the more equitable sharing of income features prominently among the political objectives of virtually all governments, it is taken much less seriously in practice than is the objective of rapid growth. Even though widespread government intervention in production and income distribution is justified largely on the grounds of reducing poverty, in fact most studies show that on balance the effects of government revenue collection and expenditure in developing countries favor the upper-income groups rather than the poor.

A few developing countries have, however, gone beyond the endorsement of equitable growth and have adopted policies designed to achieve it. Notable examples include the People’s Republic of China, Cuba, India, Israel, Sri Lanka, Tanzania, and Yugoslavia. Although their social goals vary with the form and extent of government control of the economy, there is a common emphasis on providing a minimum level of income to the poorest groups. In the more extreme socialist formulations, greater equality is considered a goal in itself, even if it is achieved with an adverse impact on efficiency—that is, lowering the incomes of the rich rather than raising those of the poor.

A pioneering attempt to reconcile the objectives of growth and the alleviation of poverty in an operational framework was made in 1962 by the Perspective Planning Division of the Indian Planning Commission (Srinivasan and Bardhan, 1974). This approach was based on a formulation in which the rate of poverty reduction in India was determined by the growth of the national income, while the extent of redistribution considered feasible was based on the experience of other countries. This approach has been refined in the concept of “Redistribution with growth” (Ahluwalia and Chenery, 1974), which forms the basis of the comparative analysis in the following section. If the idea of a feasible limit to the redistribution that can be achieved with a given set of institutions is accepted, the conflict between growth and distribution is reduced.

A further refinement in the concept of poverty alleviation has been achieved by shifting from the use of income as a measure of poverty to physical estimates of the inputs required to achieve minimum standards of nutrition, health, shelter, education, and other essentials. These indicators of basic needs provide a way of evaluating the effectiveness of any set of policies designed to reduce poverty (Streeten, 1979). The “basic needs” approach focuses particularly on the distribution of education, health, and other public services as a necessary element of policies designed to raise productivity and to alleviate poverty. This is an area in which some of the more effective socialist societies, such as the People’s Republic of China, showed marked improvement.

Formulating social objectives. The social goals of developing countries—and of international bodies representing them—tend to be stated in political terms that confuse ends and means and ignore the different dimensions of progress. For example, the goal of catching up with more advanced countries is a poor proxy for improving welfare because it often leads to emphasis on heavy industry and other policies that concentrate growth in the modern sectors of the economy. Similarly, many of the goals announced by international agencies, such as the attainment of given levels of nutrition, education, shelter, or industry, are misleading because they ignore the need to achieve a balance among the several dimensions of social progress.

The economist’s answer to this problem is to replace a set of separate objectives by a social welfare function that defines the goal of a society in utilitarian terms as the increase in a weighted average of income or consumption of its members over time. Although the national income is one such average, the typical income distribution gives a weight of over 50 per cent to the rich (the top 20 per cent) and less than 5 per cent to the poor (the bottom 20 per cent). If the growth of aggregate national income is used as a goal, it therefore implies giving 10 to 20 times as much weight to a 1 per cent increase in the incomes of the rich as to a 1 per cent increase in those of the poor (Ahluwalia and Chenery, 1974).

In principle, any set of weights could be applied to the income or consumption of different groups to remedy this bias. One possibility is to give equal weight to a given percentage increase in the income of each member of society, which is the equivalent of weighting by the population in each group. A more extreme welfare function, which corresponds to the announced goals of a few socialist societies, concentrates entirely on raising the incomes of the poor and gives social value to increasing other incomes only to the extent that they contribute to this objective.

Although there is no scientific way to determine the appropriate welfare function for any given society, the concept is useful in bringing out potential conflicts in the idea of progress and in deriving alternative measures of performance. It will be used for this purpose in the following section.

Experience with distribution

Perceptions of the nature of progress have evolved considerably as a result of the varied experience of the postwar period. Many of the early postcolonial governments set forth optimistic objectives that now seem highly oversimplified. However, there has also been a notable willingness to learn from experience in countries with varying ideologies. Equity-oriented countries such as the People’s Republic of China, Cuba, Sri Lanka, and Tanzania have found it necessary to give greater attention to economic efficiency and growth, while some of the leading exponents of rapid growth—Brazil, Mexico, Thailand, Turkey—are now taking poverty alleviation more seriously.

Although scholarly interest in these relations has expanded rapidly in recent years, the statistical measures needed to test and refine hypotheses are only now becoming available. Twenty-five years ago, Simon Kuznets addressed the question: “Does inequality in the distribution of income increase or decrease in the course of a country’s economic growth?” Although his answer was based on evidence for only a handful of countries and was labeled “perhaps 5 per cent empirical information and 95 per cent speculation,” it has provided the starting point for empirical work in this field (Kuznets, 1955). Kuznets hypothesized that the distribution of income tends to worsen in the early phases of development and to improve thereafter. This “U-shaped curve” hypothesis has been subsequently verified in several crosscountry studies based on samples of 50 or 60 developing countries (Ahluwalia, 1976).

There are several reasons for the earnings of middle-income and upper-income groups to rise more rapidly than those of the poor in the early stages of growth. Development involves a shift of population from the slow growing agricultural sector to the higher-income, more rapidly growing modern sector. In this process inequality is first accentuated by more rapid population growth in rural areas and ultimately reduced by rising wages produced by more rapid absorption of labor in the modern sector (Frank and Webb, 1977). The more capital-intensive type of development strategy followed by Mexico or Brazil absorbs less labor and produces greater concentration of income, while the more labor-intensive forms of Taiwan and Korea distribute the benefits of modernization more widely. A number of other factors, such as the greater demand for skilled than for unskilled labor and the concentration of public expenditure in urban areas, also contribute to growing inequality in many countries.

The Kuznets curve with country observations

Source: Ahluwalia Carter and Chenery (1979).

The units are 1970 U.S. dollars of constant purchasing power, which in the poorest countries are between 2.5 and 3 times the per capita income converted at official exchange rates.

My present concern is with the broader aspects of the relations between growth and distribution. How universal is the tendency toward less equal distribution in developing countries? Does it lead to an absolute decline in welfare for some groups? What kinds of policies have served to offset these tendencies? Is social conflict an inevitable concomitant of economic advance? Although none of these questions can be answered with great confidence, the average relationships and the variety of individual experience can be brought out by combining the available cross-country and time-series evidence for the postwar period.

The average relationship between rising income and its distribution is best shown by estimates of the Kuznets curve from data for all countries having comparable measures in some recent period (Ahluwalia, Carter, and Chenery, 1979). Although the variation in income shares was computed separately for each quintile, the general phenomenon is depicted in the chart by considering only two groups: the rich (upper 40 per cent) and the poor (lower 60 per cent). As national income rises from the lowest observed level to that of the middle-income countries, the share received by the poor declines on average from 32 per cent to 23 per cent of the total. In a hypothetical country following this average relationship, 80 per cent of the increase in income would go to the top 40 per cent of recipients.

The relationship between the income growth of different groups and that of the whole society can be brought out more clearly by expressing it in terms of the per capita income of each group. This is done in the chart, which plots the per capita income of the poor against that of the rich. Since the income level ˝Y˝ of the society is a weighted average of the two groups ˝a and b˝ (Y = .4Ya + .6Yb), the downward sloping straight lines define given levels of per capita income. Points on these lines indicate different distributions, and a growth process with a constant distribution is represented by a straight line through the origin, as in the case of Yugoslavia. A line deviating toward the vertical axis indicates growing inequality, as in the case of Mexico or Brazil. Growing equality is shown by Sri Lanka and Taiwan.

The Kuznets curve shown in this chart consists of two segments: a phase of worsening distribution up to an income level of about $800 (of constant purchasing power) and a phase of improving distribution thereafter. In the first phase the per capita income of the rich grows from about $300 to $1,600 while that of the poor increases from about $100 to $300. For the poorest 20 per cent, the rate of growth is considerably less. Since an increase in national income of this magnitude may take 40 or 50 years even with the relatively rapid growth rates recently experienced in developing countries, in the typical country the very poor cannot look forward to an annual increase of much more than 1 per cent—even though the economy is growing at two or three times that rate. Furthermore, there is nothing automatic about the improvement in distribution above $800, as shown by Mexico and Brazil.

Tradeoff between growth, equity. Although acceptable time-series data are only available for a dozen or so countries, they indicate a considerable variation around this average relation. The table gives selected measures of overall growth and of the share going to the lower 60 per cent for countries having observations for a decade or more.

Changes in income and its distribution
CountryIncome level1DistributionGrowth rates


share of bottom 60 per cent
(In per cent)

40 per cent

60 per cent




60 per cent
Ratio of

bottom 60

per cent

to total
Good performers
Taiwan (1964-74)56250875834136.938.539.
Yugoslavia (1963-73)1,00351882231635.736.
Sri Lanka (1963-73)388845810127.435.451.
Korea (1965-76)36254093827534.932.331.
Costa Rica (1961-71)82531145921223.728.433.
Intermediate performers
India (1954-64)226581132131.
Philippines (1961-71)336831553524.724.825.
Turkey (1963-73)56624341712820.824.
Colombia (1964-74)64823242210619.
Poor performers
Brazil (1960-70)6152144903124.820.615.
Mexico (1963-75)97444694411421.719.718.
Peru (1961-71)8342124356317.917.917.
Source: Ahluwalia, Carter, and Chenery (1979), Table 5.

Measured by per capita income expressed in 1970 U.S. dollars of constant purchasing power.

Source: Ahluwalia, Carter, and Chenery (1979), Table 5.

Measured by per capita income expressed in 1970 U.S. dollars of constant purchasing power.

They are divided into three groups according to the share of the increment in income going to the poor. The five good performers show over 30 per cent of the increment going to the bottom 60 per cent, while the three poor performers show less than 20 per cent. Whether distribution is getting better or worse is indicated by comparing these increments to the initial distribution and by the ratio of the growth of the per capita income of the poor to the national average in the last column.

This information, together with less complete data on other countries, provides a basis for describing the following patterns of growth and distribution observed in the developing world:

  • Growth-oriented pattern, illustrated by Brazil and Mexico.

  • Equity-oriented, low growth, illustrated by Sri Lanka.

  • Rapid growth with equity, illustrated by Taiwan, Yugoslavia, and Korea.

These cases illustrate the main types of deviation from the average pattern that can be observed in the 12 countries of the table; India, Turkey, the Philippines, and Colombia follow the average relations of the Kuznets curve.

These examples suggest the following observations on the relationship between income growth and social welfare in developing countries. First, a small group of countries has achieved rapid growth with considerable equity. In addition to Taiwan, Korea, and Yugoslavia, this group includes Israel, Singapore, and perhaps the People’s Republic of China. The policies underlying this successful performance vary from primary reliance on market forces in Taiwan, Korea, and Singapore to substantial income transfers and other forms of intervention in Yugoslavia and Israel. Second, substantial tradeoffs between growth and equity are illustrated by the other cases. Although Sri Lanka has grown much less rapidly than Mexico or Brazil, the poor have done considerably better in the former case. Cuba presents an even more extreme trade-off, since the welfare of the poor has risen despite a continuous fall in the nation’s per capita income since 1960 (Seers, 1974).

Only in the few cases where economic growth has been both rapid and fairly equitably distributed is it possible to make unambiguous comparisons among countries—or among different development strategies for a single country. In other cases it is necessary to define some properties of a social welfare function to make such comparisons. To take two extreme cases from the table, the incomes of the poor have grown nearly four times as fast over a decade in Sri Lanka as in Brazil, while the opposite is true of the incomes of the rich. Since the latter receive greater weight in the national income, per capita income has grown 50 per cent faster in Brazil; conversely a population-weighted index of welfare increases 50 per cent faster for Sri Lanka. Even this limited sample therefore demonstrates that judgments about economic progress cannot be separated from social and ethical postulates.

Reducing world poverty

Attempts to extend the concept of material progress to a global scale run up against more acute problems of equity than the national issues described above. Although most governments recognize their national income as one dimension of national welfare, no one has suggested that global income has much relevance to an assessment of global welfare. Instead political and economic efforts of international institutions are increasingly focused on the reduction of poverty and other aspects of equity as objectives that command the support of people of widely varying political views.

In recent years considerable efforts have been made to establish measures of poverty based on standards of nutrition, health, shelter, education, and other essentials. Conservative estimates set the proportion of the world’s population that falls below a poverty line based on such minimum standards at between 20 and 25 per cent. Although this proportion has declined somewhat in the past 30 years, the overall increase in the world’s population has meant that the absolute number of people below this poverty line has continued to grow and is currently of the order of 800 million.

In technical terms the reduction or even elimination of world poverty seems deceptively easy. If resources could be shifted to satisfying the needs of poverty groups efficiently, it would only require a reallocation of 2 to 3 per cent of the world’s output per annum from 1980 onward to meet the identifiable costs of eliminating poverty by the year 2000 (Streeten and Burki, 1978). Since three fourths of the world’s poor live in very poor countries, however, the annual cost of eliminating poverty in these countries is more meaningfully stated as equal to about 15 per cent of their gross national product (GNP), even if expenditures could be designed to serve only the target groups. In the light of the distributional experience outlined in the previous section, the problem is seen to be vastly more difficult.

Some of the principal constraints to a more realistic attempt to reduce global poverty include:

(1) The multiple objectives of nation states, among which the alleviation of poverty is usually subordinated to a variety of nationalistic goals.

(2) The limited scope for resource transfers in the existing international economic order. Official development assistance from the industrialized countries has declined from 0.50 per cent of their GNP in 1960 to 0.35 per cent or less since 1970. Transfers from the Organization of Petroleum Exporting Countries (OPEC), while substantial, do not offset the negative effects of higher oil prices on the growth of the oil importing developing countries.

(3) Rapid growth of population, which will double in the next 35 years even though the rate has started to decline.

What are the possibilities of more rapid progress in the face of these and other constraints? In an attempt to compare approaches to poverty alleviation, Ahluwalia, Carter, and Chenery have simulated income growth and the numbers of absolute poor over the next 20 years for a large sample of developing countries (Ahluwalia, Carter, and Chenery, 1979, Tables 3 and 9). If the trends of the past 20 years—a period of relatively rapid growth of income—continue, the number of absolute poor in 2000 would be at about the same level as in 1960. This represents rapid progress in one sense, since the proportion of the poor would fall from 50 per cent to 20 per cent of the population of developing countries. However, since this result would be achieved only by a reduction in absolute poverty in middle-income countries that offsets the rising numbers in the very poor countries, it is not a long-term solution.

The reduction in poverty will have to come from one of three sources: improved distribution, accelerated growth, or a more rapid decline in population growth. Improved distribution is particularly important in many middle-income countries, such as those in Latin America (where income is quite unequally distributed), but some acceleration of growth is essential in the poor countries of Africa and South Asia. Although there are some short-term tradeoffs between growth and distribution, in the longer term it is more likely that all three types of policy will be mutually reinforcing. Even within restrictive limits to capital transfers, the industrial countries can considerably improve the outcome by giving greater priority to poverty alleviation in allocating aid among countries (Edelman and Chenery, 1977).

These projections lead to the conclusion that although the elimination of poverty is much more difficult than is sometimes suggested, it remains a plausible goal for international policy. One of the principal means to this end would be accomplished if the tendency of the poor to lag behind the higher-income groups in the process of development could be eliminated. There is increasing acceptance of the idea that international efforts should be more directly focused on reducing poverty in order to offset this tendency of the international system. Enough examples of how this result can be accomplished have been cited in economic systems ranging from socialist to free enterprise to suggest that it is a feasible objective.

This conclusion leaves several fundamental issues unresolved. To what extent should poverty alleviation replace the principle of self-help as a guide to international action? To achieve this objective, will it not be necessary to establish enforceable standards of performance to assure that the benefits actually reach the poverty groups? The new emphasis on poverty alleviation does not resolve these old dilemmas in the field of international economic cooperation. It may even accentuate them.

Related reading

Other Resources Citing This Publication