Main Features of the Employment Problem in Developing Countries

International Monetary Fund. Research Dept.
Published Date:
January 1973
  • ShareShare
Show Summary Details

IN THE PAST FEW YEARS there has been an increasing concern about the present and prospective inadequacy of employment opportunities in developing countries. It emanates partly from a retrospective view of the course of events in the years since the end of World War II. The economic performance of the developing countries in this period as measured by their rate of economic growth has, by historical standards, been quite creditable, and yet their predicament with regard to the employment problem apparently has been getting progressively worse. Dissatisfaction with the past and disquiet about the future are serious enough to raise doubts regarding the whole present orientation of economic policy in developing countries and have resulted in a channeling of research efforts to employment issues. There is also increasing emphasis on employment as an objective of economic policy in the developing countries, and it appears that emphasis on employment issues will continue in the future. This paper surveys some of the recent work pertaining to this general subject and alludes to some general implications for economic policy directed toward these problems.

The paper is in five parts. Section I presents data about recent labor-force trends in the developing countries, puts them in a historical perspective, and points to the crucial role of population growth in shaping the employment problem of the developing countries. Section II discusses briefly the conceptual difficulties in coming to grips with the problem and suggests that it has two components—open unemployment and disguised unemployment. Section III, dealing with open unemployment, cites some of the available quantitative information, an exercise that is not possible in Section IV, which is about disguised unemployment. In fact, much of Section IV is concerned with the most appropriate definition of disguised unemployment, with the final choice leaning toward a definition based on an income test. The choice of an appropriate definition is of interest because of its unavoidable influence in determining the focus of economic policy in its approach to the problem. Finally, Section V summarizes the main conclusions regarding the overall approach of economic policy to the employment problem in the developing countries.

I. Labor-Force Trends

The dimensions and important features of the employment problem in the developing countries are revealed by the estimates and projections of economically active population1 in the main regions of the world (Table 1). First, of the total economically active population in the world in 1970,2 about two thirds lived in the developing countries and one third in the developed countries. Thus, the employment opportunities needed in the developing countries were twice as many as those needed in the developed countries. Second, the projected increase in the labor force of developing countries in the 1970s in absolute terms is strikingly large. At 169 million it is more than half of the total labor force of the developed countries in 1970 and over five times as large as the projected increase in them. Such an absolute increase means that there is need for the creation of that many additional employment opportunities by the end of the decade. This large absolute increase stems partly from the large base and partly from a more rapid rate of growth, the projected rate of growth of the economically active population in the developing countries in the 1970s (26 per cent for the decade) being more than twice that for the developed countries (10.2 per cent). A third noticeable feature is the acceleration in the rate of growth of the labor force of developing countries during the period under review; the annual growth rate is projected to rise from 1.6 per cent in the 1950s to 2.3 per cent in the 1970s. This contrasts with the experience of the developed countries, where it is expected to decline marginally, from 1.1 per cent in the 1960s to 1 per cent in the 1970s.

Table 1.Economically Active Population in the World,1 1950-80
Active Population



(Per cent)
Compound Growth Rate

(Per cent per annum)
World 16988029531,15310315120014.818.821.
Developing regions 24525296498177712016917.022.726.
Africa 38810312615715213117.521.825.
Asia 4308356435547487911215.522.325.
Latin America 556708811414182524.325.928.
Developed regions24627330433527313110.911.410.
North America66779010511131516.216.317.
Other 6911131622321.822.422.
Source: International Labor Office, Labour Force Projections, 1965-1985 (Geneva, 1971).

Excluding countries that are members of the Council for Mutual Economic Assistance, mainland China, North Korea, and North Viet-Nam.

Africa (excluding South Africa), Latin America, Asia (excluding Japan), and Oceania (excluding Australia and New Zealand).

Excluding South Africa.

Excluding Japan but including Oceania with the exception of Australia and New Zealand.

Including the Caribbean.

Australia, New Zealand, and South Africa.

Source: International Labor Office, Labour Force Projections, 1965-1985 (Geneva, 1971).

Excluding countries that are members of the Council for Mutual Economic Assistance, mainland China, North Korea, and North Viet-Nam.

Africa (excluding South Africa), Latin America, Asia (excluding Japan), and Oceania (excluding Australia and New Zealand).

Excluding South Africa.

Excluding Japan but including Oceania with the exception of Australia and New Zealand.

Including the Caribbean.

Australia, New Zealand, and South Africa.

These main features of the labor-force projections show forcefully the extent to which the impending employment problem in developing countries is attributable to the high rates of growth of the labor force. These are high not only in comparison with the rates of growth experienced by the developed countries at present but also, and especially, when they are put into an appropriate historical perspective. When the developed countries of the present day3 were at the stage of economic organization and development comparable to that of the developing countries today, their labor force grew at rates in the range of about 1-1.5 per cent a year.4 Furthermore, they could rely on the escape valve of emigration to the empty regions in the New World that is no longer available. The present rate of growth of the labor force in the developing countries is unique in human history. Until now, much of the concern with the issue of population growth has been focused on the Malthusian problem of the adequacy of food supplies. What is now being realized is that even when the Malthusian specter is warded off, the burden of productively absorbing a fast-growing labor force into a relatively weak economic system can also be unbearable. Training and effective utilization of a labor force increasing by about 25 per cent a decade would pose a difficult challenge even to the flexibility and resiliency of a developed economy; such a challenge may well be insuperable for economies already in the midst of the difficult task of initiating economic development. Hence, in any long-term strategy for dealing with the employment problem of the developing countries, an effective population policy must have a central role. It is true that the entrants to the labor force in the 1970s were already born before 1965; therefore, population policy can help in the long term but can make little contribution to solving the problem within the immediate time perspective of 10 to 15 years. Other measures must attempt to do this. However, this consideration in no way reduces the high priority of adopting appropriate population policies in the short term.

The word “problem” was used in this section in describing the size, growth, and geographical distribution of the labor force in developing countries; but, in fact, the presence or absence of a problem should depend on the balance between the growth of the labor force and that of employment opportunities—in other words, on the level and trend of the rates of unemployment. However, an attempt to present a systematic summary of the facts regarding rates of unemployment runs into several snags.

II. Nature of the Employment Problem

For a host of reasons, the data difficulties common to the analysis of the economic problems in developing countries are particularly pronounced with regard to unemployment—and they are further compounded by the circumstance that the conceptual framework into which such information is presented is ill suited for a clear understanding of the situation in the developing countries.5 This conceptual framework was evolved in response to the unemployment experience of the industrial countries in the 1930s. Essentially, it is centered on two ratios—the “participation rate” and the “unemployment rate.” The participation rate is the ratio of the economically active population to total population, economically active population (or labor force) being defined as the sum of those at work and those not at work but seeking it. The unemployment rate is simply the ratio of those not at work but seeking work to the total labor force. The construct based on these two ratios is found seriously wanting when applied to the labor-market phenomena in developing countries.

The main problem with the notion of the participation rate is the predominance of household enterprises and nonwage employment in the developing countries. A substantial part of the labor in such enterprises is usually provided by the members of the household. As the demand for labor fluctuates seasonally (especially in agriculture), members of the household participate in economic activity (i.e., enter the labor force) or slip back into household work. Such participants in economic activity, and they are in large numbers in many countries, have only a partial attachment to the labor force as the concept is formulated, because when they are not at work, they are usually not seeking work either. Moreover, as a general rule, they are unlikely to be available for work outside the household.6 Historically, one of the concomitants of economic development has been a secular decline in this pattern of labor utilization, a process that is now being initiated in the developing countries. But in their present situation, the approach that classifies everyone as being either economically active or economically inactive results in a concept of a labor force that is ambiguous when sought to be applied and therefore cannot be quantified with a reasonable margin of error. Notwithstanding these problems and ambiguities, on balance a quantifiable labor-force concept does provide useful information regarding the magnitude of the potential demand for employment opportunities, as illustrated, for instance, by the discussion in Section I.7 However, the difficulties besetting the concept of an unemployment rate are more complex and in some ways insurmountable.

One major problem is simply a corollary of the ambiguity attaching to the concept of the participation rate. It arises from the situation in which a substantial part of the working force is at times employed but at other times is greatly underemployed or even not employed, and yet is not quite unemployed in the sense of being freely available for employment outside the household. In this environment, a slowdown in economic activity is accompanied partly by work spreading within the enterprise and partly by shortened working hours, but rarely would the owner regard himself or his dependents as being available for alternative employment. This type of adjustment pattern is a natural reflection of the limited geographical/occupational mobility of labor as well as the omnipresence of slack labor markets. Natural as these phenomena are, for they are found to some extent in developed countries also, their widespread existence and overwhelming importance in the overall problem of labor utilization in developing countries render the notion of an unemployment rate of doubtful relevance. Under the usual survey procedures, “the unemployed” are defined on the basis of answers to these two questions: (1) Are you employed? If the answer to this is in the negative, the second question is (2) Are you seeking employment? An affirmative answer to this question then identifies the respondent as an unemployed. In the circumstances of the developing countries—and particularly in their traditional/rural/agricultural sectors (which, although they overlap substantially, are not identical)—the answers to these two questions too frequently fail to reveal the existence of an employment problem and generate statistics of doubtful policy relevance, since slight variations in survey procedures can cause large differences in recorded unemployment rates.8 Another possible effect on the supply of labor that may be lost sight of in the standard labor-force surveys is the so-called discouraged worker phenomenon. It refers to the discouraging effect of a continually slack labor market on the readiness of the unemployed to search for jobs. Those who cease looking for jobs would not be included in the economically active population or in the unemployed, thereby imparting a downward bias to the recorded unemployment rate. Because of the universality of slack labor markets in the developing countries, this phenomenon may have a significant, although unknown, effect on the measured unemployment rates.

These strictures must not be carried too far. The construct of an unemployment rate does retain considerable relevance for the urban/ modern sector in developing countries, and the policy significance of the data based on this construct is likely to depend on the size of the labor force in the urban/modern sector in relation to the total labor force in the economy.9 The conclusion to be drawn is that “open unemployment” as reflected in unemployment rates estimated in accordance with the usual procedure is one salient aspect of the overall employment problem in the developing countries and that its policy significance can be expected to vary widely from country to country; other aspects of the problem can, for analytical purposes, be brought under the heading of “disguised unemployment.”

III. Open Unemployment

The prospects of objective measurement and usable quantitative information can be expected to be greater for the open unemployment portion of the problem than for disguised unemployment, both because the necessary conceptual framework is clearly established and because available data have been collected largely within this framework. So it comes as something of a surprise to discover how limited is the available quantitative information, especially in the form of time series from which the existence of any trends could be ascertained.10Table 2 presents 12 series for the decade of the 1960s (including 3 series for urban areas) for 11 developing countries. These series do not record any rising trend in the rates of open unemployment. Only one of the series, that for Gran Santiago, registers an increase in the period 1963-70 and even that is hardly dramatic, the unemployment rate in 1970 being less than that in 1960. On the other hand, three of the series (Chile, China, and Korea) reveal a definite declining trend in the rate of unemployment; two of these three countries, however, experienced exceptionally high rates of economic growth during this period and so are somewhat unrepresentative of developing countries as a whole. The remaining 8 series in the table cannot be said to display any trend whatsoever. The country coverage of these data is too limited to permit any firm conclusion. But their acceptance as relevant although partial evidence tends to support a Scottish verdict of “not proven” on the hypothesis of rising rates of open unemployment in the developing countries. If it is indeed true, as so many perceptive observers are saying, that the employment situation in the developing countries deteriorated steadily during the 1960s, then the evidence marshaled here tends to corroborate the idea that recorded open unemployment rates are inadequate indicators of the overall employment problem of the developing countries.

Table 2.Open Unemployment Rates in Selected Countries Based on Sample Survey Data, 1960-701(Percentage of labor force)
Developing countries
Argentina (Gran Buenos Aires) 35.7* 45.4
Chile (Gran Santiago)
Colombia (Bogotá)
Trinidad and Tobago14.
Developed countries
Sweden1.5 61.7
United States5.
Sources: International Labor Office, Year Book of Labour Statistics (Geneva, 1970 and 1971). For the Colombia series, see [4], p. 361.

Asterisks refer to different months from elsewhere in the series.

Arithmetic average of the years for which there are data.

Only the urban area.


Sampling design revised.

Slightly different coverage from the rest of the series.

Sources: International Labor Office, Year Book of Labour Statistics (Geneva, 1970 and 1971). For the Colombia series, see [4], p. 361.

Asterisks refer to different months from elsewhere in the series.

Arithmetic average of the years for which there are data.

Only the urban area.


Sampling design revised.

Slightly different coverage from the rest of the series.

There is relatively more information11 available regarding open unemployment rates in urban areas based on occasional sample surveys, and some of it is presented in Table 3. The variation in the unemployment rates among different countries in this table (as in Table 2) needs to be interpreted with caution, since it is likely to have been influenced considerably by differences in definitions and survey procedures, and an impressionistic approach is the most appropriate in interpreting them. Two broad conclusions can be drawn. First, the urban unemployment rates in Table 3 generally range considerably higher than the country-wide rates in Table 2. Whereas only 1 observation out of 12 in the “average” column of Table 2 exceeds 10 per cent, 8 out of the 16 in Table 3 do so, confirming the belief that the employment problem manifests itself as open unemployment to a much greater extent in the urban area than in a country as a whole. Second, notwithstanding all the definitional problems, these recorded rates are generally high enough to be a matter for serious concern. For instance, 9 out of the 12 average rates for the countries in Table 2 and 12 out of 16 observations in Table 3 exceed the highest average rate among the developed countries in Table 2, viz., 5.2 per cent for Canada. The existence for any length of time of such high open unemployment rates is bound to be a source of social strain anywhere.

Table 3.Open Unemployment Rates in Urban Areas, Based on Occasional Sample Surveys, 1960-69(In per cent)
Survey YearUnemployment Rate
Argentina (Buenos Aires)19654.2
Iran (Teheran)19664.6
Korea (nonfarm households)196612.6
Malaysia (metropolitan towns in Malaya)19659.8
Thailand (Bangkok)19663.4
Source: [17], Chapter III, Table 2, pp. 58-60.
Source: [17], Chapter III, Table 2, pp. 58-60.

Certain other facets of the urban open unemployment problem, as revealed by various survey inquiries, are of some interest. The evidence, however, is fragmentary, and all generalizations can be only very tentative. It appears that the incidence of unemployment among age groups is highly uneven and tends to be substantially greater among the young. As a corollary of this, the proportion of the unemployed who are “inexperienced” workers or “new entrants” to the labor force tends to be high. Of course, it is natural for there to be some tendency for the young, new entrants to the labor force to have a relatively higher rate of unemployment; the point is that this tendency is strikingly strong among the urban unemployed in the developing countries.12 Another finding that contrasts with the experience of developed countries is that the unemployed as a group, and particularly the young, tend to be better educated relative to the average for the labor force as a whole. Also, there is a preponderance among the urban unemployed of individuals with a dependent status, that is, who are not heads of households. This is hardly surprising. For, in societies lacking any extensive unemployment insurance or income-support programs and for households with low net worth, open unemployment of the breadwinner is a disaster to be avoided at all costs. If an employment opportunity at a desired rate of remuneration is unavailable, the head of the household must accept any employment or enter any service or trade that will enable him to eke out a living. He cannot afford to be in the ranks of the openly unemployed for any length of time. He may work or be “at work” for very long hours indeed to earn very little, but he will not be unemployed.

These characteristics of urban open unemployment and the existence of a substantial differential between real earnings in the modern sector and the traditional sector of the developing countries suggest that urban open unemployment should be viewed against the perspective of the overall employment problem. Historically, the process of economic development, which includes increasing specialization, a more elaborate division of labor, harnessing of scientific knowledge to economic uses, growth in the scale of enterprises, and industrialization, among other things, has also required a gradual transfer of labor from the rural/ agricultural sector to the urban/modern sector. This transfer can be thought of as being mediated by a pool of the unemployed in the modern sector from which the modern sector’s growing labor requirements are recruited and which in turn is replenished by the migration of labor from the rural areas.13 The size of this pool (i.e., the rate of urban unemployment) will be influenced by, among other factors, the gap in the rate of real earnings in the two sectors. The wider this gap, the longer the period, ceteris paribus, for which people would be willing to be a part of the waiting pool of the unemployed in the expectation of getting a well-paid job in the modern sector, and therefore the higher will be the recorded unemployment rate.14 To the extent that this hypothesis is valid, and there is some evidence to support it,15 a faster growth of employment opportunities in the urban/modern sector would not make much of a dent in the observed rates of open urban unemployment as long as the rural/urban real wage differential remains unchanged; the migration from the rural/agricultural sector will simply accelerate to maintain the relative size of the unemployed pool. Incidentally, this hypothesis also provides a plausible explanation of the strikingly higher incidence of open unemployment among the urban young. Many new entrants to the urban labor force may find it rational to remain unemployed for considerable periods in the hope of getting a highly paid job in the modern sector. Over the years, they either succeed or, when they finally give up hope, are absorbed into the traditional sector, so that the observed unemployment rates for higher age groups are at all times significantly lower than those for younger age groups.

The implication of this line of reasoning is that if the aim is to limit open unemployment in urban areas, a policy designed to moderate the rural/urban real wage differential may be more effective than a policy designed to create more jobs in the urban/modern sector, although the latter policy will reduce the disguised unemployment in the rural sector.16 Our understanding of these processes, however, is so imperfect and superficial that this line of reasoning must not be pressed too far. It is obvious that the rates of rural/urban migration and of urban open unemployment during any period will be influenced by a host of economic and noneconomic circumstances. Nevertheless, the significance of the insight provided by this viewpoint is that the urban open unemployment problem is seen as a part of the overall employment problem and the focus of policy is enlarged beyond the urban/modern sector. Restraint on real wages in the modern sector would have to be an important part of such a policy. In addition, policy measures that raise rural/ agricultural real incomes and that make rural living more congenial by improving infrastructure facilities and generally redirecting public sector expenditures can also be seen to have relevance for open unemployment in the urban areas.17

IV. Disguised Unemployment

In dealing with open unemployment, the experience and practices of the developed countries have provided a readily available and clearly articulated conceptual framework that could be applied in the collection and organization of data. This advantage is not available when attention is turned to the concept of disguised unemployment, which is here intended to be a category that would encompass those aspects of the employment problem of the developing countries that are not caught in measured open unemployment rates.18 The first important task therefore is to define disguised unemployment in an operationally useful way. To be so useful, the concept should permit maximum coverage of those groups among the nominally employed who figure most prominently in the thinking of policymakers when they allude to the employment problem in the developing countries; second, the segments of the labor force covered by it should in principle be identifiable and, as far as possible, the concept should be measurable. In the literature of employment problems in developing countries, three approaches to the notion of disguised unemployment are discernible: one based on the idea of a “labor surplus” in the traditional sector, another based on explicit underemployment in terms of hours worked, and finally the so-called productivity, or income, approach. These are considered in turn.

The notion of disguised unemployment as a labor surplus in the agricultural sector of developing countries has a long genealogy going back to Rosenstein-Rodan’s paper on post-World War II problems of eastern and southeastern Europe, and subsequent recurrence in many growth models, including those of Ragnar Nurkse, W. Arthur Lewis, and others.19 However, in all these formulations it is not regarded primarily as an employment problem nor is it part of a theory of employment; rather, it is incorporated in a theory of economic development and regarded as a crucial feature of the whole development problem. At the heart of this theory is the hypothesis that the agricultural sector in the developing countries contains a significant amount of surplus labor that makes no contribution to output, so that it is in principle possible to remove this labor without any reduction in production,20 even though no change is made in methods of production or organization. The theory then points out that this labor surplus is a potential resource for capital formation and economic development. In order to realize this potential, two steps are necessary: the surplus labor must be transferred to the task of capital formation, and the wage goods that the transferred labor previously consumed on the farm must be extracted from the agricultural sector and be made available for their sustenance in their new activity. This, if done successfully, initiates the process of growth in capital, output, and employment.

The disguised unemployed according to this approach are in effect those who are the least productive among the agricultural labor force and therefore are the most “available” for transfer outside. Hence, the labor-surplus concept of disguised unemployment appears to satisfy the first criterion of operational usefulness, viz., the coverage of those portions of the labor force about which the policymakers in the developing countries are concerned when they refer to the employment problem. But it performs poorly in relation to the second criterion of identifiability and quantifiability. There are several difficulties. The hypothesis of zero marginal productivity of labor in traditional agriculture is very difficult to validate empirically. In fact, some evidence contradicts it. Empirical studies in several developing countries have shown that per acre output on smaller farms is greater than that on larger farms. The most plausible explanation for this is that smaller cultivators can extract more from a unit of similar land by tilling it more intensively, which implies a positive marginal product for labor input. If this is indeed so, the surplus labor, which is the basis of the theory, tends to become elusive. And even if one were to assume that the marginal product of labor is zero, the crucial practical consideration is how much of an organizational change is to be permitted in order to prevent a decline in output when the surplus labor is removed. If literally no change is made, then output will inevitably go down, because the existence of surplus labor does not mean in practice that while some people are at work, others are sitting idle. Rather, it means that everyone is at work but labor is being used wastefully. If output is to be maintained after the removal of some labor, additional effort is called for by those who remain. But once changes of this sort in the production process are accepted into the argument, it becomes totally unclear as to where the line should be drawn: Should the changes be limited only to working hours and work patterns? Should consolidation of fragmented holdings also be permitted? Is it not natural to permit some changes in cultivating techniques called for by the altered land/labor ratio? and so forth. In fact, it is not meaningful to talk of a labor surplus independently of the other measures that are to be taken to maintain output. The concept has no meaning apart from specific supporting agricultural policies. Once this is recognized, it is logical to define it only in the context of a specific package of policies, the contents of which need not be limited by arbitrary restrictions on the degree of organizational change but should in fact be determined directly by whatever is practicable during a given period of time. Second, the possibility of mobilizing the labor surplus also depends on preventing an increase in the consumption of those who remain behind on the farm, so that there would be means available for sustaining those who leave the farm. Another necessary condition is that this transferred surplus labor should be willing to carry on the new capital formation activity without receiving a higher level of real consumption. Thus, the size of the available surplus labor for re-employment is very much a function of a matrix of policy measures. It follows that it has no independent existence apart from a specific development strategy covering agricultural as well as industrial policy.21 Moreover, it is not possible even in principle to identify the disguised unemployed at any given time. Hence, this notion of disguised unemployment, valuable though it is in a conceptual sense, cannot be converted into a definition that is operationally useful.

Another approach to disguised unemployment simply focuses on those who are literally underemployed, viz., they are at work for a certain number of hours but would prefer to work for longer hours. This concept is quantifiable, especially in the modern sector, because it is easy to adapt the open unemployment construct for this purpose. In addition to covering those who are entirely unemployed, a survey would cover those whose hours of work are less than a certain specified number, for example, less than 25 hours a week. Thus, it is also possible to identify those whom it covers. However, it suffers from the same difficulties as the open unemployment concept because it is not capable of covering most of those who are regarded as constituting the employment problem in the developing countries. In the urban areas, it leaves out those who, for a bare pittance, may nominally be working long hours indeed. In the traditional/agricultural sector it can, with appropriate adaptation, come to grips with the problem of seasonal unemployment; apart from this, it is no more helpful than the open unemployment concept and therefore should be regarded as merely one facet of some interest to the overall employment problem.

More recently, there has been a tendency to favor the so-called productivity, or income, approach to defining the employment problem. It has been argued that the focus on hours worked is misplaced in the context of the developing countries. The employment problem there is a disguised one, not because the hours worked are too few but because in an overwhelmingly large number of situations labor utilization is extremely unproductive. A deterioration in the overall employment situation may be reflected partly in the growth of open unemployment, but its principal manifestation will be in expanding the ranks of those who are in low productivity occupations and activities. The focus, therefore, should be on the productivity of labor, and the definition of the problem should be made coterminus with situations of extremely low labor productivity. Moreover, at the lower end of the income distribution, and particularly for small farmers, real earnings are a good proxy for productivity. Approached this way, the employment problem becomes synonymous with poverty. The disguised unemployed would then be defined and identified as those whose annual real earnings are below a certain bench mark. The annual rate of earnings is selected rather than for some shorter unit of time because of the substantial variation in the rate of real earnings resulting from seasonal fluctuations in the intensity and pattern of labor utilization in the developing countries. Such a low-income definition of disguised unemployment satisfies the two criteria of operational usefulness. It performs well by way of covering those who are most in the minds of policymakers in their concern about the worsening employment problem in the developing countries. It also makes the coverage of the concept identifiable and, in principle, quantifiable. The number of those included would depend on the real income level at which the bench mark is established, and this decision would be governed by economic as well as social considerations. In each country the line may be drawn in a way that is suitable to the peculiar conditions and circumstances of that country.

The acceptance of such an income approach toward the definition of disguised unemployment in developing countries means that for policy purposes there is a shift of emphasis from the creation of jobs as such to the improvement in productivity, and thus in real earnings, on the lower rungs of income distribution. Such a shift of emphasis can be regarded as appropriate and desirable for several reasons. After all, what individuals and households ultimately desire is higher real income; employment is not desired for its own sake but as a means to higher real earnings. It is argued occasionally that employment is generally desired not merely as a means of subsistence but also because it yields certain nonmaterial benefits. While this is undeniable, it is quite right to emphasize that the desire for employment arises primarily from the fact that it is a source of income. For were it not so, it would be easy to find solutions to the employment problem by way of creation of employment opportunities on an honorary basis. The “income-cum-productivity” approach to the employment problem also focuses attention on the fact that any solution must involve either a redistribution of the given gross national product (GNP), or a faster growth of it, or some combination of the two. There are three possible ways of increasing the real income of the disguised unemployed: by increasing their productivity in their current occupations, by creating employment opportunities for them in more productive occupations, or by augmenting their real income through some redistributive device. The first two approaches simultaneously augment GNP and channel the increment in the desired direction, whereas the last one alters the distribution. Looking at possible solutions to the employment problem from this particular viewpoint reveals the basic conditions under which employment and output objectives are complementary and those under which they may be in conflict. Insofar as the increase in the real earnings of the disguised unemployed, or of the openly unemployed for that matter, is associated with an equivalent increase in their productivity, employment and output are both higher. However, if the increase in real earnings exceeds the increase in productivity, there may be an increase in aggregate consumption at the expense of investment. To the extent that investment is forgone, growth in GNP will be sacrificed. The conflict here is, in essence, the unavoidable one between current consumption and economic growth. Finally, the shift of emphasis toward income distribution issues resulting from this approach is in line with the growing concern that the incremental real GNP has not percolated to the lower reaches of the income distribution pyramid. The income-cum-productivity definition of disguised unemployment in effect integrates the employment and the income distribution problems for policy purposes.

The foregoing discussion indicates that the employment rate and its obverse, the unemployment rate, have different meanings and policy implications for developed and developing countries.22 It is meaningful to define full employment for a developed country as being equivalent to a certain rate of unemployment and to adopt that rate as a target variable for macroeconomic policy.23 However, the situation is different for a developing country for two broad reasons. The first, discussed in Section II, is that the whole conceptual framework is of limited applicability. The participation rate concept becomes ambiguous when applied in the developing countries, and the rate of open unemployment is of limited significance as an index of the overall employment problem. Consequently, there is no operationally meaningful definition of full employment applicable to the developing countries. Second, unemployment in the developed countries is a cyclical phenomenon, and its solution lies primarily in demand management policies designed to achieve full employment.24 In contrast, many of the developing countries are frequently operating under pressures of excess demand but are making hardly any dent in the employment problem, which is basically a problem of productive capacity.25 The solution lies not so much in measures to activate employment opportunities in the short run as in measures to increase the productivity of labor through improvements in the pattern of its utilization.

In this context, full employment as an objective of economic policy and unemployment as an economic problem are both relatively recent developments. In fact, the first appearance in the English language of the word “unemployment” to describe a social problem occurred only about the 1880s, and the notion that full employment is a goal to be achieved by economic policy dates back to only the 1930s. In a sense, these developments were a natural by-product of the emergence of developed industrial societies in which a large part of the labor force had come to be employed on a contractual basis. Full employment as a policy objective and unemployment as a problem would have been totally alien to the preindustrial societies in Europe. And it is to those societies that the developing countries of today are most similar in the essential nature of their economies.

V. Conclusion

Our survey of the employment problem in the developing countries indicates that it can be thought of, from the point of view of devising ameliorative policies, as having three aspects, which concern population growth, open unemployment, and widespread low-productivity utilization of labor in the traditional sector. Usually the population issue is posed as a Malthusian food scarcity issue, or more recently as an ecological issue at the global level. In fact, it is no less important as an employment issue. The prevailing high rates of growth of the labor force in the developing countries are historically unprecedented and create the difficult challenge of productively absorbing this labor into the economic systems. Of course, many complex social, political, and economic considerations govern national approaches to demographic policy. However, where the need for effective population policies is accepted, the existence of a long time lag between their initiation and the slowdown in the growth of the labor force does not reduce the high priority that such policies should have. Second, open unemployment should be regarded as only one highly visible, but not necessarily the most important, manifestation of the overall employment problem and as resulting mainly from the urban drift of the populations in the developing countries. The proper policy response to it is to be found not only in the creation of jobs geared to the characteristics of the unemployed but also in the adoption of measures to reduce both the pull from the urban areas and the push from the rural areas. In the economic sphere perhaps the most important means to these aims would be the moderation of the existing differentials between real earnings of labor in the rural/agricultural sector and the urban/modern sector. Finally, the bulk of the employment problem consists in low-productivity utilization of labor. Since real income is a good proxy for the productivity of labor in these situations, a definition of disguised unemployment can, for practical purposes, be based on an income test. At present, more than two thirds of the labor force and the bulk of the disguised unemployment are in the rural/ agricultural sector of the developing countries. Moreover, any feasible acceleration during the next decade in the productive absorption of labor by the urban/modern sector cannot be expected to make a significant reduction in the proportion of the labor force in the rural/agricultural sector; the absolute numbers will, in any case, increase. Hence, ameliorative measures would need to accord a high priority to increasing the productivity of labor in agriculture and ancillary activities. Wherever such policies are successful, they would also moderate the strength of the urban drift.


Mr. Bhagwat, Senior Economist in the Stabilization Policies Division of the Exchange and Trade Relations Department, is a graduate of the University of Bombay and of Yale University. He formerly taught at the University of Pennsylvania (Philadelphia).

Used synonymously with labor force.

Excluding countries that are members of the Council for Mutual Economic Assistance, mainland China, North Korea, and North Viet-Nam.

Developed countries in Europe.

See Turnham [17], Appendix A, pp. 175-90, for a discussion of the historical experience of employment problems. Numbers in square brackets refer to items listed in the Bibliography, pp. 98-99.

This flat statement is too simplistic. In fact, there are, of course, considerable dissimilarities within the developing countries. At one end of the spectrum are some countries in which agriculture typically absorbs a relatively small portion (30-40 per cent) of the labor force and in which a substantial part of the economic activity takes place in a labor-market environment similar to that in the developed countries. To them the usual framework is, on balance, relevant. At the other end of the spectrum there are countries whose economies are almost entirely traditional, the modern sector being small and largely unintegrated with the rest of the economy. To such economies, the “unemployment” framework does not apply at all. However, notwithstanding these dissimilarities within the developing countries, the dichotomy with the developed countries drawn in the text is a useful one. The insight that it provides substantially outweighs the violence it does to reality.

See Dantwala [1] for some empirical confirmation of this.

See Turnham [17], Chapter II, pp. 21-46, and the report by the Indian Committee of Experts [12], Chapters III and V, pp. 14-29, for further discussion. Both Turnham and, implicitly, the Indian Committee arrive at similar conclusions. Even Myrdal, after his severe and effective criticism of the usual conceptual apparatus [10], Chapter 21, Vol. II, pp. 961-1027), retains the participation rate concept. However, it is fitted into a different overall framework that excludes the unemployment rate concept.

These points will be intuitively convincing to most people with firsthand experience in developing countries. For their extensive validation, see Myrdal [10], Appendix 16, Vol. III, pp. 2203-21, and the report by the Indian Committee [12], especially Appendix I, pp. 33-41. See Ridker and Lubell [13], Chapter 2 (especially p. 11), for instances of extreme variations in reported unemployment rates that are attributable mainly to unavoidable methodological problems and that emphasize the need for extreme caution in using such data.

In developing countries as a whole, roughly 70 per cent of the labor force is estimated to be in agriculture, and those in paid employment generally account for less than 10 per cent. However, these proportions vary substantially from country to country.

The Year Book of Labour Statistics[3] presents four types of unemployment statistics: (a) labor-force sample surveys; (b) compulsory unemployment insurance statistics; (c) statistics of trade union benefit funds; and (d) employment office statistics. The Year Book for 1971 includes series for the general level of unemployment for 37 developing countries that are members of the International Monetary Fund. Of these, data for 10 countries are based on labor-force sample surveys and the remainder are employment office statistics. It is now well established that employment office statistics in developing countries are not usable as reliable indices of unemployment. There are several reasons for this, including the absence of any identifiable relationship between the number of registrants and the number of unemployed, as the number of registrants is influenced by such factors as the location and number of employment offices, how widely their existence is known, their effectiveness as employment exchanges for different categories of labor, and the literacy and knowledgeability of the unemployed. Furthermore, registrations often include people who are employed but desire another or a better job; and there can also be double counting because the same person has registered at different offices. Hence, Table 2 includes only the series that are based on sample surveys; similar data for five selected developed countries are also included as a basis for comparison. Generally, a well-designed and well-executed labor-force sample survey can be expected to yield reliable estimates of open unemployment.

Compared with the paucity of time series. Table 2 just about exhausts time-series data on national open unemployment rates.

See Turnham [17], Chapter III, pp. 47-95, for a survey of the evidence.

The rural migrants do not necessarily join the unemployed pool directly but may be initially absorbed in the urban traditional sector. In fact, there is some evidence that the incidence of open unemployment among the rural migrants is proportionately less, perhaps because they can afford it less. If so, the unemployed pool may be replenished mainly by the urbanites. All this, however, does not alter the essential nature of the process sketched in the text.

See Todaro [16] for a rigorous formulation of this hypothesis. Todaro assumes that the percentage change in the urban labor force as a result of migration during any period is governed by the differential between the discounted streams of expected urban and rural real income expressed as a percentage of the discounted stream of expected rural real income.

See, for instance, Knight [7], for a careful analysis of the experience in Ghana.

However, there is some presumption that given the constraint of available real and financial resources, a policy that moderates the rural/urban real wage differential would also make possible the creation of more jobs in the modern sector.

This is also Todaro’s conclusion [16]. See also Frank [2] for a discussion of urban unemployment problems in Africa and a similar conclusion.

There is as yet no established terminology in this area, and some writers prefer the term “underemployment” to “disguised unemployment.” Both alternatives have been used widely and have been given interpretations of varying degrees of precision. The underemployment usage focuses attention on the distinction between those who are wholly unemployed and those who are only partly unemployed, whereas the disguised unemployment usage brings to the fore the fact that, in addition to open unemployment, there are others who are seemingly employed but who in fact constitute a part of the overall problem of labor utilization in the developing countries. We prefer the latter terminology because it provides a better indication of the complexities inherent in regarding the problem of labor utilization in the developing countries as an unemployment problem.

See [14], [11], and [9].

Sometimes this is put in the form of the statement that the marginal product of labor in agriculture is zero.

See Myrdal [10], Appendix 6, Vol. III, pp. 2041-61, for a detailed, critical discussion of the labor-surplus approach to disguised unemployment and for further development of the point that the concept is meaningful only within a dynamic context.

Again, the diversity among the developing countries must be emphasized. The greater the divergence of the labor utilization patterns in a developing country from those common to the developed countries, the more valid is the statement in the text for an individual country.

See Tobin [15] for a discussion of alternative definitions of full employment in the context of a developed economy.

The Phillips curve problem of a conflict between full employment and price stability can be ignored here.

However, in particular developing countries at particular times, cyclical unemployment can be a matter for concern.

Other Resources Citing This Publication