II Labor Markets, External Developments, and Unemployment in Developing Countries
- International Monetary Fund
- Published Date:
- January 1986
Developments in the world economy since 1979 have heightened concern for the economic well-being of the poorest groups of the population in the developing countries. Widespread payments difficulties and a consequent import compression since 1982 have set back the growth process in many of these countries. One of the chief concerns arising out of the slowdown in growth has been the impact on employment, and this study focuses on three aspects of this issue: the evidence of changes in unemployment (or underemployment) in developing countries in recent years; the functioning of labor markets in these countries and the nature of policies that affect these markets; and, finally, the factors (including labor market efficiency) that affect the transmission of external economic disturbances to labor markets in developing countries.
The concepts of employment and unemployment used in general economic analysis have been formulated in the context of labor markets in the industrial countries. But there are certain structural differences between the industrial and developing countries that create special problems if these concepts are applied to the latter countries.
In the typical industrial country, for the great majority of the population, employment means having a job for which a wage is paid. An unemployed person, then, is one who has lost a job and is looking for another, or who is seeking employment—either for the first time or after a period of voluntary withdrawal from the labor force. In an industrial country, self-employment is for the most part limited to those with enough capital to own a business or with the qualifications to enter a profession. Opportunities for casual employment are rather restricted, with pay that is generally below that in regular employment. The industrial countries are also monetary economies; barter and subsistence transactions represent only a minor fraction of economic activities and actions.
The relevance for many developing countries of employment and unemployment concepts as they have emerged for the industrial countries is arguably limited.1 In a number of developing countries—particularly the low-income ones in sub-Saharan Africa—most people do not usually work for money wages; and even those who do generally have alternative opportunities for self-employment and casual work. In these economies, barter and subsistence transactions are still widespread and constitute an important proportion of economic activity. In the typical developing country, 40 percent or more of the labor force is in agriculture, compared with less than 10 percent for the typical industrial country.2 In a substantial number of developing countries (again particularly in sub-Saharan Africa), subsistence, family, or cooperative farming dominates agricultural activities. Nevertheless, typical data do mask a wide variation in individual situations, as Table 21 shows. The share of the labor force in agriculture can be fairly low in a number of middle-income (semi-industrialized) developing countries, such as Argentina and Venezuela.
|Germany, Fed. Rep. of||4||46||50|
In some developing countries (particularly in parts of Asia and Latin America) the relative shortage of land or the importance of plantation agriculture has meant that a substantial part of the rural work force are wage laborers. But even in these countries, crop-sharing arrangements have sometimes assumed great importance in relation to working for a money wage. Furthermore, some relatively densely populated countries have substantial subsistence and family farming arrangements.
Labor markets in a typical developing economy may be conveniently analyzed by first distinguishing between the rural sector and an urban sector, and then dividing the latter into a formal and informal sector. The informal sector is characterized by small privately owned enterprises using paid or own labor; it is distinguished from the formal sector by the absence of explicit rules for wage contracts, job security, and the pension, insurance, and health obligations of employers. Labor unions tend not to operate in this sector and legal minimum wage laws are not strictly applied in it; instead, wages are highly flexible, as are other conditions of service. Some 30–60 percent of urban employment in many parts of Asia, the Middle East, and sub-Saharan Africa, is in the informal sector (Table 22), while the share of informal sector employment in total employment is also sizable for Latin American countries (Table 23).3
|Country||Area||Year||Share of Urban|
|Asia and Middle East|
|Malaysia||Urban areas in West Malaysia||1970||35|
Although wage employment in the typical developing country is a much smaller proportion of total employment than in the typical industrial country, this proportion varies considerably by country. Table 24 shows that it can range from as low as 5–15 percent for many low-income countries (particularly in sub-Saharan Africa) to as high as around 70 percent in many middle-income countries of Latin America. However, wage employment in most developing countries is in the 15–50 percent range, compared with 70–90 percent for the industrial countries. In general, the proportion of the labor force in wage employment is positively related to the share of the population in urban areas, the size of the public sector, and the share of manufacturing, construction, and mining in total employment.
|Germany, Fed. Rep. of||88.2||3.2|
When an economic structure includes a large informal urban sector, a low share of money wage employment, and substantial subsistence and barter activities, there tends not only to be significant self- and own-account employment but, more important, the fraction of unpaid family workers in the labor force tends to be quite sizable. This is the case for many sub-Saharan African countries (see, for example, Burundi, Mali, and Rwanda in Table 24). The application of the concepts of employment and unemployment, as developed for the industrial countries, requires particular caution when applied to developing countries with large shares of unpaid labor.
If the concepts of employment and unemployment are to be meaningful for developing countries, they must be used and measured in ways that take into account the particular structural features of these countries. For instance, in addition to taking account of wage employment and open unemployment, such measures must also take into account casual work, self-employment, and subsistence activities.4
Measures of unemployment yield insights into two facets of economic performance: the degree of under-utilization of economic resources and the extent of poverty and economic inequality. For unemployment data to give reliable signals to policymakers, however (as was argued in the preceding section), statistics must be defined and collected in a way that is sensitive to the structure of the economy, in particular to the nature of labor markets.
There appears to be some agreement that several indicators of “unemployment” are necessary to assess the gravity of unemployment as an economic and social problem.
Five measures of unemployment have been identified as being economically meaningful for developing countries (Table 25).
|U1||Open unemployment||Job-searching is full-time activity.|
|U2||Open and passive unemployment (U1 plus passive unemployment)||Full-time job-searchers plus persons working in informal sector and passively (i.e., on a part-time and often informal basis) seeking jobs with greater income and security.|
|U3||Open unemployment plus discouraged job searchers (U1 plus discouraged job searchers)||Full-time job searchers plus previous job searchers, still not working, but dropped out of job-search activity because of discouragement.|
|U4||U2 minus those passive job searchers whose earnings in informal sector are above some minimum level.|
|U5||U3 minus discouraged job searchers whose means of support from own resources, relations, and friends are above some minimum level.|
The most widely used notion of unemployment involves dividing the total population into the economically active, who are considered as constituting the labor force, and those who are inactive or outside the labor force. The inactive population comprises students, homemakers, such income recipients as pensioners and rentiers who are not working, and other unemployed who are dependents or recipients of public assistance. The active labor force is then divided into the employed and the unemployed. An employed person can be either in paid employment, in self-employment, or can be an unpaid family worker, but in each situation, such a person is “at work.” The unemployed, according to this criterion, comprise all those who, during some relevant period, were without work but “currently available for work” and “seeking work.” But some differences remain in the definition of the important concept of “seeking work.” In the framework used by the International Labour Office, “seeking work” implies that the person took specific steps in a given reference period to obtain paid employment or self-employment.5 The steps may include registration at a labor exchange; filing an application for employment; checking at employment sites; placing or answering newspaper advertisements; obtaining the help of relatives and friends; looking for land, a building, or equipment to establish a business; or arranging for finance or for permits and licenses.
Ambiguities arise when, according to these definitions, people are both “at work” and “seeking work,” and when people use only informal methods of search, or they would like to work but for various reasons are not actively seeking work during the reference period. Three basic unemployment measures, to be identified as U1, U2, and U3, can, therefore, be distinguished according to whether the criterion employed is active job search (U1), active and passive job search (U2), and active job search plus discouraged job searcher (U3).
The U1 definition captures persons who are openly unemployed—in the sense that their full-time activity is job searching, involving both formal market institutions and informal methods such as enquiries from relations and friends. People searching for work often find that they cannot afford to stay without work beyond a very limited period, as financial support from relations and friends is meager and resources quickly dry up. In developing countries such persons often enter informal labor markets and do odd jobs, or they become engaged in subsistence activities. This situation has given rise to the “luxury unemployment hypothesis” that open unemployment in developing countries can only be afforded by the reasonably well-to-do without financial responsibilities and therefore tends to predominate among young educated people who come from reasonably well-off families that can afford to support them during long periods of search for jobs that meet their aspirations.6 The available evidence in developing countries supports this view. Descriptive studies of urban labor markets and labor migration in a number of countries (including Chile, Colombia, The Gambia, Panama, Senegal, and Tanzania) indicate that open unemployment rates tend to fall substantially with age; that such rates tend to be higher in middle-income groups than among the lowest-income groups; and that outside support, from immediate family members and friends, is virtually universal among the unemployed.7 Furthermore, the studies indicate that when demand conditions change for the worse, unemployment rates tend to increase more for the workers born in the city than for migrant workers from other parts of the country. Although the native-migrant unemployment differential is partially explained, in some cases, by the relatively young average age of migrant, compared with that of native laborers, differences in the support system available to each, in conjunction with the costs of moving between rural and urban areas, also affect the results; migrants are willing to return to the countryside, or, more easily (than natives) they accept “whatever they can find” when the moving costs to the countryside are too high.
Because of the opportunities existing in informal sectors, less fortunate job seekers (relatively low-income natives and migrants) are able to enter such markets while “passively” (that is, on a part-time, and often informal basis) seeking jobs with greater income and security. Such persons often do not consider themselves “employed” when the question is explicitly posed.8 Viewed in a wider sense, they can be considered unemployed and engaged in passive job search. From a policy point of view, an unemployment concept that includes such people (U2) is very important, because it captures the element of disguised underemployment in the labor force.
Some persons who desire work and have searched for it may, for a number of reasons, drop out of the job search activity for a while and, according to the usual definition, be regarded as “inactive” or outside the labor force. Among the reasons for such inactivity, one that has attracted particular attention is discouragement. Policymakers have been interested in identifying discouraged job seekers in order to get a more accurate picture of the degree of underutilization of resources in the economy, and a measure of unemployment (U3) that adds to U1 the discouraged job seekers has been of practical interest.
In developing countries, concern with surplus and underemployed labor has induced attempts to compute “hidden unemployment,” particularly among the employed in the urban informal sector. Underemployment and hidden unemployment are, of course, not confined to the informal urban sector. Indeed, much of the concern with employment and unemployment issues in many developing countries is related to the apparent existence of surplus labor in the public sector as well as underemployment (“disguised unemployment”) in the agricultural or rural sector; indeed, the measurement of rural surplus labor has posed conceptual problems that have interested development economists for decades.9
To the extent that apparent underemployment is genuine disguised unemployment, and does not simply reflect voluntary labor-leisure choices, it is revealed by a measure of passive job search or discouraged job search, at least as far as the non-government sectors are concerned. Such a measure is not generally intended to include passive job searchers who are in adequately paid jobs but are nevertheless searching for improved opportunities; nor is it intended to include discouraged job searchers with adequate means of support from relatives, friends, or own resources. From a welfare point of view, it has seemed logical to some to distinguish these last two groups of persons from passive job searchers and discouraged job searchers whose earnings or financial support are a specified minimum level. The specified income (support) level to be designated as adequate is open to ambiguity, thereby rendering the measure of unemployment under this approach somewhat arbitrary. But a theoretically appealing method is to include among the unemployed only those passive and discouraged job searchers whose earnings or other means of support are less than they could earn or produce in the rural area.10 Persons unemployed (or rather underemployed) under this criterion can be added to U1 to generate the measure of unemployment designated as U4 or U5 in Table 25 depending on whether the criterion is applied to employed passive job seekers or discouraged job seekers, respectively.
Despite the strides that have been made in recent decades in the classification of employment, unemployment, and underemployment in developing countries, problems of measurement still remain. Reliable statistical series for these indicators are not available for many countries; in addition, data available are often not comparable among countries, making aggregation virtually meaningless.
Countries generally compile their data on employment and unemployment from a variety of sources. The best information available on a regular basis is usually from labor force surveys or household surveys, which are generally sample surveys. Less regularly, countries will derive information about the labor force from a complete census. The definitions of terms used in such surveys and censuses usually conform to internationally recommended standards. The surveys and censuses generally distinguish among different geographical regions, and contain information on different occupations and industries. Survey data are especially useful in providing information on new entrants into the labor force, seasonal unemployment, and, depending on the survey, on hours worked. Labor force surveys are potentially the best available source of comprehensive measures of unemployment.11
Other sources of labor statistics include the employment offices (or labor exchanges). Data from these sources tend to be partial and misleading. The offices themselves are generally located only in major urban areas, and often only certain types of labor—especially those in transport and construction—will register at the exchanges. At best such statistics yield information on movements in open unemployment.
Another source of primary information on employment, unemployment, and related data is social insurance statistics, which can take many different forms in different countries, and with coverage that also varies. Such statistics, essentially providing information on changes in open unemployment (U1), can be more useful than employment office statistics, especially in countries with compulsory unemployment insurance schemes. Data from this source, however, cannot usually be used as a basis for estimates of unemployment or underemployment in the informal labor sector.
From what has been said so far, and as will be further elaborated below, to assess the gravity of the unemployment problem in developing countries it is important to look not only at the level and changes of unemployment but also at other indicators, such as the growth of the labor force and of real output. This is particularly necessary in countries where unemployment data are scanty or unreliable. Other things being equal, the unemployment rate, no matter how it is measured, tends to increase if labor force growth increases in relation to the growth of real output. Comparing the underlying labor force growth with real output growth can be especially useful in assessing likely changes in unemployment and underemployment among passive and discouraged job seekers. Hence information on the direction of change in unemployment, where unemployment data are scanty or limited in coverage, can be obtained by looking at the growth of the labor force and output. But the relationships among labor force growth, output growth, and unemployment are complex. After reviewing available unemployment data in the following subsection, the remainder of the paper analyzes the crucial factors—structural and nonstructural—that affect unemployment and underemployment in developing countries.
Table 26 presents data on unemployment, output, and labor force growth for a selected number of countries. Among the countries for which unemployment data are available, it is shown that, except for Korea, Portugal, and Singapore, the rate of unemployment in all countries increased between 1979–80 and 1982–83 (although the unemployment rate for Portugal rose in 1984 whereas it continued to fall in Korea and Singapore). The table also shows that the underlying labor force growth (approximated by growth during 1970–82) was greater than real output growth during 1983–84 for 24 out of the 44 countries concerned. Thus, other things being equal, underemployment and unemployment during 1983–84 was far worse than in 1981 when almost three quarters of the countries had real output growth rates that were greater than the trend labor force expansion rate. The crude indicator being used here shows that, during the four-year period 1981–84, underemployment and unemployment probably were greatest in 1982.
|Change in Real GDP1||Average|
Growth Rate 2
Unemployment Rate 3
|Syrian Arab Republic||10.25||3.24||–1.29||3.3||3.8||5||…||…|
|Exporters of manufactures|
|Other primary product exporters|
|Service and remittances countries 6|
Additional information on open urban unemployment is available for a number of Latin American countries. Such information is quite meaningful for these countries, since two fifths to three quarters of the labor force is in wage and salary employment and a similar proportion of the population lives in urban areas. The available evidence is that, for Latin America, the recent developments in the world economy, taken in conjunction with the operation of labor markets and domestic policy responses, have increased open urban unemployment, reduced hours worked by the employed, and increased the relative size of the urban informal labor force. For instance, between 1980 and 1983 the proportion of people working fewer hours than they desired increased from 4 to 8 percent in Buenos Aires (Argentina), from 10 to 18 percent in Santiago (Chile) and from 3 to 7 percent in San José (Costa Rica).12 The data in Table 27 show that, apart from Brazil, the open urban unemployment rates in 1983–84 tended to be much higher than in 1980–81, although for seven of the twelve countries in the sample, the unemployment rate is shown to have dropped somewhat in 1984 from the 1983 levels. A comparison between Tables 23 and 27 suggests that the size of the urban informal sector tends to increase, relatively, when open urban unemployment is increasing. But the case of Chile, where the relative size of informal sector employment shrank between 1979 and 1982 while open urban unemployment increased sharply, is evidence that the relation between the relative size of the informal sector and open unemployment is somewhat complex; not only do the formal and informal sectors tend to be affected by general macroeconomic developments, but much movement of labor also occurs between the two sectors.13
|Costa Rica 6||5.3||6.0||9.1||9.9||8.5||6.6|
Labor Market Efficiency and Economic Policy
In the transmission of external disturbances to output and unemployment in a particular developing country, the role of the labor market is crucial. It is important, therefore, to outline what is known about the functioning of labor markets in developing countries before analyzing the specific factors involved in the transmission process of external disturbances to unemployment in developing countries.
In ideally functioning goods and labor markets, in which all wages and prices are perfectly flexible, open unemployment and underemployment would tend toward zero, and smooth and timely adjustments would be made to changes in economic circumstances. But even where labor earnings are flexible in the short run in the face of disequilibria, and where unemployment tends to be voluntary, there will usually be some involuntary frictional unemployment and underemployment due to information and transaction costs.14
Another feature of an ideal labor market is that it should facilitate the efficient allocation of labor in the economy, so that each unit of labor is employed in that activity in which it is most productive. In such a case, positive net migration of labor between the rural and urban areas would result not only in better living conditions for the migrants but also in higher national output (GDP) than would have been achieved with zero net migration.15
A third feature of an ideal labor market is that it should provide clear signals for investment in training and education. Differences in education, training, and experience should be reflected in the structure of earnings, so that real hourly earnings of marginal workers would differ among sectors, industries, firms, and occupations only with differences of marginal productivity. It should be noted, however, that the marginal productivity of labor in low-income and middle-income countries has been hypothesized to increase with the real wage (up to a point) given the general level of real wages. In low-income countries this hypothesis takes the form of a nutritional theory, according to which the positive relation between marginal productivity and real wages is the result of the worker being able to afford better nutrition from higher wages which (again, up to a point), will increase his (marginal) productivity. In middle-income countries, the hypothesis takes the form of the incentive or efficiency wage theory—that the worker achieves varying degrees of efficiency in the performance of any task and that, up to a point, his level of efficiency will (through his motivation and desire to keep his job) increase with the real wage paid.16
Policy-Induced Distortions in Labor Markets
In developing countries, the ideal properties described above have often not existed because of domestic economic policies that have tended to create or aggravate distortions in the labor markets. (It is not intended to imply here that labor markets in industrial countries are efficient, or even more efficient than those of developing countries.) The resulting inefficiencies in the allocation of labor tend to depress real GDP. These domestic policy sources of distortions can be grouped under five headings: (1) urban bias in the supply of amenities; (2) artificially high and rigid wage levels in the formal sector; (3) adverse internal terms of trade for agriculture owing to relatively low administered producer prices and inappropriate foreign exchange rate policy; (4) government codes, laws and regulations with respect to labor relations, wages, and other conditions of service in the private sector, some (not all) of which have resulted in distortions from the viewpoint of efficient labor allocation and use; and (5) government taxation policies.17
Urban bias. Development policy in the majority of developing countries has had a decidedly urban bias. The resulting disproportionate increase in the supply of amenities has in turn greatly augmented the non-pecuniary net benefits of residing in urban areas. Hence there has resulted the so-called “bright lights” phenomenon, plus the fact that education and health facilities are generally more easily available in the cities than in rural areas.18 In many countries this situation is aggravated by customs and traditions in rural areas that many of the younger and more educated rural residents find onerous and seek to avoid by migrating to the towns and cities. Despite the great differences among developing countries, certain similarities permit a generalized description and analysis.
An implication of the existence of nonpecuniary (net) benefits of living in urban areas is that every unit of real income earned in a rural area is valued somewhat less than a unit of real income earned in an urban area (even after taking into account the higher cost of living in the latter), where the extent of undervaluation would tend to increase with distance from the nearest city. Furthermore, many individuals will move into a city even though they risk remaining openly unemployed for a while if they seek employment only in the formal sector. In the formal sector, then, the economic variable that is of interest to the migrant becomes the expected wage, which is a product of some objective estimated probability of finding a job during a relevant reference period and some representative wage rate. A migrant laborer who is unable to get a job in the formal sector of the urban area is generally able to obtain some work in the urban informal sector; indeed, this is the minimum that the migrant usually expects before migrating.
To work in the formal sector, the migrant laborer must often undertake some training that raises his potential earnings by some fraction above what it would be in the rural or urban informal sectors; for the marginal worker this return to his training should, in principle, just cover the cost of that training. Because of urban bias, in equilibrium, when net rural-urban migration is zero and the marginal laborer is indifferent among the rural, urban formal, and urban informal sectors, the earnings differential between the marginal rural and the marginal urban worker will differ by less than productivity differentials. The urban worker is thus willing to accept the equivalent of a tax on his earnings for the benefits of residing in the urban area. Real earnings, it may be noted, should be compared, to take account of the fact that the cost of living in the urban area is usually greater than that in the rural area. The consequence of the urban bias is that more migration takes place than would otherwise occur and the stock of urban labor relative to the stock of rural labor is greater than the level that is efficient from the viewpoint of maximizing GDP.19
Migration serves as one of the important equilibrating forces in the labor markets. As labor moves out of the rural into the urban sector, marginal rural earnings tend to increase, under normal circumstances, whether the farm or nonfarm activity concerned is a family enterprise, a cooperative, or a capitalist enterprise. The influx into the city puts downward pressure on marginal earnings in the informal urban sector and on the expected hourly earnings in the formal urban sector. The downward pressure on these expected formal sector earnings is reflected in some combination of falling wages and of rising incidence and duration of open unemployment. Net migration continues until expected earnings and informal urban earnings are sufficiently lowered and rural earnings sufficiently raised to the point where the incentive to migrate disappears.
Public sector wages and salaries. Distortions also arise in the labor markets in developing countries from a suboptimal structure of real wages in the public sector relative to those in the private sector. This tends to be true especially for public enterprises, where salaries are often unusually high in relation to private sector salaries. In government administration and services, in the narrow sense, the problem is sometimes unusually low salaries, which tend to result in loss of, or the inability to hire, properly qualified employees. Low wages and salaries in government, which have become a more widespread problem in the last several years, are often a budgetary consequence of overemployment in this sector.
Nevertheless, a problem that has attracted considerable concern in many countries is public sector wages that are higher than those of private sector wage earners with comparable skills. Public sector wages that are high in relation to labor productivity in the economy become an institutional constraint on the level of wages for the whole formal sector; real wages in the economy end up being higher than socially efficient. One consequence is that labor markets may become segmented, not only into formal and informal sectors, but also into a disequilibrium high-wage sector, which may or may not engulf the whole of the formal sector, and a relatively low-wage sector where earnings adjust more flexibly.20 Another consequence is that the public sector tends to experience sizable overemployment, a situation that is often a serious fiscal problem. Such a problem is exacerbated in certain countries where the public sector has been obligated to employ all college graduates requesting work.
Suboptimal relative commodity prices. In many low-income developing countries a serious cause of labor market distortion is distorted relative commodity prices. In particular, government policies have resulted in low agricultural prices and an overvalued exchange rate, which often combine to produce artificially high prices of nontraded goods relative to the prices of exportable goods. In general, price distortions result from official price controls in commodity markets and inflexible exchange rate policies. The effect of the distortions and, in particular, of the suboptimal terms of trade of agriculture, is to induce an overallocation of labor to the urban area. The greater the dependence of the economy on the rural sector for exports, government revenues, and employment, the greater is the potential economic cost of this misallocation of resources.
Laws relating to conditions of service. In many developing countries, especially those that are major exporters of manufactures and of minerals and where wage employment absorbs a substantial part of the labor force, various government codes, laws and regulations affecting labor relations, wages, and other conditions of employment in the private sector, are often vigorously enforced, influencing resource allocation, employment, and output. Family workers and a substantial share of workers in the agricultural and urban informal sectors, particularly in the low-income countries, are not generally covered by such legislation. When effective, such measures tend to raise labor costs above market-determined levels; to limit the ability of firms to adjust relative prices to changing demand and other factor supply conditions; and to increase disparities in the ratios of wages to marginal product. The principal types of these measures are minimum wages, wage indexation, labor tenure laws, and restrictions on labor mobility.
Although minimum wages may apply legally to an entire country, in fact they usually affect only certain cities or regions.21 Even when they are set for the whole country, they are often set at different levels for different cities and regions and for different occupational categories, in order to take account of differences in labor supply conditions, and in the cost of living. Moreover, minimum wages do not usually cover all industries and firms; they are most frequent in mining, manufacturing, transport, and construction, and usually relate only to establishments of a certain minimum size. Finally, coverage generally extends only to certain occupations and skill levels.
The economic impact of minimum wages is strongly influenced by both the frequency of their adjustment and the degree of their enforcement. In situations where the minimum wage is adjusted only infrequently, it can become quite low in real terms as prices rise; in these cases, the actual nominal wage paid to the specific groups to which the minimum wage law applies has often become much higher than the minimum wage itself.22 In the case of enforcement, a problem arises when the minimum wage is set much higher than the laborers concerned would be willing to accept; the incentive to both employers and employees to avoid the law is then quite great. In such cases, employment, especially of those without work experience, is promoted by lax enforcement. Available evidence reveals that governments in developing countries have exercised considerable caution in making minimum wage adjustments in the 1970s and 1980s. The International Labour Office, for instance, reports that, in contrast to the experience of the 1950s and 1960s, during the 1970s minimum wages in most developing countries increased less rapidly than average wages or per capita national income (I.L.O., World Labour Report (1985) Vol. 1, pp. 144–5). Moderation in real minimum wage increases has also been evident in the 1980s (see, for example, the case of Latin America in Table 28 where indices of real minimum wages fell in nine out of fourteen countries between 1980 and 1984).
|Minimum Wages||Industrial Wages||Construction Wages|
|Country||1980||1981||1982||1983||1984 1||1980||1981||1982||1983||1984 1||1980||1981||1982||1983||1984 1|
In middle-income developing countries that are also high-inflation countries (such as Brazil and Israel) and in which wage employment is a large part of total employment, it is common to find wage indexation as an established feature of the labor market. Full indexation helps to insulate output and employment from monetary shocks, though not from real shocks.23 Indexation procedures differ among and within countries over time in three main respects: the interval between wage adjustments, the degree of indexation, and the nature of adjustments for productivity changes.24
The frequency of wage adjustments has tended to increase (as in Brazil) with the rate of inflation; indeed this frequency itself becomes one of the structural elements in the inflationary process.25 In some countries (notably Brazil), the degree of indexation to inflation has been made a function of the wage level, with overindexation at certain wage levels and under-indexation at others, while the average degree of indexation has been used in some countries (for instance, recently in Argentina) as a means of altering inflationary expectations and reducing the inertial element in inflation; sometimes countries even go as far as to impose an official freeze on wages.26 Indexation procedures, under normal circumstances, allow for adjustment of wages for productivity changes as well as for past inflation. In some countries, the wage law permits the productivity adjustment to be negotiated freely between labor and management; in others, the adjustments are specified by the government.
Clearly, the manner in which indexation operates is important for the transmission of shocks to output and employment. Where indexation hinders real wage adjustment to adverse shocks, quantity adjustments in output and employment must occur. Where indexation freezes relative prices, it hinders changes in output in response to shifts in the underlying structure of demand and supply. Indeed, even wage and price adjustments for productivity changes can operate quite differently as part of an indexation scheme from what occurs in a competitive market setting, where the relative prices of commodities experiencing faster productivity growth would tend to fall, under normal demand conditions, relative to commodities that are experiencing slower productivity increases. As compared with a competitive price-setting system, an indexation system can, therefore, hinder the relative expansion of output in activities with higher productivity growth.
Apart from minimum wage laws and wage indexation, governments in many developing countries (such as Algeria, Cameroon, Congo, Côte d’Ivoire, Egypt, Gabon, Mexico, Panama, Peru, and Portugal) enforce various labor tenure laws. Such laws limit employers’ freedom to dismiss or lay off workers with contracts of indefinite duration. The right of employers to reduce their work force through collective dismissals or layoffs is also regulated in a number of countries.27 In general, labor-tenure laws have been aimed at preventing unjustified dismissals while balancing the concern of workers for employment security with that of employers for efficiency.28 It is clear that if such laws were applied in ways that proved burdensome to employers, the latter would become cautious about engaging new labor with contracts of indefinite duration, and fixed-term contracts would become more frequently used than otherwise. In situations where demand conditions tend to fluctuate, the average level of wage employment in the affected industries or occupations could then become lower than optimal.
Regulations and controls on geographical and occupational mobility are intended (a) to influence migration patterns and levels and (b) to enforce discrimination against certain groups of laborers.29 Migration policies could be designed to foster movements from certain regions or cities considered overpopulated to others considered underpopulated, relative to the employment opportunities available; alternatively, the policy could be designed to restrict or reduce migration into certain (usually urban) areas. For instance, in Indonesia, there has been, for many decades, a “transmigration program” from densely populated islands, especially Java, to lightly populated islands, especially Sumatra. India has also had rural settlement projects in the recent past. Land reform programs in Korea and Sri Lanka were designed to improve farmers’ incomes and security of tenure as part of a policy to reduce the flow of migrants into urban areas.30
Leaving aside nonpecuniary considerations, the efficiency of regulating migration patterns and levels can be judged on the basis of whether or not the policy results in raising the average productivity of labor and whether or not it raises the net value added of labor after deduction of the moving costs. Segmentation of the labor market as part of a government-enforced discrimination policy will, in general, tend to prevent equality of ratios of wages to marginal products of different laborers, with a resulting output loss.
The system of taxation in a number of developing countries has implications for capital-labor substitution and employment. In this connection, two types of tax policy are relevant: taxes that increase the unit labor cost to the employer substantially above the wage rate, and special tax incentives for investment.
Taxes that increase the effective cost of labor to employers are generally variations on payroll taxes (usually for social security contributions, such as those used in Israel and the Philippines).31 The evidence seems to be that, in most circumstances, employees end up bearing the burden of such taxes (through reduced real wages); otherwise, the resulting increase in unit labor costs adversely affects employment by inducing relatively capital-intensive techniques or by encouraging capital outflows (or reduced capital inflows).32
The idea that tax incentives for investment can raise employment is based on the expectation that the adverse substitution effect, resulting from the encouragement of increased capital-labor ratios in production, will be more than counterbalanced by the output effect of greater aggregate investment and larger firm size.33 To limit the substitution effect, developing countries that have used such incentives have utilized additional measures to induce an increase in employment, especially of unskilled labor, accompanying the additional investment. In some cases, tax credits are given for additional employment of unskilled labor; in others, tax incentives for investment are tailored to favor machinery and equipment that employ relatively large amounts of labor; or special tax advantages (such as larger-than-normal depreciation allowances) are given to multiple-shift operations.34
While it may be acknowledged that some substitution of labor for capital is always possible, and probably occurs, as a result of tax incentives, the extent of this substitution remains uncertain, particularly since incentives actually utilized have been relatively modest.35 While such incentives may have positive effects on overall employment, they may also encourage overemployment in one part of the economy and underemployment in another, with a loss in real output. Another negative effect of these incentives is, of course, their budgetary cost.
Other Sources of Labor Market Distortions
Aside from the impact of policy, earnings differences among individuals arise chiefly from four sources: institutional constraints that tend to segment the labor market; productivity differences due to education, training, nature of cooperative factors of production, and work effort; hours of work, which are determined by both supply and demand factors; and noncompetitive wage determination because certain segments of the market are closed to competition—from either the demand or supply side.
As a practical matter, among the most widespread causes of labor market distortions are labor market segmentation and noncompetitive wages. Important factors that contribute to segmentation and noncompetitive wage determination are union activity, social and ethnic biases, and market imperfections that result in especially high wages in the formal urban sector.
Labor unions. Labor unions face serious organizational problems in developing countries.36 In some countries freedom of association does not exist. In many others, as has been seen, the wage earning component of the work force tends to be rather small in relation to the total, and low incomes limit union contributions and the ability to use the strike weapon in labor disputes. In some low-income developing countries in Africa and Asia, unions have been rather successful in their organizational efforts when they have benefited from government assistance and labor legislation, particularly as part of a political process to mobilize workers and peasants in the development effort. Where there are large industrial and commercial undertakings—including large mining enclaves—and where wage labor is a sizable portion of the labor force, as in many Latin American countries, union organization has been more successful even with little or no governmental assistance. Rough estimates by the International Labour Office on the degree of unionization (defining union membership in terms of dues-paying persons) are shown in Table 29.
|Austria, Luxembourg, Norway||60–70|
|Australia, Ireland, Italy, United Kingdom||50–60|
|Germany (Fed. Rep.), New Zealand||40–50|
|Canada, Japan, Netherlands||30–40|
|France, United States||15–30|
|Algeria, Tanzania 1||40–50|
|Argentina, Barbados, Chile, Ethiopia, Fiji, Greece, Guyana, Singapore, Sri Lanka, Venezuela||30–40|
|Colombia, Egypt, Gabon, Jamaica, Mexico, Trinidad and Tobago||20–30|
|Botswana, Brazil, Costa Rica, Ecuador, Kenya, Malaysia, Nigeria, Panama, Peru, Philippines, Zambia||10–20|
|Bangladesh, Dominican Republic, Haiti, India, Indonesia, Nepal, Pakistan, Papua New Guinea, Senegal, Sierra Leone, Thailand, Togo||Below 10|
Unions are organized for various reasons other than merely for raising wages of their members.37 Hence unions, by improving working conditions for their members, can have positive effects on labor productivity. The impact of unions on the real wage structure has always been of particular interest. The limited evidence available seems to indicate that labor unions have not had a significant impact on real wages in developing countries, except where they have been able to bring political pressure to bear on the government to intervene in the wage-setting process—for instance, by establishing independent wage tribunals, as in certain sub-Saharan African countries.38 But it has also been argued (for example, for some Latin American countries), that unions have caused wages to rise above the opportunity cost of labor through a combination of union pressure, minimum wage legislation, and wage policies in the public sector.39 However, a World Bank staff study of labor earnings in Bogota indicates that union membership tended to give greater job stability and only marginally higher earnings (3–6 percent at most), especially for workers aged 35 years and older, for the less skilled, and for blue-collar workers; unions also had a greater effect on earnings in the public than in the private sector.40
Social and ethnic biases. It is well known that social and ethnic biases segment labor markets in developing countries, usually by restricting the access and mobility of certain groups into certain jobs or by encouraging different pay for the same jobs based on social rather than productivity considerations. The most common and clear-cut case is that of sex discrimination; but tribal, caste, and other factors are often even more significant, although their empirical importance is not easily measured.41 It is well known, however, that discrimination based on such social and ethnic biases tends to reduce allocative efficiency and, therefore, imposes an output loss on the economy.
High real wages in the formal sector. The reasons for relatively high real wages in the formal private urban sector, despite a highly elastic supply of labor fed by migration from the rural areas, continue to be debated. Some explanations have revolved around minimum wage laws and public sector wage policies that tend to place a floor on the wages that the private sector can pay. But even apart from this, it has been argued that the high wages in the formal sector may be due to the high supply prices of workers in that sector, reflecting the high pecuniary cost of living in towns. According to this hypothesis, high wages are therefore the price that must be paid to ensure a stable and permanent work force.42 It has also been hypothesized that non-market-clearing wage rates in the formal sector arise from imperfect information about the heterogeneous abilities of newly hired workers, that is, they are used as a device to induce the high quality workers to enter the pool.43
Notwithstanding such arguments, there is evidence that wages, when they are not officially controlled, are flexible in response to demand changes in the market as a whole.44 In the end, then, it would seem that the argument as to possible inefficiency of the private labor market boils down to the issue of whether labor markets are segmented by the internal labor markets of different firms.45 But the different economic circumstances of firms (the heterogeneity of production structures and/or output for example) makes it difficult to conclude anything about market efficiency on the basis of wage differences of apparently similar types of labor.46
Transmission of Disturbances to Unemployment
The data reported earlier on unemployment levels indicate that during the period 1979–84, a number of developing countries, in varying degrees, experienced increases in measured unemployment. In most of these cases, this growing unemployment was associated with shocks to output imparted by the global economic disturbances of that period. These disturbances affected countries in various ways, because of differences in the composition of their foreign trade and in the levels and structure of their external debt. For instance, the rise in oil prices in 1979–80 favorably affected output and employment in oil exporting countries while adversely affecting the oil importers, but income and demand in the former group also began to decline as oil prices began to slide after 1981. High interest rates, in general, have had an especially unfavorable impact on the output and employment prospects of market borrowers, especially those that have rather high ratios of private external debt to GDP. The recession in the industrial countries tended to worsen the terms of trade of developing countries that are primary product exporters and tended to reduce the growth (and sometimes level) of export volumes of developing countries generally, with obvious consequences for output. The data shown in Table 26 reveal that the unemployment situation worsened especially for certain primary product exporters of Latin America, Europe, and Asia. The growth rate of exporters of manufactures seemed to have been least adversely affected, even though only Korea and Singapore were able to reduce their unemployment rate effectively. These conclusions must be tempered, of course, in light of the fragmentary nature of the data.
An important factor influencing the way in which employment is affected by exogenous shocks is the functioning of labor markets. But the change in unemployment in response to external shocks depends on a number of other factors as well. Before discussing these factors, it may be useful to recall that an external shock that reduces a developing country’s foreign exchange inflows or raises its debt service payments will tend to reduce its capacity to import, with adverse effects on investment and growth of output. Normally, the result of the import (and investment) cutback will be a decline in output and an increase in unemployment. The extent of these changes in output and unemployment will depend on whether the initial shock is felt by the formal sector or the rural sector, on the degree of flexibility of real wages, the response of labor supply to demand changes (given real wages), and on the nature of structural and policy factors that operate in the labor market. For instance, if real wages are not flexible, the short-term impact of an adverse shock initially felt by the formal sector will be a fall in output and employment in that sector; the informal sector and rural sector employment will then tend to absorb the excess labor. How much aggregate output falls depends on the elasticity of output with respect to employment in the informal and rural sectors. Wage flexibility and reduction of labor supply would mitigate the adverse unemployment effect.
These considerations indicate that the efficiency with which the labor market works (discussed in the previous section) is extremely important in determining the unemployment effects of external shocks to the economy. Labor market efficiency affects labor mobility between sectors, real wage flexibility, and labor-output ratios at any given level and structure of real wages. Nevertheless, the change in unemployment in response to shocks also depends on a number of other factors, some of which are already evident from the above analysis, and which are discussed in this section. Apart from general macroeconomic policies, which do not receive detailed treatment in this paper (see the conclusion), these factors include: the workings of domestic financial markets; the presence of foreign workers or residents working abroad; the nature of various social and customary factors that affect output per worker and labor force participation rate; public sector employment policy; the trend labor force growth; the diversity of the economy; employment structure; and the importance of subsistence activities and barter transactions.
Labor market efficiency. In a reasonably efficient labor market, an adverse shock to the economy as a whole—such as a fall in demand for its major export—that tends to reduce output or its rate of growth, would also tend to increase open unemployment. The increased unemployment, however, would be moderated by declines in real wages (or their rates of growth); the immediate consequence would be increasing ratios of labor to output. In fact, the downward tendency in real wages would in turn permit some moderation of prices and therefore also stem the decline in real output growth. Rudimentary information available on real wages in some Latin American countries indicates that, on the whole, real wages in the early 1980s did decline in the face of adverse external shocks (Table 28); the downward trend in such wages contributed to abating open unemployment in these countries. The notable exceptions are Argentina and Colombia; real wages in these two countries rose between 1980 and 1984 in the face of high rates of open unemployment (see Tables 27 and 28). Scanty and rough data for a few African and Asian countries also indicate that, with the possible exception of Singapore, where output growth rates have been much higher than labor force growth, real wages tended either to decline (or to increase only moderately), in the early 1980s (see Table 30).
The adverse employment effects of an exogenous shock to the economy can, in general, be reduced by improving the efficiency of the labor market. A number of countries have made some progress in this regard. The policies pursued have included: more appropriate producer pricing and exchange rate policies (in a number of low-income countries, notably Ghana and Zaïre); relaxation of minimum wage rules (in virtually all developing countries); and, at least for a while, reduction in the percent of indexation of wages to past inflation (for instance, in Brazil and Israel). In addition, labor union activity has been controlled in many sub-Saharan African and Latin American countries, and public sector wages have been allowed to decline in real terms, particularly in sub-Saharan African countries, in line with declining real GDP growth. Furthermore, also in a number of sub-Saharan African and Asian countries, there has been a tendency to correct urban bias by giving greater budgetary priority to rural development programs rather than urban improvements. Such developments, increasing the relative attractiveness of rural employment, also cause exogenous shocks to have a larger impact on real wages than on output and employment.
Domestic financial markets. The more sophisticated and efficient are domestic financial markets, the smoother will tend to be the long-term and structural changes in output and in labor allocation. In particular, because financial markets are fairly well developed in many middle-income Asian and Latin American countries, industries there that have an incentive to expand are able to obtain loans at reasonable costs to build or expand their plant and equipment, and loans are generally also available to finance the training and moving costs of workers.
The development and efficiency of the financial market will be an even more important factor in the adjustment process when wages are sticky in real terms and shocks result in relatively large increases in unemployment; in such circumstances, there is a greater cost to the economy as a whole to delays in the reallocation of labor. In general, when there is a demand or supply shock requiring labor reallocation, the degree to which wages are flexible will help determine the extent of unemployment during the adjustment period. With sufficient flexibility, wage rates will rise in the favorably affected industries and fall in the adversely affected ones. These changes in relative remuneration will induce a shift (over time) of labor from the falling-wage to the rising-wage industries, until relative wages are once again restored to stable equilibrium levels consistent with the new efficient structure of output. Pressures on unemployment become more intense, however, when, as has recently occurred in some Latin American countries, the domestic financial system is itself under pressure to reduce its exposure.
Government adjustment assistance. Whatever the degree of flexibility of wages and prices, or the state of development of financial markets, official adjustment assistance can help abate the unemployment associated with reallocation of labor among industries and sectors. Such assistance can take many different forms, including explicit subsidies, provision of training in public facilities, guaranteeing private loans, or making direct loans to workers and business establishments. The high urban unemployment rates in Latin America in the 1980s, and particularly among youths, have, for example, encouraged increased efforts in these countries (including Brazil, Colombia, Ecuador, and Venezuela) to provide publicly supported training programs for the unemployed, and especially those between 14 and 24 years of age.47
Foreign workers and workers in foreign countries. In the transmission of exogenous shocks to unemployment in developing countries, migrant workers have often played an important role. There are some 3½ million to 4 million migrant workers in Latin America, for example, mainly from other Latin American countries, especially Bolivia, Colombia, and Paraguay; Argentina and Venezuela receive the most in-migration of any Latin American countries. In 1980, South Africa registered some 287,000 workers from neighboring countries; by June 1983, this figure had risen to 358,000. It is also estimated that in 1980 migrant workers in West Africa totaled about 1.3 million. Both the United States and the European Community have foreign workers from the developing world numbering in the millions.48
In recent years, there are two groups of countries for which the economic significance of international migration has been most notable (see Table 31). On the one hand there are fuel exporters (around the Persian Gulf) with extremely high oil revenue per capita; these countries have generally relied on immigrants for a substantial part of their labor supply. The most striking examples are Kuwait and the United Arab Emirates where about 70 percent and 85 percent, respectively, of the work force are estimated to be non-nationals.49 Immigrants in general have borne a disproportionate share of the burden of labor supply adjustment to the recent slowdown in these economies, which have therefore shown rather low elasticities of national unemployment to output.
|Arab Region 2||Percent of|
On the other hand, there are the service and remittance countries (especially Egypt, Greece, Pakistan, and Portugal) and a number of other countries (such as Turkey and Yugoslavia) that, because of high rates of population growth (relative to growth in other factors of production), or generally low per capita income in relation to neighboring (or easily accessible) countries, have experienced substantial emigration of their nationals. These emigrants have generally gone to the Gulf States or to the major industrial countries of Western Europe. Easily the greatest economic significance of such migration is that migrants’ remittances tend to be a very important foreign exchange earner for the out-migration countries.
Migrants from the service and remittance countries tend to leave and return to their native countries according to whether economic conditions are improving or worsening in the host countries. Accordingly, the repercussions of recent global economic shocks have forced these out-migration countries to adjust to returning migrants, or reduced real remittances, or both; the effect of such pressures on their unemployment has depended not only on the migrants’ response but also on the other factors discussed above, including the efficiency of the labor market and the policy responses that affect such efficiency, and the nature of official adjustment assistance. The rise in the unemployment rate of Yugoslavia in 1984 was partly attributable to these migration factors.
Social and customary factors. There are certain social and customary factors that tend to lower output per worker or lower labor force participation rates when capacity utilization rates drop. The importance of such factors varies across countries and affects the transmission of exogenous shocks to unemployment.
In certain sub-Saharan African and Asian countries employers may tend to hoard labor and reduce man-hours during downturns, even in the absence of officially-imposed tenure laws.50 Such labor tenure constraints, common especially in large manufacturing and mining concerns, do not necessarily reduce the average profits of the firm over the cycle but may be part of an implicit contract between employers and employees, whereby the employee, in return for a degree of security of tenure during unfavorable economic circumstances, accepts a lower wage than he would otherwise demand.51 In addition, governments in some countries during the recent past have borne some of the cost of extra workers through special tax relief and other contributions—especially when the companies are also partly government owned.
Social and customary rights to land, as well as social work-sharing arrangements in the rural subsistence subsector or in the urban informal sector, also influence open unemployment following a demand or supply shock. For instance, in many sub-Saharan African countries, family rights to land increase the accessibility of individuals to an important employment opportunity while work-sharing arrangements are quite common. These arrangements have helped hold down the elasticity of open unemployment to output changes in these economies during the last five years; despite the absence of reliable statistical series, casual, self-employed, and unpaid family workers appear to have increased as a share of the work force.
In many countries the variance of unemployment rates over the cycle tends to be reduced by procyclical labor force participation rates, while in farming communities (particularly those with a sizable number of family workers) the labor force participation rates vary with the seasonal demand for labor. This procyclical variation in participation rates, caused chiefly by women and young people entering or leaving the labor force in response to the availability of employment, has tended to hold down unemployment rates in many African and Asian countries with large rural work forces.52
Public sector employment policy has greatly influenced the transmission of adverse shocks to unemployment (particularly open unemployment). In particular, in countries with a large proportion of wage employment in the public sector, the adverse effects of shocks on their open unemployment tend to be mitigated by the retention of public sector employees during adverse economic circumstances. In such countries, the unfavorable world economic developments of recent years have resulted in reduced real output growth, an increased share of public sector employment in total formal sector employment, and an increased budgetary deficit in relation to nominal GDP, with only a moderate worsening of open unemployment.
Apart from labor tenure policies, the governments of certain countries (such as Chile, Jamaica, and Panama) have instituted special employment programs in order to abate unemployment.53 Experience shows that such programs can be quite flexible and can be geared toward specific regions or groups of persons as required. Their major disadvantage is that they use up budgetary resources and, for that reason, cannot be used to substantially abate a serious unemployment crisis.
Labor force growth. The underlying (trend or normal) labor force growth depends crucially on population growth and normal labor force participation rates. In general, the more rapid the growth of the population, the more rapid that of the labor force tends to be. The elasticity of the labor force with respect to population is greatly influenced by the “dependency ratio,” which, for international comparisons, is generally measured as the number of persons aged less than 15 (“children”) and over 64 (“elderly people”) as a percent of persons between these two ages. Of course, the true “dependency ratio,” in the sense of the ratio of those in the dependent ages to those in the working ages, could well differ from this international norm; the ages at which people normally enter and leave the work force vary across countries.
The portion of the working age population that enters the labor market, the labor force, depends on a variety of factors, some of which have been mentioned earlier.54 The higher the trend growth rate of the labor force, the more adverse, ceteris paribus, the impact of an unfavorable external shock on the unemployment rate (however it is measured). For instance, the adverse unemployment effects on Mexico and Venezuela of the need to curb imports and undertake domestic retrenchment have been aggravated by the high trend in labor force growth in those countries (about 3.2 percent and 4.1 percent, respectively, during 1970–82). (The labor force growth of Venezuela slowed down in the 1980s probably, in part, as a result of declining immigration. During 1981–84 the annual growth rate of Venezuela’s labor force averaged only 3.5 percent.)
Employment structure and diversity of economy. Open unemployment tends to be a phenomenon of wage employment, which is especially prevalent in industry, mining, and urban services. For instance, recent economic developments in the world economy have affected directly a higher fraction of the work force in Latin American countries than in African countries; industry, mining, and services employ a higher fraction of the work force in Latin America and a substantial fraction of this employment is wage employment. This does not mean that the output effects of the recent slowdown have not been severe for Africa as well. The point is that the importance of self-employment and family work has greatly moderated open unemployment, as well as the fall in output, in sub-Saharan Africa compared with Latin America.55
In particular, the adverse output and employment impact of worldwide shocks can be mitigated by subsistence and nonmarket activities. Subsistence activities, although important mainly in the primary sector—agriculture, hunting, fishing, and animal husbandry—are not necessarily confined to that sector. For instance, in many areas of rural sub-Saharan Africa a good share of construction activity is of a subsistence (nonmarket) nature. Similarly, in African and many other developing countries, many persons are engaged partly in wage and market activities and partly in subsistence activities, depending on relative opportunity and returns. The easier it is to obtain subsistence employment—the easier it is to obtain cooperative factors, such as land, to engage in subsistence activities—the shorter the duration of unemployment for any such individual.
Another relevant factor affecting the employment impact of exogenous shocks is the extent of diversification of the monetized sectors. Many oil producing countries, and certain other primary exporters with one or two crops or mineral products that dominate exports, government revenue and output, have found it difficult to adjust to large output shocks emanating from the world economy; other domestic products could not compensate for the large drop experienced in the output of the dominant crops or minerals. The effects of such output declines on unemployment has been influenced by the factors discussed earlier in this section.
Findings and Concluding Remarks
Developments in the world economy since 1979 have tended to worsen the employment situation in developing countries. But intercountry differences in the operation of labor markets have influenced both the response of output to the external shocks and the employment response to falls in output. This paper has examined how these responses have been influenced by such factors as the degree of real wage flexibility, labor force growth, policies that affect labor mobility between sectors, government employment and wage policies, immigration and emigration of labor, government controls on private sector employment and on commodity prices, and the employment structure of the economy.
Where the growth of real output has been sustained at a rate higher than the trend rate of growth of labor force, rising unemployment during the years since 1979 has been avoided. A number of countries have been able to reduce their unemployment rates by pursuing demand management, supply-side, and exchange rate policies that have contributed to sustaining high growth rates. Demand management policies have provided a stable financial setting for investment by keeping fiscal deficits, domestic credit expansion, and external indebtedness under control. Supply-side policies have included: movement toward positive real interest rates; improvements in the efficiency of public investment; modification of the levels and structure of taxes and subsidies (so as to encourage private investment, increase capacity utilization rates, and improve the structure of output); and reforms in the pricing policies of public enterprises and of state marketing boards (to reduce budgetary subsidies while encouraging a more appropriate production structure). Exchange rate policies have been designed to maintain, and where necessary raise, the international competitiveness of traded goods sectors. When these macroeconomic policies have been superimposed on adverse exernal developments, as they tend to have been in recent years, they have not, of course, always achieved the goal of efficiently allocating labor without some short-to medium-term costs in the form of increased unemployment; smooth adjustment without increased unemployment has necessitated price and wage flexibility and other favorable labor market conditions that have not, in fact, been sometimes operating. Chile and Turkey are only two examples where an appropriate policy framework has in some respects had disappointing results because of such institutional constraints.
Unemployment in developing countries in the first half of the 1980s has, therefore, reflected the interplay of external developments, general macroeconomic policies, factors (including domestic policies) that operate directly in labor markets, and various other autonomous factors. As discussed in earlier sections, the data show that the effect of all these forces on unemployment have been particularly severe for Latin America, where between 1979 and 1984 open unemployment rates increased and real wages declined. In general, there was some compensating role played by the informal sector, which increased its share of total urban employment. Countries like Bolivia, Colombia, and Paraguay, for which, in the past, migration has acted to relieve unemployment pressures at home, have been less fortunate in recent years, since Argentina and Venezuela, the traditional recipients of migration, have also experienced serious adverse output and unemployment pressures in the 1980s. In Latin America, in addition, serious underemployment has occurred and between 1979 and 1984 the number of hours worked by those employed, in the urban formal sector, relative to the number of hours desired, dropped appreciably.
A similar picture exists in other parts of the developing world, although data are scantier for the majority of the countries; the countries with reasonable amounts of data tend to be those with a substantial fraction of the labor force in wage employment. In any event, between 1979/80 and 1984 the evidence is that for these other areas as well, there were increases in open unemployment and in the relative size of the labor force in the urban informal sector, while there were declines in real wages, hours of work in wage employment, and the growth rate of real GDP. Against this general background, certain specific developments are noteworthy.
First, as already stated, in some countries (such as Korea and Singapore) general macroeconomic policies successfully forestalled a worsening of open unemployment by sustaining an underlying output growth rate greater than the trend labor force growth rate.
Second, the example of countries in sub-Saharan Africa shows that when a large proportion of total employment is in the informal and rural sectors, open unemployment can be alleviated. Where employment in the formal sector is dominant, such mechanisms were not effective. Even where effective, however, there may be a drop in real growth rate and an increase in poverty.
Third, the public sector has helped alleviate open unemployment in sub-Saharan Africa and in certain other countries but the consequence has been increased surplus labor in the public sector, coupled with rising budgetary deficits in relation to GDP. These deficits have in some countries led, in turn, to credit expansion, domestic inflation, and continuous devaluation of the national currency.
Fourth, countries such as Egypt and Pakistan that have depended on international migration to oil exporting countries to absorb some portion of their labor force were adversely affected by the fact that price declines and slowdown in demand for oil have depressed real GDP growth in the oil producing countries since 1981/82.
Finally, social conditions and institutional mechanisms already operating in labor markets before the output shocks, as well as policy measures affecting labor market efficiency, have clearly played an important part in the employment response to the exogenous shocks experienced during the early 1980s. The lack of reliable and comparable statistical data across developing countries has made a more complete analysis of the role of individual factors impossible. The systematic collection and analysis of quantitative data on the operation of labor markets in developing countries would seem, therefore, to be of some importance for the understanding of the effects on real output and employment of general macroeconomic policies and external shocks to the economy.
Ahmad, Ehtisham and NicholasStern, “Employment and Wages in Pakistan” (Discussion Paper, Report No. DRD 158, World Bank, December1985).
Bairoch, Paul, Urban Unemployment in Developing Countries: The Nature of the Problem and Proposals for its Solution (Geneva: International Labour Office, 1973).
Banerjee, Biswajit and J.B.Knight, “Caste Discrimination in the Indian Urban Labour Market,” Journal of Development Economics (Amsterdam), Vol. 17 (April1985), pp. 277–307.
Bardhan, Pranab K., “Wages and Unemployment in a Poor Agrarian Economy: A Theoretical and Empirical Analysis,” Journal of Political Economy (Chicago), Vol. 87 (June1979), pp. 479–500.
Bardhan, Pranab K., Land, Labor, and Rural Poverty: Essays in Development Economics (New York: Oxford University Press, 1984).
Ben-Porath, Yoram, “A Profile of the Labor Market in Israel” (Discussion Paper, Report No. DRD 154, World Bank, August1985).
Berg, Elliot J., “Wage Structures in Less-Developed Countries,” in Wage Policy Issues in Economic Development, edited byAnthony D.Smith (London: Macmillan, 1969), pp. 276–91.
Berry, R. Albert, “Open Unemployment as a Social Problem in Urban Colombia: Myth and Reality,” Economic Development and Cultural Change (Chicago), Vol. 23 (January1975), pp. 276–91.
Berry, R. Albert and R.H.Sabot, “Unemployment and Economic Development,” Economic Development and Cultural Change (Chicago), Vol. 32 (October1984), pp. 99–116.
Blomquist, A.G., “Urban Job Creation and Unemployment in LDCs: Todaro vs Harris and Todaro,” Journal of Development Economics (Amsterdam), Vol. 5 (March1978), pp. 3–18.
Brittain, John A., The Payroll Tax for Social Security, Studies of Government Finance (Washington: The Brookings Institution, 1972).
Bronstein, Arturo S., “Collective Bargaining in Latin America: Problems and Trends,” International Labor Review (Geneva), Vol. 17 (September1978), pp. 583–96.
Bruno, Michael and Jeffrey D.Sachs, Economics of Worldwide Stagflation (Cambridge, Massachusetts: Harvard University Press, 1985).
Cain, Glen G., “The Challenge of Segmented Labor Market Theories to Orthodox Theory: A Survey,” Journal of Economic Literature (Nashville, Tennessee), Vol. 14 (December1976), pp. 1215–57.
Drazen, Allan, “Unemployment in LDCs: Worker Heterogeneity, Screening, and Quantity Constraints,” World Development (Oxford), Vol. 10 (December1982), pp. 1039–47.
Fields, Gary S., “Rural-Urban Migration, Urban Unemployment and Underemployment, and Job-Search Activity in LDCs,” Journal of Development Economics (Amsterdam), Vol. 2 (June1975), pp. 165–87.
Fields, Gary S., “Public Policy and the Evolution of the Labor Market in Sri Lanka,” (Discussion Paper, Report No. DRD 161, World Bank, March1986).
Goode, Richard, Government Finance in Developing Countries, Studies of Government Finance. Second Series (Washington: The Brookings Institution, 1984).
Gray, Jo Anna, “Wage Indexation: A Macroeconomic Approach,” Journal of Monetary Economics (Amsterdam), Vol. 2 (April1976), pp. 221–35.
Hansen, Bent, “The Egyptian Labor Market: An Overview” (Discussion Paper, Report No. DRD 160, World Bank, June1985).
Harris, John R. and Michael P.Todaro, “Migration, Unemployment and Development: A Two-Sector Analysis,” American Economic Review (Nashville, Tennessee), Vol. 60 (March1970), pp. 126–42.
Herrick, Bruce and Charles P.Kindleberger, Economic Development, Economics Handbook Series (New York: McGraw-Hill, 1983).
House, William J. and HenryRempel, “The Impact of Unionization on Negotiated Wages in the Manufacturing Sector in Kenya,” Oxford Bulletin of Economics and Statistics (Oxford), Vol. 38 (May1976), pp. 111–23.
International Labour Office, Thirteenth International Conference of Labour Statisticians, Resolution I: “Resolution Concerning Statistics of the Economically Active Population, Employment and Underemployment (Geneva: I.L.O., 1983).
International Labour Office, Yearbook of Labour Statistics (Geneva: I.L.O., 1984, 1985).
International Labour Office, “The Crisis in the North and the South: The Impact of the World Recession on Employment and Poverty in Developing Countries” (Geneva: I.L.O., May1985).
International Labour Office, World Labour Report (Geneva: I.L.O., 1985), 2 vols.
Kannappan, Subbiah, “Urban Employment and the Labor Market in Developing Nations,” Economic Development and Cultural Change (Chicago), Vol. 33 (July1985), pp. 699–730.
Kilby, Peter, “Industrial Relations and Wage Determination: Failure of the Anglo-Saxon Model,” Journal of Developing Areas (Macomb, Illinois), Vol. 1 (1967), pp. 489–520.
Knight, J.B. and M.D.McGrath, “An Analysis of Racial Wage Discrimination in South Africa,” Oxford Bulletin of Economics and Statistics (Oxford), Vol. 39 (November1977), pp. 245–71.
Knight, J.B. and R.H.Sabot, “Labour Market Discrimination in a Poor Urban Economy,” Journal of Development Studies (London), Vol. 19 (October1982), pp. 67–87.
Kritz, Ernesto and JosephRamos, “The Measurement of Urban Underemployment: A Report on Three Experimental Surveys,” International Labour Review (Geneva), Vol. 113 (January1976), pp. 115–27.
Leibenstein, Harvey, Economic Backwardness and Economic Growth: Studies in the Theory of Economic Development (New York: Wiley, 1957).
Lewis, W. Arthur, The Theory of Economic Growth (Home-wood, Illinois: Irwin, 1955).
Lopes, Francisco L. and Edmar L.Bacha, “Inflation, Growth and Wage Policy: A Brazilian Perspective,” Journal of Development Economics (Amsterdam), Vol. 13 (August1983), pp. 1–20.
Macedo, Roberto, “Wage Indexation and Inflation: The Recent Brazilian Experience,” in Inflation, Debt and Indexation (Cambridge, Massachusetts: M.I.T. Press, 1983), edited byRudigerDornbusch and Mario HenriqueSimonsen.
Macedo, Roberto, “The Brazilian Labor Market: An Overview” (Discussion Paper, Report No. 151, World Bank, January1986).
Mazumdar, Dipak, The Urban Labor Market and Income Distribution: A Study of Malaysia (New York: Oxford University Press for the World Bank, 1981).
Mazumdar, Dipak, “The Rural-Urban Wage Gap, Migration and the Working of Urban Labor Market: An Interpretation Based on a Study of the Workers of Bombay City,” Indian Economic Review (Delhi), Vol. 18 (July1983), pp. 169–98.
Meier, Gerald M., “Development Without Employment,” Quarterly Review, Banca Nazionale Del Lavoro (Rome), No. 90 (September1969), pp. 309–19.
Mohan, Rakesh, “The Determinants of Labour Earnings in Developing Metropoli: Estimates from Bogota and Cali, Colombia,” World Bank Staff Working Paper, No. 498 (October1981).
Oberai, A.S., “State Policies and Internal Migration in Asia,” International Labour Review (Geneva), Vol. 120 (March1981), pp. 231–44.
Prest, Alan R., “The Role of Labour Taxes and Subsidies in Promoting Employment in Developing Countries,” International Labour Review (Geneva), Vol. 103 (April1971), pp. 315–32.
Rosen, Sherwin, “Implicit Contracts: A Survey,” Journal of Economic Literature (Nashville, Tennessee), Vol. 23 (September1985), pp. 1144–75.
Rosenberg, Richard E., “Payroll Taxation in Developing Countries: The Philippine Case,” Journal of Development Studies (London), Vol. 13 (January1977), pp. 144–61.
Sabot, R.H., “The Meaning and Measurement of Urban Surplus Labour,” Oxford Economic Papers, New Series (London), Vol. 29 (November1977), pp. 389–411.
Sen, Amartya Kumar, Employment, Technology and Development, Economic Development Series (Oxford: Clarendon Press, 1975).
Sethuraman, S.V., ed., The Urban Informal Sector in Developing Countries: Employment, Poverty, and Environment (Geneva: International Labour Office, 1981).
Simonsen, Mario Henrique, “Indexation: Current Theory and the Brazilian Experience,” in Inflation, Debt and Indexation, edited byRudigerDornbusch and Mario HenriqueSimonsen (Cambridge: M.I.T. Press, 1983).
Souza, Paulo R. and Victor E.Tokman, “The Informal Urban Sector in Latin America,” International Labour Review (Geneva), Vol. 114 (November1976), pp. 355–65.
Squire, Lyn, Employment Policy in Developing Countries: A Survey of Issues and Evidence (New York: Oxford University Press for the World Bank, 1981).
Standing, Guy, Labour Force Participation and Development (Geneva: International Labour Office, 1981a).
Standing, Guy, “The Notion of Voluntary Unemployment,” International Labour Review (Geneva), Vol. 120 (September1981b), pp. 563–79.
Stiglitz, Joseph E., “Alternative Theories of Wage Determination and Unemployment in LDC’s: The Labor Turnover Model,” Quarterly Journal of Economics (Cambridge, Massachusetts), Vol. 88 (May1974), pp. 194–227.
Stiglitz, Joseph E., “The Efficiency Wage Hypothesis, Surplus Labour, and the Distribution of Income in L.D.C.s,” Oxford Economic Papers, New Series (London), Vol. 28 (July1976), pp. 185–207.
Thirlwall, Anthony P., Growth and Development, with Special Reference to Developing Economies (London: Macmillan, 1978).
Tokman, Victor E., “An Exploration Into the Nature of Informal-Formal Sector Relationships,” World Development (Oxford), Vol. 6 (September1978), pp. 1065–75.
Tokman, Victor E., “The Employment Crisis in Latin America,” International Labour Review (Geneva), Vol. 123 (September1984), pp. 585–97.
Turnham, David and IngeliesJaeger, The Employment Problem in Less Developed Countries: A Review of Evidence, Employment Series No. 1 (Paris: Organization for Economic Cooperation and Development, 1971).
Udall, Alan T. and StuartSinclair, “The ‘Luxury Unemployment’ Hypothesis: A Review of Recent Evidence,” World Development (Oxford), Vol. 10 (January1982), pp. 49–62.
Watanabe, Susumu, “Minimum Wages in Developing Countries: Myth and Reality,” International Labour Review (Geneva), Vol. 113 (May1976), pp. 345–58.
Webb, Richard C., “Wage Policy and Income Distribution in Developing Countries,” in Income Distribution and Growth in the Less-Developed Countries, edited byCharles R.Frank and R. C.Webb (Washington: The Brookings Institution, 1977), pp. 215–58.
Weeks, J., “Does Employment Matter?” in Manpower and Unemployment Research in Africa (Montreal: Centre for Developing Area Studies, McGill University), Vol. 4, No. 1.
World Bank, World Development Report (New York: Oxford University Press, 1984).
World Bank, Policies for Growth and Employment—Indonesia (Washington1985a).
World Bank, World Development Report (New York: Oxford University Press, 1985b).
Yap, Lorene Y.L., “Internal Migration and Economic Development in Brazil,” Quarterly Journal of Economics (Cambridge, Massachusetts), Vol. 90 (February1976), pp. 119–37.
For instance, World Bank 1985(b), Table 11, pp. 214–15. On the whole, only a few countries or territories with special circumstances—such as oil-rich Kuwait (which has virtually no agricultural sector), Singapore (a city state), and Israel (a small relatively industrialized nation)—have the same percentages of the labor force in agriculture as the typical industrial country. Argentina, Lebanon, and Uruguay have percentages that are relatively low and similar to those for Japan and Italy, which have the highest percentages among the major industrial countries.
This conclusion is endorsed by others. For instance, Kritz and Ramos (1976), after conducting surveys of low-income neighborhoods of Managua (Nicaragua), of the whole of Santo Domingo (Dominican Republic), and of Asuncion (Paraguay) in the 1970s, concluded that underemployment was far more pervasive in developing countries than was open unemployment.
See, for example, International Labour Office (1983).
Ibid. The paper by Udal and Sinclair, particularly, is an excellent introduction to these studies.
See, for example, the discussion in Sen (1975).
See Sabot (1977) for discussion and application of this “allocative efficiency income” criterion.
For instance, in the labor force surveys of Pakistan the “unemployed” include those “looking for work, or not looking for work because of illness, or not looking for work believing jobs are not available, or temporarily or indefinitely laid off, or waiting to report to a new job, or willing to work if a job were provided, or apprentices with no guaranteed job, or had some usual occupation but were doing nothing during the reference period.” Quotation in Ahmad and Stern (1985).
See Tokman (1978) for an introduction to the issues raised in the debate as to whether the relationship between the informal and formal sectors is one of complementarity or subordination of the former to the latter.
For a critique of the attempt to distinguish voluntary from involuntary unemployment see Standing (1981b).
Yap (1976), for instance, finds that rural-urban migration has had a positive effect on Brazil’s postwar development; the rate of real GDP growth in Brazil would have been lower with lower migration than actually occurred.
The extensive literature on rural-urban migration in developing countries has incorporated the above sources of distortions into the various models. A good introduction to this literature can be obtained by reading Harris and Todaro (1970), Fields (1975), Blomquist (1978), Yap (1977), and Mazumdar (1984). For a good introduction to the issues of urbanization (including urban bias) and urban economic development, see Herrick and Kindleberger (1983), pp. 388–407. For an overview of the state of labor relations, of international labor standards, and of working conditions in the world today, see Volume 2 of the important World Labour Report recently (1985) published by the International Labour Office.
A notable exception to this general tendency to neglect the rural areas, relative to urban areas, that has attracted the attention of researchers, is Sri Lanka; see Fields (1986).
Because the structure of production changes with economic development toward industry and certain service activities where economies of scale tend to be present, and because technological developments in agriculture tend to be labor saving, urbanization tends to increase with economic development. Hence the adverse allocative consequence of urban bias cannot be assessed merely by looking at how fast a country is urbanizing. Some adjustment must be made for efficient urbanization that would be predicted by efficient changes in the structure of production.
According to the International Labour Office, virtually all developing countries have minimum wage laws whose scope has been expanded to cover more and more workers (see World Labour Report (1985), Vol. 1, p. 143). See also Squire (1981) chapter 8, for discussion and for reference to particular country studies.
See Watanabe (1976) and I.L.O. ibid., p. 144.
See Simonsen (1983) for discussion of the link between the level of wages and the degree of indexation.
See, for example, the discussion in International Labour Office, World Labour Report, Vol. 2, chapter 6, of national law and requirements regarding protection against unjustified dismissal.
See, for example, Knight and McGrath (1977), for the South African case.
See Oberai (1981) for an introduction to these examples.
See, for instance, the case of Israel as discussed by Ben-Porath (1985).
See Brittain (1972) for analysis and for evidence that in industrial countries the incidence of the social security tax is borne by labor. Rosenberg (1977), in discussing the Philippine case, argues that, because of the high elasticity of the labor covered by the social security tax, the burden is mainly on labor, either through forward shifting, which, by raising product prices, leads to a fall in real wages, or directly through wage restraint.
See Goode (1984), p. 254, for a brief reference.
See Goode op. cit, p. 253 and references cited there, as well as Prest (1971).
See Goode op. cit.
Some discussion of these difficulties can be found in a number of places, including the International Labour Office, World Labour Report (1985), Vol. 2, chapter 1.
See, for instance, Bronstein (1978).
See Meier (1969).
See Mohan (1981).
But see Knight and Sabot (1982), and Banerjee and Knight (1985). One of the most interesting findings of the Banerjee-Knight study is that job discrimination between the castes was greatest for manual (“operative”) jobs where contacts are important for recruitment. See also Mazumdar (1981) chapter 10 for some discussion of the Malaysian case.
See, for example, Mazumdar (1984).
For instance, see Drazen (1982).
Segmented labor market theories—which have relevance for developing countries—of course cover a broad range of issues, both theoretical and policy related. See the survey paper of Cain (1976).
These notions are consistent with incentive-efficiency and similar theories. See Stiglitz (1974) and (1976), for instance. Hansen, for example, tends to believe the incentive version of the efficiency wage theory in discussing why the rural labor market in Egypt does not clear outside the peak seasons despite considerable downward flexibility of wages (see Hansen (1985)).
International Labour Office, World Labour Report (1985), Vol. 2, chapter 8.
International Labour Office, World Labour Report (1985), Vol. 2, chapter 4 and Fund country data.
International Labour Office, World Labour Report (1985), Vol. 1, p. 101.
This sort of behavior is not unfamiliar even among industrial countries; see, for example, Bruno and Sachs (1985), p. 221.
For a survey of implicit contract theory as it has been developed in the industrial country context, see Rosen (1985).
See, for instance, Bardhan (1984), chapter 2.
See Tokman (1984).
See, for example, Standing (1981a), for an extended analysis on the major determinants of labor force growth and labor force participation rates.
In addition to the difficulty of properly accounting for the impact of these developments in employment data, there are also well-known problems of properly tracking changes in output in the non-monetized sector.