Journal Issue
Share
Article

Trade strategies and employment in developing countries: Results of a study of the link

Author(s):
International Monetary Fund. External Relations Dept.
Published Date:
June 1984
Share
  • ShareShare
Show Summary Details

A recent study provides a better understanding of the link

Anne O. Krueger

The growth of employment, and in particular nonagricultural employment, is a major concern in virtually all developing countries. In the early stages of growth, when a large fraction of the labor force is engaged in agriculture with low productivity, successful development must inevitably be accompanied by increasing total agricultural productivity and the simultaneous creation of nonagricultural employment opportunities for a rising fraction of the labor force.

The rate of increase in nonagricultural employment opportunities in many countries has been sluggish, however. Although the typically rapid overall rate of growth of the labor force would, in any event, result in difficult development problems, an important question is how much more rapidly employment might grow under different policy regimes. While there is some presumption that a higher rate of growth will lead to a more rapid increase in the demand for labor, there is also evidence that some other policies have positive effects on labor demand.

Earlier research showed a strong association between trade strategies and growth rates, and growth performance seemed better for countries favoring export-oriented policies than for countries encouraging import-competing industries. (The link between trade strategy and the rate of economic growth was discussed in an earlier article in Finance & Development, “The effects of trade strategies on growth,” June 1983.) The links between trade strategies and development, however, were not entirely understood. The project, set up under the National Bureau of Economic Research, studied the relationship between alternative trade regimes and employment, and the main findings are reported in this article.

In the project, several levels were recognized, at which trade regimes could affect employment and its rate of growth. First, one trade strategy may result in a higher rate of growth of the overall economy, owing to superior resource allocation, and faster growth would entail higher growth in employment. Second, different trade strategies imply different compositions of output at each point in time. Under an export promotion strategy export industries grow faster, while under an import-substitution policy they grow more slowly. If employment per unit of output (or value added) is higher in one set of industries than in the other, it would increase more rapidly under a strategy that allows labor-intensive industries to grow relatively faster. Finally, trade policies could influence the choice of techniques and the capital/labor ratio in all industries. If such policies lead to greater capital intensity and fewer jobs per unit of output in all lines of economic activity, then employment opportunities will grow more slowly as there is continued capital deepening (i.e., increased intensity).

Factor proportions

The basis for the analysis of the effects of trade strategies on employment was an extended version of the Hecksher-Ohlin-Samuelson (HOS) model. For countries where markets seem to function fairly efficiently, the model posits that differences in factor endowments will be a major source of cost differentials between countries and thus will determine the patterns of trade. That is, a labor-abundant country will have a labor-intensive manufacturing sector; and capital-intensive goods will generally be imported rather than produced domestically under an efficient allocation of resources. Given the country’s overall relative capital/labor ratio in manufacturing, the commodities produced domestically will require inputs approximating the ratio, while commodities imported and not domestically produced will have factor proportions further away from the country’s endowments.

This article is based on conclusions from the author’s volume, Trade and Employment in Developing Countries, Volume 3: Synthesis and Conclusions (University of Chicago Press for the National Bureau of Economic Research, Chicago, 1983).

The pattern of trade derived from the HOS model has the following properties: (1) manufactures whose production functions are more capital intensive will be imported from countries with higher capital/labor ratios in their manufacturing sectors, and goods that are extremely labor intensive from the countries with lower capital/labor ratios in their manufacturing sectors. (2) Insofar as the country’s manufacturing exports go to different destinations, the more capital-intensive exports will go to the more labor-abundant areas, and conversely. (3) In the process of successful growth, countries’ comparative advantage will move toward less labor-intensive manufactures as their real wage rises.

However, trade patterns can vary from this efficiency model in countries where distortions in the goods market are such that domestic prices diverge from international prices by more than transport costs, or where distortions exist in the factor market so that domestic factor prices do not reflect their opportunity cost. When factor market distortions are significant, a labor-abundant country that would export labor-intensive commodities under an efficient allocation may actually export capital-intensive commodities, reversing its true comparative advantage.

When reversals do not take place, however, this does not imply that factor market imperfections are nonexistent or insignificant. Important questions would still require analysis: How sizable are factor market distortions? How great is their influence upon factor proportions within industries? What is the effect of product and factor market distortions upon the output mix? And how much would the employment implications of alternative trade strategies differ in the absence of those distortions? What sorts of policies to implement the choice of trade strategy are likely to increase the demand for labor or to adversely affect it? Are they integral parts of the trade regime or unnecessary appendages?

The NBER project could not hope to provide definitive quantitative estimates to answer these questions given the paucity of the data. Nonetheless, some evidence emerged with which to estimate the degree to which wage and capital costs diverged from those under well-functioning factor markets in the countries covered by the project. These estimates had limited comparability across countries, but they indicated that in countries with labor market regulations, import substitution trade regimes, credit rationing, and social insurance tax systems, these all contributed to raising the domestic cost of hiring labor relative to the cost of using capital. There was no single pattern of relative importance among these factors across all the countries, but what did seem clear, even from these impressionistic data, is that the export-promoting territories and countries—Hong Kong, Korea, Brazil, and, to a lesser extent, the Ivory Coast—had relatively lower factor market distortions than the import-substitution countries, with the possible exception of Argentina.

A second conclusion, probably not unrelated to the first, was that currency overvaluation and the favorable treatment of imports of capital goods were also important sources of cheaper capital for firms eligible to import. Credit rationing at subsidized interest rates constituted another major source of underpricing of capital services. Finally, social insurance taxes drove a wedge of about 20–30 percent in the price of labor between firms and sectors subject to the taxes and other activities within the economy.

Each of these sources of pricing disparity between firms and sectors by itself could have significantly affected incentives, but when they were all found together the effects must have been fairly powerful. All of them induced lower capital costs and higher labor costs than would have prevailed if efficient patterns of resource allocation had occurred.

Empirical evidence

It is difficult to generalize about the employment experience in each of the project territories and countries, but several patterns seemed to appear. First, in some countries, urban employment opportunities grew relatively slowly for a period of time, and real wages paid to urban workers were either very high or rising fairly rapidly. This group of countries included Brazil, Chile, and Colombia in 1963–68, Thailand in 1960–68, and Uruguay in 1963–75.

In a second group of countries, real wages did not rise very much, and urban employment grew at moderate rates. This group included Argentina and Brazil in the late 1960s and early 1970s, Colombia after 1968, and Thailand between 1968 and 1973, as well as Indonesia, the Ivory Coast, Pakistan, and Tunisia in the 1960s and early 1970s. The third group, Korea and Hong Kong, appears to have had relatively full employment with rising real wages and expanding urban employment opportunities in the 1960s and 1970s. Except for the latter two, every country in the project experienced a significant “employment problem” during part of the period under review. Either unemployment rose along with real wages, or real wages were stagnant but urban employment grew only moderately.

At the outset of the project, the intent was to ascertain the extent to which domestic factor market conditions inhibited the labor-abundant developing countries from realizing their comparative advantage in international trade. The appropriate trade data were assembled and developing countries’ factor markets analyzed to determine the links between trade strategies and employment. These data were then assembled to permit a direct comparison of labor inputs, per unit of output, in export and import-competing industries.

The table provides the results from individual country studies and Hong Kong. The numbers are not comparable across countries, although the ratios of export and import use are. For Indonesia, for example, exports to developed countries used an average of 2,176 man-days, compared to 994 man-days used to generate equal domestic value added in industries competing with imports from developed countries. In general, for most of the countries in the study, export production was, by and large, considerably more labor intensive than production of import-competing goods. Despite factor market conditions, developing countries’ manufactured exports tended to exhibit the factor intensity consistent with their endowment.

Regarding the composition of employment, especially skilled employment, estimates of the coefficients for inputs of skilled, unskilled, and managerial labor showed large and systematic differences for export and import-competing industries. There was strong evidence that the production of exports used more unskilled labor and less skilled managerial labor than import-competing industries. This suggested that trade enabled developing countries to substitute their relatively abundant factor—unskilled labor—for their relatively scarce factor.

The HOS model suggests a priori grounds for expecting differences in factor intensity between exports to developed (and more capital-abundant) countries and those destined for other developing (and more labor-abundant) countries, as well as between imports by source. The model predicts that gains from trade will be large between countries with dissimilar factor endowments. If all developing countries had a comparative advantage in goods that use unskilled labor relatively intensively, it is unlikely that they would be able to penetrate each other’s markets significantly. The coefficients in the table are not necessarily perfect indicators of what would happen with an alteration in trade strategy, but the orders of magnitude suggest the importance of two factors: which sectors tradables originate in; and which countries they go to.

Direct labor coefficients per unit of domestic value added by direction of trade1(In man-years)
HOS exportablesHOS import-competing products
Country/territoryPeriodDeveloped countriesDeveloping countriesTotalDeveloped countriesDeveloping countriesTotal
Argentina1973164147
Brazil1959115141115128
197089798771
1972109788771
Chile1966–68612934434343
Colombia1970

1973
28

32
21

24
24

29






Hong Kong1973756773625560
Indonesia219712,1762,1492,1759941,1171,038
Ivory Coast319722,4881,5201,7431,652
Pakistan1969–709088887012071
Thailand1973222022112211
Uruguay1968441239366238
Source: Anne O. Krueger et al., Trade and Employment in Developing Countries. Vol. 1: Individual Studies. University of Chicago Press for NBER, Chicago, 1981.

Indicates data not available.

Labor coefficients based on estimates in country studies in Krueger et al., Volume 1.

Data for total man-days.

Figures show total man-hours per million FCFA (Francs de la Communauté Financière Africaine).

Source: Anne O. Krueger et al., Trade and Employment in Developing Countries. Vol. 1: Individual Studies. University of Chicago Press for NBER, Chicago, 1981.

Indicates data not available.

Labor coefficients based on estimates in country studies in Krueger et al., Volume 1.

Data for total man-days.

Figures show total man-hours per million FCFA (Francs de la Communauté Financière Africaine).

For Indonesia and Hong Kong, the factor proportions of exports destined for developed countries and those for other developing countries were not significantly different. However for other cases, the Ivory Coast and Pakistan, production competing with imports from other developing countries was far more labor intensive than production competing with imports from developed countries. Similarly, Chile’s exports to developed countries had a labor coefficient of 61, contrasted with 29 for exports to other developing countries (almost entirely in the Latin American Free Trade Association (LAFTA)). If Chile reduced export production for developed countries by one unit of domestic value added and replaced it with one unit of domestic value added of import-competing production, it would entail a new “loss” of 18 jobs, or a 28 percent reduction in employment. By contrast, contraction of export production for other developing countries by one unit of domestic value added and replacement with a unit of domestic value added of domestic import-competing production would result in a change from 29 jobs to 43 jobs, an increase of almost 50 percent. This implies that a strategy to promote exports in the LAFTA market would probably result in a shift toward less labor-using industries, while promoting exports to developed countries would do the opposite.

Conclusions

Perhaps the most interesting findings emerging from the project were, first, that despite factor market imperfections, developing countries’ and territories’ manufactured exports tended to exhibit the factor intensity consistent with their endowments. Moreover, these exports tended to be intensive in unskilled labor—also consistent with their endowments. Second, the scope for further increasing their demand for labor through both trade policies and realignment of domestic factor market incentives was sizable.

The evidence from the studies of individual countries and territories and the project strongly suggested that the extent of the potential gain from shifting to an export-oriented trade strategy very much depended on the degree to which factor markets functioned appropriately. The greatest potential for increased employment was where the shift in policy realigned incentives in both the domestic factor market and the trade regime.

As regards income distribution, the results suggested that import-substituting trade strategies and measures that encouraged capital-intensive techniques probably contributed to a less equal income distribution than might otherwise have been observed. The very fact that export industries consistent with comparative advantage tended to use unskilled labor relatively intensively suggested that additional employment in those industries would increase the rate at which the urban sector could absorb new entrants to the labor force.

Another significant finding was the degree to which manufactured exports, supplied in response to regional arrangements, turned out to be relatively high cost and apparently uneconomic. Potential gains from regional trade in manufactures, if undertaken behind a common wall of protection, were probably relatively small, if not negative. Exports emanating from the ad hoc, specific, and widely varying incentives created under an import-substitution regime do not usually provide the resource allocation gains that can result from uniform, across-the-board incentives. While there is undoubtedly scope for gainful intradeveloping country trade, it seems clear that the type of trade in manufactures that has been encouraged under regional trading arrangements has generally been more the outcome of the import-substitution type of incentives than of the incentives that accompany a genuinely export-oriented trade strategy.

The NBER project could not hope to provide definitive quantitative estimates to answer these questions given the paucity of the data. Nonetheless, some evidence emerged with which to estimate the degree to which wage and capital costs diverged from those under well-functioning factor markets in the countries covered by the project. These estimates had limited comparability across countries, but they indicated that in countries with labor market regulations, import substitution trade regimes, credit rationing, and social insurance tax systems, these all contributed to raising the domestic cost of hiring labor relative to the cost of using capital. There was no single pattern of relative importance among these factors across all the countries, but what did seem clear, even from these impressionistic data, is that the export-promoting territories and countries—Hong Kong, Korea, Brazil, and, to a lesser extent, the Ivory Coast—had relatively lower factor market distortions than the import-substitution countries, with the possible exception of Argentina.

A second conclusion, probably not unrelated to the first, was that currency overvaluation and the favorable treatment of imports of capital goods were also important sources of cheaper capital for firms eligible to import. Credit rationing at subsidized interest rates constituted another major source of underpricing of capital services. Finally, social insurance taxes drove a wedge of about 20–30 percent in the price of labor between firms and sectors subject to the taxes and other activities within the economy.

Each of these sources of pricing disparity between firms and sectors by itself could have significantly affected incentives, but when they were all found together the effects must have been fairly powerful. All of them induced lower capital costs and higher labor costs than would have prevailed if efficient patterns of resource allocation had occurred.

Large potential gains in real income, growth rates, employment, and income distribution are attainable largely through an export-oriented trade strategy that would increase exports of manufactures to developed countries. Since major gains from an export-oriented strategy are likely to result in trade with developed countries because of their very different factor endowments, access to developed country markets becomes crucially important in permitting developing countries to achieve maximum gains from appropriate trade policies. While there would be benefits from changing trade strategies, the magnitude of the potential gains is greater when access to markets is unrestricted and those markets are growing rapidly. Although this is not the place to assess the prospects for protectionist pressures in the developed countries, two points are worth mentioning. First, developing country exporters have been able to increase their shares of world markets despite protectionist pressures. Second, in the past, those who cautioned against an export-oriented trade strategy because of some form of export pessimism have been proved wrong.

Clearly, maintaining access to markets of the developed countries is one of the major contributions developed countries can make to the growth prospects of developing countries, especially middle-income developing countries. While appropriate policies in the developing world with respect to domestic markets, the trade regime, education, and much more are necessary for increasing prospective rates of growth of real income and living standards, those same policies clearly have a larger potential payoff in a liberal international economy.

Other Resources Citing This Publication