The structure of labor markets can have a major impact on the transmission of adjustment policies to investment and growth, and the policies themselves also influence labor market developments. This section examines the links between adjustment policies and labor markets by first identifying some salient labor market features and then by examining, to the extent that the limited data permit, the impact of adjustment policies on real wages and employment in the eight countries.87
Structure of Labor Markets
Three labor market features, shared more or less by all eight countries, are especially relevant for the impact of adjustment policies (Appendix V, Table 22 summarizes the labor market characteristics and reforms in each of these countries). First, labor markets are often segmented—between regulated formal markets and unregulated, often much larger, informal markets; between urban and rural sectors; and between public and private sectors.88 Although segmentation may arise from many factors, government regulation and unionization play a critical role. Laws against dismissal of workers or regulations requiring generous severance payments, mandatory wage increases, minimum wages, or other requirements that imply large nonwage labor costs not only create substantial fixed costs to hiring and reduce wage flexibility but also discourage the employment and real-location of labor in the formal sector.89 In contrast, regulations are either nonexistent or poorly enforced in most informal markets. Second, some labor markets are characterized by the extensive role of the public sector, as reflected by the large share of public sector employment in formal markets and/or high public sector wages (India, Bangladesh, Ghana, Morocco, and Senegal). Public sector wage awards that are not linked to productivity gains exert pressure on private sector wages in the formal sector and have an important leverage effect on wage costs. Third, labor markets can be characterized by a high degree of indexation that tends to introduce rigidity into the setting of real wages (for example, Chile before 1982).
One striking observation is that the basic regulatory structure and other institutional features of labor markets have changed little in most of the eight countries during the course of adjustment. Moreover, with the exception of public sector wages and employment, there was typically limited mention of such issues in most IMF-supported programs in these countries. Chile is the only case where major labor market reforms were undertaken. Although public service reform was recognized as important in Bangladesh, Senegal, and Ghana, actions generally fell short of original goals. In Mexico, labor market reforms were limited and lagged behind other structural reforms. In contrast to the other seven countries, Thailand’s labor markets remained relatively free of institutional restrictions from the outset of adjustment.
The Impact of Adjustment Policies on Real Wages and Relative Prices
The response of real wages influences how adjustment policies affect investment, output, and employment. Real wage flexibility (measured in terms of traded goods prices) is needed for a nominal devaluation to result in a real exchange rate depreciation and a transfer of resources to the tradable goods sector. A broad indication of relative real wage developments can be obtained through two measures of changes in “wage-gaps.” The first captures the difference between the growth of actual real wages and that warranted by productivity growth; a positive value thus has negative implications for employment and output. The second is based on the growth of unit labor costs relative to that of trading partners; a positive value indicates a decline in competitiveness (Chart 21).90
Wages, Productivity, and Unit Labor Costs
(In percentage points)
Source: IMF staff estimates.1 Difference between growth of unit labor costs in each country and in trading partner countries (U.S. dollar terms).2 Difference between the rate of growth of actual real wages (deflated by CPI) and warranted real wages (labor productivity multiplied by the ratio of the GDP deflator to CPI—as a proxy for the relative price of output and consumption). See Sachs (1979).3 Total economy.4 Manufacturing.5 Manufacturing and services.Wages, Productivity, and Unit Labor Costs
(In percentage points)
Source: IMF staff estimates.1 Difference between growth of unit labor costs in each country and in trading partner countries (U.S. dollar terms).2 Difference between the rate of growth of actual real wages (deflated by CPI) and warranted real wages (labor productivity multiplied by the ratio of the GDP deflator to CPI—as a proxy for the relative price of output and consumption). See Sachs (1979).3 Total economy.4 Manufacturing.5 Manufacturing and services.Wages, Productivity, and Unit Labor Costs
(In percentage points)
Source: IMF staff estimates.1 Difference between growth of unit labor costs in each country and in trading partner countries (U.S. dollar terms).2 Difference between the rate of growth of actual real wages (deflated by CPI) and warranted real wages (labor productivity multiplied by the ratio of the GDP deflator to CPI—as a proxy for the relative price of output and consumption). See Sachs (1979).3 Total economy.4 Manufacturing.5 Manufacturing and services.The effectiveness of exchange rate policy in bringing about a reduction in real wages and a depreciation of the real exchange rate depends on many factors, particularly the adoption of consistent macroeconomic policies. As noted in Section IV, the indexation of nominal wages can be especially important. For instance, the indexation of wages to past inflation in Chile (1978-82), together with a fixed exchange rate, contributed to a sharp increase in real wages and an unsustainable real appreciation that eventually led to a severe contraction of output and investment in 1982-83 and massive unemployment. However, following the elimination of mandatory backward wage indexation in 1982 and the shift to a flexible exchange rate regime, Chile was able to lower real wages sharply and effect a real effective exchange rate depreciation of some 50 percent by 1985-87. At the same time, it maintained low inflation and generated substantial growth in output and employment. The persistence of considerable unemployment for several years must also have put downward pressure on real wages. In line with these developments, Chart 21 shows that, prior to 1982, real wage growth exceeded that of warranted wages while relative unit labor costs increased substantially; this relationship reversed in the mid-1980s.
Mexico’s experience suggests that forward-looking wage agreements can be useful in breaking inflation inertia, but it is also indicative of the difficulties that may arise to adequately assess a country’s external competitiveness especially during periods of considerable structural transformation; in such circumstances, labor market rigidities are likely to be of particular importance. While real wages grew broadly in line with warranted wages, Mexico’s relative unit labor costs picked up appreciably from 1988 onward (Chart 21). To a considerable extent this reflected the reversal of an earlier substantial depreciation in the real exchange rate.91 The surge in capital inflows during the early 1990s was clearly a major factor underlying this development and, in practice, it would have been difficult for exchange rate or wage policies alone to fully offset the effects of these inflows. Nevertheless, even though productivity growth was strong in the manufacturing sector, labor market reforms, aimed at addressing structural rigidities in the labor market at an early stage, could have enhanced competitiveness.
Of course, the degree of indexation itself is likely to be conditioned by past experience with inflation. A history of relatively low inflation and stable macroeconomic policies, as in Thailand, is typically associated with relatively flexible labor markets so that nominal exchange rate policy can be effective in bringing about a sustained real depreciation. Thus, the nominal exchange rate depreciations of the mid-1980s brought about a substantial improvement in Thailand’s relative unit labor costs—an improvement that was maintained in later years as real wage growth after 1987 was generally at, or below, the warranted rate suggested by productivity improvements.92 In Morocco, which had a history of generally moderate inflation, adjustment measures, including devaluation and fiscal tightening, reduced real wage growth below productivity growth and thereby improved competitiveness especially during 1984-86. However, real wages and relative unit labor costs rose subsequently, particularly after the exchange rate was pegged in 1990.93
The prominence of public sector employment in the formal sector can mitigate the impact of adjustment policies on real wages. In Senegal, real wages in the formal sector remained uncompetitive after adjustment, largely because of high public sector salaries. Similarly, in Ghana, the compensation policies of the Government, which is the largest employer, have tended to limit the responsiveness of formal sector wages to changing market conditions and dampened the output response to adjustment policies (Husain and Faruqee (1994)).
Nevertheless, even in situations where wages in formal markets are rigid, a large informal sector may make it easier for adjustment policies to bring about substantial changes in relative prices, including the real exchange rate. Thus, in India and Bangladesh, where the formal sector is small relative to the total labor market, most of the adjustment to nominal exchange rate changes takes place through a reduction of real informal sector wages. Similarly, in Ghana, real wage flexibility that reflects the relatively large informal sector has facilitated the sectoral reallocations sought by the Economic Recovery Program (Horton and others (1994)). This does, however, raise questions of efficiency and equity since the burden of adjustment placed on informal sector wages is likely to be even greater.94
The Impact of Adjustment Policies on Employment and Unemployment
Even when adjustment policies succeed in lowering real wages and shifting relative prices, the speed and extent of the employment response depends on the elasticities of employment with respect to output and real wages.95 These, in turn, would be influenced by the mobility of labor across sectors and the substitutability of labor for other factors of production. The limited evidence available indicates a fairly strong relationship between employment and output both during the course of adjustment (Chart 22), particularly if the formal sector is relatively large (Chile, Mexico, and Morocco) or if labor market segmentation is limited (Thailand); in these cases, policies that have fostered (or impeded) a strong output response have typically had a similar impact on employment. By contrast, in Bangladesh and, even more notably, in India, formal sector employment has been relatively static, even at times of substantial output fluctuation, mainly because of institutional rigidities. In these latter countries, labor market adjustment appears to have taken place largely through a reduction in real wages in the informal market.
Employment, Unemployment, and Output
(Annual changes; in percent unless otherwise specified)
Source: IMF staff estimates.1 Manufacturing sector employment.2 Formal sector employment excluding central government.3 Formal sector employment including central and state governments.Employment, Unemployment, and Output
(Annual changes; in percent unless otherwise specified)
Source: IMF staff estimates.1 Manufacturing sector employment.2 Formal sector employment excluding central government.3 Formal sector employment including central and state governments.Employment, Unemployment, and Output
(Annual changes; in percent unless otherwise specified)
Source: IMF staff estimates.1 Manufacturing sector employment.2 Formal sector employment excluding central government.3 Formal sector employment including central and state governments.The elasticity of employment with respect to output can be affected by the composition of output as well as relative prices. Although data limitations do not permit a more systematic analysis, the evidence from some of the country case studies is indicative. In Chile, the reduction in real wages eventually (that is, by the mid-1980s) generated an expansion of output in relatively labor-intensive sectors and firms and also induced the use of more labor in the production process. The lag in employment expansion is partly attributable to the fact that a large proportion of exports was based on natural resources and dominated by relatively capital-intensive large firms. Similarly, in Morocco, the large decline in real wages during the early and mid-1980s led to a marked rise in employment after 1985, aided by the expansion of labor-intensive, export-oriented manufacturing following trade liberalization.96 In Senegal, because of the rigidities in the formal sector, the manufacturing supply response shifted to the informal sector in the latter half of the 1980s; consequently, the decline in formal employment appears to have been offset by job creation in the informal sector. In Ghana, agricultural employment increased markedly with the Economic Recovery Program. Finally, the employment response in Bangladesh was dampened by the prominence of the public sector in the production of cotton and jute textiles.97
Unemployment data, and even the notion of unemployment, have to be interpreted with care in many of the eight countries. The absence of formal social safety nets does not permit prolonged periods of open unemployment in informal markets. Under-employment and sporadic employment are much more common and are not accurately captured by the data. Available information suggests that labor market segmentation and the relative size of the formal market influences the increase in open unemployment during adjustment.98 In some situations, as in Mexico (1981-82) and Chile (1982-83), where formal labor markets are relatively large, a sharp increase in unemployment appears to have precipitated the fall in real wages (see Chart 22). By contrast, in Thailand, where labor markets are relatively flexible and segmentation is quite low, and in India and Bangladesh, where informal markets are relatively large, the reduction in real wages was achieved without a sharp increase in unemployment. In Morocco, although employment increased from 1986 to 1993, the entry into the labor force of increasing numbers of young, relatively more educated workers resulted in an increase in open unemployment.
