Economía 5.2 (2005) 197-242
[Access article in PDF]
What Happens to Wages after Displacement?
David S. Kaplan
Gabriel Martínez González
Economic shocks and policy reforms can induce large changes in establishment-level employment. Since wage losses from displacement can be large and long-lasting, policymakers often express a desire to support displaced workers. When resources are limited, policymakers need to target support to the workers who need it most.
But the academic literature offers little guidance on how to do this. There is little agreement on how wages change after displacement. The influential works of Jacobson, LaLonde, and Sullivan document large adverse effects of displacement on workers in the United States.1 Subsequent studies also find that displacement has significant long-term adverse effects.2 More recent international comparisons, however, find zero or positive wage changes following displacement. Abbring and others find no change in wages in the United States, and Bender and others find positive wage changes following displacement in France and Germany.3 [End Page 197]
The literature also offers conflicting explanations of why these estimates vary. Kuhn suggests that differences in inequality and institutions in France, Germany, and the United States can explain the different estimates for these countries.4 Alternatively, Howland and Peterson, Carrington, Jacobson, LaLonde, and Sullivan, and Farber suggest that labor market conditions can affect postdisplacement wages.5 Since a wide variation in displacement costs makes targeting aid difficult, the efficiency gains from identifying determinants of postdisplacement wages are potentially significant.
This paper studies the Mexican labor market to contrast various explanations for differences in postdisplacement wage changes. We hope to identify patterns that may help policymakers target aid to displaced workers. An environment with varying temporal and regional economic conditions and with economic conditions and institutions substantially different from those in the countries previously studied is ideally suited to identify such patterns. If institutions vary little across regions, then the institutional hypothesis would be an unlikely explanation of differences across regions in postdisplacement wages.
Mexico meets these conditions. Differences between Mexico and other countries, as well as differences within Mexico over time and space, can help us identify these patterns in postdisplacement wage changes. First, wage dispersion is higher in Mexico than in France, Germany, or the United States.6 If inequality drives differences in postdisplacement wages, then Mexican workers should have much more negative postdisplacement experiences than observed in these countries.
Second, institutions such as workers' separation costs, the legislated costs of displacement (to the firm), and unions are very different in Mexico than in other countries. Mexican workers are much less likely to leave firms voluntarily than workers in other countries, which suggests that they have [End Page 198] higher separation costs. Mexico's mandatory severance pay is higher than what is legislated in France, Germany, and the United States (the latter two have no legislated severance pay). Unions also have much less wage-setting power in Mexico than in the comparison countries, which can lead to negative union-wage differentials.7 If unions explain the difference in results across countries, then Mexican workers should have much more negative postdisplacement experiences than is the case elsewhere.
Third, Mexico's geographic regions exhibit little variation in unionization and inequality, but large differences in economic conditions.8 Mexican labor laws regarding severance payments, strikes, prohibitions against nominal wage reductions, the legal recourses of workers in case of unfair treatment, and guarantees of profit sharing are determined at the federal level and therefore do not vary across regions. Moreover, no important labor market reforms occurred in the period we study.
In short, we use this variation in economic conditions (but not institutions) to compare local labor market conditions and postdisplacement wages over time and space. Since inequality and institutions vary less within Mexico than across the countries previously studied in the literature, heterogeneity in postdisplacement wages within Mexico probably cannot be explained by inequality and institutions. This points instead toward an important role for labor market conditions. At the same time, our results help explain the variation...