Working Paper No. 10-12
Trade and Wage Distribution Dynamics:
When Does Trade Cause the Selection Effect?
This paper investigates when trade could cause the selection effect. Since the increased average real wage induced by trade triggers the selection effect in Melitz (2003), the main issue is the labor market conditions under which trade raises the average real wage. To identify the labor market conditions for the selection effect, this paper employs worker heterogeneity with respect to abilities in Blanchflower, Oswald, and Sanfey’s (1996) rent-sharing framework. This simple model plays a crucial role in building estimation equations that use the residual wage in order to reflect worker heterogeneity. According to the results of regressions of the average and 10th percentile of residual wages, this paper shows that with high union density, low job destruction, and low job creation, the effect of trade on the average residual wage is likely to be negative because the impact of imports exceeds that of exports. Moreover, the impact of trade on the average wage must work through the residual wage because this study does not find a significant impact of trade on the average predicted wage. As a result, the more rigid the labor market is, the less likely trade is to raise the average industrial wage and the less likely the selection effect in Melitz (2003) is to occur.
JEL classification: F16; J31; C23
Keywords: Trade, Residual wage, Selection effect, U.S.