The top panel of Table 4 reports the relevant coefficients from the convergence regressions. The simplest specification (reported in the first column) reveals a positive, though insignificant, correlation between the unadjusted rate of wage growth and the log entry wage of immigrant cohorts. This weak correlation is consistent with the raw data summarized in Table 1 more recent cohorts, who have much lower entry wages, experience roughly the same rate of wage growth as earlier cohorts. Therefore, there is little reason to expect that the earnings of immigrants who belong to different national origin groups and arrive at different times will converge as they assimilate in the United States. If we take the positive point estimate of 0 at face value, the data, in fact, suggest that there might be some divergence over time: the immigrants with the highest entry wages are also the ones who experience the most rapid wage growth. In the context of the model, there seems to be some weak relative complementarity between the skills that immigrants bring into the United States and the skills that they acquire in the post-migration period. This result resembles Mincer’s (1974) finding of complementarity between investments in school and investments in on-the-job training.

As the remaining coefficients reported in the top panel of Table 4 show, however, a simple change in the specification of the regression turns the weak positive coefficient into a significant negative one. Consider the regression model that estimates the rate of conditional convergence:
where Hijk(t) gives the effective human capital of cohort (i,j, k) at time t. The parameter 0* estimates the rate of conditional convergence, the rate at which the earnings of different immigrant cohorts converge if we hold the initial human capital endowment of the cohorts constant.