i (4)
Although the Census data do not offer precise measures of the cohort’s effective human capital at the time of entry, we have information on the average educational attainment of the cohort at time t. Table 4 reveals that the coefficient 0* is strongly negative when the regression adds the cohort’s educational attainment, and becomes even more negative if the regression includes country-of-origin fixed effects (which can also be interpreted as determining effective human capital). Holding initial human capital constant, therefore, there is convergence among the various immigrant groups. Moreover, the rate of convergence is economically significant. The regression coefficient of suggests that wage differences among the various immigrant groups (holding initial skills constant) narrow by 32.2 percent within the first decade. If this rate of convergence remained constant over the immigrant’s working life, over two-thirds of the initial wage differential would vanish within 30 years. This finding, of course, mirrors the well-known conditional convergence result in the economic growth literature (Barro, 1997).

The conditional convergence result is also related to the recent work of Duleep and Regets (1997a, 1997b), who use an alternative way of controlling for education in the analysis. Duleep and Regets define the immigrant cohort not only in terms of country-of-origin, year-of-arrival, and age-at-migration (i.e., a cell in i,j, k), but also in terms of educational attainment (s). In particular, let vjjks(t) be the log wage of an immigrant cohort originating in country i, arriving in calendar year j, migrating at age k, and with s years of schooling. Similarly, let A viJks(t, f) be the rate of wage growth experienced by this cohort over the time interval (t, t’). To simplify the exposition, suppose that all immigrant cohorts arrive in the same calendar year j, and the wage growth is observed over the same time interval (t, t’). Consider the regression model:
where (£>iks is an i.i.d. error term. Duleep and Regets (1997b) show that A is strongly negative, and interpret this finding as implying that the decline in skills across successive immigrant cohorts is not as strong as suggested by the trend in entry wages. A negative A, seems to suggest that more recent cohorts will experience faster wage growth in the future, and the present value of the age-eamings profile might not differ much across cohorts.