A new NBER working paper by Randall Akee, Jimmy Chin, and Daniel L. Crown examines trends in the earnings assimilation of immigrant workers in the United States from 1981 to 2021, drawing on administrative longitudinal data that tracks individual earnings from each immigrant's first year of residence onward. The study contributes to a long-running debate in labor economics over how quickly immigrants close the wage gap with native-born workers—a question with direct implications for assessing the fiscal and economic returns of U.S. immigration policy.

The dataset allows the researchers to distinguish between immigrants who remain in the United States permanently and those who eventually leave—a group they refer to as return migrants. This distinction has been difficult to capture in prior studies that relied on repeated cross-sectional survey data, which cannot follow the same individuals over time and may conflate the experiences of those who stay with those who depart. The paper represents the first effort to examine earnings assimilation trends over a 41-year period while estimating assimilation separately for these two groups.

Several key findings emerge for immigrants who arrived between 1981 and 2010:

Return migration is substantial. Roughly one-fifth to one-third of immigrants leave the United States within 10 years of arrival.

Return migrants start at similar wages but grow more slowly. While return migrants have entry earnings comparable to those of permanent migrants, they experience slower rates of earnings growth over time.

Recent cohorts converge faster. For arrival cohorts since the mid-1990s, permanent immigrants see their earnings converge—or come close to converging—with native-born workers within 10 years. Earlier cohorts experienced significant earnings growth but generally did not reach parity with the native-born population.

The authors also discuss how two factors—the labor market quality of incoming immigrant cohorts (measured by relative earnings upon entry) and selective return migration—play an important role in determining whether repeated cross-sectional data overestimate or underestimate earnings assimilation. This methodological insight matters because much of the prior literature on immigrant assimilation has relied precisely on such cross-sectional comparisons, meaning earlier estimates of how quickly immigrants catch up economically may have been shaped by compositional changes in who remains in the country.

The paper is available via the National Bureau of Economic Research. Author profiles: Randall Akee, Jimmy Chin, and Daniel L. Crown.

The blog Marginal Revolution, which surfaced the paper, noted that while the common distinction between high-skilled and low-skilled, or "desirable" versus "undesirable" immigrants, remains a fixture of policy debate, the current administration's experience demonstrates the difficulty of selectively restricting immigration flows. The post argued that while open borders remain untenable, a broadly liberal immigration policy serves the national interest.