The Impact of Migration on Earnings Inequality 
Using 1940–2015 data from the US census and American Community Survey, the author first documents national, regional, and state-level trends in earnings inequality and migration. He then looks at the effect of state-level flows of immigration (arriving from another country), in-migration (arriving from another state), and out-migration (departing for another state) on state-level earnings inequality. The paper explores two mechanisms for the impact on earnings inequality of migration flows into or out of states. One channel is through migrants, whose movements alter the population and associated earnings with which inequality is measured; the other channel is through non-migrants, whose earnings may be affected by migration due to, for instance, changes in productivity.
Key Findings
- State-level immigration has a small, significantly positive effect on the earnings share of the top decile in the state, but neither in-migration nor out-migration has a significant effect on such earnings inequality.
- Immigration accounted for 5.8 percent of the observed rise in US earnings inequality from 1950 to 2015, and 94 percent of that contribution (or 5.4 percent of the rise) resulted from changes in the earnings of non-migrants in the labor force.
- Education appears to be a key non-migrant trait affected by immigration, with non-migrants in the labor force more likely to have fewer than four years of high school in the presence of higher immigration rates.
Implications
Immigration as a whole is not the source of most of the growth in earnings inequality over the past several decades. Therefore, discussion designating it as a substantial reason for increases in earnings inequality is misguided. This finding may be of interest to policymakers in states hoping to mitigate inequality through legislative action.
Abstract
This paper examines the impact of migration on earnings inequality using 1940–2015 data from the U.S. census and American Community Survey. Despite measurement challenges, I successfully replicate existing findings regarding national trends in earnings inequality and migration, and subsequently analyze regional and state patterns. Using 1940 birthplace information to instrument for migration, I find that recent immigration mildly increases the top decile earnings share, while recent in-migration and out-migration have no significant effects on such inequality. I estimate that immigration contributed 5.8 percent to the observed rise in U.S. earnings inequality from 1950 to 2015, primarily through a non-migrant channel.