Understanding the Age of Mass Migration

Fascinating article by Abramitzky, Boustan, and Eriksson in the new number of the AER. They conclude:

We know surprisingly little about how migrants during the Age of Mass Migration were selected from the European population and about the economic return from their journey. In this paper, we construct a new dataset of Norwegian-to-United States migration to estimate the return to migration in the presence of selection into migration. We compare the occupation-based earnings of men who moved to the United States and their brothers who stayed behind in Norway. This approach eliminates the component of migrant selection that takes place across households. We gather further evidence about the nature of migrant selection by comparing the economic outcomes of fathers of migrants and nonmigrants. 

We estimate a return to migration from Norway to the United States of around 70 percent, which is substantially smaller than the 200–400 percent return for migra- tion from Mexico to the United States today (Hanson 2006). The contemporary return to migration may be higher than in the past because of the sizeable cost of migration—both the bureaucratic costs of legal immigration and the cost of evading detection for the undocumented—which together reduce the supply of immigrants to the country. In the late nineteenth century, the border was open to almost all prospective migrants and, therefore, the return to migration was relatively low. We note, however, that the decision to migrate in the nineteenth century (as today) may have entailed other nonpecuniary considerations that would have increased the total return to migration (Bertocchi and Strozzi 2008). For example, at the time, Norway was under Swedish control and limited its franchise to men with wealth, while the United States offered the opportunity (for white men) to participate in the demo- cratic system even to new migrants.

The authors add in the conclusion:

We find mixed evidence on selection for rural migrants, with some methods sug- gesting positive selection and others suggesting negative selection. In contrast, we consistently find that men from urban areas who faced poor economic prospects in Norway, as measured by occupation, were more likely to migrate to the United States. One result of this negative occupational selection is that the naïve return to migration underestimates the true return by 20 to 30 percent for the urban sam- ple. The fact that migrants to the United States appear to have been drawn from the lower end of the occupational distribution is consistent with a standard Roy model of migration, as in Borjas (1987), which predicts that men at the lower end of the occupational distribution would have more to gain by moving from relatively unequal European countries to the New World. Furthermore, the fact that European migrants from urban areas, when unhindered by entry restrictions, were negatively selected from the sending population may explain why the United States eventually closed its border to the free flow of labor in 1924 (Hatton and Williamson 2006) and why some countries explicitly select for more skilled applicants in their immigration policies today. 

See the complete article here.