Labour Market Flooding? Migrant Destination and Wage Change during America's Age of Mass Migration
Published on: Mar 12, 2007
Figure 3: Immigrant Magnet Counties, 1900-19107

Figure 3 identifies what we call “Immigrant Magnet Counties.”  These are the 318 counties that experienced a net increase of 1,000 or more immigrants over the decade.  Bubble size is proportional to the numerical increase.  It is clear from the figure that immigrants disproportionately settled in the industrial counties of the Northeast and Midwest and, to a lesser extent, along the Pacific coast.

Of the 318 “Immigrant Magnet Counties” depicted in Figure 3, 72 percent experienced a positive inflow of both foreign-born immigrants and native in-migrants.  Figure 4 presents a log-log scatter diagram that plots each of the magnet counties in a manner designed to illustrate the relationship between the total net in-migration (native and foreign born) and the net increase in the foreign-born population.  The 45-degree line represents the locus of points where the increase in the foreign-born and the total in-migration are the same.  Counties plotted above the 45-degree line experienced a net in-migration of both the native- and foreign-born.  Counties below the 45-degree line experienced a net in-migration of foreign-born and an offsetting net out-migration of native born.

Figure 4: Total Net In-Migration and the Increase in the Foreign-Born Population, 1900-19108

We suggest that the pattern displayed in Figure 4 is not consistent with an interpretation that the counties below the line exhibited crowding out.  The sizes of the bubbles in the diagram indicate the magnitude of each county’s immigration impact index. Where the impact of immigration was the greatest (indicated by large bubbles), we find that both natives and immigrants had poured into the county.  In all of the counties below the line, the immigration impact index is quite low (the bubble sizes are small).  Only four counties below the line had an impact index greater than the average index of those above the line.  These patterns suggest that the reason for the native outflow from the counties below the line was unlikely to have been due to a proportionately heavy foreign-born inflow.  We conclude that the crowding-out mechanisms did not operate during the Age of Mass Migration in any general way.  Instead, natives and immigrants were both moving to the same high-wage regions.  These findings suggest that the immigrant arrivals were not reducing the wages of resident workers in this period.

An Alternative Model of Immigration’s Impact

As we noted, switching the focus to a period a century ago prompts us to adopt a dynamic perspective that we believe may resolve the contradiction between the elementary models of immigration which predict a decline in wages and the empirical studies that find little or no evidence for such an effect.  The simple models assume that the only change in a labour market experiencing immigration is the exogenous outward shift of the labour supply curve.  This assumption is systematically violated in the real world.   It is scarcely an exaggeration to say that it is never the case.  Immigrant entry not only responds to market conditions; it also stimulates market change.

We also suggest that it is impossible to imagine a plausible sequence of events in which immigrant entry occurs in isolation from other related changes.  To take only the most obvious consequence, the arrival of newcomers will increase the demand for final products.  Immigrants have to eat and find shelter and they can be counted on to purchase many other goods and services ranging from necessities to indulgences.  But, when immigrants purchase these items, output will respond and thus the demand for labour will shift out.  Because the shift in the labour supply curve inevitably is accompanied by a shift in the labour demand curve, it is extremely difficult to measure a “pure” wage impact of immigration on resident wages holding “other things equal.”  The inherent technical and conceptual difficulties of this problem, heightened by the political implications of the results, are what make the issue of immigrant impacts so contested.  We suggest that immigration is interesting and important precisely because of those features that make measurement of its wage impact problematic.

In addition to the labour demand shifts that occur pari passu with the arrival of immigrants, history suggests that immigrants are generally drawn to localities, occupations, and industries experiencing innovation, growth, and evolutionary change.  In other words, they do not locate in stagnant textbook labour markets.  Our first suggestion then is that immigrants differentially select destinations with high and/or growing wages.  The reasons why this might be true are easy to understand.  Immigrants have already made the decision to leave home; they select their final destinations using economic criteria.  Indeed, immigrants’ selection of high wage cities was Goldin’s explanation for the positive correlation between the city wage and the foreign born share that she found about 1900 (Goldin 1994: 247).