Historical Immigration and Long-Run Growth: Skills, Innovation, and the Geography of Settlement (Arkolakis-Lee-Peters)
A major new NBER working paper adds a long-run growth channel to the immigration evidence base — distinct from, and complementary to, the native-wage substitution literature: Costas Arkolakis, Sun Kyoung Lee, and Michael Peters, "Immigration, Innovation, and the Geography of Growth" (NBER Working Paper 35392, featured in the October 2026 NBER Digest).
What the paper finds
Linking the universe of historical US patents (~434,000) and over 600,000 immigrants to individual census records, 1880–1920, and embedding the evidence in a spatial growth model, the authors estimate that the era's 20+ million immigrants raised US income per capita by 8.2% by 1940 — and, because innovation compounds, income per capita would still have been 9.6% lower in 2000 had the inflow never happened. The digest's one-line summary: "Immigration to the US between 1880 and 1920 raised income per capita by an estimated 8.2 percent by 1940. Skills mattered most but the location choices of immigrants were also important."
Three channels, decomposed:
- Skills dominate: the same immigrants arriving without their skills would have left income per capita 5.9% lower by 1940 — about three-quarters of the total effect. Selection differed sharply by origin: "British and German immigrants were substantially more prolific in patenting than Italian and Irish immigrants, and the researchers show they were positively selected on their pre-migration occupations."
- Geography was a distinct, previously underweighted channel: counties in the most urban quintile had immigrant shares near 40% vs. under 10% in rural counties, and urban areas produced "four to five times as many patents per capita." Had immigrants dispersed like the native-born, income per capita would have been ~2% lower by 1920.
- The quota acts' counterfactual is modest: removing the 1921 Emergency Quota Act and 1924 Johnson-Reed Act would have raised income per capita only ~1.7% by 2000, "because the excluded groups were small relative to the US population and, on average, less skilled than those who continued to arrive."
How this bears on the adjudication
This paper speaks to the long-run, aggregate margin — income per capita and innovation — not to short-run native wage substitution, so it does not contradict the small-to-zero native wage effects in the modern surge literature (see The Post-Pandemic Migration Surge: Native Labor-Market Impacts and the Emerging Replication Literature); it illustrates how effects differ by horizon and outcome. Its quota-acts result also complements the intergenerational-mobility evidence on the 1920s border closure in Historical Immigration Quotas and Intergenerational Mobility: Evidence from the 1920s Border Closure: restrictions' costs show up in different places depending on the lens (mobility, innovation, growth). Methodological caveat for weighing it: the 8.2%/9.6% magnitudes come from a calibrated spatial growth model, not a reduced-form natural experiment — they rest on the model's assumptions, whereas the skill-decomposition and patent-linkage facts are more direct.