Summary
Companion to "Adding Immigrants to Microsimulation Models," this paper traces the history of research on immigrant earnings trajectories and synthesizes the empirical regularities that motivate the microsimulation modifications described in the companion piece. The central finding is a strong inverse relationship between immigrant entry earnings and subsequent earnings growth — a pattern that holds over time, across national-origin groups, and in longitudinal Social Security Administration (SSA) administrative data. The paper develops the Immigrant Human Capital Investment (IHCI) model as the theoretical explanation and shows that source-country economic development level is the master predictor of this trajectory.
Key Claims
- Chiswick's (1978) cross-sectional assimilation model assumed constancy of entry earnings across cohorts — an assumption Borjas (1985, 1987) invalidated by tracing cohorts across censuses, revealing substantial decline in entry earnings from 1965–70 to 1985–90 (65% → 50% → 41% of native median).
- Borjas attributed declining entry earnings to negative immigrant selection driven by rising income inequality in source countries (the Roy-model argument). This explanation predicts stable or declining earnings growth rates for recent cohorts.
- The IHCI model (Duleep and Regets 1992–2002) offers an alternative: declining entry earnings reflect declining skill transferability, not declining ability. Low transferability → low opportunity cost of human capital investment → higher earnings growth. This predicts an inverse entry-earnings-growth relationship.
- Evidence strongly supports IHCI: the 1975–80 cohort (entry 50% of native median) achieves nearly the same ten-year relative earnings (83.9%) as the earlier 1965–70 cohort (entry 65%, ten-year 85.4%), because the more recent cohort's lower entry earnings were offset by higher earnings growth (Table 5).
- Confirmed longitudinally with matched SSA/Current Population Survey (CPS) data: post-1969 immigrant men have earnings growth rates consistently exceeding native men; for women the break occurs post-1980 (Table 6). Adjusting for age and education barely affects the relative growth rate differentials.
- Source-country economic development is the master predictor of entry earnings and growth: Western European and Japanese immigrants start at or above native earnings with modest growth; Asian and Latin American immigrants start at 36–57% of native earnings with rapid growth.
- The English proficiency paradox: Filipino and Indian immigrants are highly English-proficient but have low entry earnings; Japanese immigrants are barely English-proficient but have high entry earnings. Consistent with opportunity-selection theory (economic conditions determine who migrates) rather than a pure skill-transfer model.
- Education amplifies the inverse relationship: among developing-country immigrants aged 25–39, the high-education/low-education earnings ratio grows from 1.27 at entry to 1.68 after ten years (Table 7), versus 1.61 for European immigrants — more educated immigrants from low-transferability source countries invest more in U.S. human capital.
Concepts Introduced or Extended
Entities Mentioned
Quotes
"Controlling for demographic and human capital characteristics, immigrants often start their U.S. lives at substantially lower earnings, but experience faster earnings growth than natives with comparable years of schooling and experience."
"A strong inverse relationship emerges between immigrant entry earnings and earnings growth."
My Take
The IHCI model is a more nuanced and ultimately more empirically supported account than Borjas's ability-decline story. The longitudinal SSA evidence in Table 6 is particularly compelling because it sidesteps the cohort-composition problem that plagues cross-sectional and pooled cross-section approaches. The English proficiency paradox (Japanese > Filipino despite lower English) is a clever test that directly challenges the institutional skill-transfer model. One caveat: the paper's policy implications assume the inverse relationship is stable across cohorts, but if source-country economic development converges toward U.S. levels, both the entry-earnings deficit and the earnings-growth premium would shrink simultaneously.