Immigrant Earnings Trajectory

Definition

The immigrant earnings trajectory is the time-path of an immigrant cohort's earnings relative to comparable U.S. natives, from arrival (entry or "assimilation" earnings) through subsequent earnings growth and convergence. It is summarized by two parameters — the entry-earnings deficit (how far below natives a cohort starts) and the earnings growth rate (how fast it closes the gap) — and is the empirical object that assimilation theories try to explain. How the trajectory is measured (cross-sectional snapshot vs. longitudinal cohort vs. matched administrative records) materially changes the estimated profile, which is the source of a decades-long methodological debate.

Key Ideas

Cross-sectional assimilation profile (Chiswick 1978): Early estimates compared immigrants and natives in a single census cross-section, finding immigrants start below natives and "catch up" within roughly 10–15 years. This rests on the assumption that entry earnings are constant across arrival cohorts — so older arrivals in the snapshot proxy for the future of recent arrivals.
Borjas's cohort critique (1985, 1987): Tracing the same arrival cohorts across successive censuses, Borjas showed entry earnings fell across cohorts (≈65% → 50% → 41% of the native median from the 1965–70 to 1985–90 arrival cohorts). Because the cross-section conflates this cohort decline with within-cohort growth, the Chiswick profile overstates true assimilation. Borjas attributed declining entry earnings to negative selection — rising source-country income inequality drawing less-skilled migrants (a Roy-model argument) — which predicts stable-to-declining growth rates.
The inverse entry-earnings/growth relationship (Duleep & Regets; Duleep & Dowhan 2008): Contra Borjas's pessimism, lower entry earnings are systematically associated with higher subsequent growth. The 1975–80 cohort (entry ≈50% of native median) reaches nearly the same 10-year relative earnings (83.9%) as the higher-entry 1965–70 cohort (entry ≈65%, 10-year 85.4%) — its lower start is offset by faster growth. This inverse pattern is the empirical signature of the Immigrant Human Capital Investment Model (low skill transferability → low opportunity cost of investing in U.S. human capital → faster growth), not of declining immigrant ability.
Longitudinal confirmation: Matched SSA/CPS administrative panels show post-1969 immigrant men have earnings growth rates consistently exceeding native men (post-1980 for women); adjusting for age and education barely changes the differential. Administrative data sidestep the cross-section's cohort-confounding problem.
Source-country development is the master predictor: Western European and Japanese immigrants enter at or above native earnings with modest growth; Asian and Latin American immigrants enter at ≈36–57% of native earnings with rapid growth — a low-entry/high-growth profile.
English-proficiency paradox: Highly English-proficient groups (Filipino, Indian) can have low entry earnings while barely-proficient Japanese immigrants have high entry earnings — consistent with opportunity selection (who chooses to migrate, given source-country conditions) rather than a pure language/skill-transfer story.

How It Works

The estimated trajectory depends on the measurement design:

Cross-sectional (Chiswick): regress earnings on years-since-migration in one census; valid only if entry earnings are cohort-invariant — an assumption later falsified.
Synthetic-cohort (Borjas): follow a fixed arrival cohort across repeated censuses to separate cohort effects (entry-earnings level) from assimilation (within-cohort growth).
Longitudinal/administrative (Duleep & Dowhan): track the same individuals in linked SSA/CPS records, directly observing within-person growth and avoiding both cohort confounding and emigration-induced selection in repeated cross-sections.

These designs can yield qualitatively different conclusions about the same population, which is why the choice of method — not just the data — drives the assimilation debate.

Why It Matters

Interpreting the post-1965 entry-earnings decline: The trajectory framework distinguishes a quality decline (Borjas: worse immigrants) from a composition shift toward low-transferability source countries whose immigrants start low but grow fast (Duleep–Regets). The two have opposite policy implications for whether recent low-entry immigrants are permanently disadvantaged.
Microsimulation and Social Security: Because immigrants are a growing share of the workforce, long-run models (e.g., Modeling Income in the Near Term, MINT) must project immigrant earnings growth, not just levels. Duleep and Dowhan (2008) use entry earnings relative to natives as the predictor of subsequent growth — operationalizing the inverse relationship — to add immigrants to such models. See U.S. Immigration Demographics.
Assimilation debates: The trajectory is the common evidentiary ground for the Chiswick, Borjas, and Duleep–Regets schools; recognizing the measurement dependence prevents over-reading any single cross-sectional snapshot.

Open Questions

Does the inverse entry-earnings/growth relationship hold for post-2000 cohorts, where admission increasingly selects on high, immediately-transferable skills (H-1B, EB visas)?
As source countries' economic development converges toward U.S. levels, do the entry deficit and the growth premium shrink together?
Can opportunity selection be empirically separated from the institutional skill-transfer mechanism beyond the English-proficiency proxy test?