Immigration economics examines how the inflow of foreign workers affects native wages, employment, and the overall economy. The simple competitive model predicts that immigration increases labor supply, depressing wages for native workers who are substitutes for immigrants and raising wages for those who are complements. Empirical evidence — particularly Card's (1990) Mariel boatlift study and subsequent area-based analyses — generally finds small negative effects on native wages, especially for the average native worker, though effects on close substitutes (particularly low-skilled natives and previous immigrants) may be larger. The immigration surplus — the net gain to the receiving economy from employing immigrant labor at below-native wages — is positive but small relative to GDP. The debate centers on distributional effects (who gains and who loses), long-run assimilation, and fiscal impacts.
Immigration is simultaneously one of the most politically charged and empirically contested topics in labor economics. The theoretical predictions are relatively clear — they follow from supply and demand — but the empirical magnitudes have been fiercely debated for decades, with methodological choices often driving conclusions as much as the underlying data.
The basic supply-demand framework is the starting point. An inflow of immigrant workers shifts the labor supply curve outward. If immigrants are perfect substitutes for native workers, this increased supply reduces the equilibrium wage and increases employment (firms hire more at the lower wage). If immigrants are complements to native workers (e.g., immigrant construction laborers complementing native construction supervisors), both groups' wages can rise. The magnitude of the wage effect depends on the elasticity of labor demand and the degree of substitutability — parameters that are empirically estimated, not theoretically determined.
Card's Mariel boatlift study (1990) provided a natural experiment that seemed too good to be true: 125,000 Cuban refugees arrived in Miami over a few months in 1980, increasing the labor force by about 7%. If immigration depresses wages, this sudden, large, exogenous shock should produce detectable effects. Using a difference-in-differences design comparing Miami to control cities, Card found no significant impact on native wages or unemployment. This finding was influential but controversial — Borjas (2017) later reanalyzed the data using a narrower definition of low-skilled workers and found significant wage effects, leading to a methodological debate that remains unresolved.
The broader empirical literature using area-based approaches (comparing immigration-heavy and immigration-light cities) generally finds small wage effects, but these estimates may be biased by native out-migration (if natives leave high-immigration areas, diluting the measured impact) and capital adjustment (if firms invest more in high-immigration areas, absorbing the labor supply shock). National-level analyses by Borjas, which avoid the geographic mobility problem by comparing skill groups (defined by education and experience) over time, tend to find larger negative effects — a 10% increase in labor supply in a skill group reduces wages by 3-4%. The disagreement between area-based and national studies reflects fundamentally different identification strategies and remains a central methodological debate.
The distributional dimension is crucial for policy. Even if the aggregate wage effect is small, specific groups may be significantly affected. Low-skilled native workers and previous immigrants are the closest substitutes for new low-skilled immigrants and bear the largest wage losses. High-skilled natives who are complementary to immigrant labor — or who benefit from immigrant-provided services (affordable childcare, food service, construction) that expand their own labor supply options — may gain. The immigration surplus accrues mainly to employers and complementary workers, while the costs fall on substitutable workers. This distributional pattern explains why the same overall finding ("small average effects") can support very different policy conclusions depending on which distributional consequences are emphasized.
No topics depend on this one yet.