CEPII Meta-Analysis (2025): The Comprehensive Wage Effects of Immigration

CEPII Meta-Analysis (2025): The Comprehensive Wage Effects of Immigration

A comprehensive meta-analysis by Amandine Aubry, Jérôme Héricourt, Léa Marchal, and Clément Nedoncelle (published in the peer-reviewed journal Labour Economics, Volume 98, Article 102815, 2025/2026; originally CEPII Working Paper 2025-07) provides a major empirical synthesis of how immigration affects native wages. Synthesizing evidence from 88 studies published between 1985 and 2023, the authors offer a systematic and rigorous assessment of reduced-form wage estimates.

The study documents substantial heterogeneity across empirical estimates and demonstrates that contextual and methodological choices play a crucial role in shaping reported wage impacts.

Methodological Choices and Systematic Bias

The meta-analysis highlights that empirical designs systematically shape reported effects. A key finding is that the choice of econometric specification and instrumental variables (IV) significantly alters the direction and magnitude of the estimated wage impact.

According to Aubry et al. (2025):

"We document substantial heterogeneity across estimates and show that contexts and empirical designs systematically shape reported effects. In particular, shift-share instrumental-variable strategies correct the upward bias seen in OLS estimates.¹"

Key Findings and Heterogeneity

1. OLS vs. IV Estimators: Ordinary Least Squares (OLS) estimates tend to be upward-biased due to "demand-pull" endogeneity, where immigrants selectively migrate to areas experiencing positive labor demand and wage growth. Shift-share instrumental-variable (IV) strategies (such as the Card-style geographic instrument) successfully correct this upward bias, leading to more accurate, smaller, and scrupulously neutral estimates of immigration's true impact on native wages.
2. Contextual Heterogeneity: The wage impacts of immigration are highly sensitive to the skill level of the incoming immigrants relative to native-born workers, the flexibility of local labor market institutions (such as minimum wages and collective bargaining), and the time horizon analyzed (short-run vs. long-run adjustments).
3. Transparency and Replicability: The authors emphasize that the substantial variation in reported results across the literature is largely driven by idiosyncratic methodological decisions rather than fundamental economic differences. They call for enhanced transparency and standardized reporting in labor economics:

"Our findings emphasise the necessity for replication and enhanced transparency in methodological reporting."

This meta-analysis serves as a crucial reference point for adjudicating the polarized debates in immigration economics, showing that the "average" effect of immigration on native wages is highly localized and method-dependent, but generally very small when appropriate causal identification strategies (like shift-share IV) are applied.