Description:
Theory suggests that cultural similarity of countries increases migration flows between them. This paper brings best practices from the trade gravity literature to migration and tests this prediction. In my preferred specification, I use lags of time-varying similarity variables in a panel of international and domestic migration flows (>200 countries, 1990-2019, 5-year in-tervals) and estimate a theory-consistent structural gravity model with origin-year, destination-year, and corridor fixed effects. The results do not show the hypothesized positive effect of cultural similarity on migration. Instead, religious similarity has a significant negative effect on migration, while WVS-based attitudinal similarities regarding individualism, indulgence, and trust are insignificant. Additional results suggest that cultural selection and sorting can explain these findings, where migrants are attracted by destinations that are culturally similar to their personal cultural beliefs rather than the average cultural beliefs of their home country. Results of a two-stage fixed effects (TSFE) procedure and a gravity-specific matching estimator, which both allow the estimation of time-invariant similarity variables, confirm that the relationship between cultural similarity and migration is more nuanced than previously thought.