Assessing the Pollution Haven Hypothesis in an Interdependent World

After proposing a simple theoretical framework to illustrate the importance of third-country e ects in empirical studies of the Pollution Haven Hypothesis, we test the model using statelevel panel data on inbound US FDI and relative abatement costs. Our analysis reveals that while own state attributes rarely have statistically signi cant e ects on own inbound FDI when aggregated over all manufacturing sectors, many neighboring state attributes do matter. Moreover, the theoretical model does well in explaining FDI in the chemical sector; we tend to nd signi cant e ects in the correct direction of variables designed to re ect market demand and production costs. Finally, we consistently nd a negative impact of own environmental stringency on inbound FDI in the chemical sector; the impact of neighboring environmental stringency is also statistically signi cant, but the impact is negative on average, contrary to our initial expectations. Nonetheless, the fact that the impact of more stringent environmental regulations spillover across states indicates that future research into the validity of the PHH must account for spatial spillovers. JEL: C31, F21, Q52