Various mechanisms could give rise to the correlations between income, race, and pollution documented by the environmental justice literature. Using a detailed county-to-county migration dataset and pollution data from the Toxic Release Inventory, we propose an approach to identify residential sorting by income as a possible source of these correlations. We find that differences in environmental quality between home and destination counties matter for households' migration decisions. We also show that households moving to "cleaner" counties are "richer" than households staying back. We interpret those results as evidence of residential sorting in the spirit of Tiebout (1956).