Wal-Mart is not an ordinary retail store—communities are impacted in significant ways by its entrance. Using various data sources and propensity-weighted multilevel modelling, this paper explores the ‘Wal-Mart effect’ on crime. Concentrating on the 1990s, results reveal that Wal-Mart is located in United States counties with higher crime rates, net of robust macro-level correlates of crime. Wal-Mart selected into counties primed for the 1990s crime decline, but, after accounting for endogeneity, growth of the company stunted crime declines when compared to matched counties. A Wal-Mart–crime relationship exists. If Wal-Mart did not build in a county, property crime rates fell by an additional 18 units per capita from the 1990s to the 2000s. A marginally statistically significant, yet stable, effect for violent crime was also observed, falling by two units per capita. These findings provide important theoretical implications regarding the influence of specific economic forces on aggregate crime trends and offer important implications for local governments faced with the prospect of Wal-Mart entering their communities.