Exploring causal relationship between Major League Baseball games and crime: a synthetic control analysis

Using the Washington Nationals case, which moved from Montreal, Canada, to Washington, DC in 2005, as a natural experiment, I examine the impact of MLB games on crime in a host city. To address endogeneity concerns, this paper applies a synthetic control method with using 21 large cities which host an MLB team as a “donor pool” and employs a triple difference-in-difference approach to estimate the change in crime before and after the Nationals coming, between MLB season and off-season, and Washington, DC and the synthetic Washington. With using monthly crime data from the Uniform Crime Report, only assaults increased by 7–7.5% annually after the Nationals moved to DC; other crimes were unchanged. This result is supported by statistical significance and in-space placebo tests, and several alternative specifications in robustness check. These increases in assaults could be associated with additional costs, annually from $20 to $35 million. Little to no evidence of a causal relationship between MLB games and other types of crime.