Kalist & Lee include an impressive survey of the literature in their paper. The survey highlights a range of reasons why crime might change on game day. It includes the incapacitation effect where those who might otherwise be engaged in crime are incapacitated by watching a game (also examined in a recent post on this blog - here). Their survey of the literature also includes a suggestion that security for games might divert police resources, thereby, making other targets more vulnerable.
Kalist & Lee group crimes into ones that are economically motivated and noneconomic crimes. Economic crimes are burglary, larceny-theft, motor vehicle theft, and robbery. The authors find that economic crime is 2.5% higher on NFL game days for host cities. They find no statistical significant difference for non-economically motivated crime. Interestingly, economically motivated crime is actually LOWER for play-off games. Unfortunately, the authors did not offer much in the way of possible explanation for this reversal.
The authors also estimate the implications of game time, upset wins and losses, and day-after effects. Early home games are associated with an increase in crime. Their findings on "upsets" support the earlier cited work of Card and Dahl (2011).
The authors argues that their study adds to the literature on the economic costs and benefits of sporting events. Their rough estimates of the cost of crime associated with an NFL home game is €86,000.