The Relationship Between Population Density and Crime: Evidence from the Government Shutdown of 2013

Abstract: The literature that investigates the determinants of crime has often used natural experiments that often exogenously change the supply of crime such as election cycles, police strikes and law changes to learn how policy may impact crime from a supply side perspective. In this paper, we are aiming to contribute to this literature using the government shutdown that occurred in the US in October of 2013. During this period, a large share of government workers in Washington DC was furloughed and exempt from having to show at work. Using crime data sets provided by the Washington DC and Baltimore police departments, we investigate whether this temporary shift inward of demand decreased crime in DC relative to its crime rates in 2012 and the city of Baltimore (located 40 miles away). The results in this research are not only informative of the shape of the supply curve of crime (elastic versus inelastic), but also detail an intended consequence of the government shutdown. Preliminary results show that there is not a statistically significant decrease in crime when our observational unit bundles all crimes together regardless of the type of crime, city area and time of the event. Once we break the data by district, type of crime and time of the event we show that crime decreased in ways that one would expect. Our findings show that crime went down in the evening (but not at night and mornings) and also for theft (but not homicide, arson or burglary). Our results are informative that crime offenders react to demand conditions through decreasing crime offenses when the number of potential victims is low and also substitution into other areas, time of the day and type of crime. These results are telling that the war on crime is as effective if fought through the demand side as fought in the supply side. Despite this conclusion, we cannot provide evidence from our natural experiment of whether supply and demand side policies are substitutes or complements.