Does private security affect crime?: a test using state regulations as instruments

Private security is employed to deter criminals from attacking specific targets, presumably not to produce general deterrence. Indeed, private security generates negative spillovers as criminals substitute non-protected targets for protected targets. Specific deterrence efforts may generate positive spillovers too, however, by raising the expected cost of committing crimes, thereby reducing crime in an area. The hypothesis that private security deters crime at the state level is tested. The demand for private security in an area is expected to be simultaneously dependent on the level of crime; so, an instrumental variables approach is employed in a panel-data fixed-effect model using state-level data from 1998 to 2010. Instruments for the amount of private security are state-wide licensing regulations for firms specializing in providing security, since these regulations should influence entry. Some state-level measures of violent and property crime are shown to be negatively and significantly related to increases in private security, suggesting that private security generates a general deterrence effect.