In this study, we use county data on private security establishments and employment for 1977–92 to test two hypotheses. First, we test whether private security deters crime. Second, we test whether John Lott and David Mustard’s estimates of the impact of shall‐issue laws on crime are biased because of a lack of controls for private security. We find little evidence that private security reduces the crime rates for assault or larceny. Some estimates suggest murder, robbery, and/or auto theft may be deterred by private security, although these results are not robust. Of all the index crime categories, only rape is estimated to have a consistent negative relationship with private security. In addition, we find little evidence that the Lott and Mustard results are biased because of a lack of controls for the private security measures employed in this study.