Testing for the Effects of Concealed Weapons Laws: Specification Errors and Robustness

Abstract

In 1997, John Lott and David Mustard published an important paper in which they found that right‐to‐carry concealed weapons laws reduce violent crime. Although Lott and Mustard appear to do all possible variations of the analysis, a closer reading reveals that the study might suffer from several possibly important errors. I reestimate the model and check for incorrect functional form, omitted variables, and possible second‐order bias in the t‐ratios. Lott and Mustard's basic conclusions are generally robust with respect to these potential econometric problems. Overall, right‐to‐carry concealed weapons laws tend to reduce violent crime. The effect on property crime is more uncertain. I find evidence that these laws also reduce burglary.