Estimating the Effect of U.S. Concealed Carry Laws on Homicide : A Replication and Sensitivity Analysis

Given the large number of possible specification choices available to researchers when estimating the causal effect of concealed-carry laws on crime, one can generate considerable variation across models in the estimated effects of these laws. However, the growing sophistication in the econometric literature has developed new tools to weed out clearly problematic models and thereby improve the quality of treatment estimates. Researchers who offer empirical estimates on matters of policy importance should follow the latest protocols to be sure that they are not inadvertently misleading policymakers by introducing spurious findings into the literature.In this article, we demonstrate sensible robustness checks that are broadly applicable to researchers in this domain. We illustrate how the recent work by Moody and Lott (2022) fails to adhere to this approach in their article estimating the causal effect of “permitless-carry” laws on homicide. While Moody and Lott claim to have found that shall-issue and permitless-carry laws reduce homicide, our event-study reanalysis demonstrates that their model both violates the critical conditional parallel trends assumption and is highly sensitive to the removal of a small number of observations. Our analysis suggests that the Moody and Lott’s findings are undermined by insufficient in-sample treatment variation in the case of permitless carry laws and model specification problems in both cases. Additionally, we probe Moody and Lott’s hypothesis that early- and late-adopting “shall-issue” states experienced different outcomes, which illustrates that their results cannot withstand basic robustness checks thereby highlighting the importance of conducting principled sensitivity checks before introducing outlier estimates into the empirical literature