Moody and Marvell’s recent article in this journal examines a regressionbased calculation in Ayres and Donohue (2003a) that indicated, based on statespecific estimates that were generated using county data from 1977-1997, that right-to-carry concealed handguns (RTC) laws were associated with higher overall crime costs. Moody and Marvell criticize that calculation for not extrapolating our trend estimates further, but since we had limited post-passage data for roughly half of our states, we were uncomfortable projecting crime trends far beyond our data. Our caution has now been validated, as the sharp crime declines of the 1990s disappeared after 2000.
Moody and Marvell now repeat our state-specific regression-based calculation using county data through the year 2000 (but with a slightly altered specification), which finds that RTC laws have increased crime costs by $3 billion (in total) for 23 of the 24 jurisdictions they evaluated (Florida is the exception). Nonetheless, they conclude that RTC laws are “generally beneficial” because they claim that the Florida RTC laws (inexplicably) reduced crime costs by $31 billion. But the one paper to focus on the impact of Florida’s RTC law – of which Marvell was a coauthor! – found that the law had no impact on crime. If Marvell’s Florida paper is correct, then the Moody and Marvell findings are reconciled with Ayres and Donohue’s Table 14 showing RTC laws increase crime costs (Ayres and Donohue 2003a).
We also show that estimating aggregate (rather than state-specific) effects of RTC laws using the same data and same specification of Moody and Marvell provides statistically significant evidence of increases in aggravated assault, and no evidence of crime decreases. Similarly, Ayres and Donohue (2003b) showed that, after we corrected some coding errors in John Lott’s data set used in Plassmann and Whitley (2003), their aggregated analysis on 1977-2000 county data also showed statistically significant evidence only of crime increases from RTC laws.