The social costs of gun ownership: a reply to Hayo, Neumeier, and Westphal

We respond to the new article by Hayo, Neumeier, and Westphal (HNW), which is a critique of our 2006 article. The principal contribution of that article was to use a greatly improved proxy for gun prevalence to estimate the effect of gun prevalence on homicide rates. While the best available, our proxy, the ratio of firearms suicides to total suicides in a jurisdiction (FSS), is subject to measurement error which limits its use to larger jurisdictions that have enough suicides to stabilize the ratio. In this response, we report estimates for four different specifications and two data sets, the 200-county data and the data for the 50 states. We develop the claim that measurement error in FSS helps explain the observed pattern of results. Adopting the assumption that FSS follows a binomial process with a number of trials equal to the number of suicides, we characterize the relationship between measurement error and size of the jurisdiction, and thereby justify our conclusion that restricting the estimation to large jurisdictions reduces measurement error in FSS and hence the attenuation bias in the key coefficient estimate. We conclude that for the county-level data, the measurement error in FSS is of greater concern than using a specification that is flexible with respect to population. HNW focus on the latter but at the cost of increasing the effects of the former. We then demonstrate that the state-level data provide a robust case that more guns lead to more homicides.