possession of guns will be affected. The flow of new guns to the illegal sector may be reduced to the extent that legal guns enter the illegal sector through resale or theft from the legal stock in the District. Theory alone cannot determine whether this handgun ban will reduce crime and violence overall. One would expect that the share of crimes in which guns are used should decline over time if the handgun ban is effective.
The empirical evidence as to the success of the Washington, DC, handgun ban is mixed. Loftin et al. (1991) used an interrupted-time-series methodology to analyze homicides and suicides in Washington, DC, and the surrounding areas of Maryland and Virginia before and after the introduction of the ban. They included the suburban areas around Washington, DC, as a control group, since the law does not directly affect these areas. Using a sample window of 1968-1987, they report a 25 percent reduction in gun-related homicides in the District of Columbia after the handgun ban and a 23 percent reduction in gun-related suicides. In contrast, the surrounding areas of Maryland and Virginia show no consistent patterns, suggesting a possible causal link between the handgun ban and the declines in gun-related homicide and suicide. In addition, Loftin et al. (1991) report that nongun-related homicides and suicides declined only slightly after the handgun ban, arguing that this is evidence against substitution away from guns toward other weapons.
Britt et al. (1996), however, demonstrate that the earlier conclusions of Loftin et al. (1991) are sensitive to a number of modeling choices. They demonstrate that the same handgun-related homicide declines observed in Washington, DC, also occurred in Baltimore, even though Baltimore did not experience any change in handgun laws.7 Thus, if Baltimore is used as a control group rather than the suburban areas surrounding DC, the conclusion that the handgun law lowered homicide and suicide rates does not hold. Britt et al. (1996) also found that extending the sample frame an additional two years (1968-1989) eliminated any measured impact of the handgun ban in the District of Columbia. Furthermore, Jones (1981) discusses a number of contemporaneous policy interventions that took place around the time of the Washington, DC, gun ban, which further call into question a causal interpretation of the results.
In summary, the District of Columbia handgun ban yields no conclusive evidence with respect to the impact of such bans on crime and violence. The nature of the intervention—limited to a single city, nonexperimental, and accompanied by other changes that could also affect handgun homicide—make it a weak experimental design. Given the sensitivity of the results to alternative specifications, it is difficult to draw any causal inferences.