distribution of types of guns used by criminals. Since criminals use new guns, some observers have taken this information to indicate that interventions targeting new guns can reduce crime. This is a possible but not a necessary consequence. If both new and old guns are available to criminals, and criminals are observed using new guns, we can infer that criminals prefer new guns to old ones, given the respective prices and difficulties of obtaining the two types of guns. But this fact provides no information about whether criminals would substitute old guns for new ones if they faced increased difficulty of getting new ones.
That different criminals get their guns from a variety of sources—and that many guns used in crime are recent guns—provides little evidence about whether interventions would affect the volume of gun use. This is information about the types of guns used, not about the volume of, or harm caused by, guns. By itself, these findings are consistent with any level of substitution. Suppose that, when local rules are loose, some criminals get guns locally while others get them from elsewhere. The locale then adopts tight rules and suppose that all guns seized thereafter turn out to be nonlocal. That is consistent with either of two contradictory stories. In one, the restriction is totally effective and those who were purchasing local firearms can find none. In the other, there is full substitution; all the local buyers are able to find nonlocal sources without much increase in cost. Any inference requires information about the change in the tendency for the targeted type of individual to purchase other guns relative to those targeted with restrictions. In the language of our framework, we need to know the effects of the restriction on costs and of own costs and other prices on the tendencies for each type of person to buy guns.
This section summarizes the existing literature on the effects of different kind of access interventions. We do not include taxes on firearms or ammunition because there are no evaluations of either kind of tax.
As already noted, criminals can acquire guns in the primary market by personally making illegal purchases, arranging straw purchases, and by finding corrupt FFLs willing to ignore transfer laws. The available research evidence reveals that a very small number of FFLs generate a large number of crime gun traces (Pierce et al., 1995; Bureau of Alcohol, Tobacco, and Firearms, 2000b). Assuming it is possible to categorize dealers by risk of diversion per weapon, this concentration of crime gun traces suggests an opportunity to reduce the illegal supply of firearms to criminals by focusing limited regulatory and investigative resources on the relatively small group