К оглавлению
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 
102 103 104 105 106 107 108 109 110 111 112 113 114 115 

Suppose for the sake of discussion that policy interventions can raise

the difficulty faced by some individuals in obtaining some types of guns.

The question of whether such interventions reduce gun use (or crime or

violence) depends on how readily the potential buyers could substitute

alternative weapons or sources for those targeted by policy.

In our framework, the existing studies, summarized in the section on

how offenders obtain firearms, describe the distribution of guns across

acquisition sources for a particular type of buyers (felons or youthful offenders).

These surveys cannot provide an estimate of the total number of

guns held by the population of offenders.

Although the studies are conducted on nonrandom convenience samples

of inmates, they show fairly consistently that many guns are stolen or

borrowed, rather than purchased in the primary market. Many guns are

obtained through informal networks. The fact that criminals acquire guns

from a variety of sources suggests substitution. Indeed, some (see, e.g.,

Kleck, 1999; Wright and Rossi, 1994; Sheley and Wright, 1995) have taken

these studies to suggest that substitution possibilities are so pervasive that

interventions cannot control the amount of gun use or ensuing harm. Some

observers draw similar inferences from the fact that many guns are stolen

from the large stock of guns available to steal.

Our framework, though simple, suggests that there are limits to what

one can infer about substitution from these data. First, the existing studies

combine survey responses of inmates from a variety of cities. The fact that

inmates from various places, taken collectively, get their guns from different

sources does not mean that any particular criminal or criminals in any

particular city have ready access to all these alternatives. Cities may differ

in terms of the sources of guns. Furthermore, even if persons of a particular

type in a locale obtain their guns through different channels, this does not

imply that each person has a variety of channels if deprived of the channel

he or she currently uses.

Another finding in the literature concerns the vintage of guns used in

crime. Vintage enters the framework through type: new and old guns may

be seen as different types with particular policy relevance because there are

different interventions for each type. In spite of the vast numbers of used

guns that could be stolen and then transferred to criminals, the trace data

suggest that a disproportionate fraction of crime guns are quite new, although,

as noted in Chapter 2, it cannot be determined how well the trace

data represent the total population of crime guns. In our framework, we

can interpret this information to mean that criminals favor new guns over

used guns, given current acquisition costs; this reflects in part the fact that

new guns lack a potential liability from use in a previous crime that is

unknown to the current purchaser. Again, it is only information on the

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.

Regulating Gun Dealers

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

of high-risk dealers. In theory, this approach would increase the cost of

guns to criminals by restricting their availability through retail outlets.

However, in order for this approach to be effective in reducing gun violence,

there must be limited substitution from regulated primary markets to

unregulated secondary markets.

In their analysis of trace data contained in BATF’s Firearm Tracing

System, Pierce et al. (1995) found that nearly half of all traces came back to

only 0.4 percent of all licensed dealers. However, the concentration of trace

data may simply reflect the very high concentration of firearms sales among

FFLs. In California, the 13 percent of FFLs with more than 100 sales during

1996-1998 accounted for 88 percent of all sales (Wintemute, 2000). While

handgun trace volume from 1998 was strongly correlated with handgun

sales volume at the level of the individual dealer and highly concentrated

among high-volume dealers, Wintemute (2000) also found that trace volume

varied substantially among dealers with similar sales volumes, suggesting

that guns sold by certain dealers were more at risk for generating crime

guns than others. However, as Braga and his colleagues (2002) point out,

Wintemute did not determine whether this variation was greater than could

be explained by chance alone. It is possible that the variation of traces

among dealers with similar trace volume was not significantly different

from what would be expected from a normal distribution of crime gun

traces among dealers.

The findings are important nonetheless. Even if only some high-volume

dealers are high risk, the fact that most crime weapons come from highvolume

dealers suggests that concentration of regulatory resources on this

relatively small population may lead to more efficient enforcement, unless

there is substitution across dealers by size category.

Due to concern that some FFLs were scofflaws who used their licenses

to supply criminals with guns, the Clinton administration initiated a review

of licensing procedures that led to their tightening (Bureau of Alcohol,

Tobacco, and Firearms, 2000b). In 1993 and 1994, federal law was

amended to provide more restrictive application requirements and a hefty

increase in the licensing fee, from $30 to $200 for three years. After these

provisions were put into place, the number of federal licensees declined

steadily from 284,117 in 1992 to 103,942 in 1999 (Bureau of Alcohol,

Tobacco, and Firearms, 2000b). With the elimination of some 180,000

dealers, BATF regulatory and enforcement resources became less thinly

spread. In 2000, BATF conducted focused compliance inspections on dealers

who had been uncooperative in response to trace requests and on FFLs

who had 10 or more crime guns (regardless of time-to-crime) traced to

them in 1999 (Bureau of Alcohol, Tobacco, and Firearms, 2000a). The

inspections disclosed violations in about 75 percent of the 1,012 dealers

inspected. While the majority of the discrepancies were resolved during the

inspection process, some 13,271 missing guns could not be accounted for

by 202 licenses, and 16 FFLs each had more than 200 missing guns. More

than half of the licensees had record-keeping violations only. The focused

compliance inspections identified sales to more than 400 potential firearms

traffickers and nearly 300 potentially prohibited persons, resulting in 691

referrals sent to BATF agents for further investigation (Bureau of Alcohol,

Tobacco, and Firearms, 2000a). This reinforces the impression that a relatively

small number of dealers systematically violate rules in ways that

allow for leakage of guns to prohibited persons.

In a recent paper, Koper (2002) examined the effects of the nearly 70

percent reduction in FFLs following the 1993 and 1994 federal licensing

reforms on the availability of guns to criminals. Using a data base of all

active gun dealers in summer 1994 and the number of BATF gun traces to

each dealer since 1990, Koper examined whether “dropout” dealers were

more likely to be suppliers of crime guns than were “survivor” dealers. He

concluded that it was not clear whether guns sold by the dropout dealers

had a higher probability of being used in crime or moved into criminal

channels more quickly when compared with active dealers. This study,

however, used national BATF firearms trace data from 1990 through 1995,

before the adoption of comprehensive tracing practices in most major cities

and prior to BATF nationwide efforts to encourage law enforcement agencies

to submit guns for tracing (Cook and Braga, 2001; Bureau of Alcohol,

Tobacco, and Firearms, 2000c). National trace data from this time period

are not representative of guns recovered by law enforcement, so it is difficult

to interpret the findings of Koper’s analysis of the impact of federal

licensing reforms on the availability of guns to criminals.

Some states and localities have imposed additional regulations on gun

dealers. In 1993, North Carolina found that only 23 percent of dealers also

possessed its required state license (Cook et al., 1995). Noncomplying dealers

were required to obtain a state license or forfeit their federal license.

Alabama also identified FFLs who did not possess the required state license;

900 claimed not to know about the state requirements and obtained the

license; another 900 reported that they were not currently engaged in the

business of selling firearms; and 200 more could not be located (Cook et al.,

1995). Alabama officials scheduled the licenses for these 1,100 dealers for

cancellation. The Oakland (CA) Police Department worked with BATF to

enforce a requirement for all licensed dealers to hold a local permit that

required dealers to undergo screening and a criminal background check

(Veen et al., 1997). This effort caused the number of license holders in

Oakland to drop from 57 to 7 in 1997. Officials in New York found that

only 29 of 950 FFLs were operating in compliance with local ordinances. In

cooperation with BATF, all local license applications were forwarded to the

New York Police Department, which assumed responsibility for screening

and inspections. The increased scrutiny reduced the number of license holders

in New York from 950 to 259 (Veen et al., 1997).

These state-level and local initiatives have not been rigorously evaluated

to determine whether they have affected criminal access to guns and

rates of gun misuse.