Gun Buy-Backs

Gun buy-back programs involve a government or private group paying

individuals to turn in guns they possess. The programs do not require the

participants to identify themselves, in order to encourage participation by

offenders or those with weapons used in crimes. The guns are then destroyed.

The theoretical premise for gun buy-back programs is that the

program will lead to fewer guns on the streets because fewer guns are

available for either theft or trade, and that consequently violence will decline.

It is the committee’s view that the theory underlying gun buy-back

programs is badly flawed and the empirical evidence demonstrates the

ineffectiveness of these programs.

The theory on which gun buy-back programs is based is flawed in three

respects. First, the guns that are typically surrendered in gun buy-backs are

those that are least likely to be used in criminal activities. Typically, the

guns turned in tend to be of two types: (1) old, malfunctioning guns whose

resale value is less than the reward offered in buy-back programs or (2)

guns owned by individuals who derive little value from the possession of

the guns (e.g., those who have inherited guns). The Police Executive Research

Forum (1996) found this in their analysis of the differences between

weapons handed in and those used in crimes. In contrast, those who are

either using guns to carry out crimes or as protection in the course of

engaging in other illegal activities, such as drug selling, have actively acquired

their guns and are unlikely to want to participate in such programs.

Second, because replacement guns are relatively easily obtained, the

actual decline in the number of guns on the street may be smaller than the

number of guns that are turned in. Third, the likelihood that any particular

gun will be used in a crime in a given year is low. In 1999, approximately

6,500 homicides were committed with handguns. There are approximately

70 million handguns in the United States. Thus, if a different handgun were

used in each homicide, the likelihood that a particular handgun would be

used to kill an individual in a particular year is 1 in 10,000. The typical gun

buy-back program yields less than 1,000 guns. Even ignoring the first two

points made above (the guns turned in are unlikely to be used by criminals

and may be replaced by purchases of new guns), one would expect a reduction

of less than one-tenth of one homicide per year in response to such a

gun buy-back program. The program might be cost-effective if those were

the correct parameters, but the small scale makes it highly unlikely that its

effects would be detected.

In light of the weakness in the theory underlying gun buy-backs, it is

not surprising that research evaluations of U.S. efforts have consistently

failed to document any link between such programs and reductions in gun

violence (Callahan et al., 1994; Police Executive Research Forum, 1996;

Rosenfeld, 1996).

Outside the United States there have been a small number of buy-backs

of much larger quantities of weapons, in response to high-profile mass

murders with firearms. Following a killing of 35 persons in Tasmania in

1996 by a lone gunman, the Australian government prohibited certain

categories of long guns and provided funds to buy back all such weapons in

private hands (Reuter and Mouzos, 2003). A total of 640,000 weapons

were handed in to the government (at an average price of approximately

$350), constituting about 20 percent of the estimated stock of weapons.

The weapons subject to the buy-back, however, accounted for a modest

share of all homicides or violent crimes more generally prior to the buyback.

Unsurprisingly, Reuter and Mouzos (2003) were unable to find evidence

of a substantial decline in rates for these crimes. They noted that in

the six years following the buy-back, there were no mass murders with

firearms and fewer mass murders than in the previous period; these are

both weak tests given the small numbers of such incidents annually.

Banning Assault Weapons

In 1994, Congress enacted the Violent Crime Control and Law Enforcement

Act, which banned the importation and manufacture of certain

military-style semiautomatic “assault” weapons and ammunition magazines

capable of holding more than 10 rounds (National Institute of Justice,

1997). Assault weapons and large-capacity magazines manufactured before

the effective date of the ban were grandfathered and thus legal to own and

transfer. These guns are believed to be particularly dangerous because they

facilitate the rapid firing of high numbers of shots. While assault weapons

and large-capacity magazines are used only in a modest fraction of gun

crimes, the premise of the ban was that a decrease in their use may reduce

gunshot victimization, particularly victimizations involving multiple wounds

or multiple victims (Roth and Koper, 1997).

A recent evaluation of the short-term effects of the 1994 federal assault

weapons ban did not reveal any clear impacts on gun violence outcomes

(Koper and Roth, 2001b). Using state-level Uniform Crime Reports data on

gun homicides, the authors of this study suggest that the potential impact of

the law on gun violence was limited by the continuing availability of assault

weapons through the ban’s grandfathering provision and the relative rarity

with which the banned guns were used in crime before the ban. Indeed, as

the authors concede and other critics suggest (e.g., Kleck, 2001), given the

nature of the intervention, the maximum potential effect of the ban on gun

violence outcomes would be very small and, if there were any observable

effects, very difficult to disentangle from chance yearly variation and other

state and local gun violence initiatives that took place simultaneously. In a

subsequent paper on the effects of the assault weapons ban on gun markets,

Koper and Roth (2001a) found that, in the short term, the prices of assault

weapons in both primary and legal secondary markets rose substantially at

the time of the ban, and this may have reduced the availability of the assault

weapons to criminals. However, this increase in price was short-lived as a

surge in assault weapon production in the months prior to the ban and the

availability of legal substitutes caused prices to fall back to nearly preban

levels. The ban is also weakened by the ease with which legally available

guns and magazines can be altered to evade the intent of the ban. The

results of these two studies should be interpreted with caution, since any

trends observed in the relatively short study time period (24-month followup

period) are unlikely to predict long-term trends accurately.

District of Columbia Handgun Ban

Bans on the ownership, possession, or purchase of guns are the most

direct means available to policy makers for reducing the prevalence of guns.

The District of Columbia’s Firearms Control Regulations Act of 1975 is the

most carefully analyzed example of a handgun ban. This law prohibited the

purchase, sale, transfer, and possession of handguns by D.C. residents other

than law enforcement officers or members of the military. Note, however,

that individuals who had previously registered handguns prior to the passage

of this law were allowed to keep them under this law. Long guns were

not covered by the ban.6

One would expect the passage of the District’s handgun ban to have

little impact on the existing stock of legally held handguns but to greatly

reduce the flow of new handguns to law-abiding citizens. Over time, the

number of legally held handguns will decline. It is less clear how the illegal

6For a more detailed discussion of the law and the politics surrounding its passage, see

Jones (1981).

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 gunrelated

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 gunrelated

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.

7Britt et al. (1996) do not report results for suicide.