The Uses And Limitations Of BATF Tracing Data For Law
Enforcement, Policymaking, And Criminological Research

by:
Paul H. Blackman, Ph.D.


BATF firearms tracing data have always provided some benefits as a tool for investigating individual crimes. And there are some clear potential benefits for using the statistics for law enforcement. There are, however, severe limitations on the utility of those data for making public policy aimed at reducing crime. Limitations include: the minimal number of guns BATF attempts to trace or succeeds in tracing, the rules for excluding guns and efforts to trace them, and the limited information on the basis for gun traces. Dr. Paul Blackman is Research Coordinator for the National Rifle Association’s Institute for Legislative Action.The views expressed in this paper do not necessarily represent the views of the National Rifle Association or its Institute for Legislative Action.



INTRODUCTION

Soon after the Gun Control Act of 1968 (GCA’68) improved the paperwork trail for guns, thereby allowing law enforcement, without too much trouble, to trace possible crime guns to their first retail sale, criminologists attempted to study statistical summaries of those traces with a view toward policy recommendations, even though such tracing “was not designed to collect statistics.” (Bea, 1992, p. 65)

To some extent, the paper trail had existed since 1938, when the Federal Firearms Act required inexpensive federal licenses for persons selling guns interstate. The low price encouraged some 100,000 such dealer licenses by the mid-1960s (Zimring, 1975, pp. 140-41), even though one requirement of having a license was to keep records of transfers, and to maintain them for a specified period of time, the length of which varied between 1938 and 1968 (Zimring, 1975, p. 144).

Proportionate to the population, there were more dealers in the late 1930s than there are now, and, while serial numbers and manufacturing information were not then required to be engraved on firearms, they have almost always been placed on quality firearms. But the paper trail was not generally employed. When the Massachusetts Commissioner of Public Safety, for example, testified on the need for additional federal gun control legislation, his assertion that 87% of the state’s crime guns came from elsewhere was based not on tracing them to other states but on failing to find them in Massachusetts’ records. (U.S. Senate, 1965, pp. 345-46) Similarly, the staff report of the National Commission on the Causes and Prevention of Violence relied on permit applications and the like to determine whether the gun originated where it was misused, rather than on tracing data. (Newton and Zimring, 1969, ch. 8)

Some of those early efforts, particularly by Frank Zimring, simultaneously attempted analyses and evaluation while recognizing the limitations of the criminological use of BATF (Bureau of Alcohol, Tobacco & Firearms[1]) tracing data, limitations also emphasized by the Police Foundation in its study, Firearm Abuse. ( Zimring, 1975, p. 183; Brill, 1977).

For example, while using BATF tracing data to support the theory that relatively new guns are disproportionately used in crime, Zimring noted “the possibility that police and federal agency sampling procedures had produced a nonrepresentative sample of guns from New York...the bedeviling problem of sample selection....” (Zimring, 1976, p. 96) He noted that older guns were more difficult to trace; that some data were not crime-specific; that some guns had been seized merely for possessory offenses rather than for use in a crime; that prescreening prevented even tracing attempts for some firearms; and that various other limitations impaired analyses based on tracing data. (Zimring, 1976, pp. 97-98, 101, 104-l06)

For the first two decades of GCA ‘68, as BATF sought to assist law enforcement, it produced annual summaries of tracing efforts, including the number of traces officially attempted and successfully completed, generally accompanied by anecdotal references to a few major crimes solved with tracing and/or the speed of tracing guns involved in prominent shootings. For example, BATF proudly noted that tracing a gun from the scene of a $3.6 million armored-car robbery “led to the arrest and prosecution of the neo-Nazi cult known as The Order,” and that the gun used in the attempted assassination of Ronald Reagan was traced to John Hinckley in 16 minutes.[2] (King, 1988)

More recently, a list of about 100 “significant” trace requests (often involving more than one firearm), of the 79,000 traces overall conducted during Fiscal Year 1994, included tracing the gun used on a Long Island commuter train to kill six persons and wound another 14. BATF noted they “learned the suspect had illegally purchased the weapon in California. The results of the trace culminated with a criminal case being made, charging the suspect with violations of the Federal firearms laws, specifically, 18 U.S.C. § 922(a)(3), 922(a)(6), and 924(c).” (National Tracing Center, 1995, p. 15)

In fact, no federal charge was brought against Colin Ferguson, who was already in custody for the violent crime itself; he was tried and convicted in the New York court system.

Other “significant” traces included a gun used in a suicide, and a gun dropped by a man who wished to avoid being found armed by a magnetometer set up near where the President was to speak. “The U.S. Secret Service does not believe the purchaser intended to harm anyone with the firearm.” (National Tracing Center, 1995, p. 1)

One possible inference is that the other 78,000 traces were less useful to law enforcement. “And BATF would generally acknowledge that its tracing is primarily for the purpose of aiding law enforcement in identifying suspects, establishing whether guns were stolen, and proving ownership” (Pierce, Briggs and Carlson, 1996, p.5), rather than for the systematic analysis of crime guns or other policy- or research-related purposes.[3]

I. LIMITATIONS RELATED TO STATE AND LOCAL LAW ENFORCEMENT

A. Unrepresentative Nature of Guns Selected for Tracing

Most guns involved in violent crimes are not traced, and those which are represent not merely a small but an unrepresentative sample. (Bea, 1992, p. 65) Some of this, of course, is unavoidable in a country with a relatively low clearance rate for violent crimes. Nonetheless, even most guns seized as a result of violent-crime investigations are not traced, and those traced are unrepresentative of firearms used, leading some scholars to suggest that confiscated firearms, while still flawed as a sample, provide a better sample of “crime guns” than traced guns. (Brill, 1977, pp. 26, 42)

As Gregore J. Sambor, then Philadelphia Police Commissioner, noted, “tracing a gun by use of serial number and proofmarks from a manufacturer, through the wholesaler, to the retailer and then the purchaser, and eventually the user, is not always necessary to prove the facts of the case or the elements of the crime....[And] when a local agency has adequate information and their own means available, they can sometimes produce their own results quicker and with less chance of error.”[4] He went on to cite a police killing where the Philadelphia police found it more expeditious to telephone the German manufacturer, and thence the Virginia dealer, leading them to the brother of the person convicted. (Sambor, 1985)

Tracing data may be unrepresentative because of the nature of criminal investigations. Jurisdictions with more thorough recordkeeping than provided by federal law, in particular, have no incentive to trace guns if their own state or local records supply all information needed for law enforcement purposes (prosecution of gun-wielding criminals, identifying stolen goods for determining whether a suspect was a thief, returning stolen property, etc.). Particularly in places with more restrictive gun laws than federal law requires, tracing through BATF could be considered both superfluous and less efficient.

For example, a Justice Department study based on surveys of police departments, reported that some jurisdictions, such as California, began with their own files on guns, and went on to note that such files existed, too, for New York, New Jersey, Iowa, Maryland, in addition to several cities, including Philadelphia and Miami.[5] (Weber-Burdin, Rossi, Wright and Daly, 1981, ch. 4, p. 9) If jurisdictions with more records first use them (Roth and Koper, 1997, p. 83) and only then turn to BATF for firearms not found, while less restrictive jurisdictions start with BATF, one result of BATF tracing would be to exaggerate the out-of-state sources of “crime guns” in restrictive jurisdictions vis-à-vis non-restrictive jurisdictions.

Even without such records, tracing might provide no particular benefit—aside from the sometimes useful one of excluding other suspects or other charges. Tracing is least needed where local resources are sufficient, or the basis for access to the gun irrelevant, as with violent gun-related crimes or burglary investigations.

Tracing should prove most useful—and thus used—where local resources are insufficient and tracing information is likely to be available and useful. Such would be the case with out-of-jurisdiction guns not used in serious felonies, particularly if the trace might suggest the possibility of a less obvious serious crime, such as gun and narcotics offenses. (For example, had tracing provided evidence that John Hinckley had broken the law in his acquisition of handguns, such tracing might have allowed prosecution for a GCA’68 violation, but tracing provided no information necessary for his prosecution for the violent crime of attempted presidential assassination.[6])

B. Law Enforcement Dissatisfaction with BATF Traces

In the 1970s, most law enforcement agencies, according to an NIJ-funded study led by James Wright and Peter Rossi, made little use of BATF and were generally dissatisfied with the results. (Wright and Rossi, 1981, p. 23)

Almost all surveyed law enforcement departments which used the National Crime Information Center (NCIC) used the NCIC for almost all firearms crime. In contrast, little more than a quarter of departments used BATF for most or every firearm either implicated in a crime or found, confiscated, or recovered. (Weber-Burdin, Rossi, Wright and Daly, 1981, ch. 4, p. 13)

And the departments which did use BATF found the Bureau much less useful than the NCIC; over 30% of departments reported the BATF experience was seldom useful or was useless. Thus, almost twice as many departments reported generally finding NCIC useful as similarly found BATF generally useful. (Weber-Burdin, Rossi, Wright and Daly, 1981, ch. 4, p. 16)

C. Crime Classification of Traced Guns

Soon after, BATF began making serious efforts to improve cooperation with local police (Vizzard, 1997, pp. 88-89), and there has clearly been a great change in the willingness of local law enforcement to use BATF’s tracing services. But two facts have remained constant over the decades: There is no standardized procedure for ensuring consistent definitions or terms for identifying the circumstances leading to a trace, and sometimes circumstances are not identified at all.

In addition, categorization might have been done hastily, because the investigation which would explain in full the reason a firearm was obtained by police had not yet been completed. (Bea, 1992, pp. 65, 70-71)

And some dramatic changes in classification figures would suggest a classification change. For example, traced military-style semi-automatics went from being traced generally for “miscellaneous” reasons (39%) in 1986 to just 1% for that reason in 1990, with disproportionate increases both in violent crimes and gun-law violations. (Bea, 1992, p. 72)

1. Most Gun Traces not for Violent Crimes

Most guns seized by police and/or traced by BATF are not involved in violent crimes. Possessory offenses constitute the most common basis for a trace, with violent crimes only a minority of the reasons. Violent crimes explained 15% of traces in 1977, and gun-law violations (federal or state) about 45%, along with 20% unspecified “other” reasons than specific sorts of crimes. (Letter and documents from Paul Mosny, BATF Disclosure Branch, to Bob Dowlut, NRA, July 21, 1980) This despite the fact that, during the 1980s, the crime codes were listed in order of BATF-perceived severity, with only one crime code to be chosen. Nonetheless, property crimes, drug investigations, and gun-law violations predominated. Homicide investigations were the most common violent crime investigation associated with a trace. “Miscellaneous” and “other” explained almost as many traces as other violent crimes. (BATF tape analysis for 1989 supplied to the NRA, Feb. 9, 1990, based on coding tables, effective Oct. 1, 1986) There was not even a specific category for burglary. (Bea, 1992, pp. 70-71)

The 1990s coding of the types of crimes associated with traces is much more extensive, with nearly three dozen categories, compared to ten or twelve in the 1980s. One new category is transportation/possession of untaxed cigarettes. Still, property crimes, gun-law violations, drug offenses, and other unspecified criminal investigations predominate. (Letter from BATF Director John W. Magaw, to Sen. Larry E. Craig, April 1, 1994, regarding “assault weapons.”)

In the largest recent study of BATF traces, roughly five-eighths were for weapons offenses, and just over one-seventh for violent crimes. (Pierce, Briggs and Carlson, 1996, Table 3) A study in Boston, where traces were to be conducted wherever possible on all seized guns, showed that about three-fourths of gun seizures were for possessory rather than substantive crimes; if other reasons for guns to come into police custody (i.e., voluntary surrender) were added, the percentage of seized guns connected to a substantive offense falls to 18%. (Kennedy, Piehl and Braga, 1996a, p. 196)

A Los Angeles area study of traced guns showed two-thirds for possessory offenses and one-sixth for violent crimes. (Wachtel, 1996, p. 12) Roughly the same figures predominate in a recent BATF study of 17 cities, with most traces (exceeding 80% in some cities) involving possessory offenses and only about one-sixth of traces involving guns associated with violent crime investigations. (BATF, 1997)

2. Additional Issues related to Crime Classification

And even traces of guns as a result of a violent-crime investigation do not indicate the role of the firearm in the crime or the investigation, regardless of the reason given for the trace. It might have been possessed by the alleged perpetrator but not involved in the specific crime. It might have been found at the scene of the crime. A firearm from a homicide investigation may have been used in the homicide, found at the scene, recovered from the body of the deceased—not likely to be uncommon where many homicides involve disputes between persons with criminal records (Kates, Schaffer, Lattimer, Murray and Cassem, 1995, pp. 579-84) where the loser is less likely than the winner to remove any of his weapons from the scene of the crime.

The most detailed statistical information from BATF simply indicates the sort of investigation associated with the trace request.[7] A firearm simply found, turned in, or otherwise recovered, might be traced to indicate whether it might have been stolen, potentially making it a property crime investigation, or the official reason given BATF for the trace might be miscellaneous or other. Nothing in the coding, or in any information collected by BATF and thus available to researchers or to others, indicates which of the guns traced were used to commit which crimes.

As BATF has made clear, with regard to the guns it traces, “ATF does not track the incidence of specific use of each one of these firearms in crimes....[T]races requested by police are not always for guns that are used in crimes. Traces are sometimes submitted for firearms recovered by police investigating crimes where the guns were found but were not necessarily used to commit a crime....We do not establish the criteria as to when State or local law enforcement agencies initiate a trace of a firearm.” (Letter from Daniel M. Hartnett [Deputy Director for Enforcement], for the BATF Director, to Rep. Richard T. Schulze, March 31, 1992)

Traces of guns to other states would not necessarily represent gun trafficking patterns as a means of avoiding restrictive gun laws, especially with the average traced gun about five years old (Pierce, Briggs and Carlson, 1996), and untraced guns presumably still older, since age is a key reason for BATF not to attempt a trace. In a mobile nation, where roughly one-fifth of the populace moves each year, guns may be brought from another state simply because persons previously lived in another state. Unsurprisingly, more guns are apt to be bought where paperwork for firearms purchases make the process not only less cumbersome but, in general, less expensive. The large proportion of traces for possessory offenses would support skepticism regarding the amount of trafficking suggested by traces, but, of course, there is insufficient information on the reason for the trace for any such inference to be well founded.

The willingness of youthful offenders to discuss gun dealers but not drug dealers with police (Kennedy, Piehl and Braga, 1996a, p.179) might suggest that gun traffickers are at least perceived by their customers to be less dangerous offenders.

D. Few Firearms are Traced

Even when BATF is encouraging tracing, as with Project Lead in New York City, relatively few firearms are traced. During the first nine months of 1992, for example, of 13,382 firearms recovered by the New York Police Department, only 1,231 (9%) were submitted for tracing, and 824 traced (6%). (Memorandum from Project Lead to Special Agent in Charge, New York Field Division, BATF, October 22, 1992) And there is no basis for believing that even that small percentage is representative. Regarding 1990, when a similar portion of New York City guns were selected for tracing, BATF indicated that “No information is available on why those 1,000 guns were selected out of the 17,000 for tracing.” (Bea, 1992, p. 67)

During the first nine months of 1992, there were about 35,000 gun-related violent crimes reported to the New York City police (Letter from Michael A. Markman, NYPD Office of Management Analysis and Planning, to Mark Overstreet, NRA, January 21, 1993); if violent crime involvement in all the traces were similar to the portion traced to Virginia, and from national BATF tracing data over the years, traces would have been completed on approximately 0.4% of New York City’s gun-related violent crimes.

A BATF study in the Los Angeles area involved only about 5,000 guns traced during an eight-year period during which an estimated 13,000 guns came into police custody annually, with the ensuing study thus based on successful traces of about 2% of the guns which came into police custody. (Wachtel, 1996, pp. 10-12)

E. Over-representation of Homicide

Traces are much more apt to involve weapons violations rather than violent crimes—roughly five-eighths of the traces analyzed by Pierce, Briggs and Carlson (1996, Table 3). Among violent crimes, homicide traces predominate, and they always have. A study based on traces from the mid-1970s found that, among violent crime-related traces, homicide investigations accounted for 45%.

Unfortunately, that study also found a greater disparity between what confiscation and tracing data suggested about the types of guns used in homicides than for other violent crimes. (Brill, 1977, pp. 61-62) A computer analysis provided to the NRA by BATF for 1989 traces suggested about one gun trace for every four gun-related homicides reported to police, compared to one for every 125 gun-related assaults and one for every 250 such robberies. (FBI, 1990) More recently, with more extensive BATF efforts to persuade local authorities to use their National Tracing Center (NTC), the figure has risen to one trace for every: two gun-related homicides, 50 gun-related assaults, and 100 gun-related robberies. (Pierce, Briggs and Carlson, 1996, Table 3; FBI, 1996) The numbers for homicides, at any rate, are clearly reaching impressive size, even though the guns are not necessarily murder weapons.

The expanded tracing efforts, however, because of BATF tracing practices, mean that for homicide there is now a large and unrepresentative sample rather than a small unrepresentative sample.[8] For other violent crimes, the traces remain relatively small and unrepresentative.

II. LIMITATIONS DUE TO BATF TRACING PRACTICES

BATF recognizes the limitations local law enforcement practices place on statistical analyses based on tracing data. The standard “data advisory” BATF’s NTC sends out with data requests reflects that. The NTC notes

· that their data only reflect trends relating to trace-requested guns, not to crime guns overall;

· that trace requests involve “trace requests submitted on firearms used in crimes, recovered from crime scenes, or suspected of being involved in crimes”;

· that BATF relies upon those federal, state, or local authorities submitting a request to ensure that guns are related to crime investigations;

· that not every gun recovered is traced, and

· that BATF does not know the extent to which recovered guns are voluntarily submitted for traces.

BATF’s NTC notes that the accuracy of their reports is dependent upon the accuracy of data submitted. That advisory is well worth respecting.

In addition to local law enforcement limitations on the representativeness of traced guns, BATF imposes restrictions on tracing all but guaranteed to make traced guns unrepresentative of crime guns—with changes in those restrictions making temporal comparisons of tracing data problematic even as they improve the efficiency and usefulness of tracing as a law-enforcement tool—even when those restrictions are used without an eye for policymaking or policy support. This is somewhat contrary to the BATF’s NTC’s more limited advisory that the data “ONLY reflects trends relating to those firearms for which a trace request is submitted and is only as accurate as the information provided by trace requestors.”

A. BATF Refusals to Trace

At least some of BATF’s restrictions on tracing, however, unlike the unknown mechanisms by which local jurisdictions decide which guns to trace, are systematic, even if changing. BATF does not like to attempt traces when success is unlikely. The motive may to enhance BATF’s tracing success rate—in the same way prosecutors pride themselves on conviction percentages—or simply because the Bureau wishes the most cost-effective use of its resources.

BATF has thus long excluded older firearms (Brill, 1977, pp. 94-95), as well as those whose serial numbers have apparently been removed (Kennedy, Piehl and Braga, 1996a, pp. 172-73), the technical efforts needed to determine the number being deemed excessively costly.

In order to enhance the apparent success rate, local law enforcement is asked to prescreen guns, and not ask for traces on those likely to be too old to be traced. (Brill, 1977, pp. 57-58)

B. Traces only to Retail Sale

The same cost-effective motivation means almost never seeking to trace a firearm beyond its initial retail transfer.[9] In the past, BATF counted a trace completed and successful “(1) where the firearm is traced to a dealer located in the same state as the requestor or (2) where the firearm is traced to an individual purchaser located in a state other than that of the requestor.” (BATF, 1978, p. 2) One explanation was that, once a gun was traced to the state where the request was made, it was no longer a matter of interstate commerce and, thus, no longer a federal responsibility. (Brill, 1977, p. 83)

BATF’s desire to make its tracing cost-effective severely limits its ability to provide useful data for analysis. In the past, the records of out-of-business dealers were less accessible than those of active federally-licensed dealers, so such traces would be scotched as not worth the effort (Brill, 1977, p. 125), with such handguns underrepresented in trace samples. (Zimring, 1976, p. 105) With computerization of those records, now over half of traces use information from federal licensees who have gone out of business (Pierce, Briggs and Carlson, 1996, p. 8), a figure likely to rise, as the number of dealers has dropped by over 60% during the Clinton administration.[10]

C. Information not Reported in Traces

Tracing data rarely give much in the way of sufficient detail for analysis. For example, the make, model, and serial number of a gun may allow a quick trace, but specific information about the cosmetics of the gun may not be on record (e.g., whether a particular semi-automatic rifle has a folding stock); other information not determined by the manufacturer will also be left out, such as the capacity of the magazine in the firearm as recovered.[11]

Information which should be readily available may be reported incorrectly or at least inconsistently. (Roth and Koper, 1997, p. 88) Tracing data for 1988, for example, list Colt’s semi-automatic versions of its M16 at least a dozen different ways—with variations on spacing, hyphenation, names, letters, abbreviations, and the like, plus others where the designation is unclear, or the name and model are totally wrong. There is even more variance for listing of the Norinco semi-automatic imitation of the AK-47.

In addition, traces rarely go beyond the simple information of who bought a gun where, to whether that same purchaser acquired other firearms within a relatively short period of time in the same or nearby stores. While some additional data could be elicited from traces, that would involve expenditures of manpower incompatible with BATF efforts to make tracing a cost-effective law enforcement tool.

D. BATF Recordkeeping Improvements

Improvements in BATF recordkeeping and computerization—some lawful and some apparently ultra vires—have enhanced the Bureau’s ability to conduct traces, particularly of recent sales and of out-of-business dealers. And BATF has made efforts to encourage more traces by law enforcement agencies, particularly urban agencies, increasing the number of traces from roughly 40,000 annually to closer to 100,000.

That effort has been seen by a friendly critic with decades of experience at BATF as partially politically inspired and based on a misunderstanding of the firearms market and the purposes of tracing. He argues that the current tracing efforts incorrectly emphasize trafficking, even though most crime guns move in individual transactions. (Vizzard, 1997, pp. 202, 210, 217-18)

Of course, the improvements make earlier tracing data chronologically incomparable to more recent data. The improvements are geared toward enhancing the speed with which successful traces can be conducted, and minimizing the need for labor intensive work by BATF agents. Yet it is precisely the sorts of information which might be elicited from in-depth investigation from which criminologists might hope to learn more detailed information about criminals and their guns and gun sources.[12]

E. Emphasis on Newer Guns

BATF has recognized that tracing older firearms to their first retail purchaser is not a cost-effective way to attempt to solve crimes, but that tracing more recent guns may help solve crimes and also provide information useful for allocating law-enforcement resources toward particular dealers, dealer types, or areas. Thus, BATF has more sharply limited its willingness to attempt traces. BATF has gone this decade from rejecting most attempts at pre-1985 guns to rejecting most attempts at pre-1990 guns. (Kennedy, Piehl and Braga, 1996a, pp. 170-71) Because traced guns are normally over five years old,[13] the pre-1990 exclusion obviously undermines any confidence that traced guns are representative of crime guns. (Pierce, Briggs and Carlson, 1996, pp. 8-9 and Table 3)

The emphasis on newer guns automatically means an emphasis on semi-automatics compared to revolvers, since they have come to dominate the newly-manufactured handgun market, going from about one-quarter to about four-fifths of new handguns between 1978 and 1993. (Thurman, 1994, pp. 102-103) To some extent, such a new-gun emphasis would also emphasize the relatively newer military-style semi-automatics and relatively inexpensive semi-autos as well. (Wintemute, 1994)

F. Tracing Failures

Trace attempts are frequently unsuccessful, even after exclusions, with a figure increasing slightly over time. In the 1970s, the estimate was that up to about 40% of traces were unsuccessful (Brill, 1977, pp. 84, 117; Weber-Burdin, Rossi, Wright and Daly, 1981, ch. 4, pp. 6-7), with a 45% failure rate with the massive tracing the guns of “youth offenders” in Boston.[14] (Kennedy, Piehl and Braga, 1996a, p. 193) And, while the data were not presented particularly clearly, it appears that a trace study by a BATF agent in the Los Angeles area achieved only about a 42% success rate, supplementing California state records checks with traditional BATF tracing procedures. (Wachtel, 1996, pp. 10-12) More recently, an extensive effort to trace guns in 17 cities resulted in a 37% success rate. (BATF, 1997, p. 6)

G. Effect of Investigations in Skewing Trace Data

Investigations can, whether deliberately with a view toward influencing policy, or by chance, affect what tracing may indicate. As was noted by Pierce, Briggs and Carlson (1996, p. 9), when they looked at dealers with the most guns traced to them, Vermont stood out disproportionately but irrelevantly because of a “sting” operation affecting the data.

Just as a sting operation would make Vermont artificially high, a serious investigation would have the same effect in evaluating Virginia as a source of crime guns—and, thus, the end of the investigation could affect the evaluation of Virginia’s gun rationing law later. When it was reported that 41% of the crime guns came from Virginia, it was variously reported that 27% of the 41% (10% of the total) (Goode, 1992), or “the vast majority” of the 41% (Hynes, 1992) came from a single gun store, which BATF was investigating in part with undercover purchases going to New York.

Regarding the one store in Virginia, there is some discrepancy between the government and the store’s owners. BATF insisted they warned the store of the problem of multiple purchasers being straw purchasers and gun traffickers and that the store was uncooperative. The owners insisted that they regularly telephoned BATF regarding multiple purchases which they thought might be suspicious and/or headed for New York, and BATF had reassured them that they were investigating the buyers and the guns should be sold. Whichever, the problem involved some guns carefully followed by BATF to New York and then traced back, not guns which just by chance happened to be traced to Virginia. (Affidavit of BATF Agent Irvin W. Moran, before U.S. Magistrate Judge David G. Lowe, August 25, 1992; letter from BATF Director John W. Magaw to Senator Olympia J. Snowe, February 23, 1996).

Speedier investigation and crackdown on the offending gunshop would have prevented the gun trafficking data from being so impressive. The Virginia data are unrepresentative because of an investigation or sting or entrapment—depending upon one’s view of the investigation.

Another clue that the investigations affected the traces is that with homicide investigations the official reason for about 8% of traces (Pierce, Briggs and Carlson, 1996, Table 3), murder investigations accounted for only 1.6% (3.7% if suicides are included as homicide investigations) of the Project Lead guns traced from New York City to Virginia. At a time period when there were over one thousand gun-related homicides in the city, three firearms were traced to Virginia as a result of homicide investigations. (Memorandum from Project Lead to Special Agent in Charge, BATF New York Field Division, October 22, 1992)

III. POLICY-INFLUENCED LIMITATIONS ON TRACING DATA

Some of the unrepresentative nature of traces may be policy related, intentionally or unintentionally. When BATF sought, in Operation CUE (Concentrated Urban Enforcement), to buy undercover the types of guns criminals buy, they had to assume or guess or calculate the types of firearms which criminals sought out. Having predetermined that criminals tended to prefer guns then described as “Saturday Night Specials”—relatively inexpensive, lower-caliber handguns with short barrels—and long guns which were NFA[15] weapons, that is what they wound up purchasing. Analysis of those guns, unsurprisingly, found that criminals use SNSs and NFA long guns.[16] (BATF, 1977, pp. viii, 20-23) The Police Foundation, about the same time, disputed both BATF’s initial evaluation and its conclusion, fearing focus on “Saturday Night Specials” could “confuse the police administrator in confronting the problem of firearm abuse.”[17] (Brill, 1977, p. v)

A. Semiautomatics

With the rise of the military-style semi-auto issue, BATF made special efforts to check out purchasers of such arms, in projects known as “forward traces” from the manufacturer or distributor to the first retail purchaser, rather than starting with a gun seized as part of an investigation. (Personal communication from gun dealers regarding BATF investigatory practices)

Special studies may influence the sort of firearm being traced, such as one in Detroit, focusing specifically on “assault weapons.” (Bea, 1992, pp. 67-68)

In addition, whether BATF made greater efforts to have local authorities solicit traces on such arms or not, rhetorical statements by politicians and higher-ranking BATF employees that such guns were the preferred choice of drug traffickers, organized crime, etc., would presumably have spurred at least some local authorities to make greater efforts to trace such guns on the presupposed and circular reasoning that the traces were more apt to provide evidence of drug trafficking, organized crime, etc.

Such an investigative reason could be the basis for the trace request, even if the ensuing investigation demonstrated that gun possession was the most serious offense involved particular cases. “If...law enforcement offices in certain regions have determined that certain types of firearms (such as military-style semiautomatics that accept large capacity magazines) should be traced because they are thought to be used by dangerous offenders, the data in the tracing system will reflect those specific concerns.” (Bea, 1992, p. 68)

If some law enforcement experts assert, and the media report, that certain types of guns are the preferred guns of terrorists, of drug traffickers, or the like, then some law enforcement authorities will be more inclined to trace those types of guns when they are seized. Similarly, if certain persons are said to be more apt to be involved in certain types of offenses—say, young black males and gangs—then guns found with the arrest of those persons are more apt to be traced, with the suspected characteristic the basis for the trace request.

It then becomes of self-fulfilling prophesy: If there is a greater tendency to trace certain types of guns, or guns found in the course of the arrest of certain types of persons, with narcotics, organized crime, or the like given as the type of criminal investigation, then those guns or persons will be found, using tracing data, to be disproportionately involved in the activity in question. The trace request cites the type of investigation; nothing in BATF tracing data indicates a negative investigative conclusion.

The unrepresentative effect of policy-related tracing was demonstrated perhaps most dramatically with the advent of the so-called “assault weapons” issue in the late 1980s, and the Cox Newspapers analysis of BATF traces.[18] While BATF tracing data indicated that military-style semi-automatic firearms constituted 19% of crime guns in Los Angeles, the highest of any of the cities studied, LAPD data indicated that such firearms constituted only 3% of crime guns seized in that city. (Cox Newspapers, 1989, p. 4; letter from Edward C. Ezell, Curator, National Firearms Collection, Smithsonian Institution, to Rep. John D. Dingell, March 27, 1989) And their actual use, two years later, in famed youthful drive-by shootings was all but non-existent, at one documented incident in 677 incidents (Hutson, Anglin and Pratts, 1994, p.326), and dismissed for their “minor role” in a study of the gang killings, for which they were supposedly a preferred weapon, during that period. (Hutson, Anglin, Kyriacou, Hart and Spears, 1995)

B. Handgun Rationing

It is possible that part of the reason for the policy goals influenced the investigation of the Virginia gun dealer and its reporting, and that the investigation was at least partly geared toward enhancing the likelihood that Virginia would impose a handgun-rationing plan, to limit handgun purchasers to one handgun purchase in a 30-day period. After all, the U.S. Attorney most actively involved in the investigation was also the Republican most outspokenly campaigning for the gun rationing measure which the state legislature was considering. (Goode, 1992; Johnson, 1992)

Similarly, more recent efforts involve the goal of nationalizing the rationing policy, by showing that Virginia’s role as a gun-supplying state has been curtailed (Weil and Knox)[19]—a goal easily achieved by determining on which states’ dealers to focus limited BATF investigatory efforts. A recent arrest suggested a single Alabama shop as the source for 2,000 guns taken to New York over a five-year period. (Associated Press, 1997)

C. Interstate Gun-running

Working with the Atlanta office of BATF, New York City authorities arranged that an “undercover officer in New York City would place an order for handguns with the defendants, who would then travel to Georgia, use an accomplice to make a seemingly lawful purchase of firearms from a local dealer, and then immediately return to New York with the guns.” Forty-eight firearms were recovered in the course of the investigation and, presumably, dutifully traced by BATF back to the place where New York authorities had arranged for many of them to be purchased. (District Attorney, County of New York, 1997) The New York authorities involved in the investigation are also actively promoting gun rationing on a national level, which is also the policy of the Clinton administration under which BATF operated.[20]

Even if policy is not the only goal, the investigators themselves helped to determine where guns would be traced to, and, in all likelihood, determined at least some of the details (caliber, action type, and price, if not also makes and models) of the sorts of guns which would be purchased and thus traced.

IV. BATF TRACING DATA USED IN POLICYMAKING AND EVALUATION

Whether due to local law enforcement practices, BATF tracing and enforcement practices, or policy-influenced decisions on what to trace, the result is that traced guns are simply not representative of crime guns.

A. Revolvers vs. Semiautomatics

Recent figures from New York City would suggest that revolvers are not used in criminal activity there. BATF’s explanation, that New York’s criminals no longer use revolvers, preferring the more modern guns (personal communication from Jerry Nunziato, BATF’s NTC) approaches the absurd.

The dramatic increase in the popularity of semi-automatics among the general public, and criminals, has led to their accounting for about three-fourths of new handguns, and an increasing portion of crime guns. But their use in crime lags behind their percentage of new guns, even where their popularity is greatest, among younger criminals; in Boston, the percentage of youths using semi-automatics was reportedly up to 63% of handguns—with near parity between the two main action-types of handguns among older criminals. (Kennedy, Piehl and Braga, 1996b, pp. 149, 155)

Notwithstanding limitations on the usefulness of tracing as a means of understanding criminal sources for and preferences in crime guns, particularly during the past decade, and often with the assistance or encouragement of BATF, those data have been used to influence policymaking on the gun issue. Curiously, the areas where tracing data use has most reasonably been related to possible suggestions regarding policymaking, the data have not influenced policy.

Nevertheless, with some encouragement from BATF (e.g., Vince, 1997, p. 207), tracing data analyses and studies are being used to influence and evaluate policymaking. BATF’s interest may involve partly a desire to support the policies of the administration overseeing its operations. But BATF tracing data do not necessarily spring to mind when considering serious efforts to evaluate gun-related legislation.

B. “Assault Weapons”

The federal ban on so-called “assault weapons”—primarily a redesign requirement since existing guns were not regulated (Roth and Koper, 1997, pp. 13-14), and virtually any banned gun could be modified to be removed from the definition of “assault weapon” (letter from John W. Magaw, BATF Director, to Sen. Larry E. Craig, April 1, 1994)—called for an evaluation on the effects of the legislation after three years. The FBI Uniform Crime Reporting Section was asked in advance of the legislation’s enactment if it knew “of any data which exist which would provide a base for determining whether these firearms are used more, less, or the same during the next three or four years, or are more of less available to criminals?” The response was, “The UCR Section knows of no existing data to provide a basis to address the question.” (Letter from Paul H. Blackman, NRA, to J. Harper Wilson, July 20, 1990; letter from J. Harper Wilson, Chief, Uniform Crime Reporting Section, to Paul H. Blackman, September 5, 1990)

Nonetheless, the legislation was enacted, including an obligation to evaluate its effectiveness, a task assigned to the Urban Institute, with assistance received from BATF, the National Alliance of Stocking Gun Dealers, Handgun Control, Inc., and a number of other researchers and organizations.

Absent other sources of information, the Urban Institute used BATF tracing data, recognizing some of its limitations, including the nonrepresentative sampling suggesting only about 10% of gun crimes and 2% of violent crimes result in BATF trace requests. The Urban Institute further noted the lack of a comparison between traces of “specific banned assault weapon models with trends for non-banned models that are close substitutes.” (Roth and Koper, 1997, pp. 8, 82) They nonetheless defended the use as “the only such national sample” although “BATF trace data should be interpreted cautiously.” (Roth and Koper, 1997, p. 83) With no reliable data on pre- or post-legislative criminal misuse of proscribed or similar guns, the caution is more advised than nevertheless proceeding with the uncertain interpretation.

C. One Gun a Month

As was noted above, BATF tracing data were used in popular literature—with “Batman” possibly making the Virginia state legislature the only such body effectively lobbied by a cartoon character, certainly the only one where violence-control legislation was deliberately inspired by a cartoon character devoted to the glamorization of violence—to adopt legislation designed to support gun rationing as a means to curb gun trafficking,[21] and then in its evaluation. (Ostrander and Giarrano, 1993; Weil and Knox, 1996; Vizzard, 1997, pp. 217-18)

Even though the effort was to place chronological limits on handgun purchases in response to allegations that gun traffickers bought substantial numbers of guns at a time, no effort was made to determine whether any of the guns involved in violent crime investigations, before or after the law took effect, involved multiple purchase. This despite the fact that purchases of more than one handgun in a business week are reported to BATF by the dealer [18 U.S.C. § 923(g)(3)], and investigations of dealers, such as that which led to the prosecution of the largest alleged Virginia source of New York crime guns, was spurred by such multiple purchase reports. (Hynes, 1992)

And the number of guns associated with violent crimes was tiny. Roughly one-quarter of one-percent of New York’s criminal homicide investigations resulted in a trace to Virginia. That figure is meaningless; Virginia-bought guns could have been involved in a small percentage of homicides, easily explained by normal American mobility, or a large number best explained by gunrunning, or by something in between—such as individual evasions of New York’s restrictions on the private acquisition of handguns.

Project CUE, at a time with similar out-of-state sources for New York’s traced guns, in its investigations, found “that the majority of the firearm movement from States is occurring on an individual basis. That is to say that an individual will acquire a firearm in another State through the actual purchase by relative or friends and then transport that firearm back” to his own metropolitan area, with self-protection the primary motive. (BATF, 1977, p. 61) That view remains the conclusion of the historian of BATF, who voices criticism of the new focus on trafficking. (Vizzard, 1997, p. 202) Project CUE went beyond simple tracing data, which provide no particular reason to suggest any particular explanation as to where New York City’s violent criminals get their guns or whether gun rationing at the state or federal level is a rational response.

BATF tracing data are nonetheless being used to support the notion that gun trafficking is widespread and requires national gun rationing. The data, although described as a “Congressional study,” with words of praise by respected criminologists (Butterfield, 1997), is simply the analysis of BATF tracing data provided to Rep. Charles Schumer—as they would be to anyone requesting them (Letter, with documents, from Averill P. Graham, BATF Senior Disclosure Specialist, to Mark Barnes, April 3, 1997)—broken down by state of retail dealer for each state of trace for a 12- or 13-month period beginning January 1, 1996.

As always, since only an unrepresentative fraction of guns involved in criminal activity were traced, and most traced guns are not involved in violent crimes, and no research was conducted to determine how the guns happened to reach the state from which the trace was initiated, there is no way to know the extent, if any, to which gun trafficking, specifically gun trafficking involving purchases of over one handgun in a 30-day period—already a serious federal felony [18 U.S.C. §§ 922(a)(1) and (5), 924(b), among others]—was involved.

D. Brady Act

BATF tracing data are also being used to demonstrate that gun trafficking—sometimes from the same source states—is diminishing due to the Brady Act. (Weil, 1997) Since the tracing data can show neither the problem of gun trafficking nor a contemporaneous solution, there is no more reason the data should not be able simultaneously to show both.

Crediting the Brady Act requires assuming that an Act aimed at encouraging local law enforcement officials to check criminal records of residents attempting to purchase handguns from licensed dealers also affects something at which it was not aimed, and at records not being checked by police—sales to non-residents. In addition with problems with data, then, there is a problem with logic. Brady was aimed not at interstate trafficking but at in-state purchases, envisioning a police check of records, not of residences.

E. Inexpensive Handguns

In addition to legislative measures addressed toward military-style firearms and gun rationing, efforts have been made by policymakers and those assisting them to use tracing data to support additional restrictions on small, inexpensive handguns—variously called “Saturday Night Specials,” “junk guns,” and “ring of fire” guns—and semi-automatic handguns,[22] on dealers, and on the frequency of handgun purchases. None of the tracing data available would allow an evaluation of the proposals or support the restrictions or, for that matter, opposition to the restrictions. Not enough is known about the sorts of crimes involved with the various small guns, nor is a sizeable enough portion of guns used in violent crimes traced by BATF.

F. Firearms Surrenders

Another study used tracing data to show, among other things, how different guns turned in during amnesties were from guns used by criminals, particularly younger criminals. So different were the guns turned in that only about one-eighth could be traced, and an effort at evaluation found that three-fourths of the guns were manufactured before the enactment of GCA’68. (Kennedy, Piehl and Braga, 1996b, pp. 156-58) The authors went on to conclude that, while tracing data gave no reason to believe turn-in programs would have crime-control value, they might be beneficial for symbolic values. (1996b, p. 165)

V. POTENTIAL POLICYMAKING AND EVALUATIVE USES OF BATF TRACING DATA AND THEIR LIMITATIONS

The improvements in tracing records and their analysis should enhance law enforcement efforts, particularly against illicit firearms traffickers, even if tracing’s role is exaggerated partly for political reasons. (Vizzard, 1997, pp. 202, 218) Certainly an avid supporter of tracing for law-enforcement and analytical purposes, David Kennedy, observed with regard to Boston, “There is very little illegal trafficking interdiction going on.” (Lattimore and Nahabedian, 1997, p. 223)

A. Problem Dealers

There is only one apparent policymaking utility for that tracing. An evaluation of which dealers are more apt to have firearms traced to them, in addition to suggesting which dealers may be breaking the law themselves, or insufficiently diligent, or simply in an area where criminal misuse by customers is more popular, might suggest the curtailment of which sorts of dealerships might disproportionately reduce illicit firearms trafficking.

Research by Glenn Pierce et al. for BATF has suggested that a tiny fraction of dealers are vastly disproportionately involved in firearms traces. (Pierce, Briggs and Carlson, 1996, Table 5) Those data could provide a basis for seeking more information about those dealers which could suggest for whom federal firearms licenses should be more difficult to obtain, or other regulations which might be appropriate.

For example, the administration, eventually with the legislative approval of Congress, has drastically reduced the number of dealers during the past few years. Data on dealer tracing could suggest whether the sorts of dealers driven out of business constitute the sort of dealer most or least apt to sell guns eventually traced to them. Those data were not used to make the policy. And there has been no post facto suggestion that the policy was warranted by the data.

Indeed, with the goal of putting out of business “convenience” dealers—those who have licensed in order cheaply to purchase firearms for themselves and their friends—in order to allow BATF to focus more regulatory attention on the remaining dealers, the sharp reduction in dealers may have been akin to rescinding the drivers licenses of persons over 40 in order to allow more effective policing of the driving habits of the remaining younger drivers.

Similarly, there are current legislative efforts to authorize BATF to mandate storage requirements for licensed dealers. Tracing data, and data on stolen firearms, might allow some evaluation of whether theoretically more susceptible dealers are, in fact, more apt to be the victims of thefts. But no such attempt has yet been made. (BATF, 1995)

B. Firearms Trafficking

There would appear to be no other obvious area where policymaking might benefit from an analysis of BATF firearms tracing data as currently collected. A trafficking study could be useful for law enforcement, but not for a study of criminals’ guns or their sources, given the small number traced, the huge number of models recovered and the resultant small numbers of traces required for a gun to make it into some city’s “top ten,” and the lack of relationship of most crime guns to violent crimes. (BATF, 1997)

And thus even in a trafficking study, traces alone would be insufficient without additional information about the types of dealerships—their conformity to local zoning and other regulations, and the like—which would make traces more time-consuming and costly. With more serious follow-up research, there would, however, be other areas where cautious use of tracing data might provide the base for more extensive research.

Similarly, if BATF traces were followed up by more extensive investigation than the simple trace, the data could prove useful in learning more about where criminals get their guns and what their preferences are. For example, if data were collected on the relation of the traced firearm to the criminal investigation (used in the homicide, recovered at the scene, etc.) or follow-up information on the criminal investigation (was the criminal investigation founded? was there drug trafficking involved, or had the gun in fact been taken in a burglary, etc.? how did the firearm come to be in the state where it was recovered? what was the path of ownership and the means of transfer?), then the potential would exist for learning more about the nature at least of relatively new crime guns or criminal preferences in guns.

Most efforts by BATF, however, have been to curtail tracing to make it more cost-effective, not to expand the information gathering with labor-intensive follow-up inquiries. Thus, while the Congressional Research Service noted the problems with the tracing system in terms of statistical analysis, it made it clear that the limitations on the system should not necessarily be rectified: “the system is designed to expedite requests from law enforcement agencies on the history of firearm ownership, there would likely be little benefit in placing additional restrictions or requirements on officers submitting the trace request. The more important accomplishment of the system design...is to minimize paperwork and administrative burdens on the requesting agency.” (Bea, 1992, pp. 65-66)

Efforts to encourage more detailed data collection by BATF and from local law enforcement is apt to be even less successful than the current efforts at more thorough data collection for the Uniform Crime Reporting Program. The currently envisioned expansion of tracing to include guns which cannot be traced, but are merely seized in cooperating cities, means less and less will be known about more and more firearms.

CONCLUSION: GIGO (“GARBAGE IN, GARBAGE OUT”)

As currently collected—a small non-random undifferentiated sample of guns about whose involvement in crimes committed by whom little is known or asked—BATF tracing data cannot be used criminologically, with the possible hypothetical exception of giving some clues regarding dealers as sources for some misused firearms.

Suggesting sharp limitations on the utility of BATF tracing for criminological research in no way undermines either the benefits of tracing as a law-enforcement tool in general, or the benefits of recent improvements in BATF’s tracing abilities. The traces were envisioned as a law-enforcement tool, not a law-making tool, and retain utility for that envisioned purpose.

To the extent it might be argued that, however weak, BATF tracing data are the only data available for certain criminological or policymaking goals, that discouraging fact would simply mean there are no data available; absence of other data does not make unrepresentative data representative. And no amount of sophisticated computer-assisted analysis changes the fact that if garbage is programmed in, garbage will be programmed out.

Analyses of tracing data, however performed, are akin to analyses of astronomical data for astrological projections. There is no need carefully to evaluate the data or the analyses; they are worthless. Tracing data can no more provide a sound basis for criminological analysis than can works of fiction.[23] Studies based on tracing data simply diminish the value of otherwise useful blank paper used for publishing the analyses.

REFERENCES

Adams, N.M. 1978. “The Tracing of Beretta A47469.” Reader’s Digest 204-34 (January).

Associated Press. 1997. “Siblings Held in Gunrunning Ring.” May 9.

BATF [Bureau of Alcohol, Tobacco & Firearms]. 1977. Concentrated Urban Enforcement: An analysis of the initial year of operation CUE in the cities of Washington D.C., Boston, Mass., Chicago, Ill.

BATF. 1995.[24] Stolen Firearms: Statistical Report, 1994-1995. For official use only.

BATF. 1997. Crime Gun Trace Analysis Reports: The Illegal Youth Firearms Markets in 17 Communities. Washington, D.C. (July).

BATF. 1978. Briefing Paper: Firearm Tracing. ATF news Release, FY 78-26, April 1.

Bea, K. 1992. “Assault Weapons”: Military-Style Semiautomatic Firearms: Facts and Issues. Washington, D.C.: Congressional Research Service (May 13; Technical Revisions, June 4).

Brill, S. 1977. Firearm Abuse: A Research and Policy Report. Washington, D.C.: Police Foundation.

Butterfield, F. 1997. “Report Links Crimes to States With Weak Gun Laws.” New York Times, April 9, p. A14.

Chichioco, T. 1989. “Computer-assisted reporting.” Editor & Publisher 20-21 (Sept. 2).

Cox Newspapers. 1989. Firepower: Assault Weapons in America. Washington, D.C.: Cox Newspapers.

District Attorney [Robert M. Morgenthau], County of New York [Manhattan]. 1997. News Release, April 29.

FBI [Federal Bureau of Investigation]. 1990. Crime in the United States, 1989. Washington, D.C.: U.S. Government Printing Office.

FBI. 1996. Crime in the United States, 1995. Washington, D.C.: U.S. Government Printing Office.

Fyfe, J.J., D.A. Klinger and J. Flavin. 1997. “Differential Police Treatment of Male-on-Female Spousal Violence,” Criminology, 35: 455-73.

Goode, R. 1992. “Co-owners of gun shop plead guilty.” Richmond Times-Dispatch, September 2.

Hargarten, S.W., T.A. Karlson, M. O’Brien, J. Hancock and E. Quebberman. 1996. “Characteristics of Firearms Involved in Fatalities.” Journal of the American Medical Association 275:42-45.

Hutson, H.R., D. Anglin and M.J. Pratts, Jr. 1994. “Adolescents and Children Injured or Killed in Drive-By Shootings in Los Angeles.” New England Journal of Medicine 330:324-327.

Hutson, H.R., D. Anglin, D.N. Kyriacou, J. Hart and K. Spears. “The Epidemic of Gang-Related Homicides in Los Angeles County From 1979 Through 1994.” Journal of the American Medical Association 274:1031-1036.

Hynes, P. 1992. Case Report, Case No. 63560-92-4031 S, Eastern Judicial District of Virginia [re: defendants Virginia Police Equipment Co., Donald Edward Weiss, and Hildegard Francia Weiss], BATF, Nov. 23.

Johnson, M. 1992. “Virginia gun-running is ‘embarrassment.’” Richmond Times-Dispatch, September 11.

Kates, D.B., H.E. Schaffer, J.K. Lattimer, G.B. Murray and E.W. Cassem. 1995. “Guns and Public Health: Epidemic of Violence, or Pandemic of Propaganda?” Tennessee Law Review 62:513-596.

Kennedy, D.M., A.M. Piehl and A.A. Braga. 1996a. “Youth Gun Violence in Boston: Gun Markets, Serious Youth Offenders, and a Use Reduction Strategy.” Law and Contemporary Problems 59:147-196.

Kennedy, D.M., A.M. Piehl and A.A. Braga. 1996b. “Gun Buy-Backs: Where Do We Stand and Where Do We Go?” Pp. 141-171 in Under Fire: Gun Buy-Backs, Exchanges and Amnesty Programs, M.R. Plotkin, ed. Washington, D.C.: Police Executive Research Forum.

King, W. 1988. “A Bureau That Battled Bootleggers Is Tough Target for Budget-Cutters.” New York Times, February 1, p. A26.

Kobayashi, B.H. and J.E. Olson. 1997. “In Re 101 California Street: A Legal and Economic Analysis of Strict Liability for the Manufacture and Sale of >Assault Weapons.’“ Stanford Law & Policy Review 8: 41-62.

Lattimore, P.K. and C. A. Nahabedian, eds. 1997. The Nature of Homicide: Trends and Changes: Proceedings of the 1996 Meeting of the Homicide Research Working Group. Washington, D.C.: National Institute of Justice.

National Tracing Center. 1995. Significant Firearm Trace Results, FY-94. Bureau of Alcohol, Tobacco and Firearms [official use only].

Newton, G.D. and F.E. Zimring. 1969. Firearms and Violence in American Life. Washington, D.C.: U.S. Government Printing Office.

Office of the General Counsel. 1981. Management review on the Performance of the U.S. Department of the Treasury in connection with the March 30, 1981 Assassination Attempt on President Ronald Reagan. Department of the Treasury, August.

Orr, J.S. 1985. “Awesome Weapon: Authorities take hard look at ‘assassination machine.’” Newark Star-Ledger.

Ostrander, J. and V. Giarrano. 1993. Batman: seduction of the gun. New York: DC Comics.

Pierce, G.L., L. Briggs and D.A. Carlson. 1996[?]. The Identification of Patterns in Firearms Trafficking: Implications for Focused Enforcement Strategies. Washington, D.C: A Report to the United States Department of Treasury, Bureau of Alcohol, Tobacco & Firearms, Office of Enforcement.

Ross, J. 1996. Unintended Consequences: A Novel. St. Louis: Accurate Press.

Roth, J.A. and C.S. Koper. 1997. Impact Evaluation of the Public Safety and Recreational Firearms Use Protection Act of 1994: Final Report. Washington, D.C.: Urban Institute (March 13).

Sambor, G.J. 1985. “Tracing Firearms.” Police Chief 52(3): 73-76 (March).

Thurman, R. 1994. “Firearms Business Analysis.” Shooting Industry 39(12):98-127 (December).

U.S. Senate. 1965. Hearings before the Subcommittee to Investigate Juvenile Delinquency of the Committee on the Judiciary to Amend the Federal Firearms Act, 89th Congress, First Session (Washington, D.C.: U.S. Government Printing Office).

Vince, J.J., Jr. 1997. “Disarming the Criminal: ATF’s Strategic and Scientific Approach to Focused and Targeted Enforcement.” Pp. 206-214 in The Nature of Homicide: Trends and Changes: Proceedings of the 1996 Meeting of the Homicide Research Working Group, P.K. Lattimore and C.A. Nahabedian, eds. Washington, D.C.: National Institute of Justice.

Vizzard, W.J. 1997. In the Cross Fire: A Political History of the Bureau of Alcohol, Tobacco and Firearms. Boulder, Colo., and London: Lynne Rienner.

Wachtel, J. 1996. “Sources of Crime Guns.” Paper delivered at the annual meeting of the Academy of Criminal Justice Sciences, Las Vegas, March.

Weber-Burdin, E., P.H. Rossi, J.D. Wright and K. Daly. 1981. Weapons Policies: A Survey of Police Department Practices Concerning Weapons and Related Issues (Preliminary Draft).[25] University of Massachusetts/Amherst; U.S. Dept. of Justice, National Institute of Justice.

Weil, D.S. 1997. Traffic Stop: How the Brady Act Disrupts Interstate Gun Trafficking. Washington, D.C.: Center to Prevent Handgun Violence (September 19).

Weil, D.S. and R.C. Knox. 1996. “Effects of Limiting Handgun Purchases on Interstate Transfer of Firearms.” Journal of the American Medical Association 275:1759-1761.

Wintemute, G.J. 1994. Ring of Fire: The Handgun Makers of Southern California. Sacramento: Violence Prevention Research Program (University of California, Davis).

Wright, J.D. and P.H. Rossi. 1981. Weapons, Crime, and Violence in America: Executive Summary. U.S. Department of Justice, National Institute of Justice, November.

Wright, J.D. and P.H. Rossi. 1986. Armed and Considered Dangerous: A Survey of Felons and Their Firearms. New York: Aldine de Gruyter.

Zimring, F.E. 1975. “Firearms and Federal Law: The Gun Control Act of 1968.” Journal of Legal Studies 4:133-198.

Zimring, F.E. 1976. “Street Crime and New Guns: Some Implications for Firearms Control.” Journal of Criminal Justice 4: 95-107.



Endnotes


--------------------------------------------------------------------------------

[1]Until its promotion to bureau status in 1972, it was the Alcohol, Tobacco & Firearms Division of the Internal Revenue Service. For the past quarter century, its name has generally been shortened to the letters BATF or ATF, the former seeming more thoroughly descriptive, and perhaps slightly more suitable as an acronym, but both have been used extensively both outside and inside government, including the Treasury Department itself.

[2]BATF offered its tracing capability to the U.S. Secret Service at 2:40 p.m. on March 30, 1981, with BATF personnel ordered to stand by for an urgent trace. The Secret Service contacted the BATF liaison to begin the tracing process at 3:20 p.m. and, following some confusion on the Secret Service’s part regarding the serial number, the trace was completed by 4:30 p.m. (Office of the General Counsel, 1981, pp. 78-79.) The General Counsel found it noteworthy that the investigative activities were initiated during normal working hours, and that the tracing capability “would assume even more importance if a suspect had not been immediately apprehended at the scene.” (Office of the General Counsel, 1981, pp. III and 81) The trace of the handgun used in the assassination attempt on Governor George Wallace took about 30 minutes. (Brill, 1977, p. 119)

[3]Many of the limitations noted in this paper replicate the introduction to BATF’s National Tracing Center (NTC) by its director, Gerald A. Nunziato, in speaking to visitors, for example, from the Homicide Research Working Group on June 10, 1997.

[4]Although Commissioner Sambor did not explain, the reduction in likelihood of an error simply reflects the fact that the fewer the intermediaries, the less the likelihood of an error in the transcription of a serial number or other information essential to an accurate trace, and that no reflection on BATF personnel in particular was intended.

[5]Similarly, because of Michigan’s records, guns originally purchased there were prescreened and excluded from BATF’s Project Identification data. (Zimring, 1976, p. 105)

[6]According to the trial transcript in United States v. Hinckley, #81-306, pp. 1489-1559, 1751-52, when he was arrested at the Nashville airport on October 9, 1980, with three handguns in his carry-on luggage, they were seized, but he was fined only a total of $62.50 for the misdemeanor, using his Texas drivers license as identification, which listed his address as 1612 Avenue Y, Lubbock, Texas. He used the same drivers license on October 13, when he purchased two inexpensive .22 caliber revolvers, to replace two of the seized handguns, from Rocky’s Pawn Shop, although the address he put on the actual federal form 4473 was 2404 10th Street, Lubbock. The address listed for him in the November 1979 Lubbock-Slaton telephone directory was 409 University Avenue. According to a Washington Post article from March 31, 1981 (reproduced as part of a fundraising package from Handgun Control, Inc.), each of the revolvers cost about $47.50 (although later advertisements from that organization have lowered the price to $29). The diversity of addresses suggests that some college students—Hinckley was a sometime student at Texas Tech, as he had been in the summer sessions of 1980 (personal communication)—move around, but was irrelevant under federal law which was only concerned with whether he was a resident of Texas. Explaining the meaning of “State of residency,” ATF Rul. 80-21 explains that “during the time the students actually reside in a college dormitory or at an off-campus location they are considered residents of the State where the dormitory or off-campus home is located. During the time out-of-State college students actually reside in their home State they are considered residents of their home State.” Had a name-check been conducted—as it likely was following BATF’s receipt of the multiple purchase form from the pawnshop—it would have found, at most, a misdemeanor record for a Texas resident. As the resident of his home state, using a Colorado identification card with an Evergreen, Colorado, address, Hinckley purchased a .38 caliber revolver from the Kawasake West gunshop on January 21, 1981, to replace the third handgun seized in Nashville. Any alleged illegalities in either purchase could have led to prosecution in Texas or Colorado until the mid-1980s, but no such charges were brought.

[7]The categories are not necessarily all that revealing: “other”; “miscellaneous.” Even “weapons” or “GCA” or “Title 1” cover a multitude of possible offenses, from trivial typographical errors to gun-trafficking and violent offenses. (Bea, 1992, p. 71) A stolen weapon trace could involve the thief or a gun found and turned in to authorities. Some traces may be of police-owned firearms. (Brill, 1977, pp. 23-25) Crimes reported as the basis for traces in BATF’s Project Lead in New York City included suicide and loitering. (Memorandum from Project Lead to Special Agent in Charge, New York Field Division, BATF, October 22, 1992)

[8]Murder weapons may differ from guns used in non-lethal assaults (Brill, 1977, p.71), regardless of one’s position in the motivation versus instrumentality debate.

[9]Such labor-intensive tracing may be possible if deemed essential to a case. The Beretta used in about half of the so-called Zebra slayings in San Francisco in the 1970s was traced, over a period of over eight months, by BATF and then the San Francisco police beyond the first retail sale through seven private transfers. (Adams, 1978)

[10]It will not necessarily rise as fast as the number of dealers has fallen, or just because of that. The vast majority of traces involve a small portion of dealers. Ninety-two percent of dealers were involved with no traces, and less than 2% of dealers accounted for over three-fourths of traces. (Pierce, Briggs and Carlson, 1996, Table 5) The records of those dealers driven out of business by increased regulations will likely not be anywhere near so useful as of those driven out of business as a result of criminal investigations of their activities.

[11]The more information which is required or provided, the more likelihood for error. The author once ran a .38 caliber revolver from California through the NCIC and was told it was a .22 caliber revolver from Alabama.

[12]BATF’s then-Chief, Firearms Division, Joseph J. Vince, Jr., told the Homicide Research Working Group’s 1997 symposium that in some cities, even though BATF could not trace all firearms, it was collecting available information on all guns seized by police. Since those data would be limited to the few cooperating cities, they would be comparable to data currently available from some cities analyzing the guns which are taken into custody. Cities can vary dramatically in a variety of data, so that there would be no reason to suppose a few cities might be representative either of cities in general or of the nation as a whole. For example, with 17 cities cooperating extensively in tracing efforts, the percentage of guns associated with firearms offenses ranged from 36% to 92%, and the number with an in-state source varied from 0-77%. (BATF, 1997) In addition, it is unclear whether departments which cooperate with outside studies differ from those more reluctant to cooperate. If differences exist in the way spousal violence is treated, the differences may or may not imply similar differences in way gun seizures, information, and gun-related cases are treated. (Fyfe, Klinger and Flavin, 1997). Even if the cities were typical, the information would still generally be useless for anything except telling about the sorts of guns were seized by police. If broken down by crime type, the data might give some information about the sorts of guns used by criminals in specific crimes in specific cities, but, since most of the firearms could not be traced, the data would still not provide information about the sources of criminals’ guns.

[13]Traced semi-automatic handguns tend to be roughly half as old as revolvers and long guns (Pierce, Briggs and Carlson, 1996, Table 3; Wachtel, 1996, Table 5), and their predominance in the marketplace is similarly relatively recent. Semiautos clearly overtook revolvers among domestically manufactured firearms only during the past decade. (Thurman, 1994)

[14]The Boston police and BATF reportedly agreed to trace every firearm seized. The figures suggest about 500 trace attempts of seized guns annually, plus an additional 120 guns recovered other than for possessory or substantive crimes. (Kennedy, Piehl and Braga, 1996a, p. 196) David Kennedy says about 700-1000 firearms came into BPD custody annually during the 1990s, with the number decreasing. The estimate on firearm confiscations in Boston in 1974, on the other hand, was over 1700. (Brill, 1977, p. 27) The annual number of violent crimes remains at roughly 10,000.

[15]National Firearms Act of 1934, also known as Title II, based on its incorporation into Title II of the GCA’68. The most common NFA weapons used by criminals are sawed-off shotguns. (Wright and Rossi, 1986, p. 95-97)

[16]Although Operation CUE was not primarily a tracing activity, tracing was a facet of the operation. (BATF, 1977, pp. v, 58-65)

[17]That conclusion was supported by the NIJ-funded survey of felons regarding their preferences in firearms. (Wright and Rossi, 1986, ch. 8)

[18]The Cox Newspapers analysis of BATF traces constitutes an odd combination of a news company’s policymaking goal and BATF’s desire for increased access to its own data, at a time when BATF was being asked by the administration to justify restrictions on military-style semi-automatic firearms. In exchange for access to BATF tracing files, which it hoped to use to show that so-called “assault weapons” were disproportionately used in crime, Cox Newspapers assisted in getting those data onto computers, to the benefit of BATF. (Cox Newspapers, 1989, pp. 31-32; Chichioco, 1989) A somewhat different approach was taken in California where, following initial indications that information on the types of guns used in crime would show low levels of “assault weapons,” collected state data indicating 1.8-2.9% use in serious crime were suppressed. BATF tracing was not involved. (Kobayashi and Olson, 1997, pp. 43-44)

[19]In addition to problems with examining changes in traces to Virginia, explaining changes based on the gun rationing law would be undermined by two factors: First, the same legislative session required proof of residency for driver’s license applicants (Virginia Code § 46.2-323). And the rationing, in fact, rarely applies; during the first three years, applications for multiple handgun purchase requests were denied to 3% of applicants, and another 2% withdrew their applications. (Personal communication from Captain R. Lewis Vass, Department of State Police, August 30, 1996) Captain Vass testified to a Virginia crime commission that the rationing law has “not significantly affected ... the number of multiple handgun purchases within the Commonwealth.” (August 29, 1995) Subtracting the single gunshop from the 1991-92 data would suggest roughly 24-28 traces from New York City to Virginia monthly. (Hynes, 1992, p. 113; Goode, 1992) BATF tracing data from January 1, 1996, through January 31, 1997, reports 372 handguns traced to Virginia (with Florida in second place at 242, and South Carolina, another gun-rationing state, third at 220, and New York sixth), which would work out to about the same, with no data on the source of actual crime guns used in New York City before or after the Virginia law took effect. Tracing data would suggest nothing much had changed in the past two decades. New York City guns traced to New York State has risen from 4% or 5% in 1973 to 8%, and those traced to the four Southern states of Virginia, South Carolina, Georgia, and Florida, has fallen from 56% to 46%. But in between, a small sample hoping to show that South Carolina’s gun rationing law might be helping, 19% of New York City’s guns were traced to New York State and 46% to the four southern states. (BATF tracing data, January 1, 1996, through January 31, 1997; Brill, 1977, pp. 83-84, 91-93; Zimring, 1975, pp. 181-82) Traces of Boston crime guns to Massachusetts has gone from 35% to 37%. (Brill, 1977, p. 84; Kennedy, Piehl and Braga, 1996a, p. 196) And Los Angeles crime guns traced to California has gone from 82% to 81%. (Brill, 1977, p. 84; Wachtel, 1996, Table 6) Tracing data, however, do not provide a reliable measure of changes in gun sources, especially relative to recent changes in federal or state laws, and a time frame involving substantial changes in BATF tracing practices.

[20]Those suspecting a possible public relations aspect of the investigation would also note that the BATF Special Agent in Charge of the Atlanta field division is Jack Killorin, who for a long time headed BATF’s public affairs office in Washington.

[21]The actual punishment for massive numbers of trafficked guns can be relatively small, with relatively short prison terms (14 months not being unusual) and/or probation. (Wachtel, 1996, Table 7)

[22]One generally invalid criticism of the use of tracing, and more localized similar, data on firearms is that those using it provide no reports on the proportionate availability of the guns, only data on the use of some guns in crime; that is, there are no data provided on non-crime guns. (Kobayashi and Olson, 1997, p. 49) In fact, those using traces to attack particular guns do provide some information on the traces proportionate to availability.

Unfortunately, those data are often combined with rhetoric and the supposed uselessness of the guns in question for legitimate purposes. Wintemute, for example, asserted that what he called “ring of fire” handguns—predominately small and inexpensive—“truly are weapons of choice for criminal use,” because they were traced disproportionate to their production. While his data support the disproportionality, they also show traces accounting for 0.33% of the guns manufactured rather than the 0.1% for major manufacturers in Connecticut. (Wintemute, 1994, p. 63)

While higher percentages of both groups of manufacturers’ products may well have been involved in crime, tracing data provide no real confidence the “ring of fire” handguns are misused relatively more than the “Gun Valley” handguns, and suggest vast numbers of the handguns in question are “weapons of choice” for non-criminal use.

Another dishonest effort at comparison, albeit not one relying upon tracing data, involved using overall domestic manufacturing data over a 20-year period to suggest disproportionate involvement of .25 caliber pistols in a big-city’s suicides and homicides—10% of manufacturing vs. 14% of reported involvement (and 13% if unknown calibers were not apportioned). (Hargarten et al., 1996)

That use of comparative data ignored the fact that protection-type calibers are more apt to be owned in big cities, where sporting uses of handguns are less available, and that a shorter time frame (more in keeping with the fact that relatively newer guns are used in crime) would similarly record that 13% of manufactured handguns are .25 caliber pistols. (Thurman, 1994, pp. 101-102)

The Cox Newspaper analysis, while asserting that “assault weapons” were ten times more likely to be misused relative to their availability, did not emphasize that this was based on four thousand traces of what it asserted to be about one million guns. In addition, the availability data may not be accurate. Cox Newspapers assertion of one million “assault weapons” (1989, p. 1), conflicted with the estimate of at least 3.7 million such arms by the Smithsonian Institution’s firearms expert, Edward Ezell (letter to Rep. John D. Dingell, March 27, 1989), and the Cox Newspapers’ elsewhere counting the M1 Garand as an assault gun and reporting the availability of 1.5 million of those. (1989, p. 10)

[23]This view is not necessarily universally shared. Actress Demi Moore told the television show “Entertainment Tonight” that her movie “GI Jane” proved that women could successfully serve in elite military combat forces. Gun control opponents, who believe some of their points about the evils of firearms registration were proven in the film “Red Dawn,” more recently have been promoting the novel Unexpected Consequences (Ross, 1996) as evidence that gun control could be dangerous to American society.

On the other side, in the mid-1980s, when then-Rep. Robert Torricelli was introducing legislation aimed at semi-automatic handguns which could be readily converted to machineguns, his source for believing such firearms were a crime problem was the “Miami Vice” television show. (Orr, 1985)

Like effective fiction, tracing data may provide rhetorical support for criminological or political views. This does not mean that fiction and tracing data are of no utility to criminology. One could use Erle Stanley Gardner’s novels to supplement his other writings to summarize and evaluate his criminological beliefs, and one could use BATF tracing data to evaluate how traces are used for crime control. Those data, however, do not provide information useful for studying how criminals obtain firearms or the firearms criminals use to commit crimes.

[24]The report has no date, but internal evidence suggests it was produced in 1995.

[25]According to Lois Mock of the NIJ, there was no final draft published, leaving the preliminary draft, paginated within each chapter, the only one available.