Commentary on “Gun control focus narrows”
Philip F. Lee, 1/1/01 (rev. 4/20/08)

The Minneapolis Star Tribune article "Gun control focus narrows" (Dec. 18, 2000) contained the following paragraph:

ATF figures  compiled  by [Senator]  Schumer  show  that nearly half

of  the  77,010  guns traced  to crimes  around the nation  were sold

legally by about 1,160 gun sellers - or about 1 percent of all federally

licensed firearms dealers.

Senator Schumer's figures and the article seem to say that many criminals buy guns  through trafficking schemes involving gun retailers and tightening restrictions on retail sales might cut criminal access by 50%.

There is good reason to be doubtful about the value of such restrictions. A recent Bureau of Alcohol, Tobacco, and Firearms (BATF) document shows nowhere near 50% of Minneapolis crime guns are obtained from retail sources -- see "The Youth Crime Gun Interdiction Initiative (YCGII), Crime Gun Trace Analysis Report," BATF, Nov. 2000 (covering gun tracing between 1/1/1999 and 12/31/1999 and available at URL: http://www.atf.treas.gov/firearms/ycgii/1999/index.htm).

The BATF does not report results from investigations of trafficking in these documents. Instead, it estimates trafficking for the 10 most frequently used gun models submitted for tracing in each community reported. For Minneapolis this estimate covers only 186 guns (23%) of the 833 gun traces submitted. Forty of those 186 guns were sold in retail within three years of being used in a crime. A three year period is a BATF threshold for estimating gun-trafficking. In Minneapolis only 22% of the 186 guns satisfy this criteria or 5% of the 833 total. It is possible that 95% of the criminal use of guns would be unaffected by tightening restrictions on the retail market.

So how could Senator Schumer conclude that at least 50% of crime guns were trafficked? A simple answer is that the BATF faulty estimation process inflates trafficking percentages and Senator Schumer uses these inflated results. To illustrate this inflation with one example, consider Table E: "Time-to-Crime for Most Frequently Traced Crime Guns" for Minneapolis (the relevant part of this table is reproduced at the end of this article). Nineteen Marlin .22 rifles are reported under the "All Crime Guns” category. The BATF obtains Time-to-Crime for 10 of these rifles with four measuring less than three years. By discarding the remaining nine rifles, it "claims" 40% (4 of 10) of the rifles in this case meet the trafficking threshold. Overall, the BATF discarded 58% of the Minneapolis data (traces were not completed on 484 guns -- 58% of 833). These discarded guns are mostly the old guns that can't be traced and, because they are old, are unlikely to be in the "fast time to crime" group.

When traces are not completed, Time-to-Crime is not obtained and the BATF does not include the gun in its statistics. Because older guns are more likely to have unsuccessful traces, they are disproportionately eliminated from consideration. Eliminating older guns makes recent gun sales appear more significant in the statistics. This distortion can make trafficking from retail appear to be a more significant problem than it is. In fact the BATF really has only 21% of the Marlin .22 rifles showing its Time-to-Crime trafficking indication. Most of the other firearm models in this table have similarly inflated trafficking percentages.

Also, important, the ATF trafficking "estimate" isn't confirmed with investigations. The Marlin .22 rifle doesn't seem a firearm particularly useful to criminals and it is unlikely that these four rifles were truly trafficked. But, ATF doesn't report results from investigations and that failure makes their estimates suspect.

Even if trafficking involves only 5% of retail sales, wouldn’t it be worthwhile imposing restrictions to eliminate that 5%? Certainly, that is a matter for society to debate. Such restrictions on firearms impose a cost on the lawful user including lives lost. In order to evaluate the cost and benefit to society, the nature of the restrictions should be defined and the impact on crime estimated. If trafficking involves only 5% of the retail market as the BATF data seems to suggest, it is possible that any restrictions could be offset by criminals easily. In that case, the cost would be to the lawful of society and there would be no benefit to society.

This article should be taken as a Commentary.

Philip F. Lee has a PhD in Mathematics from Georgia Institute of Technology, 1970. He has more than 28 years experience in the application of statistics to defense and industrial applications. He completed an analysis of the "Crime Gun Trace Analysis Reports: The Illegal Youth Firearms Market in 27 Communities", Department of the Treasury and the Bureau of Alcohol, Tobacco and Firearms (ATF), February 1999 and this analysis is available from URL

http://www.mcrkba.org/YCGII99/YouthCrime.html

Table E: Time-to-Crime for Most Frequently Traced Crime Guns by Manufacturer, Caliber, and Type

All Crime Guns in this Jurisdiction****
Manufacturer	Caliber		Type of Crime Gun	Number of Crime Guns				Median Time- to-Crime* in Years	Time-to-Crime    of 3 Years or less				Fastest Case (in days)***
				All	With Time-to-Crime				Number		Percent**
Remington Arms		12 GA	Shotgun	27		12	6.9		2	16.7		29
Ruger		9mm	Semiauto Pistol	23		18	1.3		11	61.1		26
Smith & Wesson		9mm	Semiauto Pistol	22		12	4.8		4	33.3		71
Smith & Wesson		.38	Revolver	20		8	10.5		0	0.0		1,994
Lorcin Engineering		.380	Semiauto Pistol	19		15	2.6		8	53.3		38
Marlin		.22	Rifle	19		10	5.9		4	40.0		198
Glock G.m.b.H.		9mm	Semiauto Pistol	14		12	1.8		8	66.7		7
Raven Arms		.25	Semiauto Pistol	14		10	10.6		0	0.0		2,473
Remington Arms		.22	Rifle	14		3	3.0		1	33.3		206
Taurus		9mm	Semiauto Pistol	14		7	4.6		2	28.6		14

* Time-to-crime can only be calculated when a trace is completed and a recovery date is submitted.  Table J provides further explanation as to why a trace cannot be completed, such that a time-to-crime cannot be established.

** The denominator used to calculate this result is the total number of trace requests where a time-to-crime was established.

*** A time-to-crime of 0 days indicates the recovery of a firearm during or immediately following a sale from a Federal firearms licensee.

**** This table is based on crime guns associated with adults, youth, juveniles, and individuals whose ages are unknown.