Understanding Guns and Gun Violence
Understanding Guns and Gun Violence
Understanding Guns and Gun Violence
Michael R. Weisser
mike@mikethegunguy.com
1. The Problem
What is the connection between the number of guns in private hands and the amount of
violence caused by guns? Scholars have examined this question for more than forty years, and
the consensus appears to be that the excessively-high rate of gun violence experienced in the
United States is first and foremost caused by the excessively-high number of privately held arms.
This argument is based on comparing rates of violence between various countries, then
comparing rates of fatal violence, then comparing per-capita rates of gun ownership.
Regarding the initial comparison, the United States is no more violent than other advanced
()ECD) countries. Even though definitions of culpable violence differ from one nation-state to
another, the United States appears to fall somewhere in the middle of rates of intentional
When a comparison is made between fatal violence rates, however, the results dramatically
change. The U.S. experiences three to seven times the rate of fatal violence found in other
advanced nation-states. Here the comparisons are also a little less than ideal, because some
jurisdictions count murder and manslaughter together as fatal violence, other penal-judicial
systems count them as separate categories, but allowing for these differences in definition, the
United States remains the murder capital for the entire OECD. 2
When we then look at other, standard SES comparisons; i.e., income, employment,
education, etc., none of these factors appear to play as salient a role as the degree to which
Americans have access to guns and use these weapons to inflict most of the heightened level of
homicides that occur within the United States. It should be noted that comparing the U.S. non-
gun homicide rate to what we find in other countries still has us sitting at the top of the heap.
Thus, it would be more accurate to say that our access to small arms doesn’t lift our fatal
violence rate above everywhere else; rather, it takes a higher rate than what is experienced in
The relationship to gun access and intentional, fatal injuries in the U.S. versus other countries
has been most recently discussed by Grinshteyn and Hemenway, in an 2015 article that updates
earlier research published by Hemenway based on data through 2003, the 2015 piece bringing
the argument forward with data through 2010.4 In this respect, the argument connecting
intentional gun homicide to the per capita number of guns follows from earlier work by
Kellerman, Rivara and others linking increased levels of homicide and suicide to gun access
within the civilian population.5 Both arguments are compelling and have been accepted by
virtually every advocacy organization trying to develop strategies to reduce violence caused by
guns.6
There is only one problem with this universally-accepted explanation, namely, that it fails to
distinguish the types of guns which are owned by civilians versus the kinds of guns which are
used to commit gun violence, either against oneself or against someone else. The fact that
someone has a ‘gun’ in the home is not the same thing as the fact that they have a ‘car’ sitting in
the driveway outside the home. Driving off in a car creates the same level of risk for every driver
regardless of the type of car being driven away. Picking up a gun creates much different levels of
The purpose of this article is to examine gun violence not from a quantitative point of view in
terms of how many guns are used to commit injuries, but from a qualitative point of view, i.e.,
what types of guns are used to commit gun injuries. The importance of approaching gun violence
from this perspective should not be underestimated because the regulatory system, developed in
1968 to reduce gun violence, presupposes that every type of gun should be regulated to the same
degree and in the same way in terms of access and use. 7 But does such a blanket regulatory
system reflect how different guns are used and therefore require the same degree of control?
2.The Data.
In fact, although there are endless studies on the number of guns that may or may not be
owned by civilians, this data and the studies which utilize it do not go beyond distinguishing the
gun stock beyond, at best, dividing all gun into two categories: hand guns and long guns, the
latter being defined as any firearm with a barrel length of 16 inches or longer, obviously the
former being defined as any firearm with a barrel of less than 16 inches in length. Where did this
distinction come from? It was embodied in the first, federal gun-control law of 1934, and it is
totally arbitrary in terms of what it says or doesn’t say about different types of guns. 8 At best, it
is a descriptive definition which tells us nothing about how a gun functions, or is used, or what it
Unfortunately, this broad and therefore basically useless definition of a ‘gun’ remains
with us to the present day and is still the only descriptive used to define the American private
gun stock from a qualitative point of view. Since 1968, every gun which enters the commercial
gun market through purchase from s licensed dealer is designated either as a long gun or a hand
gun on the ATF-devised registration form (4473) filled out at the point of transfer; the monthly
numbers for each type have been collected and published by the FBI since 1999. 9 At the point of
manufacture, when gun makers have to report the number of guns they produce upon which the
payment of excise tax is based, they are required to break out the types of guns that leave their
factories based on whether the gun is a pistol, revolver, rifle, shotgun, or semi-automatic
weapon, as well as caliber.10 But this information, which is aggregated and published yearly by
the ATF, only speaks to the manufacturing of guns, not their ultimate movement into the civilian
gun stock.11
The result is that we not only rely on estimates for how many guns are owned by
Americans, we rely even more on estimates for how this gun stocks breaks down in qualitative
terms. The most recent study claims that Americans today have an arsenal which is composed of
the following gun types and percentages: pistols (26%), revolvers (12%), ‘other’ handguns (4%),
rifles (33%), shotguns (20%) and ‘other’ guns (4%.)12 The overall percentage of hand guns
versus long guns is not substantially changed from a previous survey conducted in 2004. 13
These survey estimates, however, tell us next to nothing about whether our current
regulatory system is doing what it is supposed to do; namely, keeping guns from falling into the
hands of people who will use them in inappropriate (self-harm) or illegal (assaultive) ways. In
the case of self-harm, i.e., suicide, we have fairly reliable data on how many people intentionally
kill themselves each year with guns, but we do not know to any specific degree what kinds of
guns they use. As for gun assaults, we do not possess any reliable, published data on either what
type of guns are used or how these guns may or may not be in unlawful hands. Even the degree
to which gun violence defined as a criminal act is committed by individuals who are not lawfully
This paper breaks new ground in three respects. First, it is based on an analysis of guns
used for intentional assaults, not just by type of weapon (hand guns versus long guns which is the
only defining category used by the FBI in their Uniform Crime Reports) but by caliber,
manufacturer, gun model and age of gun. This analysis allows us, for the first time, to understand
not just the relationship between guns and violence per se, but between different types of guns
and gun violence. Second, the paper also analyzes the types of guns utilized for suicide, again
allowing us to understand this behavioral phenomenon from the perspective of the type of
instrument utilized, rather than just the behavior and mental status of the victim him/her self.
Third, we also present details on the types, caliber and brand-name of stolen guns, again, data
which has never previously been available within the public domain. What we learn from a
specific analysis and comparison between these three ‘crime gun’ categories will then lead us to
a more realistic view of the degree to which the current regulatory system can meet its intended
goals.
The data for this study comes from police departments which, in certain jurisdictions, are
required to collect, store and inventory all firearms received in the course of investigations for
both criminal and self-injury events. It does not include guns that are received for reasons other
than criminal investigations for injury (or threatened injury) such as guns that are abandoned or
‘found,’ or guns which are simply turned in by individuals who no longer feel the need to keep a
gun in their homes. Many of these guns have been traced through the ATF National Tracing
Center, but there is no requirement that any law-enforcement agency, other than federal agencies,
are required to conduct a trace on any gun received. Some police departments conduct such
The jurisdictions whose data has been analyzed are: Chicago IL, Baltimore MD, Denver
CO, Sacramento CA and Chandler AZ. To maintain consistency, these five jurisdictions were
examined for all guns collected and inventoried in 2014. The total guns studied for purposes of
this article was 9,356, of which 8,714 were for legal violations involving guns, and another 642
were for guns recovered following suicides. Unfortunately, this disparity between crime guns
and suicide guns does not reflect the degree to which two-thirds of fatal gun violence is driven
by self-inflicted, fatal injuries. On the other hand, our overall data is not just based on gun
deaths, but aggregates all investigations involving the criminal use of a gun. Since such events
(aggravated assaults, threats, armed robbery) easily exceed more than 200,000 on an annual
basis, constructing an analysis of gun violence based on criminal gun use should show an
overwhelming proportion of guns being used in interpersonal crime. The latest report from the
National Crime Victimization Survey covering 2016, for example, shows that more than 480,000
3. Suicide Guns.
Insofar as our data sample for gun suicides is much more limited than the data we have
analyzed for gun crimes, we begin first with a discussion about gun suicide. This section is based
on suicide reports in 2014 from Denver, Baltimore, San Diego, Austin and Charlotte, for a total
of 277 cases.15 Of the other suicides contained in our overall data collection, the specific
information on types of guns was not exact or complete enough to be used for understanding the
In these five cities, we were to identify specific guns in 233 cases. We divided the guns into
four categories: new hand guns, old hand guns, new long guns and old long guns. We classified
a hand gun as new if it was sold on the commercial market any time after 1980; an old hand gun
was any hand gun that was not sold on the commercial market after 1980, although such guns
could be purchased as used guns from a licensed dealer. We followed the same designations in
In 2014, of the five states whose data was used for this part of the study, only Maryland
required registration of private hand gun transfers beyond the initial sale by dealer to consumer.
This requirement did not apply to long gun transfers and there was not any registration
requirement for hand gun transfers in the other four states. The Maryland registration
requirement may explain why only Baltimore had a 50-50 division between hand guns and long
guns used for suicide. Totals for all five cities are presented here:
OH NH OL NL
Austin 13 4 2
Baltimore 10 18 20 7
Charlotte 25 35 28 1
Denver 3 19 8 1
San Diego 4 26 7 2
42 111 67 13 233
Note that of the 233 suicide guns, two-thirds (n=153) were hand guns. Of these guns, nearly
three-quarters were ‘new’ hand guns; i.e., they might have been purchased from a dealer either
new or used. The 42 ‘old’ handguns would probably have been transferred privately or were
purchased prior to the registration requirements of GCA68. As for the long guns, they were
overwhelmingly pre-1980 models, and of the 80 total long guns, most were either military
surplus weapons or obsolescent shotguns. Of the new long guns, we could only identify 3 guns
In terms of understanding access to weapons by individuals who use guns to commit suicide,
this data has certain limitations. Most important, it covers guns recovered by police agencies
operating in major, urban centers, whereas suicide is much more a function of small towns and
rural populations. In those latter jurisdictions, we would expect to find a much larger proportion
of long guns as being connected to suicide events, particularly in western and mountain states
like Montana and Idaho whose suicide rates are far above national levels.
On the other hand, this data reveals and substantiates the idea that placing greater restrictions
gun suicide. The fact that only the city of Baltimore had an almost equal use of hand guns and
long guns may support this hypothesis, but it doesn’t necessarily mean that the suicide rate in
Baltimore was less than it would have been if city residents had unlicensed access to hand guns
4. Crime Guns.
The data on gun crimes comes from two sources. One source is a listing of 846,353 guns
inventoried by 1,054 law enforcement agencies in 36 states and Washington, D.C. between 2010
and 2016. Suicide data analyzed in Section 3 (above) comes from the same dataset. The other
source, again published in The Trace, is a detailed dataset covering all ‘crime guns’ picked up by
the Chicago P.D. over the course of 2014.16 Of the total numbers, more than 90% were guns
picked up by agencies investigating a crime to which the particular weapon was connected, or in
police parlance, ‘recovered.’ The remaining guns were either reported as stolen or came into
agency possession for some administrative reason not directly connected to a specific crime.
The data analyzed for this section of the article only covered ‘recovered’ guns; i.e., guns used
in specific crimes; we will analyze stolen guns in the following section We did not break down
gun types by crime categories because those listing were created at the time the gun was seized
and could not reflect the ultimate disposition of any particular criminal case. Nevertheless,
virtually every gun recovered by an agency during a criminal investigation was connected to a
serious crime. The important exception in this respect were guns listed in inventories as being
seized for a firearm ‘violation,’ i.e., illegal possession of a gun which may or may not have been
connected to another, specific crime. Such guns represented xx% of all inventoried weapons.
The crime guns that we specifically identified as to name, date of manufacture, type, model
and caliber, were the 7,926 recovered in Chicago, Baltimore and St. Louis over the course of
2014. In that year, these three cities represented and still represent major concentrations of gun
violence, with Chicago registering a per-100K murder rate of 15, Baltimore at 34 and St. Louis
Not surprising, when we compute the number of crime guns relative to the overall size of
each city’s population, an interesting correlation appears. The number of crime guns recovered in
Chicago 3,160
Baltimore 1,876
St. Louis 1,539.
But when we compute crime guns per-100,000k population, the numbers look like this:
St. Louis 50
Baltimore 34
Chicago 15
In other words, as the murder rate in a particular jurisdiction goes up, so does the rate of
crime guns recovered by the police. Which means that the more guns floating around, the more
guns become the method of choice for ending someone else’s life. This is a very important
finding because it introduces a significant nuance into the usual argument about more guns
equals more criminal violence. The nuance is as follows – more guns equals more gun violence
in a particular jurisdiction depending on how many guns can be accessed from place to place.
Of the total 7,926 crime guns recovered in these three cities, 5606 (70%) were center-fire
handguns, of which 3,408 were guns commonly found in retail gun shops, the brand breakdown
Of the remaining 2,300 handguns, 1,376 would have to be classified as old, used junk, with
the traditional, racially-tinged ‘Saturday-night Special’ sobriquet applying to most of these guns.
The junk guns included nearly 700 RG-Rohm guns which used to be imported from Germany
but stopped coming into the U.S. following the import prohibitions imposed by GCA68. The
continued distribution of these ‘junk’ guns then shifted to domestic manufacture, a story partially
explained by Wintemute in a study covering some of these companies which operated (and still
operate) in California.18
The fact that nearly one-quarter of all recovered crime handguns were manufactured either
by foreign or domestic gun makers, most of whom ceased operations more than twenty if not
thirty years ago, should alert us to a second, very important finding that emerges from the data
under review. While the lack of serial numbers for all recovered guns from Chicago (most of the
St. Louis and Baltimore guns did include serial numbers) makes it impossible to determine, with
any degree of accuracy, the average age of every crime gun, it is possible to estimate the date of
manufacture through an analysis of the type and model name of each gun.
The issue of date of manufacture is of primary concern to gun regulators and gun-control
advocates because a significant amount of gun regulation activity involves tracing of crime guns
and the development of what regulators call ‘time to crime,’ or TTC. The TTC numeric is used
by law enforcement to identify both individuals and gun dealers which show up more than an
average amount as the sources of guns whose TTC is brief. It is assumed that any gun connected
to a crime that was initially brought into the commercial market in less than to years may be a
gun whose initial transfer was a ‘straw’ sale; i.e., it was purchased for the express purpose of
According to the most recent TTC report issued by the ATF, the average TTC time for
141,476 guns traced in 2016 was roughly ten years, for our three states it was twelve years for
Maryland, nine years for Illinois and six years for Missouri. 19 There are two problems with this
data, of which the ATF appears to be unaware of both. First, the TTC number is calculated by
determining the amount of time which has passed from the first date a gun was sold by a dealer
to the date that the trace request was made. But for at least 40% of all guns, which is the average
percentage of used guns found in retail gun shops, the first transfer date cannot be used to figure
out when a particular gun began the movement from legal to illegal hands. This time-span can
only be figured out if the specific trace started on the date that the gun was last sold in a retail
shop. In other words, the overall TTC average is probably longer than it should be for at least 40
On the other hand, what comes out from the Baltimore-Chicago-St. Louis crime gun list is
the fact that it is impossible to determine the initial date of manufacture or sale of at least one-
quarter of all the guns connected to criminal investigations in those jurisdictions; one would
probably find this percentage to be constant in every police jurisdiction within the United States.
This is because the same cheap, older guns not only appear with the same degree of frequency in
all three cities whose crime guns we analyzed in detail but are also found in virtually every other
Some of the older guns which appeared in police inventories have not been manufactured for
a century or more. For example, Mossberg manufactured a 4-shot derringer called the Brownie,
from 1920 to 1932. More than 40 of these guns turned up in the crime gun listings in Baltimore,
Chicago and St. Louis (most of them in Chicago) but we found them in other cities as well. Both
Colt and Browning manufactured 25-caliber ‘baby’ pistols, the former first appearing in 1908,
the latter in 1906. The Colt guns were serialized right from the first guns, but the Browning guns
didn’t get serialized in any organized fashion until 1954. Both of these guns appeared in all of
While centerfire pistols were 70% of all crime guns in the Baltimore-Chicago-St. Louis lists
(and were roughly a similar proportion in other jurisdictions that were sampled) the remaining
30% were divided roughly in half between 22-caliber guns and long guns; i.e., rifles and
shotguns. In the rifle category, less than 2% of all recovered guns were assault rifles (AR and
AK) and the only hunting rifles were a handful (6) of lever-action guns in 30-30 caliber. There
were virtually no bolt-action or semi-automatic hunting rifles recovered in any of those cities,
and even in more rural Western jurisdictions such as cities in Colorado, hunting and sporting
In addition to understanding what types of weapons account for gun violence, we also need
to look at the ammunition which accounts for deaths and injuries from guns. Somewhat
arbitrarily, we have decided that crime guns which cause gun violence are those weapons
chambered for any center-fire round. The reason for this is because the rim-fire cartridge family
(22 short and 22 long rifle) do not show up in great numbers in counts for gun injuries, nor are
they prevalent in the types of ammunition carried in the crime guns picked up by Chicago P.D.
The most popular street caliber was 9mm, which comprised 25% (n=1,037) of the 4,113 calibers
identified in the 2014 list of crime guns. The next most common caliber was 38 Special with
10.7% (n=443) 10of all identified calibers, followed by 40 S & W, which was 10.6% (n=438) of
all crime guns. A total breakdown of all handgun calibers looks like this:
Of the remaining 655 identified calibers, a few additional handgun calibers show up (44
Magnum, 45 Long Colt) of which 40% were shotgun calibers and 60% were 22 calibers, but half
of the latter came from guns that were either no longer manufacture or more than thirty years old.
Data from the National Violent Death Survey confirms the degree to which gun violence is
basically a function of centerfire ammunition.20 For 2015, the percentages of gun deaths from
9mm 31.1%
22 cal. 8.7
357 3.6
38 6.1
40 S&W 16.8
45 10.4
25 – 32 8.2
This adds up to 80% of all bullet wounds, the total percentages from assault-rifle categories
was 2.2% and hunting calibers represented exactly two bullets out of 2,118 identified rounds, the
latter also possibly assault rifle rounds.21 In other words, ammunition calibers associated with
crime guns validates that most gun violence is committed with guns that fire centerfire, handgun
ammunition.
5. Stolen Guns
For this section of the report, we rely on the 651 stolen guns reported to the St. Louis
Police Department in 2014. It should be noted that while we have estimates that somewhere
between 200,000 and 400,000 guns are stolen from privately-held locations each year, these
numbers are based on public opinion surveys, they do not represent any hard data at all. 22 This is
because: a) very few jurisdictions require that civilians report stolen guns to the police; and b) we
assume that many guns are stolen from individuals who are not legally allowed access to guns in
The latter issue about stolen gun data, namely, the illegal ownership of guns, imposes
severe limitations on making any definitive judgements about the quantity or types of stolen
guns. There is probably not a single law enforcement agency in the entire United States that will
enter information on a stolen gun without receiving and validating the name and address of the
individual who calls in the report. Which means that probably a majority of guns stolen every
year remain totally outside the purview of the law, nor will the estimates on gun thefts, all of
which come from public surveys, pick up thefts of ‘illegal’ guns as well.
That being said, an analysis of the guns reported stolen in St. Louis in 2014, still gives us
some important information about the types of guns which end up contributing to gun violence
and violent crime. Of the 466 handguns whose calibers were identified in the report, the
Note that the percentages of guns stolen on a caliber basis conforms to calibers identified
as being the most frequently used in gun crimes: 9mm, 38 and 40 S&W. This breakdown of
stolen gun calibers compared to gun-crime calibers is also similar to what was found for both
assault rifles and shotguns, because neither category registered more than five percent of all
stolen guns, with assault rifles accounting for 4% and shotguns accounting for 5% of stolen
calibers, respectively. Furthermore, of all the guns reported stolen, only 12% were long guns of
any type, the other 88% were classified as handguns, with pistols accounting for more than 80%
As for older guns, only 17 specific junkers were found on this list, of which half were
made by Jimenez Arms, the remainder being Davis, Bryco and the ubiquitous RG. But we
suspect that the percentage of the older guns would have been significantly higher were it not for
the fact that of the total number of 561 stolen guns, 153 were reported without any brand
identification at all. More than 90% of the unidentified guns by brand name were classified as
handguns, leading us to believe that many of them might have been older guns as well. But this
Not surprisingly, of the handguns which could be identified by manufacturer, the most
common brands were Smith & Wesson, Taurus, Ruger and Glock, together accounting for 52%
of all handguns on the stolen list. Again, this data corresponds to what we found for the
6. Summary
Putting all the data together, we arrive at the following conclusions about guns that are
used in every type of gun violence – intentional injury against oneself or someone else. This
paper does not contain any data on unintentional gun injury, first because gun accidents are not
considered to be any kind of violence; second, because the number of unintentional fatal and
non-fatal shootings (approximately 17,000 each year) can hardly be considered a serious health
issue, given the existence of more than 300 million guns in private hands. 23
If the data presented above tells us anything about the connection between guns and gun
violence, it is that this entire discussion should be revised to focus not on the number of guns in
the civilian arsenal, now estimated between 265 million and 395 million, but should focus on the
number of handguns owned legally and illegally by Americans. Putting together the three
categories that we have examined – suicide, crime, theft – centerfire handguns account for 70
percent of guns in all three categories, with the percentage of centerfire handguns in the crime
we actually experience is a gun violence problem because we own centerfire handguns, which is
a much more exceptional situation in comparative, global terms, than the overall size of the
civilian arsenal itself. If we could gather reliable data on handgun ownership within other
advanced countries, we would probably find that the per capita comparisons currently made
between the size of the U.S. civilian arsenal to the size arsenals in other countries would be
The reason that we need to focus more directly on handgun regulation, as opposed to
regulating guns in a generic sense, is that the only way we can craft effective and meaningful
regulatory policies is if our regulatory strategies aim (pardon the pun) not at solving the problem
of gun violence in a generic sense, but the violence caused by the misuse of handguns. Because
we do not focus on regulating the guns which cause gun violence, we are spending time, energy,
physical and financial resources on regulatory procedures which cannot reduce gun violence at
all.
The five rifles above (top to bottom) are the Remington 700, the Ruger 77, the
Winchester 70, the Marlin 1894 and the Savage 11. Together, these guns represent at least 40
million of the total number of guns owned by Americans today, which is somewhere between
10% and 15% of the entire pile of guns that exists throughout the United States. 26 How many of
these five models were found in a word-search of the 846,353 guns inventoried by 1,054
It may be the case that a specific word search did not pick up every, single instance in
which on of those guns was listed as a stolen, criminal or suicide gun by the police. Cops are
human, they make mistakes, and as our analysis on St. Louis stolen guns makes clear, perhaps as
many as 20% of all the stolen guns listed in the national inventory could not be identified as to
which company manufactured the gun. But if any of those models had constituted even a fraction
of the guns analyzed for this effort, the examination we conducted on thousands of the guns
In addition to guns that virtually never contribute to gun violence for reasons having to
do with model and types, there is also the issue of using regulatory procedures which cannot
begin to yield any information which would aid law enforcement agencies in understanding how
a particular gun ended up in the ‘wrong’ hands. Our detailed analysis of crime guns in Baltimore,
Chicago and St. Louis clearly indicates that at least one-third of all guns connected to criminal
violence cannot be traced beyond the individual who used the gun to commit a particular crime.
Either the guns were manufactured by gun makers who went out of business prior to GCA68, or
they were imported before GCA68, or they were manufactured before the law required
serialization of all small arms. In fact, even the ATF admits that although they processed 408,000
traces in 2017, they were only able to correctly identify the type of gun and its first place of sale
in less than 325,000 trace requests, a failure rate of 20 percent. 26 Recall that we described ‘junk’
guns as representing perhaps as many as one-third of all crime guns recovered each year. The
ATF trace data confirms the degree to which old, untraceable handguns circulate in large
Much of the reluctance on the part of gun-control researchers and advocates to avoid
pursuing strategies that would restrict handgun ownership is based on the idea that the 2008
Heller decision giving Constitutional protection to guns commonly found in the home takes a
handgun ban, de facto or de jure, off the table.27 Thus, avoiding the policy implications of what
this paper demonstrates not about gun violence, but about handgun violence, continues to be an
handguns, a careful reading of Scalia’s majority opinion in Heller, along with an understanding
of handgun function and design renders the 2nd Amendment argument not just inappropriate but
completely wrong. Scalia differentiates between weapons that are ‘commonly owned,’ as
opposed to ‘unusual’ weapons; i.e., military weapons or what we usually refer to as weapons of
war. And basically what he is arguing is that someone who keeps a Glock or a Sig in their house
for personal defense wouldn’t take this weapon with them if they were called up to serve.
What Scalia didn’t understand, nor did anyone else on the Court have a clue, was the fact
that virtually all the handguns which are used to commit gun violence happen to be designed for
military use and are carried by troops in every battle zone today. The current pistol used by the
U.S. Army is a Sig whose design is incorporated into every handgun manufactured in the
SigArms factory and sold to military and civilians worldwide. Gaston Glock developed his first
gun (and every subsequent model is designed in exactly the same way) in response to an
Austrian Army RFP. That we have decided, for purely political reasons, to allow gun makers and
gun importers to sell weapons of war to the general public doesn’t change the fact that not just
the assault rifles, but virtually all the handguns now ‘in common use’ (to quote Scalia) do not
This is the reason why any discussion about gun violence is at best incomplete, at worst
simply wrong, if the discussion does not contain clear references and understanding not just to
the perpetrators and victims of gun violence, but to the types of guns used to commit this kind of
violence every day. If it were the case that crime guns were a random assortment of all the guns
held by civilians, then we would have no choice but to continue building and refurbishing a
regulatory system that attempts to impose the same restrictions on every type of gun and every
individual who owns any type of gun. But this paper clearly demonstrates that the individuals
who commit gun violence are making very conscious and deliberate choices about what kinds of
guns they should use, and what types of guns won’t meet their requirements for the violence they
choose to impose. If our aim is to reduce gun violence, knowing the guns is an important aspect
NOTES
1
https://www.worldlifeexpectancy.com/cause-of-death/violence/by-country/.
2
http://www.unodc.org/unodc/en/data-and-analysis/crime-and-criminal-justice.html.
3
M. Weisser, “Is Gun Violence an Example of American Exceptionalism?”
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3158214.
4
E. Grinshteyn & D. Hemenway, “Violent Death Rates: The U.S. Compared with Other
High-Income Countries, 2010,” The American Journal of Medicine, 129 (2016) 266-273.
5
A. Kellerman, F. Rivara, et. al., “Gun Ownership as a Risk Factor for Homicide in the
Home,” New England Journal of Medicine, 329 (1993), 1084-1091; Kellerman & Rivara, et. al.,
“Suicide in the Home in Relation to Gun Ownership,” NEJM, 327 (1992), 467-472.
6
Cf., website content from Brady, Everytown, etc.
7
The qualifying information required by a prospective gun owner prior to purchasing a
legal small arm can be found in the ATF 4473 background-check form, in particular Questions
11a – 11i. https://www.atf.gov/firearms/docs/4473-part-1-firearms-transaction-record-over-
counter-atf-form-53009/download. This qualifying information is the same for all legal firearms
transferred from a licensed dealer to a gun purchaser.
8
https://www.atf.gov/rules-and-regulations/national-firearms-act.
9
https://www.fbi.gov/file-repository/nics_firearm_checks_-
_month_year_by_state_type.pdf/view.
10
The annual manufacturing report can be accessed here:
https://www.atf.gov/about/docs/undefined/afmer2016webreport508pdf/download.
11
The annual firearms commerce report can be accessed here:
https://www.atf.gov/news/pr/atf-releases-2017-report-firearms-commerce-us. It does not give
any data on the locations to which imported or manufactured firearms are shipped for consumer
resale.
12
D. Azrael, et. al., “The Stock and Flow of U.S. Firearms: Results from the 2015
National Firearms Survey,” The Russell Sage Foundation Journal of the Social Sciences, 3, 5
(October, 2017). https://www.rsfjournal.org/doi/full/10.7758/RSF.2017.3.5.02.
13
Hepburn, L., et. al., “The U.S. Gun Stock: Results from the 2004 National Firearms
Survey,” Injury Prevention, 13, 1 (2007), 15 – 19. DOI: 10.1136/ip.2006.013607.
14
Rachel E. Morgan, Ph.D., and Grace Kena, “Criminal Victimization, 2016,” U.S.
Department of Justice, Office of Justice Programs, (December, 2017), p. 5.
15
The information on suicides comes from the dataset that can be downloaded from The
Trace (https://www.thetrace.org/missing-pieces-data/) which was collected and analyzed by
Daniel Nass and Bryan Freskos. We discuss the dataset in the Crime Guns section.
16
https://www.thetrace.org/2016/01/chicago-crime-guns-raw-data/.
17
National murder rate from CDC-WISQARS. City rates from local media reports on
murder numbers announced by law enforcement, then factored against U.S. Census population
estimates.
18
G. Wintemute, Ring of Fire - The handgun makers of Southern California : a report
from the Violence Prevention Research Program (Santa Barbara, 1994.)
19
https://www.atf.gov/news/pr/atf-releases-2017-us-firearms-trace-data-report.
20
We wish to thank Catherine Barber of the Harvard University School of Public Health
Injury Control Research Center for this data.
21
Frank Zimring, “The Medium is the Message: Firearm Caliber as a Determinant of
Death from Assault,” I. J. Legal Studies, 97 (1972) 97 – 123, who found that half of the 1970
fatal shootings in Chicago were committed with gun chambered for either 22 or 25-caliber
ammunition.
22
https://www.americanprogress.org/issues/gunscrime/reports/2017/07/25/436533/stolen-
guns-america/.
23
In 2016, according to the CDC, there were 58,854 unintentional deaths, the largest
category for all deaths in that age cohort. Of these deaths, 234 were from firearm injuries; i.e.,
.003%.