Hubris in the North:
The Canadian Firearms Registry[1]
Gary A. Mauser[2]
Abstract
This paper is a preliminary effort
to evaluate the effects of the 1998 firearm registry on public safety. The Federal Government saw the firearm
registry as crucial for reducing criminal violence and for saving lives. The government’s approach to public safety
relied upon an unscientific analysis of firearms and violence that, as a result
of its acceptance of public health research, greatly exaggerated the dangers of
lawful firearm ownership. In this paper I criticize public health research
methods as being moralistic and pseudoscientific. The Federal Government’s
approach to public safety is compared with a provincial program that is more
consultative.
The results show that since the firearm registry was implemented, the number of firearm owners has significantly declined, as well as the number of firearm crimes and the number of firearms-related deaths. Nevertheless, public safety cannot be said to have improved because overall criminal violence and suicide rates remain stubbornly stable. The violent crime rate has declined by only 4% since the registry was implemented, but the homicide rate has actually increased by more than 3%. Perhaps the most striking change is that gang-related homicides and homicides involving handguns have increased substantially. Overall suicide rates have declined by just 2% since the registry began. Despite a drop in suicides involving firearms, hangings have increased nearly cancelling out the drop in firearm suicides. No persuasive link can be found between the firearm registry and any of these small changes. In comparison, the provincial hunter-safety program has more modest goals, i.e., to reduce hunting and firearm accidents, but limited evidence suggests that it is effective in actually saving lives.
In conclusion, no convincing empirical evidence can be found that the firearm program has improved public safety. For the first seven years, the firearms registry had virtually unlimited annual budgets, but violent crime and suicide rates remain virtually unchanged. Despite costing an estimated C$ 2 billion to date, the firearms registry remains notably incomplete and has an error rate that remains high. As a result of its many failures, particularly its failure to reduce gang violence, the firearms registry has failed to win the trust of either the public or the police.
This paper is an invited keynote presentation to: In the Right Hands - an international firearm safety
seminar, to be held at Christchurch,
New Zealand, 21-23 February 2006. This seminar is co-hosted by the New
Zealand Police, New Zealand Mountain Safety Council, and New Zealand Council of
Licensed Firearm Owners.
Of all tyrannies, a tyranny sincerely exercised for the good of its victims may be the most oppressive. It would be better to live under robber barons than under omnipotent moral busybodies. The robber baron's cruelty may sometimes sleep, his cupidity may at some point be satiated; but those who torment us for our own good will torment us without end for they do so with the approval of their own conscience.
– C. S. Lewis,
C. S. Lewis, God in the Dock (Grand Rapids: W.B. Eerdmans, 2002), p. 292.
The 1995 firearms legislation was a bold attempt to improve public safety in Canada. In the words of Allan Rock, the Justice Minister responsible for shepherding the legislation through Parliament, the goals were sweeping indeed: “the primary objective of regulating firearms should be to ensure that Canada remains a peaceful and civilized country”.[3] In the same address to Parliament, the Justice Minister assured Canadians that firearm registration would save lives by reducing criminal violence, domestic violence, suicide, and firearm accidents. The legislation had three interlocking components: increased penalties for criminal misuse of firearms, tighter border controls, and universal firearms registration. The firearms registry was viewed as the lynchpin of the government’s plan.[4] Ten years later it is now possible to empirically evaluate its effects on public safety. It is particularly important to know whether the firearm registry has lived up to the promises made at the time of its introduction.[5] If legislation is to be more than symbolic then politicians must be held to account for their promises. Legislation must be empirically assessed in order for legislators as well as citizens to learn what works and what does not.
This evaluation is necessarily preliminary, as there are several serious impediments. First, evaluating any legislative effort is an exceedingly difficult challenge, especially legislation as complex as this package. Second, the 1995 act has been phased in over time. Third, it is difficult to obtain the necessary information with which to evaluate the legislation as the government refuses to release crucial data. Nevertheless, it is important that an attempt be made to assess whether this expensive experiment is worth the cost. To simplify the task, I will focus on the firearm registry and leave to others further evaluation of this legislation. I will attempt to draw inferences from the best, albeit flawed, information available. I will inform the reader when evidence is weak or inconsistent, but I will not ignore information that refuses to fit comfortably with other information or with my tentative conclusions.
The government’s approach to public safety relied upon an unscientific
analysis of firearms and violence that exaggerated the dangers of firearm
ownership. Unfortunately, the government’s analysis was based upon
public health research that had ignored basic scientific principles in favour
of advocacy. Advocacy research is unscientific because it uses research
methodology to prove conclusions rather than to test hypotheses. Public health
researchers draw conclusions that are not supported by their research
methodology and compound their errors by recommending legislative solutions
that fall outside the boundaries of their research. Such studies are not
properly scientific but sagecraft, i.e., the use of the scientific trappings of
research in order to prove rather than to test claims.[6]
The reliance upon scientific trappings
to consecrate their policy proposals gives the public health approach to public
safety a heavy-handed moralistic style that may be contrasted with more
consultative approaches, such as “community oriented policing” in North America
or the “crime reduction”
approach as exemplified in the UK Crime and Disorder Act and in Prime Minister
Tony Blair's drive to engage local authorities to become involved in crime
reduction (Home Office, 1998, 2005).
As with most controversies, opponents and supporters seem to compete with each other in emotional hyperbole and in cherry-picking facts to support their claims. It is time for a rational assessment. The key questions to ask are whether the 1995 firearms legislation has met the goal originally set for it, that of improving public safety, and whether the benefits exceed the costs. These are of course difficult questions. In this paper, I will briefly discuss public safety, present the history of firearm legislation in Canada, assess the problems faced by the Federal Government in designing and implementing the information system, make a preliminary evaluation of the effectiveness of the firearm registry, and, in passing, evaluate provincial efforts at firearm safety.
Before we can begin our evaluation, we must identify the original goals of this legislation.
Access to firearms and improving public safety
“Success has many fathers, but failure is an orphan”.[7]
Given the gift of hindsight, few currently admit to the hopes they initially held for this legislation back in 1995. Nevertheless, it is important to evaluate the legislation in light of its original goals, which were to improve public safety. Before we go too much further we will have to assess the meaning of “public safety”. Unfortunately for efforts at evaluating the legislation, there has been some drift in the meaning of this term.
At the time the legislation was introduced, Allan Rock, then Minister of Justice, emphasized that its primary goal was to improve public safety. In his testimony before Parliament and in his public speeches, he stressed that firearm registration would “enable us to achieve the objectives of a safe and peaceful society, a more effective response to the criminal misuse of firearms and enhanced public safety”. Underlying these claims is the assumption that reducing firearm violence would “deal with the scourge of domestic violence”, cause overall criminal violence to decline, as well as reduce impulsive acts such as suicide and accidents.[8]
For Allan Rock, the key to improving public safety lay in controlling the availability of firearms, and he believed firearm registration was the way to achieve this. According to this view, registration would control the availability of firearms, which would reduce firearm misuse, which in turn would reduce criminal violence, not just gun violence, and in addition total suicides, as well as domestic abuse. By focusing upon the ordinary firearm owner, Rock’s comprehensive approach to public safety rests squarely upon the “weapon instrumentality hypothesis” or the “weapons effect” (Zimring, 1968). According to this hypothesis, firearms are inherently violence enhancing, and their possession acts only to increase the likelihood of the victim’s injury or death (Newton and Zimring, 1969). Two distinct empirical claims are being made here: the first is that firearms are inherently more “dangerous” than other weapons, i.e., an analysis of victims of violence would show that gunshot wounds are more likely to lead to “serious injury or death” than are injuries from other weapons.[9] The second is that a sizable proportion of aggressors have “ambiguous motives,” so that the availability of a firearm can transform a relatively minor confrontation into one where the victim is seriously injured or dead.[10] Based on these premises, some researchers advocate restrictions on firearms that would decrease their general availability in order to reduce the number of victims who receive serious injury or death (Cook, 1991; Gabor, 1994).
The Federal Government also argued that restricting firearms availability would reduce the human cost of impulsive acts such as suicide and accidents.[11] This is particularly important for suicide, as it makes up such a large proportion of firearms deaths. Rock found support for his claims in a Justice Department study that purported to show that firearms availability is associated with total suicide rates, not just firearm suicide (Gabor, 1995, p203). However, the assertion that firearms availability is linked with total suicide rates has not been corroborated. Gary Kleck, one of the most widely respected researchers in criminology, has shown that Gabor’s claim was based on a misrepresentation of the research literature on guns and suicide. In his review, Gabor: (a) padded the list of supportive studies by including studies that were irrelevant, (b) mischaracterized studies as providing support for his thesis that were in fact unsupportive, and (c) omitted studies from the review that were not supportive of the conclusions he desired (Kleck, 1997, p. 49 – 53). Gabor’s fallacious review has proved to be quite influential outside of Canada: it has been cited as important in Lord Cullen’s inquiry into the Dunblane shootings (Mayhew, 1996) and in the Australian government’s response to the Tasmanian shootings (Chapman, 1998).
The general goal of improving
public safety has been repeated by subsequent Canadian Justice Ministers, most
recently by Anne McLellan (Naumetz, 2004).
This goal is reiterated in the recently released 2003 report of the Canadian
Firearms Centre (CFC) (Baker, 2004). However, the goal has become much narrower
over time: “The objective of the [Canadian Firearms] Program is to keep
firearms from those who should not have them while encouraging safe and
responsible firearm use by legitimate firearms owners. Through firearms control
measures, such as screening and licensing of gun owners, registration and
tracking of firearms, and safety training, the Program aims to prevent firearm
crime while reducing the number of firearms-related deaths, accidents and
threats.”
Note how the goal has evolved since 1995. The original objective for
the firearms legislation was much broader, to reduce criminal violence and to
save lives; now the goal focuses exclusively on “legitimate firearm owners” and
even specifies a single criterion, “firearm deaths”, to evaluate their goal. To
be sure, this is reasonable in one sense, because the Firearms Program is only
charged with controlling firearms and is not responsible for enforcing criminal
sanctions, nor for guarding the country’s borders. However, this narrower goal
is only justifiable to the extent that the firearms program fits into the
overall plan. Reducing gun violence is important in order to reduce overall
criminal violence; reducing “gun deaths” is important in saving lives.
Originally, the gun registry was seen as the key to a general improvement in
public safety, not as an end in itself. Whether reducing “gun deaths” will
actually result in saving lives is an empirical question, as is whether
reducing gun violence will reduce criminal violence overall. By introducing the
criterion of “gun deaths” the scope of the program has now been defined too
narrowly. The original, more general goal has been lost because there is no
longer any way to measure possible increases (or decreases) in public benefit.
This change exemplifies “mission creep”. Mission creep is seen in many
organizations. When organizations discover their original goals are too
ambitious, some react by simply changing the goals. In 1995, the goals were the
reduction of total criminal violence, domestic violence, and suicide; firearms
controls were merely the means of effecting this reduction. The goals were not
just reducing firearms violence, nor just keeping lives from being lost through
the misuse of firearms. The government wanted to reduce criminal violence
overall and to save lives from any kind of violence. The reduction of firearm
violence was seen as the key to reducing overall criminal violence, not as an
end in itself. The new goal ignores the big picture to focus myopically upon
guns.
Used in this way, “gun deaths” are a red herring. Given the many ways
available to kill, a drop in “gun deaths” does not necessarily imply that any
lives have been saved in total. If a gun is unavailable, a murderer can find a
sharp knife or heavy object that may be used as a bludgeon in virtually any
home. For anyone tempted by suicide, it is not difficult to find a length of
rope or a tall building. The question of substitution (or displacement) remains
open. There is no compelling evidence that general gun control laws can reduce
overall murder or suicide rates.[12]
In Canada as in Australia, the current drop in ‘gun deaths’ has not been
accompanied by a fall in suicide rates (Australian Bureau of Statistics,
2001, and Ozanne-Smith, et al, 2004).
Nor has the homicide rate fallen in conjunction with the decline in gun
homicides. It is difficult to argue that the declining trend in homicide rates
has been caused by the 1995 firearms legislation. Focusing upon ‘gun deaths’
diverts attention away from the original objective, which was an overall
improvement in public safety. The Minister has deliberately set a goal for the
firearms registry that is easily attainable and has ignored the more difficult
challenge of reducing suicide and homicide rates, that is, of actually saving lives
overall.
To understand the current legislation, we must view it in context with previous firearms laws. The next section briefly reviews Canada’s gun laws.
History of firearms legislation in Canada
Canada has long had strict firearm legislation. Criminal law is a federal responsibility, and it includes the misuse of firearms. The 1995 Firearms Act is not the first time firearms were included in the criminal code, nor the first time guns were required to be registered in Canada. Legal restrictions have been imposed on handguns since the 1890s and registration has been mandatory since 1934.[13] During World War II, there was a temporary requirement to register long-guns (rifles and shotguns) (Smithies, 2003). Prior to the current firearms legislation, passed in 1995, the firearms law was extensively amended in 1969, again in 1977, and further amended in 1991 (Kopel, Chapter 4, 1992, Chapter 4).
Canada is a federal system, with provinces constitutionally responsible for natural resources and licensing. Hunting regulations fall under provincial jurisdiction, and since the bulk of firearms owners are hunters, provinces have been traditionally responsible for regulating the normal usage of firearms. For example, provinces are responsible for regulating alcohol, motor vehicles, and aircraft licences. Firearm safety training is included under provincial hunting legislation. Prior to the 1960s, firearm safety courses were voluntary and were offered by gun clubs and other non-profit groups. Starting with Ontario in 1960, firearm safety training was made mandatory in order to obtain a hunting licence. By 1981 all provinces (except Prince Edward Island) required applicants to pass a provincial examination on hunter safety in order to qualify for a hunting licence.
At the federal level, the basic framework for modern Canadian firearm legislation was established in 1969 when three classes of firearms (restricted, non-restricted, and prohibited) were defined. The 1969 legislation was amended in 1977 as part of a Parliamentary agreement that ended the death penalty.[14] The 1977 firearms act required a police permit for the first time in order to purchase a firearm (the Firearm Acquisition Certificate), introduced a requirement for safe storage, and banned certain types of firearms, including fully-automatic weapons.
In 1991 the government of Canada amended its firearm law in reaction to a horrific shooting at the University of Montreal (Dixon, 2003). The 1991 law banned a number of “military-style” semi-automatic rifles as well as “high-capacity” magazines, i.e., those that held more than 10 rounds. In addition, stringent new requirements were added to the process of purchasing a firearm, including a firearm safety course, a mandatory 28-day waiting period, two character references, one of which must be from the applicant’s spouse, a passport-type photograph, and a long series of personal questions.[15] In addition, specific regulations were introduced covering safe storage, handling, and transportation of firearms.
A new government came to power in 1993 determined to extensively amend the firearms laws. In their view, much tighter control of firearms was needed in order to improve public safety. The 1995 legislation made sweeping and radical changes and focused on firearm registration. Allan Rock, then the Minister of Justice, summed up the bill before Parliament:
The components of Bill C-68 [the 1995 firearms amendments] that we will
be focusing on today are as follows: firstly, strict measures to counter the
criminal use of firearms; secondly, specific penalties to punish those engaged
in the smuggling of firearms; and thirdly, broad measures to define what
constitutes the lawful use of firearms in a manner that poses no threat to
public safety.
In the case of each component, universal firearms registration is a
fundamental requirement for achieving the stated objectives.[16]
In addition to owner licensing and the registration of all rifles and shotguns, this legislation also banned more than half of all currently registered firearms (also scheduling their confiscation), and introduced a framework for detailed regulations covering all aspects of firearms in Canada.
The firearm prohibitions were introduced immediately in February 15, 1995; the increased penalties for criminal misuse of firearms became operative when the bill was proclaimed into law, on January 1, 1996, after the bill received Royal Assent in December 1995. The provisions for the registry took longer to implement. Owner licensing and long-gun registration were implemented on January 1, 1998 (Niemczak, 2004).
Starting in 1998, Canadian firearm owners were required to acquire a licence, to keep it current, and register all firearms – including long guns.[17] This paper continues to focus on evaluating the firearm registry, although this legislation includes a wide range of other draconian provisions that make sweeping changes to the legal status of firearm owners.
These
draconian new laws were brought in without a review of existing legislation or
a cost-benefit analysis.[18]
Allan Rock is even reported to have said in an unguarded moment that no one
should own firearms except the police.[19]
The government’s heavy-handed approach generated intense controversy,
particularly the proposed firearm registry. The government did not engage in
any meaningful consultation with user groups as had been done with previous
firearm legislation. Nor did the government accept amendments in Parliament.
Three of the four opposition parties (Reform, Progressive Conservatives, and
New Democrats) – despite their mutual antipathy – joined together to fight Bill
C-68.[20]
Several provincial governments (spanning the political spectrum, from NDP to
Conservative) actively opposed the legislation. Almost all provinces (including
Ontario, the largest province in Canada) backed a constitutional challenge to
the legislation. When the challenge was finally rejected by the Supreme Court
of Canada in 2000, eight of ten provinces declined to cooperate with the
Federal Government in enforcing the new law (Lindgren and Naumetz, 2003). The
government’s approach to firearm registration proved to be extremely divisive,
and as other governments have discovered with other divisive issues, such as
abortion or alcohol prohibition, good intentions alone are rarely a sound basis
for criminal law.
The government’s heavy-handed approach to firearm legislation amounted
to a moral crusade against firearms – even firearms used for purposes
beneficial to society. As I will show in the next section this heavy-handedness
exemplifies the public health approach to public safety. Public health
researchers consider firearms to be a “disease vector,” and they see themselves
as crusading against these pathogens. Unsurprisingly, when Allan Rock
introduced the firearm registry he relied heavily upon public health activists.[21]
The next section compares the public health approach to public safety with more
consultative approaches, such as community oriented policing or “crime
reduction”.
Public health and
“public safety”
Public safety is one of those phrases that can take on a wide range of meanings. The term originated with the Committee of Public Safety during the French revolution in 1793, where the phrase justified the identification and elimination of opponents as enemies of the revolution. In the past century public safety has been typically viewed as “the protection of the general population from all manners of harm, danger, injury or damage” (Wikipedia , 2005).[22] In practice, this primarily means traditional policing efforts as well as emergency services such as fire, rescue, and ambulance.[23] Public safety also includes preparations for surviving severe weather conditions, e.g., wild fires, floods, hurricanes, or snowstorms. In the 21st century it has also come to include protecting against terrorism.
Historically, crime has been seen as one of the most important continuous threats to the community. Policing is at the heart of the traditional approach to public safety. Professional policing got its start in London of the 1820s with Sir Robert Peel who stated his basic premise that “the police are the public and the public are the police” (Braiden, 1992). Sir Robert Peel saw that responsiveness to the community was as important for effective policing as maintaining professional standards. Over the past two centuries since Sir Robert Peel introduced the Bobbies in London, policing in the British Commonwealth has varied tremendously in its approach to community responsiveness, from a traditional adherence to a “law enforcement approach”, emphasizing the prevention and detection of crime and the security of life and property, moving to the more modern commitment on the part of the police to working closely with the public to attain common goals. In the past thirty years “community oriented policing”, or “crime reduction”, has become the predominant philosophy of policing (Oliver, 2001). An important example of this approach is the UK Crime and Disorder Act (Home Office, 1998). It is demonstrated by Prime Minister Tony Blair's drive to engage local authorities in a formal requirement to become involved in crime reduction as a focused goal. In this approach, the police work closely with the community not only to enforce the law, but also to engage proactively with community leaders to identify their priorities and their needs. Ideally, this approach goes beyond “community relations” to actively engage the community in working together to attain common goals in reducing crime. As exemplified by the Royal Canadian Mounted Police, it involves three main goals: taking repeat offenders off the streets, identifying crime hot spots, and coordinating with community leaders (Lepine, 2005).
An alternative approach to public safety has recently emerged led by public health professionals. The public health approach grew out of a concern about traumatic injury as a public health problem and has shifted concern more towards suicide and accidents and away from the traditional focus on criminal violence. While this broadened mandate has brought new data and new funding sources into criminology, it has had several unfortunate side effects: first, it diverts scholarly attention away from the perpetrator and focuses more on the instrument, and second, it tends to shift policing away from community consultation and more towards paternalistic prescriptions.
Perhaps the most negative consequence is the encouragement of moralistic reasoning (rather than scientific) that has accompanied such advocacy research.[24]
Public health activists identify themselves as advocates, not objective scientists. Both the Canadian Public Health Association and the Public Health Association of Australia take this position.[25] “The Public Health Association of Australia (PHAA) is a forum for the promotion of the health of the public ... The Association … actively undertakes advocacy for public health policy, development, research and training” (PHAA, 2004). Perhaps the most widely known public health advocate in Australia is Simon Chapman at the University of Sydney. His book, Over Our Dead Bodies, illustrates the pseudo-scientific nature of the reasoning of public health researchers. This is not a scientific publication; it is a passionate argument against guns as a public health risk. He only selects studies that support his position, and he ignores or denigrates any that do not. This is advocacy or “sagecraft”, not science. He revels in the trappings of scientific studies (as long as they are supportive), but he does not respect the logic of science. If he cannot ignore a criticism, because it was made by a reputable researcher, he dismisses the critic as “relied upon by the gun lobby” (Chapman, 1998, p. 57). Ad hominem arguments are not scientific. Kleck is a very respected researcher in criminology. His first book on firearms, Point Blank, won the Michael Hindelang Prize which is awarded by the American Society of Criminology for the best book in criminology. An honest discussion of these issues would give his criticism more extensive discussion. Chapman’s treatment of the old lethal-gas argument is similarly unscientific. This claim, at the very least is controversial, and many reputable researchers believe it has been refuted. Chapman is not writing as a scientist but as an advocate.
Chapman frames the debate as if it were between “science” and “the gun lobby”, when the truth is that criminology is riven with many methodological debates, and respectable criminologists find themselves on diverse sides of the sometimes acrimonious debates over firearms and gun control. Science rarely speaks with one voice. Chapman’s moralistic world is black and white: public health advocates are good and scientific; his opponents, the gun lobby, are evil and unscientific. This is not just advocacy; this is demagoguery. I do not wish to deny that academics, as well as ordinary citizens, have the right to take positions on the political issues of the day. But academics, unlike citizens, have a duty to base their arguments upon the best available research, and to conduct their research using scientific principles. Chapman illustrates the all-too-frequent tendency in public health research to ignore basic scientific principles in favour of advocacy.
In their zeal to lobby government, public health researchers have oversimplified the epidemiological model.[26] In this model, guns are treated as if they were a disease vector and “gun deaths” as a disease. However, epidemiologists have long recognized that the same agent may be a disease hazard, a protectorant, a cause, or a preventative, depending in part upon the susceptibilities of particular hosts. In the hands of public health, guns are treated as if they were wholly evil. For example, criminologists have found that the manner of introduction to guns – family vs. peers – has a strong influence upon how likely teenagers are to engage in violent behaviour (Lizotte and Tessoriero, 1991). This suggests a direction for research to seek to explain why firearms in one host neighbourhood are linked with criminal violence, while in another firearms may act as a “protectorant”. Public health researchers have ignored the factors of “hosts” or users, in preference to repetitious studies of the dangers of the “availability” of firearms.
Epidemiological studies are useful in identifying the susceptibility or immunity of segments of the population to morbidity and mortality from particular causes (Lilienfeld and Stolley 1994, p3). Public health research has not attempted to determine which groups or individuals may be made more or less susceptible to homicide or violence because of the presence of a firearm. Textbook epidemiology recognizes that differences in frequency and severity of diseases vary importantly among racial groups, but public health researchers frequently ignore important differences in homicide rates that are related to ethnicity.
This failure is illustrated by one widely cited study that compared homicide rates in Vancouver, BC, Canada, and a neighbouring city, Seattle, Washington (Sloan et al 1988). Despite the glaring differences in their ethnic profiles, the two cities are described as having a “comparable … ethnic makeup” (Cotton, 1992, p 1172). The Black population of Seattle was 30 times larger than Vancouver’s and had a disproportionately high homicide rate. If the comparison of two cities is limited to more closely matched populations, e.g., non-Hispanic whites, no significant difference can be found between the homicide rates for the two cities. Rather than recognizing the problems of inner-city Black youths, the study concluded that stricter Canadian gun laws were responsible for Vancouver’s lower overall homicide rate (Sloan et al 1988). Even more perplexing is the failure to pursue the differences in exposure to guns. In the epidemiological model, “exposure” refers to the extent to which populations are in contact with a pathogen (Mausner and Kramer, 1985, pp 188-190). Non-Hispanic white households are much more likely to have firearms in the home, than are Black households, particularly urban Black households, and non-Hispanic whites are much more likely to grow up with firearms in their home[27]. Thus inner-city Blacks are much less likely to be exposed to guns than are non-Hispanic whites (Kleck 1991, pp 56-57; Wright, Rossi, & Daly, 1993, pp. 87-89).
Much of the public health research into firearms relies upon the case-control method. The case-control method is a legitimate research methodology for identifying risk factors, i.e., generating hypotheses about what factors might increase the risk of catching a disease. Epistemologically, this is a method for discovering hypotheses, not testing them (Campbell and Stanley 1963, p 12). It is not a confirmatory methodology, i.e., intended to test hypotheses that certain conditions cause the disease under study (Lillienfeld and Stolley, 1994, p. 227). Hypothesis testing is properly reserved to experimental methods. In public health, this typically means subjecting risk factors to clinical trials. All too often public health researchers uncritically rely upon the case-control methodology as if it were a confirmatory methodology. This tendency is particularly evident when firearms are at issue.
It is critically important in the case-control method to match subjects in the experimental and control conditions. It involves the comparison of “case” subjects, who have been exposed to the test stimulus, with a “control” sample that has not been randomly assigned, who have not been so exposed. The primary threat to validity in case-control studies is selection bias. Logically, to draw the inference that the test stimulus caused the observed differences between the groups on the dependent variables, the control group must be identical (not “similar”) to the case group except for exposure to the test. This is best achieved by random assignment of subjects. However, this design is critically weakened if the researcher selects the control group. All that can be done is to match the case subjects with those in the control sample as closely as possible on the variables the researcher believes are the most important. Matching cannot ensure that the groups are equivalent. In social science research, matching on background characteristics has all too often been found to be ineffective and misleading (Campbell and Stanley 1963).
The problems inherent in the case-control design may be illustrated with a well-known paper that found that firearm availability increases the risk of homicide victimization (Kellerman et al, 1993). For the “case” sample, Kellermann and his colleagues selected households in three urban counties in the USA where people had been murdered in their own homes. They excluded any instances (a) where intruders were killed by the homeowner, (b) where people were killed away from home, or (c) where any children were killed. To find out information about the conditions of the homicide, Kellermann and his colleagues interviewed, “persons who were close to the victim”, whom they refer to as “proxies”. The researchers did not ask the victim's proxy (from whom they derived their information about the victim and his or her household) whether or not the victim had previously defended himself or herself with a gun.
In order to approximate a “control group”, these researchers selected other households from the neighborhood of the same sex, race, and age group as the victim. The “controls” were asked the same questions as had been asked of the victim's proxy. Respondents often find it easier to admit socially unacceptable practices about their friends or relatives than about themselves. It follows that there would be a significant amount of under-reporting in the control group. This is particularly problematic with firearm ownership. To the extent that firearm ownership was underreported in the control group, the odds-ratio that is the objective of the study would have been undermined.
As noted earlier, matching cannot ensure that the groups are equivalent. The control group differed markedly from the victim group. While matched on the demographic variables, the control group was distinctive on behavioural measures. Compared to the control group the victim group was more likely to: rent rather than own, live alone, drink alcoholic beverages, have problems in the household because of drinking, have trouble at work because of drinking, be hospitalized because of drinking, use illicit drugs, have physical fights in the home during drinking, have a household member hit or hurt in a fight in the home, have a household member require medical attention because of a fight in the home, have a household member involved in a physical fight outside the home, have any household member arrested, and be arrested personally (Kellermann, et al, 1993, p 1086-1088). In sum, the victim (or “case”) group and the control group reported very different lifestyles, with the homicide victims living a very high-risk lifestyle. If the groups are not equivalent, as they demonstrably are not in this study, then the odds-ratios are of doubtful validity.
Another crucial threat to the validity of case-control studies is nonparticipation. Kellerman reports that 30% of the people who were initially contacted to act as controls refused to participate. Epidemiological research has found that there is a tendency for less healthy respondents to refuse to participate (Austin, 1994). The utilization of healthier controls exaggerates the differences between the controls, and the victims thus may contribute to an overestimation of the odds ratio.
Kellermann’s study cannot be considered externally valid either because households were selected in only three urban counties in the USA where people had been murdered in their own homes. Since the households were not randomly selected, the sample cannot be considered representative of all households in the USA nor even urban American counties. For example, 53% of the case subjects had a household member arrested, 25% had alcohol-related problems, 31% had a household history of illicit drug abuse, and 62% of the case sample were Black, compared with 25% of the households in the urban counties where the study was conducted, and 12% of all US households. Thus, the results may not logically be generalized to any target population.
Finally, Kellermann over-interpreted his findings. Even though the case-control methodology is not designed to determine causality, Kellermann asserted an unambiguous causal result. Moreover, Kellermann claimed his findings of an odds ratio of 2.7 was a “strong” result, but such an odds ratio falls below the well established threshold set for identifying potential risk factors for disease (Lillenfeld and Stolley, 1994). Because of his inability to control for the confounding factors already discussed, his results are most likely spurious. Despite all of these methodological problems, Kellermann’s results are widely accepted in the public health field.
Public health researchers have also tended to exaggerate the importance of the “weapons hypothesis” so that firearms availability is equated with death or injury. This is a misrepresentation of criminological research findings. By focusing myopically upon “firearms death” researchers gloss over important distinctions between suicide and homicide, as well as ignoring violence from other types of weapons. The narrow research focus of public health researchers amounts to a refusal to even consider theoretically the diverse “susceptibilities” to firearms of particular hosts. As with any “disease vector”, firearms theoretically may act as a disease hazard, a “protectorant”, a cause, or a preventative, depending in part upon the susceptibilities of particular hosts.
Such a basic misunderstanding of criminology might be charitably ascribed to the failure of public health researchers to be familiar with the research literature in criminology. This is irresponsible because good epidemiological methodology requires researchers to learn as much as possible about a disease they are attempting to understand. Nevertheless, public health researchers are often woefully ignorant of even the most basic research in criminology. On the other hand, if one wished to be uncharitable, these lacunae could be seen as due to efforts to promote their a priori agenda through pseudoscientific studies.
Sound epidemiological research requires establishing research protocols that conform to known biological and other important factors. This implies that when epidemiologists study firearm-related violence, they should become familiar with criminological studies. Unfortunately, public health researchers appear to be ignorant of much of the basic research in sociology and criminology. For example, despite their intellectual importance, few public health studies cite work by Gary Kleck, Gwen Nettler or Jim Wright (e.g., Kleck, 1991, 1997; Nettler, 1982; or Wright et al, 1983).
In summary then, public health researchers frequently ignore basic scientific principles in favour of advocacy. The epidemiological model is oversimplified to justify moralistic campaigns, basic research findings in criminology are ignored, and the case-control method is misapplied. These failings lead too many researchers to draw conclusions that are not supported by their research methodology and to compound these errors by recommending legislative solutions that fall far outside the boundaries of their research. Such studies are not properly scientific but sagecraft, i.e., the scientific trappings of research are used to win arguments rather than test propositions.[28] Many of the magazines where these studies are published, e.g., Canadian Medical Association Journal, are not proper scientific journals because they do not subject manuscripts to blind review by academics qualified in research methodology. These pseudo-scientific studies provide a very flimsy foundation for public policy.
Organizational Problems
Large
sprawling government programs are invitations to waste and inefficiency.
It is not an easy task to create a large information database.[29]
Creating and managing the firearms registry posed particularly challenging
problems that were underestimated by the Canadian Government. The Department of
Justice failed to develop a clear understanding of the project’s scope and to
plan for the level of inter-governmental and inter-agency cooperation that
would be needed. Apparently, no one in
the Department of Justice had experience with designing and implementing an
information technology project of this size or scope. Another reason is that
firearms are uniquely complex, and this complexity is reflected in the
different agencies’ widely differing information needs. Perhaps the best
example of mismanagement is the department’s failure to understand that the
standards for data quality varied across the agencies involved, and this
created virtually insurmountable obstacles to the development of an accurate
and common database.
Freedom-of-Information requests have revealed that the Royal Canadian
Mounted Police (RCMP) continue to have serious doubts about the validity and
usefulness of the information it contains (Breitkreuz, Sept 23, 2003).
The originally modest information
technology project grew rapidly in the face of numerous demands for change.
Five years after the contract for the project was awarded, the development team
had dealt with more than 2,000 orders for changes to the original licensing and
registration forms or to the approval processes. Many of these changes required
extensive additional programming rewrites. As the public learned about its
problems, the Quixotic nature of the firearms registry was revealed.[30]
The cost-overruns were caused by the
failure of the Government to anticipate the complexities of creating and
maintaining the firearm registry. The Canadian Government was aware of the New
Zealand government decision to abandon a firearm registry, but these warnings
were ignored. Unwilling to admit its failure, the government resorted to
financing the ever-growing project through “supplementary estimates” that
avoided reporting requirements (Stanbury,
2003).
The problems in the Department of Justice became widely known in Canada when Auditor General Sheila Fraser released a scathing report in December 2002. Despite promises at the time that the firearms program, including the registry, would not cost over C$ 2 million, the Auditor General estimated that the firearms program would cost taxpayers at least C$ 1 billion by 2005.[31] She summarized her report by saying, “This is certainly the largest cost overrun we’ve ever seen in this office” (Fraser, 2002).
This is a staggering cost overrun, but it is necessarily an underestimate. First, the scope of the Auditor General’s inquiry was formally limited to examining the Department of Justice, but several other federal ministries are involved in administering the firearms program including the Solicitor General, Canada Border Services Agency, Department of International Trade Canada, Department of Foreign Affairs, and Native Affairs and Northern Development; other ministries incur costs due to the firearms registry, such as Parks, Fisheries, Environment, that have not yet been reported; the firearms in the hands of the police have yet to be registered; the cost of destroying the guns confiscated since 1998 have yet to be accounted for; in addition, the federal government partially reimburses the program expenses of cooperating provinces and territories.[32]
When the full scope of this sprawling program is included, specifically those of other federal departments as well as the provincial expenditures reimbursed by the federal government, Garry Breitkreuz estimated that if all governmental costs are considered, the total will exceed two billion dollars by 2005 (Breitkreuz, May 26, 2003). This is 1,000 times more than was originally budgeted, but this too may be an underestimate. Unfortunately, the total costs remain unknown as many program expenditures related to this program remain hidden.[33]
The Auditor General also complained that the registry audit was the first time her office had had to discontinue an audit because she could not obtain the necessary information.[34] The Auditor General had to end her audit precipitously, and leave her financial analysis incomplete, because the government either could not or would not cooperate with her by revealing all of the program’s expenditures.[35]
The
Auditor General saved her strongest criticism for the way the Government
deliberately misled parliament: “The issue here is not gun control. And it's
not even astronomical cost overruns, although those are serious. What's really
inexcusable is that Parliament was in the dark.” The government knew about the
mismanagement problems in the firearm registry years ago, but ignored questions
from MPs such as Garry Breitkreuz whose requests for financial information were
repeatedly refused on the grounds of “cabinet secrecy”.
In response to this independent
assessment, in February 2003 the registry was relocated in the Ministry of the
Solicitor General along with the Royal Canadian Mounted Police (RCMP). Paul
Martin became Prime Minister early in 2004, and the firearm program got another
Minister.[36] Because of
heightened concerns about budgetary concerns, the firearms program is in an
awkward position with regard to managing the firearm registry. Tight budgetary
restrictions have led to complaints that the program has reduced the quality of
service. Long waits are normal, and errors frequent. Nevertheless, the registry
is ineffective in tracking stolen firearms, because of duplicate serial numbers
and inadequate descriptive information (Naumetz,
Sept 25, 2003). This again reflects the inherent difficulty of the task.[37]
Budgetary restrictions also
compromise data quality. Unfortunately, one of the cost-cutting decisions was
to reduce efforts to verify descriptive information submitted about firearms.
Many applicants for firearm permits appear not to have been as thoroughly
screened. One imaginative Canadian even managed to register a soldering gun
without the Canadian Firearms Centre knowing that it was not a “firearm” under
the Canadian criminal code (CNEWS
(March 21, 2002). This example not only illustrates the level of screening
given by the firearm registry, but it also demonstrates the contempt that many
feel for the registry. Few now take seriously claims that the registry has any
real use.
The 2004-2005 budget eliminated
funding for firearms safety programs altogether, even though it maintained the
public relations budget.[38]
Despite the huge expenditures, the firearms registry is plagued with errors.
Millions of entries are incomplete or incorrect. Fraser (2002) also reported
that the RCMP in 2002 announced that it did not trust the information in the
registry (Fraser, 2002, Chapter 11). As the New Zealand Police discovered
decades ago, it is exceptionally difficult to maintain a firearm registry.[39]
If police are to trust the registry, to protect police lives, to enforce court
orders, or to testify in court, the data contained in the registry must be both
accurate and complete. An inaccurate registry becomes a self-defeating exercise
and cannot be useful in aiding the police protect the public.
The principle of the registry
demands an exceptionally high level of accuracy to guarantee to the police
officer knocking on a door that the information in the registry is correct
about the number and nature of the firearms owned in the residence. If any
sizable percentage of the firearms remain unregistered, it is very likely that
the firearms in the hands of the most violent criminals are not registered. If
this is the case, police officers cannot trust the information that there are
no firearms in a residence. Failure to register a firearm does not mean no
firearm exists. Or, if there is only one firearm registered, the officer cannot
infer there are no other firearms. Practically speaking, the registry is not
useful to an investigating officer since it must be assumed that a firearm is
available when the officer knocks on a door, whatever is reported in the
registry. Trusting the information in the registry could get police officers
killed.[40]
Despite its current cost of over one billion dollars, the police still have
grave reservations about the usefulness of the firearms registry (Breitkreuz, Dec 9, 2004).
In summary, the Canadian firearm registry has had numerous action plans, a series of ministers, and thousands of changes made to the computer system. This is not a recipe for effectiveness or efficiency. In other words, as New Zealand discovered decades ago, a firearms registry may not be worth the effort, as such a database is exceptionally difficult to maintain, outrageously expensive, and any benefits are all but impossible to demonstrate. In a subsequent section of this paper I evaluate the success of the firearm registry in including all the firearms in the country. An incomplete registry is a guarantee that it will include only those firearms that are the least likely to be used in crime.
Large sprawling government programs are invitations to waste and inefficiency, even corruption. It may be comforting to dismiss the Canadian failure by imagining that Canadians are uniquely incompetent or corrupt. This may be true, but it is more salutary to recognize the universal implications for any large-scale national program. Organizationally, firearm registration is essentially similar to a mass inoculation program. It is intrinsically expensive to deliver a service to an entire population dispersed across a continent.[41] If any country did decide to inoculate its entire human population, citizens would be wise to keep close control over the program in order to minimize waste and fraud. In the case of firearm registration in Canada, it is even more expensive because a new organization had to be created in order to accomplish it.
-----------------------
Tables 1 & 2. Hunters and hunting in Canada
-----------------------
Firearms and Firearm Owners in Canada
As may be seen in Table 1, the primary reason (73%) people in Canada own a firearm is for hunting. The second most popular reason given is target shooting (13%). In contrast to the USA, few Canadians own firearms for protection. Only 4 - 6% volunteer that protection is the principal reason for owning a firearm; in the US, surveys find between 15 - 22% who give this reason (GPC Research 2001; Mauser and Margolis, 1992, and Mauser, 2001b). It is likely that many of respondents who do so are employed in security or enforcement.
Table 2 compares Canadian firearm owners with the general population. Firearm owners are predominantly male; in comparison to the general population, they are older, somewhat less well educated, but have a higher annual income.
Handguns have been tightly controlled in Canada since the 1890s. The handgun registry, which began in 1935, had just over 1 million guns in 1995. The current registry, which includes both long guns and handguns, is reported to have approximately 7 million firearms in 2005 (CFC, Feb. 12, 2005).
Estimates of the number of firearm
owners in 2004 range from the CFC’s estimate of 2.2 million to the National
Firearm Association’s estimate of 7 million. My best estimate is that in 2004
there are between 3 and 3.5 million firearm owners. If this is correct, this
implies that there are more than 1 million unlicensed gun owners who have
refused to cooperate with surveys or the CFC. Nevertheless, this is a
significant drop from the early 1990s, when I estimated that there were between
4.5 and 5.5 million firearm owners at that time (Mauser 1995a).
The decline in the number of firearm
owners in Canada has corresponded with a drop in the numbers of hunters. Such a
decrease bodes ill for wildlife conservation in Canada, as hunters are the
mainstay of provincial budgets for wildlife management. The implications are
substantial. Canadian hunters pay
provincial governments almost $70 million per year for hunting opportunities;
this roughly equals what the provinces spend on wildlife management (Mauser, 2004).
In fact, hunters are the driving force behind conservation in Canada as they
are in North America. Canadian hunters,
in addition to what they spend on licence and fees, also voluntarily contribute
over $33 million annually for habitat protection and conservation projects
(Powers, 2000). More important, the Canadian Wildlife Service reports that
hunters spend almost half ($2.7 billion) of the $5.6 billion the Canadian
public spends on wildlife-related activities each year on hunting-related
tourism, and hunters contribute over $10 billion annually to the Canadian
economy (Filion et al, 1993).
Hunters
are also important in controlling problem wildlife. Not only do wildlife cause
economic losses for farmers and orchardists through crop damage, but wildlife
pose roadway dangers for motorists (Cotter, 2005). In just one province,
British Columbia, there were over 4,700 wildlife-related accidents in 2000.
That year, it was estimated that wildlife accidents cost the province over $18
million in accident claims, $600,000 in highway accident clean-up costs, and
$300,000 in lost provincial hunting licence revenue. In addition, collisions
between wildlife and motor vehicles kill at least two people each year in
British Columbia alone (BC Ministry of Transportation, 2000). A decline in
hunter numbers would mean governments would have to pay to reduce wildlife
numbers rather than having hunters pay to kill and eat them. A recent study of
another Canadian province, Manitoba, found that if hunting were eliminated,
problems in wildlife damage would increase substantially because deer and bear
populations would increase over 200 percent and waterfowl would increase over
300 percent. Manitoba, with a population of 1.1 million people experienced
10,475 wildlife collisions in 2003. As a result, Manitoba Public Insurance paid
out $20.1 million in insurance claims (IAFWA, 2005). Hunters provide a vitally
important tool for wildlife management.
Cooperation
It is difficult to know the level of
nonparticipation because there is no agreed number of firearm owners in Canada,
nor are there accurate counts of the number of firearms in private hands. Canada
is not unique in this. Similar difficulties have been encountered in Britain,
the Federal Republic of Germany, and Australia[42].
In 1997, estimates of the total number of firearms in Australia ranged from 2.5
million to over 10 million, and estimates of the number of firearms to be
prohibited ranged from 800,000 to 3.35 million (ILA Report, 1997). It is essentially impossible to conduct
accurate counts of items, such as firearms, that the population considers
important. As of November 11, 2004, the
Canada Firearms Centre (CFC) reported that almost 2 million firearm owners had
been licensed, out of a total of 2.2 million owners, and that nearly 7 million
firearms had been registered. By their own figures, this means that, as of July
1, 2004, the official deadline, there were 406,834 holders of long gun possession
licences who had failed to register any long guns, and in addition there are a
further 316,837 handgun owners who have failed to re-register or dispose of
their handguns (Naumetz, Aug. 25, 2004; CFC, Nov. 1, 2004). Thus, according to
the CFC, over 90% of owners have taken out a licence and registered at least
one firearm. However, if we accept my best estimate that there were
between 3 and 3.5 million owners in 2004, then the participation rate is much
lower, between 60% and 65%. The large drop in the number of firearm owners
since the registry was introduced indicates that a substantial number of former
owners either have divested themselves of their firearms or they have simply
not registered their firearms.[43] The proportion of ‘refusers’ is unknown.
Estimates of the participation rates
among Aboriginal Canadians are even lower. Many bands have refused to comply,
while others have only partially cooperated.[44]
The government of Nunavut, one of Canada’s three northern territories, has a
court injunction forcing the Federal Government to halt registration in Nunavut
since 2002. The most optimistic estimate is that fewer than 25% of residents of
First Nation communities are estimated to have complied with the firearms act
(Breitkruez, Aug. 2003; Naumetz, July, 2003).
One band in BC has even decided, in defiance of the Federal Government,
to issue its own firearm licences (Canadian News Service, Feb 2003).
How complete is firearm
registration? Estimates of the actual gun supply range from 7.7 million (the
government's preferred number) to over 25 million, plus an unknown number of
air guns (GPC Research,
2001, and Smithies, 2003). Estimates based on import/export
figures tend to be higher than survey-based estimates. Based on government
surveys and import/export figures, Garry Breitkreuz, MP, estimates that there
are approximately 16.5 million firearms in private hands in Canada (Breitkreuz , Dec 13, 2001). Independently, I estimated
there were between 12 and 15 million firearms in private hands in Canada
(Mauser, 1995b).
Thus, the best estimate is that
approximately half (between 44 percent and 56 percent) of the private firearm
stock is registered. However, since the
estimate of the total firearm stock is based upon telephone surveys, this
probably underestimates the number of firearms and firearm owners, and so
overestimates the share that is registered. Surveys must rely upon voluntary
compliance from respondents. It almost certainly excludes any weapons in the
hands of criminals, as violent criminals are extremely unlikely to be contacted
in a telephone survey, or if contacted, to respond honestly.[45]
In summary, it is difficult to know
the level of nonparticipation among Canadians because there is no agreed number
of firearm owners in Canada, nor are there accurate counts of the number of
firearms in private hands. Estimates of the number of previously law-abiding
firearm owners who do not have a firearm licence range from approximately
700,000 up to 2.5 million. However, the
number of former owners who have divested themselves of their firearms is
unknown. Surveys show that the number of firearm owners has been decreasing
since 1995, so the most optimistic estimate is less than 65% of firearm owners
have licences and approximately half of all firearms are registered.
A Preliminary Evaluation
In evaluating public safety, we need to avoid being misled by simplistic and emotional concepts like “gun deaths”. Because of the problem of substitution, a rise or fall in “gun deaths” does not necessary imply that any lives have been saved or lost overall. There is no necessary link between “gun deaths” and suicide or homicide trends. The most appropriate measures of public safety are meaningful measures, such as homicide rates and violent crime rates because they tell us whether more or fewer human beings are actually dying or being hurt.
“Gun deaths”
At first glance, the concept of “gun deaths” may appear plausible. Gun
laws are supposed to stop gun misuse, so an obvious measure of success would be
a drop in “gun deaths”. However, because of the problem of substitution, it is
a fallacy to imagine that a reduction in “gun deaths” implies that any lives
have been saved.
Japan has exceptionally tight firearms laws, but a high suicide rate.
Mass killing in Japan has often involved poisoning. In many countries, arson
and bombing have been used in multiple killings, whether or not they involve
suicide, and are obviously unaffected by firearm laws. Because of the variety
of alternative ways to kill, the only useful measures of public safety are more
meaningful ones, such as rates of violent crime, suicide, and homicide.
In Canada, trends in “gun deaths” are not a good indicator of suicide
trends, in part because firearms are involved in only a small fraction (18%) of
suicides. An inspection of the trend in Canadian suicide rates since 1990 shows
that they have stayed remarkably stable even though firearm suicides have
fallen precipitously. See Figure 1.
-----------------------
Figures 1 & 2. Trends in Suicide Methods in Canada and Australia
-----------------------
It is time to stop using the term “gun deaths”. The term distracts from understanding what firearms have to do with homicide or suicide. “Gun death” is a pot pourri of suicides, homicides, and accidents. The supposed link is that these deaths share a common cause: a gun was accessible. But the mere availability of guns does not make ordinary people commit murder or suicide or accidents. This term may be useful to frighten the public into thinking that expensive measures taken to reduce the number of guns held lawfully will automatically spill over into crime reduction but it does not aid understanding any more than it reduces the overall number of people killed.
The concept of “gun deaths” is a grab-bag that confuses rather than
clarifies. As may be seen in Table 3, “gun deaths” are largely suicides, some
homicides, and a very small percentage of accidents. We could easily make up
our own grab-bag for further examination of this concept. Consider, say, “bed
deaths,” and let us say this combines any cause of death as long as it happened
in bed.
No policy goals are served by mixing such disparate phenomena. Such terms just muddy the water. Bed deaths illustrate this analytical impotence. For good reason, hospitals do not announce that “bed deaths” have decreased – or increased. They do not do so because the concept of “bed death” is not helpful in understanding the actual numbers of people being killed by diseases. One would think doctors would be embarrassed to use such a vague term. It could be used, though, by an unscrupulous hospital registrar who decided to puff up the hospital figures for more funding, using part of the cancer deaths and part of the heart attack figures in a bid to create a third overlapping category to show how many new beds the hospital may want. The number of people actually dying would, however, not be honestly represented. I suggest that this is exactly what the public health advocates have done in their conflationary “gun death” concept. And we can see the result of it in that when gun laws are introduced, often very restrictive ones indeed including total, countrywide bans such as those of Jamaica in 1974, or in the Republic of Ireland in 1972, the number of people actually killed did not fall.
Focusing on “gun deaths” diverts attention from the original goal of the Canadian legislation
– indeed, the only sensible goal – which was to improve public safety.
----------------------------------
Table 3. Canadian ‘Gun Deaths’
----------------------------------
One of the assumptions underlying the use of the artificial term “gun death” is that general firearm availability increases the overall suicide rate. Comprehensive reviews of technically sound studies have not found strong empirical connections between firearm ownership and either overall homicide or suicide rates.[46] Comprehensive reviews of the literature have been published by both Gary Kleck (1997) and Hemenway (1999 and 2004). Gary Kleck (1997) reviewed previous studies that examined the link between gun ownership and both violent crime and suicide. In his review of studies of the link between ownership and violent crime, he only included studies that actually measured gun availability rather than simply assuming it, and only studies that measured rates of criminal violence, not percentages. The results were mixed for the entire set. After screening out those studies that did not meet rigorous methodological standards, none of those remaining reported a significant relationship between level of firearm ownership and homicide rate (Kleck 1997, p. 248-251). In his review of 13 previous studies examining the link between guns and suicide, he reported that only two had found a significant association between firearm prevalence and total suicide, including the Kleck & Patterson study (Kleck, 1997 p. 284-285). However, both of these studies had methodological problems that undermined their results.
Nor has convincing empirical support been presented that gun laws have lowered the overall suicide rate or homicide rate (Kleck, 1997, pp. 49-53, pp. 286-288, p 377; Marvell and Moody, 1995). See Table 4. In contrast, research consistently shows that the availability of one means of committing suicide has a strong influence on the frequency of suicides using that method (Kleck, 1997, p. 285). Miller and Hemenway (1999), while evading a direct answer to the question about firearms and suicide, say that the “evidence…. is currently less compelling”. However, concerning the link between guns and homicide, Hepburn and Hemenway (2004) disagree with Kleck’s pessimistic assessment.
The use of “gun death” as a dependent variable merely muddies the water
as there is no logical link between increases (or decreases) in “gun
deaths” and suicide or homicide trends.
It is frequently claimed that guns are uniquely more lethal than other
methods of attempting suicide, but this is misleading and wildly exaggerated.
Kleck (1991, p 258) reported findings that the fatality rates for hanging,
carbon monoxide poisoning, drowning, and shooting oneself were all in the
75-85% range. Similarly, Sayer et al (1996) found that hanging (82%) had a
higher fatality rate than firearms (75%) in New South Wales. Moreover, researchers frequently confound
method lethality with the user’s determination. Anyone serious about
committing suicide would be expected to choose an effective method. Given the
ubiquity of ropes and motor vehicles, there would appear no shortage of methods
available that are highly lethal.
The key question is method substitution, or displacement, and on this question, the research is becoming clearer. There is considerable research support that limiting the availability of one suicide method reduces the frequency with which that method is used; but support is conspicuously lacking for the assertion that reducing firearm availability also reduces the overall suicide rate.[47] While some public health researchers claim that the unique deadliness of firearms means that substitution effects are not important in suicide (eg, Gabor, 1994), this is belied by the empirical evidence. Unfortunately, the public health literature generally ignores the relevant criminological research. As shown above, the evidence is consistent with strong substitution effects.
----------------------------------
Table 4. Kleck’s Table 8.4
----------------------------------
In summary, “gun death” is not an appropriate measure for evaluating firearms laws. Certainly gun laws are expected to reduce homicides with guns and violent gun crime. But reducing gun violence should be considered as only part of any master strategy to improve public safety by reducing homicide and violent crime overall. Evaluating the firearm registry means assessing its effectiveness in improving public safety. The government’s original goals were clear: “to reduce criminal violence generally, and more specifically, to reduce domestic violence and homicide”. The reduction in firearm violence was viewed primarily as a means of reducing total criminal violence, not as an end in itself. The Federal Government also thought that the gun laws would reduce loss of life due to impulsive acts such as suicide and accidents. Again, controlling firearms was seen not as just a way to reduce deaths from guns but as a way to save lives overall.
Evaluating the firearm registry
The firearm registry should be evaluated by asking how well it has accomplished the goals originally set for it. At the time the legislation was introduced, the Federal Government asserted that the key to improving public safety lay in controlling the availability of firearms, and that universal firearm registration was the best way to achieve this. At that time it was argued that restricting firearms availability would reduce homicide and criminal violence, as well as domestic violence, and, in addition, save lives by reducing suicides and firearm accidents. Each of these goals will now be examined in turn. Thus, the important questions are whether public safety has improved since universal firearm registration became mandatory in 1998, and if so, whether it is worth the cost.
The approach is preliminary. For the present, it is possible to look only at broad trends in the overall national rates. Although a few trend analyses have been conducted, clearly, further work needs to be done to confirm these observations.
The first step in answering this question is the most fundamental: has the firearm registry reduced access to firearms? For if the registry has not, then logically according to its proponents’ premises it cannot be expected to have any impact on firearms violence, or total criminal violence, or suicide. As Table 5 shows, the answer to this question is clear: the number of lawful firearm owners has declined substantially. The best estimate is that there were between 4.5 and 5.5 million firearm owners prior to the 1995 legislation; this number had dropped to between 3.5 and 4 million firearms owners by 1996 (Mauser, 2001a; Mauser and Buckner, 1997). And it has continued to drop. After firearm registration was introduced in 1998, it dropped still further. I estimate there were between 3 and 3.5 million firearm owners in Canada by 2002. There are probably fewer in 2004, although it is likely that the rate of decline has slowed.
-----------------
Table
5. Estimated Number of Firearm Owners
-----------------
Surveys
show a continuous drop in people who admit to owning firearms since the early
1990s. Surveys during the early 1990s found that an average of 29% of Canadian
households reported owning one or more firearms; surveys conducted between 1995
and 1998 found an average of only 21% of households reporting firearms. Later
surveys have found still lower percentages. In 2000, only 17% of households
admitted to owning one or more firearms (GPC Research, 2001). But how much of
this apparent drop is real?
The
estimates presented here are based upon surveys, but surveys have well-known
limitations. It is highly probable that some owners would deny that they have
firearms. Moreover, given the demonization of firearms over the past decade,
the share of deniers may well have increased. To what extent this is true is
unknown. In the GPC survey 6% of the respondents admitted to owning firearms
but claimed they planned to divest themselves of their firearms because of the
new legislation. Government estimates of the number of gun owners in 2001
assume that all of these respondents acted upon their declared intentions. How
many actually did so is unknown. Considering the way people tend to retain
heirlooms, and the way guns are very strongly seen as valuable personal items
(very good examples of, for instance, fine hunting shotguns have historically
cost up to a year’s wage for a working man), it is highly likely that a
substantial proportion of these have just been pushed back deeper into people’s
closets.
Another
reason for part of this drop is the change in the definition of “firearm
owner”. The advent of licensing meant that a firearm could officially only have
one unique legal owner; prior to licensing, firearms were frequently treated as
family property, similar to other commonly owned goods, such as computers, lawn
mowers, refrigerators, antiques, or household furniture, and hence they could
have more than one legal owner. This was particularly true in rural homes.
Given that licensing costs are significant, some families reportedly have
decided to limit the number of “official” owners.
However
true any of these considerations may be, I judge that there has been a real and
significant drop in the number of law-abiding firearm owners after the
legislation was passed in 1995, and then again after the introduction of the
firearm registry in 1998. It would appear that both events reduced the
availability of firearms. As noted in a previous section of this paper, the
declining numbers of firearm owners and hunters has strong negative
implications for wildlife conservation and public safety, particularly road
safety.
The
drop in the number of firearm owners is in part due to demographics. Firearm
owners are older than the general population, so their numbers would be
expected to shrink faster than the general population through differential
mortality. However, the size of the
differential death rate is indirectly due to firearm registration. The
increasingly strict requirements for owning firearms have reduced the
recruitment of younger people to the shooting sports, so the average age of
firearm owners has increased relative to the general population.
Some
researchers have used proxies for estimating firearm ownership because survey
estimates are not always available. Unfortunately, present research suggests
that none of the widely used proxies are valid measurements for time series.
That is, none of the proxies can accurately track survey estimates over time.
The most widely accepted proxy for cross sectional studies for the number of
firearm owners is the proportion of suicides that involve firearms although
Kleck argues that, while this proxy is valid in cross sectional studies, it
cannot be relied upon for time series.[48]
------------------------------------------
Figs. 3 - 4. Homicide Trends in Canada and E&W vs. USA
------------------------------------------
Now that we have shown that the firearm registry has restricted firearm availability, we can now turn to evaluating whether it has improved public safety. Specifically, we can frame our questions this way: has the registry reduced homicide, criminal violence, and specifically domestic violence, and has it saved lives by reducing suicides and firearm accidents since firearm registration became mandatory? We will examine each of these questions in turn.
First, consider homicide. As may be seen in Figures 3 and 4, since the firearm registry was introduced in 1998, the Canadian homicide rate has remained