“Everything secret degenerates, even the administration of justice; nothing is safe that does not show how it can bear discussion and publicity.” — Lord Acton, Letter, Jan. 23, 1861
A commonly-heard boast is that Canadians are safer than Americans as a result of our more restrictive approach to firearms. Government policy assumes that restricting civilian access to guns saves lives and reduces criminal violence. And even if our gun laws are not actually beneficial, for advocates of gun control, cracking down on guns just feels right, regardless of the facts. But is gun control effective?
Despite claims that the current firearms regime is “evidence based,” the government has not made public any studies that support its effectiveness. On the contrary, scientific research challenges the orthodox belief that restricting law-abiding citizens’ access to firearms improves public safety. The stage must be set by briefly reviewing the history of gun laws in Canada before describing the current situation with respect to firearms and violence.
Gun laws and immigrants
Canadian gun legislation has a dirty secret. Hiding behind government pronouncements about protecting the public from violent crime is often an unstated fear of recent immigrants, though the specific group depends upon the era.
Canadian gun laws in the 19th century were in part driven by concerns about Irish navvies and Métis uprisings. Starting early in the 20th century, additional restrictions on firearms were sparked by immigration from Southern and Eastern Europe, as well as from Asia.
The Winnipeg Strike of 1919, and similar events around the world, sparked fears about communists and anarchists in the 1920s and 1930s. Handgun registration in the 1930s was prompted by gangsters during Prohibition in the United States and by labour unrest (including concerns about Scottish agitators, of all things). The Second World War brought in universal firearm registration in Canada as well as confiscating firearms from Oriental immigrants (including Chinese) before interning Japanese Canadians, though these restrictions were eased after the war.
The ostensible justification for introducing new gun laws has typically been violent crime in major cities, especially Toronto and Montreal. But the government often privately had other concerns, sometimes unrelated to firearms at all. Thus Bill C-51, passed in 1977, was arguably enacted to distract public attention from, and persuade MPs to support, the government’s proposal to abolish capital punishment.
Bill C-51 was a potent measure. It required all firearms purchasers to undergo a criminal record check and obtain a Firearms Acquisition Certificate (FAC) prior to buying or otherwise acquiring a firearm. Mandatory minimum sentences and increased penalties were enacted, search and seizure powers granted, fully-automatic firearms were re-classified as prohibited instead of restricted, safe-storage was mandated, and people were no longer allowed to register handguns at commercial addresses, though the legal owners of recently prohibited firearms were grandfathered and allowed to keep their firearms.
The controls enacted under Bill C-51 remained in place for a decade until a multiple homicide at the Polytechnique Institute in Quebec prompted the Progressive Conservatives under Kim Campbell completely to redesign Canadian gun laws (Bill C-17) in 1991. Campbell’s Bill C-17 added the requirement for two personal reference checks from people familiar with the applicant including a spousal endorsement, photo ID, safety training involving written and practical testing, a psychological questionnaire (designed to screen applicants for a past history of mental health problems), and a mandatory waiting period prior to obtaining a FAC. A large number of “military-looking” firearms were reclassified as restricted or prohibited (grandfathering some legal owners), minimum sentences were also added and the Bill also required detailed regulations for safe storage, transportation to and from shooting ranges and gunsmiths, and set restrictions on magazine capacity.
Not to be outdone, and having campaigned in the 1993 election to “strengthen gun control,” the Liberal government in 1995 under Jean Chrétien brought in Bill C-68, which prohibited over one-third of all legally registered handguns (again grandfathering legal owners), mandated the registration of rifles and shotguns, and required a firearms licence to purchase ammunition, while the requirements for authorizing the transport of restricted firearms (such as handguns) to shooting ranges and gunsmiths was further restricted and screening was centralized in Miramichi.
C-68 also introduced the firearms licence (Possession and Acquisition Licence or PAL) to replace the FAC. The crucial difference is that the FAC was only required to purchase a firearm, whereas the PAL was mandatory to legally possess it. Importantly, the PAL required renewal every five years, and holders were checked daily to ensure that they were not subject to court orders or any new criminal charges. Under those rules, around 3,000 licences are revoked annually, and the firearms confiscated. The licensing and registration system was originally estimated to cost only $2 million, but soon, because of mismanagement, the cumulative cost passed $2.7 billion by 2012.
Given the bureaucratic nature of Canadian firearms regulations, they are only implemented years after their adoption by Parliament. For example, the FAC requirement came into effect in 1979 (two years after passage of C-51), and the PAL became mandatory in 2001 (six years after C-68). The psychological questionnaire was first implemented in 1994 (three years after C-17). The registration of all rifles and shotguns became mandatory in 2003 (eight years after passage of C-68), known as the “long-gun registry”; handguns have been registered since 1934.
Registration and confiscation
In 2012, the Harper government scrapped the long-gun registry. And the regulatory scheme has yet to be implemented for Bill C-71, passed in 2019. But on May 1 of last year the Liberal government banned “over 1,500 models and variants of assault-style firearms by Order-in-Council”, ostensibly to “remove dangerous firearms designed for military use” from the community. The government has not clarified if banned firearms must be surrendered, or if current legal owners will be able to keep them, but only under tighter limitations. The ban targets all legal owners, except Indigenous Canadians, who are exempted so they can use newly-banned firearms for hunting. The ban also includes bolt-action hunting rifles, typical shotguns, and anti-tank guns sitting in museums. The OIC has not stated whether museums must surrender these historical pieces or adhere to stricter controls. This OIC is currently being challenged in federal court. But the RCMP has continued to add more types of prohibited firearms since May 1, 2020 without directly notifying owners. In February 2021, the government finally tabled legislation (Bill C-21) in Parliament, partially to support the OIC. But details remain unknown as regulations have yet to be drawn up. The government has issued confusing announcements about a buy-back program.
When Bill C-21 was tabled in Parliament on Feb. 16 this year, the government predictably claimed it would combat gun violence. But the measure does not address violent crime directly. Instead among other measures, it invents a variety of administrative crimes, creates a “red flag” law (Section 117.0101 ff.) that allows police to seize property without a warrant, and allows municipalities to ban handguns from legal owners (Section 19; 58.01 ff.).
Firearms and violence
With that background in mind, let us turn to the facts about guns and crime. According to the RCMP, more than 2.2 million Canadians possessed a firearms licence in 2019 after having been vetted by police. And the vast majority of Canadians who own firearms are solid citizens who do not just contribute substantially to their communities, they are much less likely to be murderous than other Canadians. A Special Request to Statistics Canada found a homicide rate of 0.67 per 100,000 licensed gun owners between 2001-2015. Over the same 15-year period, the average national homicide rate (including gun owners) was more than twice as high, 1.76 per 100,000 people.
According to surveys, between 17% and 29% of Canadian households possess firearms. The primary reason (73%) Canadians give for owning a firearm is hunting. The next most popular reason given for owning a firearm is sport shooting. Pest control is also mentioned. In 2012, more hunters (55%) lived in urban Canada than in rural Canada (45%), according to one survey. Between hunting and sport shooting, gun owners contribute $5.9 billion to the Canadian economy.
Criminal violence involving firearms is an important and controversial subject even though, according to Statistics Canada, firearms were used in less than 1% of violent crimes in 2019, the most recent year for which statistics are available. In 2019 firearms were involved in about 40% of homicides, but most of those (51%) involved gang activity. The bulk of homicide victims (74%) were male, as were those accused (87%). Almost two-thirds (63%) of accused killers were known to have a criminal record; as did over one-third (35%) of their victims.
Multiple-victim murders (i.e., with four or more victims) that dominate the media, they account for less than 1% of homicides. And they need not involve firearms; a recent multiple public killing in Toronto involving a van killed 10 people and injured 16 others, some quite seriously.
A tiresomely familiar pattern
But recent events follow a tiresomely familiar pattern. After every headline-grabbing (but extremely rare) gun murder spree, the media are filled with emotional calls for more restrictions on firearms. Rather than analysing the facts, the media amplify the emotion, in most cases taking it for granted that more gun control is needed. The same applies to much of the professoriate that interests itself in firearms: facts and logic are absent. Government then reacts by promising “to do something” about guns. The most recent firearm restrictions (Bill C-21 as well as the prohibitions of May 1, 2020) were prompted by a murderous rampage in Nova Scotia in April 2020 that claimed 22 victims: nine who died in house fires and 13 who were shot to death.
The Nova Scotia killer was armed with two semi-automatic rifles and two pistols. In reaction, the Trudeau Liberals issued an Order-in-Council banning “assault-style weapons.” The killer’s guns were in fact already obtained illegally (most had been smuggled), while of course the OIC ban includes only firearms legally owned by people without criminal records. It is a perfect example of the lack of logic in framing laws that do nothing to curb criminal activity but instead impose bureaucracy on, and stigmatize, law-abiding people.
Where’s the evidence?
There is no compelling evidence that gun control legislation is effective at reducing violent crime or suicide. It may appear hard to believe given the inundation of media stories. A search for “gun control studies” reveals millions of entries, virtually all of which are supportive of gun legislation. By and large, the studies are American and were conducted by public health activists. But closer inspection reveals that few of these studies are methodologically sound; their conclusions do not follow logically from their research methods or data.
Methodology is important. Social science research is complex, and it is easy to make mistakes or fail to correct for the researcher’s biases. Variables that are included, or left out, can drastically shape results. Some factors are obvious, but others are more subtle. As any data-monger knows, with a little ingenuity data can be contorted to yield whatever result is desired. Conversely, ferreting the truth out of data is much more challenging.
It is important to distinguish between cross-sectional studies and time-series of gun control. Cross-sectional studies compare two or more jurisdictions over a single time period to discover patterns. It can be particularly illuminating to analyze a large number of jurisdictions simultaneously (such as all 50 US states). Such an approach can suggest hypotheses for further examination and help understand a puzzling question, but logically cannot provide a conclusive result because cross-sectional studies may be suggestive are not designed to assess whether an intervention caused a result.
In contrast, a time-series study takes one jurisdiction (state, province, country) and follows events in that location over a number of years. In this way it is possible to see if changes in one variable (the dependent variable) follow from earlier changes in another variable (the independent variable). The first step is to collect measurements for a sufficiently long period both before an intervention and afterwards. If homicide or suicide rates decrease after a legislative initiative that is intended to make this happen, then it is logically possible that the change influenced the decline. However, comparing average rates before and afterwards is not sufficient.
Demography and economics
Legislation is not the sole influence on criminality. Socioeconomic factors including demographics can be much more powerful, if less amenable to control. It would for instance be illogical to attribute a change in crime rates to a legislative intervention in the face of the powerful effects of an economic downturn that began before the passage of the gun law and continued afterwards. Any decrease in violent crime might be better explained by an improvement in the economy, not the new legislation.
Researchers normally account for alternative explanatory causes by including them as “covariates” in the statistical model. A time-series analysis will be vacuous if it does not include the most likely variables as covariates, such as family income, unemployment rates, police effectiveness, immigration levels, or ethnic differences. Previous research has found these independent variables influence homicide, suicide, and violent crime rates generally, so it makes sense to assume that they influence the rates involving firearms as well.
As noted these studies are overwhelmingly American. Given the importance of homicide and suicide rates, surprisingly few studies have attempted to evaluate the effectiveness of firearms legislation on either in Canada. The following section briefly reviews the research, showing that later studies corrected the weaknesses of earlier studies by employing more sophisticated statistical methodologies. Since all researchers have biases, it is important for those who are searching for answers, rather than gaming statistics, to employ research designs that guard against their own biases as well as those of others.
Back in the 1980s, a small number of attempts were made to evaluate the effects of the 1977 Canadian gun-control legislation (Bill C-51) that introduced the Firearms Acquisition Certificate (FAC), requiring potential firearms buyers to submit to vetting by the police, in addition to other amendments to the Criminal Code dealing with firearms. Unfortunately, the methodologies of these early studies were less than rigorous and the results equivocal.
Two studies are instructive. Both relied upon simple before-and-after comparisons to evaluate the effects of the 1977 legislation and were limited to examining brief time periods (three and five years, respectively) before and after passage of the bill.
One study concluded that the gun control law was effective because firearm use in homicides and suicides declined after 1977. But it focused on firearm misuse and did not examine overall homicide or suicide rates. The second study examined total homicide rates, employing an analysis-of-variance approach, and found that the effect of the law was statistically insignificant.
Both studies restricted their analyses to comparing average rates before and after 1977. Neither study explored time trends nor included potential confounding variables. The results are equivocal because important socio-economic variables, such as economic conditions, and police effectiveness, and an aging population, were ignored.
It is illogical to conclude that an intervention caused changes that began before the intervention and continued afterwards. A pooled cross-section, time-series model is a more powerful approach to evaluate the effect of firearms legislation than a simple comparison of before-after averages. In order to investigate the 1977 Canadian firearms legislation, Gary Mauser and Richard Holmes (1992) used such a model to analyze the provincial homicide rates over twenty years (between 1969 and 1989), including ten years both before and after the 1977 legislation.
This type of model was selected because of its ability to capture variation across space as well as time. The hypothetical explanatory variables included in this model, measured at the provincial level, were: unemployment rate, percentage Status Indian, percentage immigrant, percentage male youth, and the clearance rate.
The results of this time-series model showed that the 1977 Canadian firearms legislation did not have a significant effect on homicide rates. Importantly, all of the covariates were found to be statistically significantly related to the homicide rate.
The economy plays an important background role in driving homicide rates. Nationally, unemployment peaked in 1979 and then rapidly declined during the 1980s, but this pattern was uneven across the country. Since economic downturns disproportionately affect young males, particularly Indigenous youth, provinces with higher percentages of either young men or Indigenous youth tend to have higher homicide rates. And consistent with previous sociological research, this study found that the percentage of Status Indian and male youth were the strongest explanatory factors.
Evaluating firearms legislation
Another study by Caillin Langmann evaluated the effect of firearms legislation on homicide as well as spousal homicide in 2012. Using data obtained from Statistics Canada from 1974 to 2008, he employed three distinct statistical methods to probe for effects of firearms legislation after the passage of three Acts by the Canadian Parliament: Bill C-51 (1977), C-17 (1991), and C-68 (1995). The statistical methods were interrupted time-series regression, ARIMA, and Joinpoint analysis. No significant beneficial associations were found between any of these legislative packages and either homicide or spousal homicide rates. Each of these statistical models included several covariates, including median age, unemployment and immigration rates, percent of population in the low income bracket, the Gini index of income equality, two variables that assessed police activity, population per police officer, and incarceration rate.
Specifically, Langmann found no statistically significant effects after the implementation of licensing in 2001 or the registration of rifles and shotguns in 2003. Firearm homicides as well as total homicides declined throughout the 1990s. But after the passage of C-68 the rate of decline for firearm homicide was found to have slowed down rather than speeding up. And after the enactment of licensing in 2001, as mandated by Bill C-68, the rate of firearm homicide was discovered to have increased.
This study failed to demonstrate a beneficial association between legislation and firearm homicide rates between 1974 and 2008. Instead, the covariates that had been included were found to be statistically associated with homicide rates.
Can restricting access stop suicide?
Suicide accounts for two-thirds (67%) of deaths associated with firearms in Canada. But hanging is the most common method of committing suicide (44%), while poisoning accounts for almost one-quarter (23%) and firearms just 16%.
In 2020, Caillin Langmann returned to the subject with a comprehensive effort to evaluate the effect of firearms legislation on both suicide as well as homicide rates. Using data obtained from Statistics Canada from 1981 to 2016, this evaluation covered a series of laws implemented in the years 1991, 1994, and 2001 that include background checks, psychological screening, owner licensing, and safety training. It is the first empirical evaluation of the psychological screening introduced in 1994, and the licensing implemented in 2001, on suicide rates.
He constructed models using difference-in-difference analysis of firearms and non-firearms death rates from 1981 to 2016. Difference-in-difference analysis is a quasi-experimental method that uses non-firearms deaths as a type of control for changes across time in firearms death rates, to see whether the legislation causes changes in the firearms death rate relative to the non-firearms death rate. The control group for firearm homicides consisted of all non-firearm homicides, but for firearm suicides, suicide by hanging was chosen for two reasons: it is the most frequent method used by males, it is almost as likely to result in death as suicide by gunshot.
Another model was built using negative binomial regression to test for an association between suicide rates and firearms prevalence by Canadian Province. Both overall suicide and firearms suicide were included, and licensing rates were used as a proxy for prevalence. The results showed that there were no associated benefits from firearms legislation on firearm homicide rates. No association was found with provincial firearm prevalence rates and suicide rates. However, consistent with previous research, this study found a strong association between suicide rates and provincial rates of low income, increased unemployment, and indigenous percentage of the population. Once again, firearms legislation had no associated beneficial effect on overall suicide and homicide rates. Prevalence of firearms ownership was not associated with suicide rates.
Failure of gun control
Canada’s firearms legislation is revealed to lack a solid foundation. Any apparent success is a mirage. The firearms regulations consist of an excessively complex bureaucracy, expensive as well as arbitrary, that cannot be shown to enhance public safety. Hunters and sport shooters do not require such a cumbersome regulatory system, and the laws do little or nothing to protect the public from violent criminals or protect desperate people from suicide.
The studies presented here provide convincing evidence that Canada’s gun laws do not reduce homicide or suicide. Each study was carefully designed to assess the possibility of a causal link between firearms legislation and homicide or suicide. Each study included only the most likely alternatives for explaining the trends in order to guard against spurious conclusions. Not one of them was able to find a statistically significant link between the firearms legislation and homicide or suicide rates. These results provide strong evidence that such a link does not exist.
The lack of a link between gun legislation and public safety would explain the government’s reluctance to conduct sound empirical studies, or, if there are such studies, to make their findings public. The Canadian government routinely denies there are any. But it is hard to believe that it has never examined the issue and in fact a few government studies are known to exist.
Access to Information requests uncovered one study conducted by Albina Guarnieri M.P. in 2004. But unfortunately, it remains under cabinet secrecy and will remain so for a few more years. Another study evaluating the 1977 and 1991 firearms legislation (ED1996-1e), conducted by the Policy Sector of the Department of Justice, received very limited circulation in 1996, but soon disappeared and is no longer available. This study was a methodologically respectable evaluation, but the government of the day wasn’t pleased with its findings that neither Bill C-51 in 1977 nor Bill C-17 in 1991 had a statistically significant effect on either homicide rates or suicide rates. So down the memory hole it went.
Why politicians pursue failed policies
If gun laws are not able to satisfy their ostensible purpose, why then does the government repeatedly advocate additional restrictions? Clearly winning electoral support is important. In 1995 Jean Chrétien was said to have stressed the strategic importance of the Liberals staking out more restrictive firearms controls policies than the Progressive Conservatives. The Liberals were concerned that the Progressive Conservatives would steal the urban women’s anti-gun vote having introduced Bill C-17 in 1991. Clearly firearms legislation is about vote-getting as much as it was ever about safety.
Focusing on guns rather than violent criminals lets Ottawa pose as a protector of public safety while maintaining progressive police and court policies. Such a strategy allows the government to court progressive activist groups and avoid charges of overt racism. Unfortunately, immigrant and indigenous communities suffer disproportionately from the failure of gun laws to reduce criminal violence.
Another factor is making promises of more gun laws as a red herring to divert attention from events that embarrass the government. Perhaps the current round of firearms restrictions is intended to divert attention from the Liberal government’s failure to successfully handle the Covid-19 pandemic or the continual high-level governmental chaos, beginning with the firing of Vice-Admiral Mark Norman for doing his job, followed by the resignation of Bill Morneau as Finance Minister, and the fallout from the SNC-Lavalin affair, the WE charity scandal, among many others, most recently, the constitutional embarrassment of asking Governor General Julie Payette to resign. Not to mention ongoing problems of mismanagement by RCMP Commissioner Brenda Lucki.
There are also public choice considerations. Once it comes into existence, any bureaucracy including the firearms bureaucracy becomes a self-perpetuating entity for reasons of financial and political self-interest. Thus when then-Prime Minister Harper scrapped the long-gun registry, he promised to maintain the federal facility at Miramichi. Political parties have long looked to government as a way to employ or promise to employ supporters and potential voters.
Finally, gun control lets advocates feel good about themselves. But they should find another less expensive and counterproductive way, because it doesn’t make Canadians safer.
Originally published in The Dorchester Review.
“The Failure of Gun Control (Again)”
Volume II, No. 1 Spring/Summer 2021, pp 25-32.