Murder, he wrote: Colby Cosh investigates Canada’s homicide files

If one were to choose a single core responsibility of the state, it would probably be the prevention of violence. Protecting people from homicide could not be more intimately related to the origins of, and the justification for, government. So how come we don’t talk much about how poorly or well we are doing at it? In the early 1960s, according to the Canadian Centre for Justice Statistics, it was typical for Canadian police to solve 90 to 95 per cent of all murders. The figures for recent years, after a long and steady decline, are generally below 80 per cent; in one year, 2008, the clearance rate dipped to slightly below 70 per cent.

Interactive graphic: Where Canadian criminals go to play

Numbers released in June by the CCJS show that Canadian investigators enjoyed a good performance in 2010 by recent standards, clearing 75.3 per cent of homicides. A homicide is normally “cleared” by laying a charge against a perpetrator, or by the mere identification of one for cases in which no arrest is possible (murder-suicides or self-defence killings, for example). An odd feature of the decline in homicide clearances is that it does not appear to bear any relationship to overall homicide rates, which peaked in the mid-1970s and have been dropping ever since. Police are simply solving slightly fewer of the homicides they are presented with every year, irrespective of how violent the social environment is.

This is not necessarily bad news. We want murder files to be closed by the police through legitimate means, with the suspects identified accurately. Perhaps the 95 per cent clearance rates of the mid-1960s reflect a pre-Charter of Rights era, in which it was easier to finagle a truncheon into the investigative process. But there is no sharp change in the data that one might associate with a more active, rights-oriented judiciary, and contemporary cops don’t like to lean on the Charter as an excuse.

“I don’t think the changes in the legal environment have hampered investigations,” says retired Toronto homicide detective Mark Mendelson. “Everyone’s accustomed to the idea of the Charter by now.” From Mendelson’s big-city vantage point, the explanation for declining clearance rates is simple. “In the ’60s, when people got killed they generally knew their killer,” he says. “You didn’t have bullets flying around the Eaton Centre. There has been a dramatic decline in the perceived value of a human life. In my early years, I had cases in which a guy might get kicked out of a bar and take a swing at a bouncer; now he might go get a gun and start shooting at doormen.”

It is the kind of thing coppers have probably been saying since classical Athens hired Scythian archers to handle crowd control. But the idea finds some support in the research. Tanya Trussler, a sociologist at Calgary’s Mount Royal University, is a specialist on the topic—“I have a head for numbers and a fascination with violence,” she notes wryly. In a paper published in the International Criminal Justice Review in summer 2010, she took a CCJS data set of all known Canadian homicides between 1991 and 2006 and performed a logistic regression to tease out the factors that made homicides easier or harder to clear. The data cover more than 11,000 killings.

One of Trussler’s aims was to test earlier sociological theories that police might pay less attention to particular kinds of underprivileged victims—women or racial minorities, for example. In fact, the homicide of a woman has 1.6 times the likelihood (expressed in odds) of being cleared as that of a man, and that of an Aboriginal victim has 1.8 times that of a non-Aboriginal one. Plug situational factors into the model, however, and the race and gender effects disappear. For persons killed at home versus those killed elsewhere, the odds ratio is a whopping 2.4. For rural homicides versus urban, it’s 1.9. And the method of killing makes a big difference. Compared to handgun homicides, long-gun killings have 2.2 times stronger odds of being cleared; beatings and stranglings, three times; and stabbings, 3.8 times.

Trussler’s figures seem to quantify what anyone might have suspected: up-close-and-personal killings performed in anger are easier to solve than ones perpetrated in cold blood, for business reasons or for contemplated revenge. Indeed, in cases where the CCJS records show that gang or drug-trade involvement was suspected, the likelihood of clearance is almost halved—even when the method of killing and the location are corrected for.

“Although the rate of homicide has been declining over the past 20 years quite significantly, the character has changed somewhat,” attests Neil Boyd, associate director of Simon Fraser University’s School of Criminology. “In the late ’80s, for example, we used to have 10 per cent of all our killings with handguns. Now that figure has inched up closer to 20 per cent. And we know of course that we’ve got relatively brazen gangland killings in many of the major cities in Canada. As a proportion, those kind of killings form a greater percentage of the total than they did in the past.”

That indicates that the police of our time have a tougher job than earlier generations did, although in Trussler’s model, all these extra factors do not quite cancel out the overall decline in clearances over time. They also do not entirely explain lower rates of homicide clearance in Quebec and British Columbia compared with the rest of the country. In general, those provinces are much worse at cracking open murders than Ontario and the Maritime provinces. And they are even further behind the Prairie provinces, which are, overall, the most homicidal part of Canada (and hardly free from gang activity or drug trafficking).

To be sure, if B.C. and Quebec have cleverer, better-organized gangs than everyone else—which seems possible, even likely—a logistic regression model that treats all gang murders the same would fail to capture that. (Regression is, in general, a finicky procedure.) The CCJS also notes that some homicides may go uncleared in B.C. and Quebec because those provinces require an extra layer of “prosecutorial screening” of evidence before charges are laid.

Trussler is now working with Calgary police—who have achieved both low homicide rates and high clearance rates, especially contrasted with other Western cities—to find out how the organization of a police force might influence clearances. “B.C.’s particular problem, if there is one, may conceivably be related to having seven different police forces in the lower mainland,” she speculates. “It’s also possible these provinces just don’t have the manpower they should.”