Monday, February 28, 2011

Crime, Part III: underreported crime

Last summer, Stockwell Day made the mistake of seeming to say that the crime rate wasn't really falling it was rising, because so many crimes were going unreported. We all had a good laugh at this -- how could the minister possibly know how many crimes go unreported to the police? -- until it was realized that Statscan does, in fact, take a very accurate count, every five years. Here's the most recent study, "Criminal Victimization in Canada 2009."

Now, I don't want to let Stock off the hook on this one: crime rates in Canada are falling. (Though as my earlier posts point out, they're falling from such high levels that our crime rates are still far, far higher than they were back in the early 1960s.) And it was a bit of cosmic justice to see the government decimating and decimating the census and belittling Statscan one day, and then being forced to lean on it the next.

But if we leave behind the politics and look at policy, the truth is, as Statscan has shown, that a lot of crime goes unreported: according to the Statscan study, in 1999, only 37% of criminal incidents were reported to police. By 2004, that figured had fallen to 34%. And in 2009, it fell to 31%. That's not a Conservative fact or a Liberal fact, just a fact.

The assumption of many critics is that the low reporting rate is due to the crimes in question being so minor. Maybe. But I started thinking about my own experience, and I realized that I've been the victim of a "minor" crime four times in the last decade and a half. And minor is very much relative, and in the mind of the beholder. And in the "broken window" theory, it is minor crimes that are the most important crimes to prevent, because they lead to an unsafe atmosphere that leads to more crime, and more serious crimes.

So, the four times I was the victim of a crime: My car was once broken in to on the street, it was twice broken in to in my building's underground garage in, and once my car was broken in to in broad daylight, while parked in the employee parking lot at the Globe. Each time, the cost to me was around $500, plus time and aggravation. Each time, I had to replace a smashed window, and each time, something in the car was stolen: once it was running shoes and gym clothes; another time, I had left a pair of unassembled Ikea chairs in the car, one of which was taken. And so on.

And I never reported any of those crimes to the police. Why not? Because I assumed that they wouldn't have been able to do anything. And, given the way our car insurance rates work, I feared that I could be doubly punished through higher insurance rates. I might have been robbed twice, once illegally and once legally. And so, like most people who are the victims of "minor" crimes, I shut up and ate the cost.

So how minor were these crimes? Let's say that the cost to me was $500 per theft, times 4 thefts, or $2,000 in total, over approximately 15 years. Total cost: call it $130/year. And on top of that, there was a cost to me in wasted time (several hours per incident) plus a hard to quantify cost in terms of aggravation and inconvenience each time my car was burgled, plus a general feeling of insecurity.

How much would I have been willing to pay to avoid all of the above? Or at least reduce the odds of being robbed? How much would you pay? If there were some insurance or policy or service that you could buy that would eliminate or at least reduce the risk of a "minor" but costly and insecurity-breeding theft, how much would you be willing to pay for that? How about, say, $200? If every Canadian coughed up $200, that would be just shy of $7 billion a year.

Keep those numbers in mind for a second.

If you're going to criticize the Tory more-criminals-in-prison-for-longer proposals, the logical place (as opposed to the politically expedient place) to start is not with the argument that the policy is too expensive, but rather with the argument that it won't reduce crime -- that it will be an expense that delivers little or no returns. Which might well be the case: After all, putting more small time drug offenders and white collars criminals behind bars isn't likely to make your neighbourhood safer. But if -- if -- the Tory strategy could reduce crime, its costs, estimated as running into the high-single digit billions per year by the Parliamentary Budget Officer, vs. a few hundred million by the government, would be counterbalanced by some benefits.

Has the government made the case? No. It doesn't appear to be even trying. But that question -- the question of costs vs. benefits -- is the question it has to answer. And the counter-argument -- that the policy's costs will outweigh its benefits -- is the one the critics should be making.

The more-criminals-behind-bars-for-longer policy will be expensive: that's kind of a given. But if the government can make the case that its an actual crime reduction strategy, that it will deliver returns, then many Canadians will support it.

The interesting thing to note is that Freakonomics professor Steven Levitt's research on all of this, which I referenced in my Globe article, suggests that higher incarceration rates can reduce crime -- or at least that, in the very particular case of the ultra-high crime US in the 1990s, crime levels were brought down in part by high and rising incarceration rates.

So does any of that apply at all to the Canadian context? I don't know. The government might want to think about making its case, and making in on this cost-benefit basis. Before you spend the money, show us the evidence.

More crime rate surprises, Part II

For anyone who wants to dig a big deeper into the crime stats I mentioned in my last blog post/Globe article, here's the background data.
Crime has been going down, across Canada and in most categories, since the 1990s. However, if you take a longer historical perspective, our current crime rates are actually quite high.
Statscan has been compiling the crime rate (measured as crimes per 100,000 people) and violent crime since the early 1960s. What's more, to allow for comparisons over time, such as I made in my article, Statscan makes adjustments to the historical data to account for the introduction of new crimes, or the movement of some crimes from the non-violent category to the violent category.

Here's the crime rate, 1962-2007, as calculated by Statscan.

Crime is "falling", but it's falling from a historically very high level. And it's nowhere near back to the low levels that were once the norm. It's sort of like unemployment in Canada, circa 2011: falling, but still historically high.

The crime rate -- what Statscan calls the "total criminal code (excluding traffic) crime rate" went from 2,771 per 100,000 in 1962 to a peak of 10,342 in 1991. It then fell, as this chart shows, to 6,899 in 2007. I called Statscan and got the updated numbers: crime rate has continued to fall over past two years, and in 2009 the rate was 6,406.

In other words, the crime rate in Canada roughly quadrupled between 1962 and 1991. And then it fell by 40%. But even after than fall, Canada's crime rate is more than double what it was in 1962.

And violent crime? The story is worse.**

 In 1962, there were 221 violent crimes per 100,000. Violent crime climbed steadily until 1992, with the rate more than quadrupling to 1,084 per 100,000 in 1992. And then, violent crime started to fall, though not as sharply as non-violent crime. By 2009, Canada experienced 1,314 violent crimes per 100,000 -- but to allow for the changes in the definition and measurement of crime that have take place since 1962, Statscan adjusts that figure to a 2009 violent crime rate of 920 per 100,000.

In other words, even adjusting for changes in the law and improvements in record-keeping by law enforcement, Canada's violent crime rate is up 316% since 1962.

Another thing I looked at: murder rates, Canada vs. the US. This was a big surprise to me. I knew that crime had been falling sharply since the early 1990s in the US, and that all US big cities were experiencing remarkably reduced murder rates. New York City, for example, has seen its murder rate fall by about 80%. What I hadn't counted on was that US murder rates could have fallen so far, so fast, that many US cities are now safer than the least safe Canadian cities -- all of which are in Western Canada.

For Canadian murder rates, see Statscan's annual tally in each of Canada's census metropolitan areas (CMAs), here:

For comparable US figures, I used the FBI crime database, and looked at what the US calls Metropolitan Statistical Areas (MSAs), which like Canada's CMAs include both the city and its suburbs. (And in the case of really big cities, the exburbs too).

The results are kind of shocking. Go through the US database for yourself. Almost every US city in the Northwestern US states -- scores of communities in 11 states, in a band stretching from Wisconsin/Minnesota/Iowa to Washington/Oregon -- has a lower murder rate that Winnipeg, Edmonton or Vancouver. Urban Western Canada is, if we take murder as our measure, more dangerous than the urban Northwestern US states.

This data upends a few stereotypes and shoots a few sacred cows. The lawless American West? The peaceable land North of 49, kept safe by Sam Steele and his Mountie friends? Not exactly. Not anymore.

One final note, which I didn't have a chance go get into in my Globe article: crime rates in Western Canada are double those Ontario and Quebec -- but crime in the North is off the charts. According to Statscan's annual Crime Severity Index, the most violent jurisdiction, Nunavut, had 11 times as much violent crime as the safest place, Prince Edward Island.

** Correction: In original version I had accidentally attached the wrong table. That's been fixed.

Sunday, February 27, 2011

Surprise: crime in Canada is surprisingly high

In my article in Saturday's Globe, I looked at crime stats. Crime is falling, right?
Yeah, well, about that.... somehow, nobody seems to have noticed that Canada's violent crime rate is 4X what it was in early 1960s.
And Winnipeg has a higher murder rate than Minneapolis. Or Boston. Or NY.
And Vancouver's murder rate is the same as Seattle, and higher than Portland.

I'm surprised. You're surprised. We should all be surprised, given that the main argument against the Tory get-tough-on-crime bills has not been "get-tough-on-crime won't fix this problem" -- a plausible argument, and one I'm partly (maybe even mostly) in agreement with -- but rather "crime problem? What crime problem?"

Link to my Globe story here, or read it below:

It wasn't long after Shawn Beauchamp was released from prison, having served a 40-month term for breaking and entering, assault and robbery, that he was shot to death in Winnipeg's North End. Mr. Beauchamp, 26, was a former gang member apparently trying to go straight. “He wanted to quit the gang,” his cousin told the CBC. “But the only way you get out of the gang is in a box.”
Mr. Beauchamp was Winnipeg's first homicide victim of 2009. That year, Manitoba's capital would go on to claim the title as Canada's deadliest big city, with a record 32 homicides in the urban area and its suburbs. It also took the crown in 2008, and though the rate dropped last year, the city is already on track to eclipse its previous record: Six people have been killed so far in 2011.
In the debate over the federal government's tough-on-crime bills, the standard response of opponents has been to point to one fact: Crime in Canada is falling. But averages hide a lot of variations. A deeper look at the country's crime statistics reveals a more complex picture – and may help to explain the appeal that a tough-on-crime stance has for the Conservatives and their core voters.
Where is crime highest in Canada? In the West, the Tory heartland. Statistics Canada's crime-severity index, a measure combining the number of offences and their seriousness, shows crime still high in the West. Saskatchewan, the province with the highest overall index score, was roughly twice as crime-ridden as Ontario or Quebec in 2009. Manitoba, which topped the index, had more than twice as much violent crime as Ontario or Quebec.
And the discrepancy is not just East-West. It is also North-South. Crime rates in the United States have been plummeting for two decades, dropping far faster than in Canada. In fact, crime has fallen to such a degree in the U.S. that a comparison of data for the most recent year available, 2009, reveals that some cities in Western Canada are now more dangerous than many in the U.S.
For example, Winnipeg's homicide rate is more than double that of the nearest big U.S. city, Minneapolis-St. Paul. Its rate is higher than Boston's. Higher than San Diego's. It's even slightly higher than New York's, if you include the Big Apple's suburbs.
And Winnipeg is not alone. You're more likely to be murdered in Edmonton than in the much larger cities of San Jose, Calif., Salt Lake City, Utah, or Austin, Tex. The homicide rate in Vancouver is higher than in Portland, Ore., and the same as in neighbouring Seattle. In 2009, Edmonton, Winnipeg and Vancouver had higher homicide rates than almost any city in the Northwestern U.S.
Crime rates are much lower east of Manitoba and, as a result, relative to population, Canada has only about a third as many homicides as the United States. (Homicide is often used as a proxy for crime when doing comparisons because, unlike with other crimes, there are no problems of reporting, definitions or fudging of data.) This is another argument that critics of the Conservative crime strategy repeatedly make: Over all, crime rates are still lower in Canada than in the U.S.
The U.S. comparison may be flattering to Canada, but it's not the ideal benchmark. Why not? Because the U.S. is still the most crime-ridden country in the developed world. Aiming no higher is setting the bar pretty low. (And, as Western Canada shows, sometimes we're not even clearing that hurdle.)
The homicide rate in Canada is higher than that of most industrialized countries, often by wide margins. We have 39 per cent more homicides per capita than Australia, 41 per cent more than the United Kingdom, 50 per cent more than Poland, twice as many as Sweden, three times as many as Norway and nearly four times as many as Japan.
And Canada has no reason to be smug about crime when compared with what could be the best standard of all: Canada. According to Statscan, the crime rate today is more than double what it was in the early 1960s. And the rate of violent crime is more than four times higher. The homicide rate? About a third higher. Yes, crime has been on the decline since the 1990s. But that's only after shooting up over the previous three decades.
All of which doesn't mean that the crime-reduction measures proposed by the Tories are necessarily the right ones. But anyone who claims that the government is going after a problem that doesn't exist – anyone who says, ‘Move along, no crime to see here' – isn't telling the truth.
Another argument you hear in the Canadian crime debate: The U.S. experience proves incarceration doesn't reduce crime. A generation ago, the U.S. decided to fight crime by imprisoning more criminals, ending up with more people behind bars than almost any country on Earth. Yet it still has high rates of violent crime. Ergo, imprisonment doesn't work. If anything, imprisonment increases crime. Right?
Maybe not. The American strategy has been exceptionally expensive and has had deeply unpleasant side effects, but when it comes to reducing crime, it appears to have worked, at least somewhat. According to University of Chicago economist Steven Levitt, who spent much of his time analyzing crime data before he became famous for co-writing Freakonomics, the sharp increase in the number of Americans in prison and the sharp drop in U.S. crime are closely related. His research says the former helped cause the latter by deterrence and through “incapacitation,” a creepy technical term for the fact that someone cannot rob your house while he is locked up in the Big House.
In a 2004 study, Prof. Levitt concluded that the growing number of Americans behind bars was responsible for about one-third of the enormous drop in crime in the 1990s. (Other factors he identified were the hiring of more police, the end of the crack epidemic and, controversially, the legalization of abortion – which ensured that fewer unwanted children were born.)
Many American scholars disagree with Prof. Levitt, and there's a real debate over why crime has fallen in the U.S. It's exactly the kind of serious debate that we should have in Canada. The government hasn't exactly been promoting thoughtful discussion, but neither have its “crime-what-crime?” critics.
Canada's crime problem is hardly an out-of-control epidemic. It's not even a crisis, except in a few places such as Winnipeg. It's a problem that has been trending, slowly, in the right direction. But it's a bit premature to be declaring victory. More than two million crimes were reported to the police in 2009, including 443,000 violent crimes, 205,000 break-ins and 108,000 motor-vehicle thefts. And those figures greatly understate the problem: According to Statscan's most recent survey of victims, only 31 per cent of criminal incidents were reported to police in 2009, down from 37 per cent in 1999.
The numbers sure don't sound like a clarion call to complacency.
Tony Keller, a former editorial page editor of The Globe and Mail, is a Toronto writer.