Today Fri Sat
It is forecast to be Chance of Rain at 11:00 PM EDT on July 31, 2014
Chance of Rain
It is forecast to be Clear at 11:00 PM EDT on August 01, 2014
It is forecast to be Partly Cloudy at 11:00 PM EDT on August 02, 2014
Partly Cloudy



Metrocide: A History of Violence

Is Toronto a dangerous city? And is it getting worse? This week for Metrocide, Torontoist is examining a sea of homicide data and trying to come up with conclusions based not in fear or fantasy but fact.
Photo from 1986 by .allen from the Torontoist Flickr Pool.
1981 was the first year for Canada’s Wonderland and NOW, the year of the Toronto bathhouse raids, and the year that Terry Fox died. That year, Toronto the (still) Good had 40 fewer homicides than there were in 2007—and the second-highest homicide rate of any year since.
In today’s edition of Metrocide, we’ll examine homicides and homicide rates, going back to 1981.

Notes on Statistics

For today and tomorrow, we’ve used both Statistics Canada and Toronto Police statistics. They don’t match, and that’s for one key reason: when it comes to homicides, both organizations define “Toronto” differently. Statistics Canada’s numbers for homicides in cities since 1981 are for those cities’ Census Metropolitan Areas (or CMAs)—so for Toronto, that means that Ajax, Aurora, Brampton, Markham, Pickering, Uxbridge, Milton, and Newmarket, among others, all count as part of Toronto (a map of Toronto’s CMA is here; Toronto’s CMA is in green). The Toronto Police boundaries are substantially smaller; their Toronto ends once you cross Steeles, for instance, or Highway 427. Toronto Police data pegged Toronto’s population as 2,724,784 in 2006; Statistics Canada counts 5,113,149 Torontonians in the same year.
Today and tomorrow, we’ll focus largely on StatsCan data, because, as the Toronto Police year-end statistical report from 2007 notes, “the statistics in this report count all offences reported to the police….a single incident reported to police may generate more than one offence.” Moreover: “it is not recommended that the crime statistics in this report be compared to crime statistics prepared by other agencies due to the fact that different methods of categorization, geographic, technical, data and time constraints may affect the outcome.” The reason we’re including Toronto Police data rather than abandoning it altogether is because, as that same report notes, “although comparing exact numbers is not recommended, the general trends and magnitude of change should be similar regardless of the counting method.” They’re also useful for examining homicides that occur in the smaller boundaries of Toronto—something we’ll be doing more of tomorrow when we examine homicides in central and downtown Toronto. Either way, take Toronto Police data with a grain of salt, but don’t rule it out altogether.
Also, though crime statistics are available for Toronto going back further than what we’ve used here, we’re only using numbers from 1981 on from Statistics Canada, and from the 90s on for Toronto Police. For Statistics Canada, the most recent data is the most reliable; we’ve spent some time looking through StatsCan volumes going back as far as the 1940s, but there seems to be no consistent treatment as far as we can see until 1981 for how to count and compare homicides in Toronto and other Canadian cities. (A StatsCan account executive advised us against using any homicide statistics before 1981 and comparing them to what we were using from that point on, and it’s probably wise to take their advice.) Second, for both Toronto Police and Statistics Canada, the recent statistics are the most readily accessible, meaning that anything we present can easily be reframed, reinterpreted, or reviewed. Statistics Canada’s data can be purchased from them for next to nothing ($6 for what we used below), or for more recent years compiled for free from yearly PDF reports, all from their website. Toronto Police data is available in annual statistical reports from 1997 on on their website—and some of those annual reports also contain data for previous years, which is why some charts we’ll have today and tomorrow go back earlier than 1997.
So, onto the data.
Overall, the number of homicides in Toronto does seem to be increasing according to Statistics Canada data. The average number of homicides from 1981–1990 was 67.2; from 1991–2000 was 80.3; and from 2001–2007 was 95. Determining any trend on a year-to-year basis, however, seems futile, since the number of homicides often rises and falls sporadically. For instance, according to StatsCan data, there were 103 homicides in Toronto in 1991, more than there would be for another fourteen years. Five years before that, in 1986, there were only 46 homicides, 15 fewer than there have been in any other year charted above. While Toronto Police reported 84 homicides in Toronto last year, if this year’s pace holds steady—which we’ll have more to say about on Friday—2008 will see a noticeable dip in the number of homicides recorded by year’s end.
What any chart that simply counts homicides ignores entirely is population. Because of Toronto’s growing population, even while the number of homicides appears to be inching up, the homicide rate has been perceptibly inching down. So while in 2007 Toronto had more homicides than it had had any other year going back to 1981—that year simultaneously had a homicide rate equal to or lower than the homicide rates in 1981, 1982, 1984, 1985, 1987, 1991, 1992, and 1994.
What the homicide rate chart shows that the homicide chart didn’t is that Toronto seems to have been more dangerous in the 1980s and early 1990s than it is now. For instance, according to Statistics Canada’s data, no year since 1994 has had a homicide rate above 2.0 per 100,000 people—but the rate hit or exceeded 2.0 per 100,000 eight times between 1981 and 1994. The average StatsCan homicide rate since 1981 is 1.89 per 100,000; in the past 12 years, only 2005 and 2007 have had rates higher than that.
In Toronto, is there anything that directly causes about as many deaths as homicides do? As it turns out, there is: motor vehicle collisions.
A fatal motor vehicle collision is defined by the Toronto Police as any death that results from a collision involving one or more motor vehicles in any way. Any car crash that results in the death of drivers or passengers counts, as would a fatal crash where a cyclist is hit by a car, or a motorcyclist crashes into a guardrail.
Over fifteen years, fatalities from motor vehicle collisions and homicides have been quite close in the numbers measured by Toronto Police: over that period, the average number of people killed a year in motor vehicle collisions has been 73, while the average number of homicides has been 63. Toronto Police statistics point to a conclusion that is at once obvious and underreported—that in Toronto, you are more likely to die because of a motor vehicle than because of a gun or knife (and that is excluding deaths caused by problems that cars indirectly contribute to, like respiratory diseases and certain kinds of cancer). And while the graph above shows the two statistics headed in different directions of late, there have been more traffic fatalities than there have been homicides this year.


Metropolitan Toronto homicides from 1981 to 2007 are from Statistics Canada. (Table 253-0004—Homicide survey, number and rates [per 100,000 population] of homicide victims, by census metropolitan area (CMA), annual, CANSIM [database].)
All Toronto Police statistics are from their year-end statistical reports.


  • kmaryan

    Nice article, but I thought it needed a bit of interpretation to put it in perspective. So, here goes…
    Although I couldn’t find comparable numbers for Toronto, for Canada wide numbers you can compare homicide rates to other causes of death here.
    The jist of it is that, while non one wants to live in a violent city and no one wants see anyone suffer from violence, you don’t need to be paranoid and start contemplating your mortality every time you step outside. You’re more likely to die by almost any other means. So don’t worry about jogging after dark, it’ll help stave things like heart attacks and lung disease which are two orders of magnitude more likely to be your cause of death.
    (Give or take the danger of other violent crimes, which I don’t mean to come across as insensitive to. I suspect these will be covered in later editions of this article series).

  • kmaryan

    Also useful in putting these numbers in perspective is where they happen. Homicides are not uniformly spread”>>spread throughout the city. The Toronto Star homicide map isn’t a lot of data to go by, but off hand it’s easy to see clusters in dense residential neighbourhoods or poorer areas. Whereas neighbourhoods that are a bit more upscale or consist of single family houses tend to be relatively murder free.

  • lisarr

    I think the traffic-related deaths could be under-reported.
    The recent death of a Toronto cyclist due to ‘dooring’ resulted in the driver apparently only being given a $110 ticket, perhaps some sort of parking infraction?
    If no criminal charges are laid, would the cyclist appear as a traffic-related death in official statistics?

  • David Topping

    kmaryan, tomorrow I’m breaking down the data a bit more to show downtown and central homicides—but you’re obviously right that homicides are not uniformly spread across the city (or even across “dense residential neighbourhoods or poorer areas”).
    lisarr, traffic-related deaths are traffic-related deaths, regardless of whether or not someone is charged or otherwise deemed responsible for them. The numbers above—for both motor vehicle collisions and homicides—do not reflect charges laid, but the actual number of deaths. The example you gave would indeed count as a fatality in 2008 statistics, even if the driver who opened their car door had no charges or fines imposed on them whatsoever.

  • matty

    This is pretty incredible Mr. Topping. Really good work.
    Oh and for some perspective in 1992 Chicago’s murder rate was 943…

  • bbpsi

    Thank you for including the per capita data.
    The media is constantly reporting the raw numbers like it means something — but Its clear from your graphs that the city is really about as safe as it’s always been (very, very safe).

  • uskyscraper

    I’m most interested in where homicides occur, and why Toronto has such a problem with public shootings.
    The odd thing about Toronto is that while the overall number of murders is quite low, the number that happen in central, public, tourist-friendly places is extremely high. A daylight shooting in midtown or downtown Manhattan is almost unheard of and club violence in the high-end entertainment districts there is rare, yet Toronto suffers murders on Yonge St, in the club district, and on other downtown streets on a regular basis. It’s very strange. Look at a homicide map of the city from CBC or The Star and you will see a U shape – concentrations in the poverty-stricken NW and NE corners, of course, but a suprising concentration of murders are in the compact downtown. This is very different from New York, where the most “downtown”, office and retail packed, tourist-friendly districts see the fewest homicides.
    What is going on with Toronto that causes this to happen? The public perception of out-of-control crime would certainly be different if the violence was not taking place in the city’s symbolic heart.

  • andrew

    How many “daylight” shootings downtown have there been? Most of them, at least as far as I know, seem to occur in the entertainment district at night. Because some jerk-offs like to carry guns. And then drink alcohol. And then get upset when people insult them, or brush past them wrong, or something.
    I don’t have the time to check but it seems to me most daylight shootings tend to be in the city’s symbolic beer gut – this ring around the city that we tend to regard as useless and try not to think about.

  • David Topping

    Andrew, I think I’d have to sit down and look at each individual death separately to determine that; neither the Toronto Police nor Statistics Canada seem to group any sort of crime by the time of day it occurs.
    But if you check out Toronto Police’s statistical reports for recent years and scroll to the very end of them, there is a map of all the shootings that took place in Toronto for that year (though not broken down into time of day). The majority of downtown shootings and majority of downtown homicides seem to typically occur east of Yonge, not west. (Yonge is the line that divides 51st and 52nd divisions.)
    Also, uskyscraper, the data I’ve gathered pretty much entirely refutes this claim, unless your definition of regular is different from mine:

    …while the overall number of murders is quite low, the number that happen in central, public, tourist-friendly places is extremely high….Toronto suffers murders on Yonge St, in the club district, and on other downtown streets on a regular basis.

    I’ll be looking at downtown and central homicides (not shootings, but homicides) later today.

  • awb

    Can I just be a jerk and point this out:
    Less refers to a general amount and fewer refers to a specific number of things. eg, Less traffic, fewer cars.
    So the second sentence should read: “That year, Toronto the (still) Good had 40 fewer homicides than in 2007″. Torontoist editors should catch that.

  • David Topping

    Fixed; that was my mistake.

  • uskyscraper

    I look forward to your analysis. What I mean by “extremely high” by the way is not relative to the overall homicide rate in the city or metro region as a whole, but relative to what happens in other cities. I could well be mistaken, but I think it is an abberation for homicides to occur evenly across a central/downtown area relative to areas of reduced socio-economic activity in the same city. i.e. you would anecdotally expect Bed-Stuy or the South Bronx to have far more murders than midtown Manhattan, and that is overwhelmingly in fact the case. There should, I think, be a doughnut hole in any city’s homicide map for the CBD, but in Toronto this hole seems to exist instead only in the well-to-do residential areas north of downtown. Therefore, too many homicides are taking place in central public places, relative to most other cities (SF seems to also be an exception).
    New York:

  • rek

    Canada: 605 homicides
    Toronto: 70 homicides
    US cities of comparable population (2006):
    Philadelphia: 406 homicides
    Houston: 376 homicides
    Chicago: 466 homicides
    Dallas: 187 homicides
    We have it pretty good here.

  • kmaryan

    Good point T. Rek on the comparison to US cities. To rephrase those numbers in terms of murder rate:
    Philadelphia 27 per 100,000 people
    Chicago 15
    Washington DC 50.82
    New York 9.32
    These varied a great deal over the years, Houston saw over 800 murders in the early 80s and is down quite a bit since then.
    Across the entire US the number is around 7 per 100000.
    I found some numbers for global cities for comparison:
    Moscow 18.2
    Belfast 5.2
    Toronto 3.1
    London 2.4
    Paris 2.2
    Tokyo 1.2
    Interesting note, guns are illegal in Tokyo (all guns) and London (handguns).

  • kmaryan

    Are there really that many shootings in public places? In recent memory, there have been a handful in the entertainment district and a couple high profile ones on Yonge. And while these are all very visible crimes, I don’t think there have been all that many. Anyone have any numbers to back this up?

  • David Topping

    As I’ve mentioned a few times, today’s post will look at downtown and central homicides. Patience!
    And Thursday’s post will look at other cities, but it’s a reasonable conclusion to draw that, compared to American cities, Toronto is enormously safe; and compared to other Canadian cities, Toronto is pretty safe.

  • uskyscraper

    Sorry that we keep jumping the gun, David (bad pun). I won’t post again until after your followup on the topic.
    Yes, US cities have far higher rates _overall_ but my interest is really aimed at what happens in the tourist/business cores of cities frequented by regular citizens. It’s sad, but for most people it really doesn’t matter how many homicides there are in Malvern or Jane-Finch or how many gangsters get shot up in back alleys behind housing projects. What they want to know is “am I safe if I go to the Eaton Centre/ride the TTC/go out to a bar”? This is where the Toronto vs. US cities breaks down a bit – Midtown Manhattan (MTS/MTN precincts) had 1 murder last year, while the equivalent area of downtown Toronto had about 5 – and I think some core-to-core comparisons may have some merit. If Toronto can still be shown to be safer, it will be a boon to tourism and public confidence. If not, then the city has a glass house to rebuild.

  • rek

    I don’t think you can make that sort of comparison in a vacuum, usky.

  • McKingford

    I would also point out that not all homicides are crimes.

  • David Topping

    The homicides that Statistics Canada counts all are, McKingford. Their glossary of terms [PDF] says:

    A homicide occurs when a person directly or indirectly, by any means, causes the death of a human being. Homicide is either culpable (murder, manslaughter or infanticide) or non-culpable (not an offence and, therefore, not included in the Homicide Survey). Deaths caused by criminal negligence, suicide and accidental or justifiable homicide (e.g. self-defence) are not included.

  • David Topping

    And ditto for the Toronto Police:

    The Homicide or Murder category includes first and second degree murder, and manslaughter. A Homicide or Murder occurs when a person directly or indirectly, by any means, causes the death of another human being. Deaths caused by criminal negligence, suicide and accidental or justifiable homicide (e.g. self-defence) are not included.

    The FBI data I looked at in today’s post is counted the same way.
    Not all homicides are crimes, true—but all the homicides included here are.

  • McKingford

    The police definition does not contemplate the small, but non-trivial, number of charges which are successfully defended at trial as being self-defence or accidental. In short, what the police consider to be a homicide remains classified as a homicide regardless of what a court ultimately rules. I know this for a fact, having represented someone charged with murder, who was acquitted at trial on the basis of self-defence; the case remains, for police purposes, a “homicide” (which is definitionally correct) even though a jury found that no crime occurred.
    Another example of this would be the Tyson Talbot case.

  • http://undefined lewyn

    I realize I’m a year late to the party but one thing that surprised me is that even the bad neighborhoods aren’t too bad by US standards. For example, the area just south of Jane-Finch had 11 murders in 2005-08, or 2.7 per year. Given a population of 30,000 according to the city that’s 9 per 100,000- way less than many major American cities (as shown by posts above). Some of the other bad neighborhoods like Parkdale, Regent Park etc average a murder a year, which is in the same ballpark.