We’ve Talked About Homicide In Chicago At Least One Million Times But I Don’t Think This Has Come Up

Here are two maps:


1990-1993                                       2008-2011

Hom90 hom20

Like the captions say, the one on the left shows homicide rates by police district in the early 90s, when crime was at its peak in Chicago, and the one on the right shows the same thing, but about two decades later.* The areas in dark green are the safest; the ones in dark pink are the most dangerous. The colors are calibrated so that green areas are safer than average for the early 90s, and pink ones are more dangerous than average for the early 90s. The 2008-2011 map keeps the same calibration: green is safe compared to the early 90s, so that you can see change in the levels of violence over time.

And, indeed, the first thing that jumps out from these maps is that there’s way more green nowadays, and it tends to be darker. The city is way safer! Some areas we might consider a bit dicey today – like, say, the Lawndale/Little Village area – actually register as light green, meaning that by early 90s standards, they would be considered relatively safe.

[For those of you craving numbers, the murder rate averaged 30 per 100k during the first period, and 17 per 100k during the second, a decline of nearly 50%.]

Of course, the other thing we notice is that there are some very distinct patterns to safety. These maps are breaking exactly no news by indicating that the more dangerous parts of the city are on the West and South Sides, but it is striking, I think, to see that nowadays, basically the entire North Side is the darkest green, which translates to a homicide rate of less than 6 per 100k. In fact, the  dark-green part of the city has a murder rate of 3.3 per 100k.

Three point three. In New York City, which is constantly (and mostly correctly) being held up as proof that urban safety miracles can happen in America, it’s 6.3. Toronto, which as far as North American big cities go occupies a fairy tale land where no one hurts anybody, had a homicide rate of 3.3 per 100k as recently as 2007. The North Side is unbelievably safe, at least as far as murder goes.

But there are none of the darkest green on the West or South Sides. There’s actually a fair amount of pink, meaning places that are relatively dangerous even by the terrifying standards of the early 90s.

This raises a question: Has the great Crime Decline benefited the whole city equally? Are the South and West Sides still relatively dangerous because they started from such a bad place, or because they haven’t seen nearly as much of a decline as the North Side has?

Here is the answer in another map:




The areas in darkest green saw the greatest decline; red means the murder rate actually increased.

So: Yes, the great Crime Decline is a fickle thing. The North Side saw huge decreases (in Rogers Park, it was over 80%) pretty much everywhere; the few areas that are lighter green were the safest in the city to begin with. The parts of the South and West Sides closest to downtown – Bronzeville, the West Loop, Pilsen, etc. – got a lot safer. But most of the rest actually got worse, including some neighborhoods that were already among the most dangerous in the city, like Englewood and Garfield Park.

This is a complicated state of affairs, and probably goes at least part of the way to explaining why, in the face of a 50% decrease in homicides citywide over the last two decades, many people persist in believing that the opposite is true: because in their neighborhoods, it is. It’s a dynamic that defies an easy narrative, and makes me slightly less angry (though only slightly) at all those journalists who have written in the last year or two about murder in Chicago without mentioning that the city is, in fact, safer on the whole than it has been in fifty years.

Here is one final pair of maps:


1990-1993                                       2008-2011



This is slightly less intuitive. These maps show the how the homicide rate in any given police district compares to the citywide average, using ratios; for example, if the homicide rate in West Town is 10 per 100k, and citywide it’s 5 per 100k, West Town’s ratio is 2 to 1. If West Town were 2.5 per 100k, its ratio would be 0.5 to 1. (Obviously the numbers in these examples are made up.) Blue areas have ratios below 1, and so are relatively safe; red ones above 1, and are relatively dangerous.

With the help of these maps, I’m going to ignore what I said about all this defying an easy narrative, and try to supply one: Over the last twenty years, at the same time as overall crime has declined, the inequality of violence in Chicago has skyrocketed. The pattern of what’s blue and what’s red in each map is mostly the same; I count only three out of twenty-five districts that switched from one color to another. But the colors are much darker in the 2000s than they were in the 1990s. There have always been safer and more dangerous areas here, as there are everywhere; but the gap between them is way, way bigger now than it used to be.

Numbers will help this case. Imagine that for each of these two time periods, we cut the city into equal thirds: one contains the most dangerous neighborhoods; another, the safest; and the last, everything else. In the early 90s, the most dangerous third of the city had about six times as many murders as the safest third. By the late 2000s, the most dangerous part of the city had nearly fifteen times more homicides than the safest third.

In addition, here are two charts:



The divergence is self-evident. The early 90s look very roughly like a normal curve: most neighborhoods are in the middle, and there’s a clear, if slightly bumpy, slope down towards the extremes.

Today, any semblance of a normal curve has been annihilated. Or, actually, that’s not quite right. Now it looks like there might be two completely separate normal curves, one with a peak at 0.2-0.4, and the other peaking at 3.1-4. Plus a few guys who got lost in the middle.

I suppose there are many, many things that one might say about what this means, but here’s the bottom line: The disadvantages and tragedies that people in “dangerous” neighborhoods experience are both absolute and relative. The death of an innocent person** is an indescribable loss no matter what. And, on that count, things are somewhat better for Chicago’s most violent areas: the homicide rate for the most dangerous third of the city declined from 51 to 39 per 100k in the time period we’ve looked at here. That is a real accomplishment, and hundreds, if not thousands, of people are still with their families and friends because of it.

But in other ways, it does matter if other parts of the city are getting safer much, much faster. When people weigh safety in their decisions about where to live, they do so by comparing: How much safety am I gaining by living in one neighborhood versus another? The same is true of entrepreneurs considering where to open their next business. The same is true of tourists looking to explore the city. The same is true of locals looking to travel to another neighborhood to eat out or go shopping.

On every one of those counts, the disadvantages that are accruing to already-disadvantaged neighborhoods in terms of lost population, investment, and connections to the rest of the city are now much more severe. The hurdles are that much higher.

That’s bad for those physical neighborhoods. It’s also terrible for the people who have good reasons to live there, like social networks, nearby family, or the affordability of real estate.

Because I don’t have the data in front of me, but who would doubt that over these same twenty years, there has also been a growing gap between how much it costs to live on the safe North Side compared to the more dangerous parts of the South and West Sides? Who would doubt that, as the North Side reaches Toronto-level peacefulness, the cost of rent has greatly diminished the number of apartments there affordable to the poor and working class?

In other words, just as the stakes have been tripled as to whether you live in Relatively Safe Chicago or Relatively Dangerous Chicago, it has become much, much harder to establish yourself on the winning side.

So: Next time you hear someone talking about “record violence” in the city, tell them that actually, murders are down almost 50% from twenty years ago. And then tell them that what’s really alarming is murder inequality.

* Why does this data end in 2011? Because I made these maps using data from the Chicago Police Department annual reports, which are available online, and which only broke down crimes by police district in the 1990s. In 2012, the police district boundaries changed, making it not quite an apples-to-apples comparison to prior years. Maybe somewhere data exists by Community Area for the early 90s, and then I could redo all of this.

** And I think reporting like that done by This American Life at Harper High in Englewood ought to challenge conventional middle-class ideas about “innocence” in the ghetto. It is very easy for those who don’t live in the neighborhood to talk about “thugs” and “gangsters” getting what they deserve. It is also very cruel, and very naive about what exactly “gangs” are, and what kind of people join one, and how, and why.

20 thoughts on “We’ve Talked About Homicide In Chicago At Least One Million Times But I Don’t Think This Has Come Up

  1. This seems informative, but hard to follow because the captions are missing. Would it be possible to include them?

  2. I knew that I wasn’t hearing gunfire for days at a time (unusual) and I wasn’t hearing my neighborhood named in the weekly shooting/dead count. So
    I’m feeling a little optimistic.

  3. Compare the poverty levels and the economic investment the city makes to each community.. its already obvious the those red and pink areas are the ones with the closing schools..

    1. Let’s all repeat, “Correlation is not causation…correlation is not causation…correlation is not causation…”

      1. Correlation can be causation, though. It is so easy to dismiss arguments for presenting definitive causative links between correlated statistics, but what exactly does that accomplish? We should not assume the correlation is the causation, but it is hardly requires a sociological study to see how school closings would increase gang membership.

  4. This is actually one case where I think the raw number of homicides, as opposed to the “rate,” would be useful. The precincts have, maybe, tens of thousands to less than 200k in population. Any change in that population is going to move the ratio rather quickly. Showing raw numbers along with actual population change will give a more accurate picture of whether crime is increasing or just that good people are being chased out of the worst neighborhoods.

    1. I’ve seen this critique a number of times now, and I have to admit I’m a little confused by it. I understand the distinction, obviously, between a crime rate increasing because the absolute number of crimes has increased, and it increasing because the population has fallen while the absolute number of crimes has remained the same, but I don’t see how this changes the fact that *for the people who live in that community*, there are now more crimes per person, and they are, as a result, statistically more likely to be, or be close to, a victim. Isn’t that what we mean when we call a neighborhood “dangerous”?

      You’re definitely right that in relatively small areas, like these police districts, small year-to-year changes can produce large swings in the homicide rate. But that’s why I used four-year averages, which cancel out *huge* amounts of variation. Not perfect, but much better.

      And I’m not sure I understand the distinction you’re making in the last sentence. If “good people” – ie, non-criminals – are leaving a neighborhood, and a higher proportion (although still, it has to be noted, a very small one) of the residents are, as a result, practicing criminals, again, how does that not result in a more dangerous neighborhood?

    2. No homicide statistic is more accurate in defining whether a geographic area is very safe or very dangerous and all else in-between than the ratio of homicides to population. This is true for a city block, a census tract, a police beat, a neighborhood, a community area, a police district or any other geographically division one uses. In other words, the crime rate per capita is the most accurate crime statistic for defining the actual crime rates of geographic areas.

  5. I would be willing to bet that the closing of the housing projects (e.g., Cabrini, Taylor) had a big influence on these stats. I suspect if you followed the migration of folks out of these historic high crime projects you would see crime rates spike in areas where folks from the projects moved and a similar reduction in the areas where the projects used to be located.

    1. 100% I’m with you on that bet. I have lived in the almost laughably safe remains of The Cabrini Green footprint (Clybourn/Halsted) for over 10 yrs, and have seen the neighborhood stay gentrified (legitimately) with mixed housing and great results, but the “worst of the worst” of the Cabrini denizens ended up on a Trail of Tears to the West and South sides. Now they are fighting and battling over street corners and streets for their respective drug lords in these new territories in Austin and Englewood. The decrease in my near north crime and increase in “theirs” is unmistakably a result of the destruction and removal of those housing projects. I’m not sure why there are statistical arguments circling around this elephant in the room.

  6. Finally. Thank you for pointing out what someone should have long ago. I feel like I am talking on deaf ears when I point this out. Now, I can refer them to your article.

  7. “Toronto, which as far as North American big cities go occupies a fairy tale land where no one hurts anybody, had a homicide rate of 3.3 per 100k as recently as 2007.”

    Why are you using Toronto’s 2007 murder rate for comparison? That was Toronto’s all-time murder peak. The rate has dropped precipitously since then. The last couple years it was around 1.8-1.9 per 100,000.

  8. New Yorker here. There seems to be squarish district in the middle, towards the bottom, that always seems to wind up in last place. Can you identify the district, and maybe speculate why it does so poorly?

    1. I think you’re referring to Englewood. It’s one of the poorest, most institutionally isolated neighborhoods in the city, and has been for a while. That said, there are some pretty cool things happening there – locals have organized a 5k run that got a lot of press, opened a really nice new cafe, started a huge greenhouse farm operation, attracted a city college with a major culinary school, started a local youth coding training program, and earlier this year, Whole Foods announced they were building a store at the heart of the neighborhood, near an L station.

  9. Great read, and great comments, thank you!
    And, as we read this and know the answers, we have a governor and mayor that keep unequally slicing school budgets and programs designed to help young people have real choices (After School Matters, Thresholds). Can you consider adding a map correlating $ school funding $ related to violence? Maybe comparing Winnetka schools, where our governor lives, or even U of C Lab School Tuition (half day Nursery $22,000) vs where our mayor sends his children vs where I teach on the far Southeast side of Chicago? The state sends almost $20,000 per student to Winnetka per pupil and 1/3 of that to the high poverty schools. http://www.ilraiseyourhand.org/statefunding

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s