Thursday, June 20, 2013

Chicago violence really down? Not quite

Apart from the Blackhawks' run at the Stanley Cup, Chicago has no bigger news story than its supposedly declining homicide rate. Aggressive Chicago Police Department strategies have cut down homicides by 34% from 2012 totals. At least, so the CPD would have us believe. Numerous national news outlets reported on the drop over the past few weeks, lauding Mayor Emanuel and Superintendent McCarthy for their tenacity and ingenuity in handling Chicago's violence. But as any longtime, or even transient, Chicago resident knows, our city's crime has a nasty habit of defying interventions and maintaining old trends. With daily headlines reminding us of the 1 or 2 dead and 10 others wounded, two such stories the last week alone, is it possible that violence has really declined as much as the CPD and the City would have us believe? Or is there another side to the story that the papers and police aren't talking about?

In this article, I offer an in-depth analysis of Chicago's 2013 violence drop. As is often the case with most social science statistics, there is both fact and fiction to the CPD reports. Although citywide homicides and shootings are indeed down, the trends vary widely from neighborhood to neighborhood. As a result of CPD strategies (and likely other factors), some communities have experienced precipitous declines in area violence - the infamous Englewood and West Englewood, for example. But other communities have not seen their violence rates move at all, such as North Lawndale and Humboldt Park. By focusing on some historically high-crime neighborhoods, I show the divergent trends of 2013 crime, demonstrating that a citywide crime drop does not necessarily mean reduced violence for all of the residents that live in violent communities. This will help explain how crime can be both down across Chicago while we simultaneously have Saturdays where 6 are killed and 26 others wounded in less than 12 hours.

All crime data is taken from the CPD Clear system, filtered by description and community area. The brunt of this study will take the form of time series analysis, looking at crime trends over time. It is important to offer a few words of caution when using any time series evaluation. Tracking any real world phenomenon over time is fraught with statistical peril. Wikipedia-statisticians across the internet are always keen to point out that correlation does not necessarily imply causation. In something as complex as crime rates, that is even more true than in other areas of data analysis. An array of factors, including weather, population, and unemployment all contribute to crime rate fluctuations. In Chicago especially, temperature and precipitation have a huge effect on crime rates.

But temperature and rain/snowfall are constant across the city. They explain citywide crime but not community-specific incidents. So when some neighborhoods enjoy crime declines while others suffer from violence spikes, we need to turn to something else to explain the change. This is especially important if crime was on an upward trajectory before reversing its course. In this case, a police intervention seems the likeliest candidate. The increased police presence and focused police strategies are notable factors that are common to all of the neighborhoods in which crime fell, and absent from all the neighborhoods in which crime increased.

Both Chicagoans and non-residents alike have probably heard of the notorious Englewood and West Englewood neighborhoods. For decades, these paired community areas have exemplified the violence, poverty, and inequality of Chicago's South and West Sides. They have also accounted for a disproportionate share of the city's violence. In 2013, the This American Life radio show, a product of WBEZ, did a two part story on West Englewood's Harper High School. The story received national acclaim, with thousands of listeners tuning in to the experience of the young men and women of one of Chicago's worst neighborhoods. This nationwide publicity, coupled with crime statistics from previous years, made these neighborhoods the perfect testing grounds for Garry McCarthy's strategies: foot patrols, overtime authorization, gang audits, and a variety of other programs. At the core of these deployments was an old police adage, one put into practice in NYC: more cops means less crime. It's a theory supported both by common sense and the evidence. 

As Frank Zimring explained in his study of the New York City crime drop of the 1990s, more police do, in fact, lead to less crime. Although there are other unintended consequences of additional officers and increased patrols (e.g. discrimination, clogged courts, increased costs, etc.), the principal effect of the policy is to decrease crime through police presence, rapid apprehension of criminals, and deterrence. More police officers taking to the streets was one of the chief causes of the NYC crime drop; from early 1990 through the mid 2000s, the City saw an 80% reduction in felony rates. A comparable police-saturation program in a violent Pittsburgh neighborhood led to a 34% reduction in shots-fired and a 71% decrease in area gunshot injuries.

McCarthy expected similar successes in the targeted Chicago neighborhoods. By deploying additional officers to these neighborhoods and authorizing millions of dollars in overtime spending, McCarthy sought to reduce their crime. The following section gives examples that are strongly suggestive of his success.

Many violent neighborhoods have experienced significant improvements in the past 5 months, likely as a result of improved police strategies. Although other factors may have contributed to these crime drops, the CPD intervention was the most high-profile, expensive, and potentially impactful candidate for the reduction.

The four tables below show a 7 year perspective on shootings/homicides in four different neighborhoods, all of which received special CPD attention in the past few months. I give 7 years of January-May data to show that 2013 violence was substantially improved from all the forgoing years. In each row, the  bolded number in parentheses indicates the cumulative number of incidents in that year to-date.

West Englewood: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 4 (4) 14 (14)11 (11)6 (6)4 (4)2 (2)4 (4)
February 1 (5) 5 (19)4 (15)4 (10)11 (15)3 (5)3 (7)
March 6 (11) 17 (36)6 (21)5 (15)7 (22)13 (18)7 (14)
April 4 (15) 7 (43)10 (31)14 (29)12 (34)14 (32)8 (22)
May 9 (24) 10 (53)16 (47)8 (37)7 (41)14 (46)7 (29)

By the end of May 2013, West Englewood had a combined total of only 24 shootings and murders. By that same time in the previous 5 years, the neighborhood had consistently had over 37 such incidents, with most years having more than 40.

Englewood: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 5 (5) 6 (6)7 (7)2 (2)4 (4)6 (6)6 (6)
February 4 (9) 3 (9)5 (12)5 (7)5 (9)4 (10)2 (8)
March 3 (12) 17 (26)5 (17)7 (14)2 (11)8 (18)8 (16)
April 6 (18) 5 (31)2 (19)10 (24)8 (19)14 (32)4 (20)
May 6 (24) 8 (39)9 (28)10 (34)16 (35)10 (42)7 (27)

The Englewood crime drop was less pronounced than that seen in West Englewood. Cumulative incidents in 2013 are more or less on track with both 2011 and 2007 totals, but still represent big improvements over other years. We might have expected West Englewood crime to displace to Englewood, but if criminals did try and move operations, their progress seems to have been halted by the CPD. The fact that Englewood violence is at a 7 year low is, on its own, encouraging.

Austin: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 8 (8) 8 (9)9 (9)7 (7)3 (3)13 (13)13 (13)
February 5 (13) 2 (10)4 (13)6 (13)10 (13)7 (20)4 (17)
March 4 (17) 13 (23)10 (23)14 (27)12 (25)13 (33)16 (33)
April 10 (27) 11 (34)9 (32)16 (43)10 (35)14 (47)9 (42)
May 8 (35) 11 (45)18 (50)17 (60)19 (54)13 (60)13 (55)

Austin's raw number of incidents might still appear high, with 35 by the end of May. Given both the population of Austin (98,000 residents) and its land area (7 square miles, one of the four largest in the city), however, the reduction to 35 incidents is a promising one. It might be expected that such a large neighborhood would be resistant to increased policing, with too many criminals and too much area to cover. The numbers suggest, however, that this was not the case. The March totals alone showcase the effect of the CPD intervention, with that month being the height of the overtime expenditure on additional officers. To some extent, this drop might be a function of weather. But this March crime drop was not common to all neighborhoods in Chicago, even though the inclement weather was. This suggests that something unique to Austin drove down violence, and that was likely the police presence. 

Woodlawn: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 0 (0) 9 (9)3 (3)5 (5)1 (1)6 (6)1 (1)
February 0 (0) 4 (13)6 (9)2 (7)3 (4)3 (9)1 (2)
March 1 (1) 11 (24)10 (19)6 (13)2 (6)3 (12)7 (9)
April 2 (3) 11 (35)4 (23)3 (16)7 (13)5 (17)4 (13)
May 5 (8) 12 (47)6 (29)7 (23)10 (23)1 (18)7 (20)

It would have been hard for the CPD to make up a success story like Woodlawn with such persuasive statistical backing. The crime drop in this neighborhood has been truly remarkable. Obviously, the 2012 spike was an anomaly, and we shouldn't judge the CPD's 2013 success on its ability to reduce 2012 violence. But even before 2012, Woodlawn had a fairly high rate of violence relative to its small population. All of that has turned around in 2013, with persistently low incident rates throughout the year. CPD foot patrols and additional hot-spot policing have undoubtedly played an impact in keeping Woodlawn violence so low, given that this is one of the only differences between the neighborhood from last year to this one.

These four neighborhoods all experienced significant drops in area violence. All four of which were targeted by high-profile police interventions. Although this is not conclusive of a direct effect (as the Wikipedia statisticians will remind us), it is strongly suggestive. A cursory examination of other factors that could reduce crime just doesn't account for the sheer magnitude of the 2013 drop. Population changes obviously might reduce crime, but it is extremely unlikely that the population changed so dramatically in the course of just 6 months. Weather patterns would reduce citywide crime (this April had the most rainfall on record), but that doesn't account for the huge crime drops in March that we see in four specific neighborhoods, drops not shared across the city. All factors considered, the likeliest explanation for the crime drop was the new CPD strategy, which is good news for both these communities and for Mr. Emanuel and Mr. McCarthy.

The problem with the "more police, less crime" philosophy is that it is a) expensive and b) limited by the number of officers in a department. With only about 13,500 officers on the budget, the CPD has one of the highest per-capita manpowers in the country (NYC has about 400 officers per 100,000 residents, whereas Chicago has 500) but also one of the lowest per-square mile staffing. Chicago, at 234 square miles, is a huge city. Although not all of those miles cover high-crime neighborhoods, the most violent communities tend to be spread across large swaths of South and West Side land. It is impossible for the CPD to deploy officers uniformly across these areas, and crime rates reflect this uneven pattern of deployment. Neighborhoods that did not receive special attention from McCarthy's expensive 2013 strategies have had unchanged rates of violence from previous years. On the one hand, the entrenched crime of these communities suggests that the improved violence rates in other neighborhoods are more than just chance (as in the previous section). But on the other hand, it shows that Chicago violence is incredibly resilient to law enforcement intervention.

The tables below show the per-month and cumulative shootings/murders for a variety of Chicago neighborhoods. These areas did not receive special CPD attention in 2013, and although officers in their districts have worked tirelessly to lower crime rates, violence remains deeply entrenched in these communities.

Humboldt Park: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 9 (9) 5 (5)6 (6)13 (13)2 (2)6 (6)9 (9)
February 4 (13) 5 (10)5 (11)6 (19)5 (7)2 (8)0 (9)
March 4 (17) 6 (16)6 (17)11 (30)4 (11)2 (10)4 (13)
April 7 (24) 10 (26)5 (22)8 (38)12 (23)11 (21)9 (22)
May 11 (35) 8 (34)12 (34)12 (50)15 (38)5 (26)11 (33)

Humboldt Park epitomizes a neighborhood that is completely resistant to policing and crime prevention strategies. Combined shootings and homicides in 2013 are virtually identical to levels in 2012, 2011, 2009, and 2007. With 50 total incidents, 2010 was a particularly bad year, but such abnormal spikes should not be benchmarks of improvement. Simply put, Humboldt Park violence is no better or worse than it was 7 years ago, with no combination of police, social service, and community intervention having any effect on crime.

Greater Grand Crossing: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 7 (7) 6 (6)6 (6)9 (9)5 (6)1 (1)4 (4)
February 2 (9) 4 (10)2 (7)4 (13)6 (11)0 (1)1 (5)
March 3 (12) 6 (16)6 (14)0 (13)7 (18)5 (6)6 (11)
April 9 (21) 7 (23)10 (24)7 (20)11 (29)11 (17)16 (27)
May 7 (28) 13 (36)3 (27)4 (24)6 (35)11 (28)11 (38)

A neighbor of Englewood and Woodlawn, Greater Grand Crossing has violence rates that have been in constant flux for the past 7 years. 2012, 2009, and 2007 appeared high. The remaining four years are all about the same, with totals hovering around 28. This is exactly the sort of pattern that you don't want to see as a police commander, because it is impossible to determine if your forces are having an effect on area violence. Greater Grand Crossing's 2013 looks to be just a repeat of three previous years, so if the CPD is doing anything new, it isn't having much of an effect.

North Lawndale: Shootings/Murders in January-May, 2013-2007

2013 201220112010200920082007
January 5 (5) 3 (3)6 (6)5 (5)1 (1)6 (6)1 (1)
February 7 (12) 7 (10)1 (7)3 (8)7 (8)1 (7)3 (4)
March 4 (16) 7 (17)3 (10)6 (14)4 (12)5 (12)4 (8)
April 2 (18) 8 (25)4 (14)12 (26)1 (13)7 (19)3 (11)
May 6 (24) 9 (34)7 (21)2 (28)10 (23)6 (25)5 (16)

Of all the neighborhoods discussed in this article, North Lawndale is the only one that appears to actually be getting worse. In every year since 2007, violence has been steadily increasing in small steps, following a zig-zagging upwards pattern. 16 incidents up to 25. Down to 23, then up to 28. Down to 21, then up to 34. 2013 looks to continue that trend. Moreover, over that 7-year period, North Lawndale's population continually declined while violence simultaneously rose. Although the sample size is not statistically large enough to draw conclusions from, on-the-ground reports from my colleagues confirm these observations. The gangs and criminals of North Lawndale have steadily grown over the last years while frightened residents have, when possible, left. It is too early to say with certainty, but if this trend continues unchecked, North Lawndale violence will continue to rise even in spite of police interventions.

These three neighborhoods are by no means the only Chicago communities that are unaffected by police intervention. East Garfield Park, Auburn Gresham, and West Pullman could easily have been added to this list, along with a half dozen others. Many of these communities are not as well-known as the infamous Englewood, but to those in the business of reducing violence, they continue to be household names. As 2013 unfolds and some neighborhoods improve, many others are likely to remain as violent as they were last year and the year before.

Overall, the Chicago of 2013 is much the same as the Chicago of 2011 or 2010. Crime may have dropped from the 2012 and 2008 spikes, but in many neighborhoods, it remains deeply entrenched. Thankfully, some neighborhoods have improved as a result of CPD presence. In some cases, especially Woodlawn, the improvement has been dramatic. When the politicians, police commanders, and pundits start to trumpet the so-called violence decline of Chicago, these are the communities that they are talking about. And to be sure, those improvements are welcome ones.

But many neighborhoods across the city are tragically resistant to police and community intervention. These neighborhoods may not have had a high-profile killing (Derrion Albert or Hadiya Pendleton), nor have they had a nationally-acclaimed radio show about their youth. A lack of public attention, however, does not reduce the violence that has become a daily reality for area residents. These are the neighborhoods that make the Tribune headlines if not in name than in consequence: Some number dead, a larger number wounded. As the summer unfolds, communities targeted by expensive CPD hot-spot policing will continue to improve (at least, until the money runs out). Meanwhile, in those neighborhoods that are not receiving special attention, violence will continue to rise. It is only May, and some of these trends are liable to change as the year goes on. But if we have learned anything about Chicago summers, it is that violence only worsens as the years goes on. If the trends so far are any indication, this summer looks to be a long one and a bloody one.

No comments :

Post a Comment