Legal Commentary

It's good to be back on Prawfs after a long layoff.  I?m glad Dan invited me back.  I guess that means I didn?t embarrass him too much the last time?either that, or he has a short memory.

Since my last guest stint, one of my new gigs has been with the Milwaukee Fire and Police Commission.  This is a six-member citizen board that oversees the city?s fire and police departments.  It?s been a fascinating experience for me, as a criminal law professor, to learn more about how a large, urban police department functions on a day-to-day basis.  

This week, I?ve been puzzling over some data on poverty and crime that the police department recently provided to the Commission.  This is comparative crime data over the years 2007-2009 for the fifty U.S. cities with the highest poverty rates. 

One would expect the data to reveal a positive correlation between poverty and crime rates, and, in a general way, inter-city comparisons are consistent with that expectation.

For instance, if one considers the ten cities that had the highest poverty rates in 2009, the rate of violent crime seems relatively high: 14.4 violent crimes per 1,000 residents.  By contrast, among the bottom ten cities of the top fifty (in other words, the cities ranked 41-50 by poverty rate), the average rate of violent crime is only 9.8 per 1,000.  The top ten have an average poverty rate of 28 percent, while the bottom ten average about 17 percent.  The data thus suggest that a higher poverty rate in a city is associated with a higher rate of violent crime. 

But this general association masks the reality that many high-poverty cities actually do much better with crime than others.  Consider, for instance, the large disparity in the poverty and violent-crime rankings of a few of the cities on the top 50 list (the number on the left is the poverty rank and the number on the right is the violent-crime rank):

 If we think about a city?s poverty ranking as setting an expected baseline for its crime ranking, some cities on the list markedly over-perform, while others just as markedly under-perform.  Here are the top ten over-performers, in terms of difference between the violent-crime rank and the poverty rank:

1.  El Paso (violent crime rank=49, poverty rate=20, difference=29)

2T.  Fresno (38-16=22)

2T.  Tuscon (35-13=22)

4.  Phoenix (45-24=21)

5.  Santa Ana, CA (47-29=18)

6T.  Milwaukee (21-4=17)

6T.  New Orleans (29-12=17)

8T.  Columbus (33-19=14)

8T.  Dallas (28-14=14)

8T.   Newark (24-10=14)

Here are the top (bottom?) ten under-performers:

1.  Oakland (poverty rank=46, violent-crime rank=4, difference=42)

2.  Kansas City, MO (49-8=41)

3.  Washington, DC (41-10=31)

4.  Nashville (45-16=29)

5.  Boston (47-22=27)

6.  Albuquerque (50-30=20)

7.  Baltimore (25-5=20)

8.  Indianapolis (28-12=16)

9T.  Tulsa (32-19=13)

9T.  Stockton, CA (22-9=13)

These numbers raise questions in my mind about what El Paso, Fresno, and Tuscon are doing to beat expectations so much better than places like Oakland and Kansas City.

I?m even more puzzled, though, by the seeming disappearance ? even reversal -- of the violence-poverty relationship if one moves from a comparison of different cities to an analysis of the same cities over time.  Twenty-two cities on the top 50 list in 2007 experienced an increase in poverty rate of two percentage points or more by 2009.  (I exclude Chicago, which would otherwise qualify, because I do not have violent crime data for the Windy City).  Yet, the rate of violent crime in 17 of these cities decreased over that time period.  In the remaining five, the rate of violent crime was essentially unchanged; none experienced an increase commensurate with the increase in the poverty rate.  On average, the 22 cities saw a decrease in their violent crime rate of 1.6 crimes per 1,000 residents.

To be sure, violent crime was decreasing generally in the cities on the top 50 list.  But the rate of decrease was actually less in the 13 cities whose poverty rate remained unchanged or improved.  These cities averaged a drop of only 1.2 per 1,000.  Thus, it seems that increased poverty was associated with better success in reducing violent crime over the 2007-2009 time period ? a very surprising pattern indeed.

A similar pattern is evident if we consider changes in poverty ranking.  Between 2007 and 2009, 17 cities on the top 50 list saw their poverty ranking increase (e.g., Milwaukee?s rank went from seven to four).  Of these 17 cities, a clear majority (11) improved in their ranking on the violent crime list.  Again, the data seem to indicate that declining performance on the poverty front is associated with improving performance on the crime front.

I can't come up with any reason to think that increased poverty actually contributes to crime reduction, but the data do at least suggest that the poverty-crime relationship is far more complex than is commonly imagined.