Comparing Police Stops of Citizens in New York and Milwaukee, Part II

As I discussed in my previous post, frequent police stops of citizens may ultimately prove counterproductive to crime-fighting objectives.  In this regard, I also suggested that who is targeted and how they are treated may actually matter more than the sheer quantity of stops.  If that’s right, then several aspects of the New York stop data are troubling.

Racial disparities are one concern.  Black and Hispanic males between the ages of 14 and 24 accounted for more than 40% of the stops made by the NYPD in 2011, even though they amount to less than 5% of the city’s population.


To be sure, this disparity must in part reflect the fact that high-crime neighborhoods — where one would want and expect to see more proactive policing — tend to have disproportionately high minority populations.  Yet, stops of blacks and Hispanics also predominate in some New York neighborhoods that have low minority populations.  In fact, in eight of the nine precincts in which blacks and Hispanics make up less than 14% of the population, stops of blacks and Hispanics constitute at least 47% of the total stops.  For instance, in the 17th Precinct, 71.4% of stops were of blacks and Hispanics, even though these minority groups constitute only 7.8% of the population.  So, it is hard to write off the city’s overall disparities as purely an incidental byproduct of stopping more citizens in high-crime neighborhoods, which just also happen to be high-minority neighborhoods.

There are also concerns that the NYPD’s stops are more intrusive and forceful than necessary.  Officers conducted frisks in 55.7% of stops in 2011, but found weapons less than 2% of the time.  Moreover, force was used in 21.6% of stops (although, to be fair, most of the uses of force were categorized as simply “hands on suspect”).  Suspects were handcuffed in 3.4% of stops and put against a wall or car in 3.0% of stops.

Racial disparities are also evident in the frisk and force data.  While 57.5% of stopped blacks and Hispanics were frisked, only 44.2% of whites suffered this intrusion.  Yet, frisks of whites were considerably more productive on average than minority frisks, with 3.8% of white frisks finding a weapon, as against only 1.8% of frisks of blacks and Hispanics.  If the NYPD considers its frisk program an effective way to recover weapons and deter individuals from carrying firearms, then it seems that the Department would do well to frisk fewer minorities and more whites.

As to force, 22.5% of minority stops, but only 15.8% of white stops, resulted in force being used.

Finally, there are concerns about what seems to be a very high rate in New York of stopping innocent people.  Fewer than 12% of the individuals stopped in 2011 were arrested or received a summons.  In the NYCLU report, the remaining 88+% are characterized as “innocent.”  This strikes me as a bit misleading, particularly insofar as the connotation is that, in retrospect, the police had no good reason to stop any of the 88+%.  I imagine that in a number of these incidents, police actually did have evidence of a law being broken, but prudently exercised their discretion based on the nature of the violation to issue a verbal warning rather than initiating legal proceedings.  I imagine, too, that a number of the 88+% stops deterred impending crimes or resulted in the police obtaining information that was helpful in preventing other crimes or apprehending other offenders.  Yet, even discounting the idea that an arrestless, summonsless stop necessarily means that the life of an entirely innocent citizen was needlessly interrupted, the 88+% number does seem awfully high and suggests that the criteria officers use for stops could be made more rigorous without much loss of crime-fighting benefits.

Put it all together, and I am left with the sense that the NYPD may indeed be at some risk of lowering its standing among minority New Yorkers through its increasingly aggressive stop-and-frisk program.

Do the Milwaukee data raise comparable concerns?  Unfortunately, the data made available by the MPD are not as comprehensive as the NYPD data, which makes it hard to compare the two cities along the various dimensions suggested by the NYCLU report.  It would be helpful to know what percentage of MPD stops resulted in frisks and uses of force, as well as the gun recovery rate and the arrest/summons rate.  However, none of this information appears to be available.

Nor does the MPD report provide an overall, city-wide breakdown of the race and ethnicity of the individuals who were stopped.  Instead, the report provides district-by-district breakdowns.  It is hard to know what to make of this data, though, because the report does not indicate the racial breakdown of each district.  In poking around the Internet, I was not able to find full information about this on my own, either, although this City of Milwaukee page does provide data from Census 2000 that permits one to calculate the black percentage of each district.  Here’s what I came up with:



Black Stops (2011)

Black Population (2000)

Disparity (Percentage Points)






























By way of comparison, about 53% of the New York stops were of blacks, as against an overall 25% of the city’s population, for a disparity of 28 percentage points.  This seems generally in line with what one sees in the MPD districts.  (To be clear, though, my population numbers must be used cautiously since they are from 2000, and the racial composition of some of the districts may have changed in significant ways since then.)

In discussing this with Chief Flynn at the Milwaukee Fire and Police Commission meeting last night, the Chief suggested that the global demographics of a district are not an appropriate benchmark for the stop data.  If I understood him correctly, his view was that the best comparison would be with the percentage of residents of each district who are involved in crime — those are the people the police should be targeting for stops, and their demographics should be the basis for deciding whether there are unwarranted disparities.  In theory, this point seems sound.  However, it is hard to know how many people in each district are criminals; this seems not to be one of the questions that people answer on their census forms.

The Chief suggested using as proxies the data on who is arrested and who is victimized in the city.  Both proxies are problematic.  Since arrests are a function of police decisionmaking, the arrest numbers would presumably be tainted by any racial bias that is tainting the stop numbers (if there is any).  The racial demographics of victims, meanwhile, seem for obvious reasons a much-less-than-perfect way of estimating the racial demographics of offenders.  The Chief justified using victim demographics because there is a substantial overlap between the victim and offender populations, but it is not clear how much overlap there is and the extent to which this overlap is consistent across neighborhoods and crime types.  It is also possible that different racial and ethnic groups are more or less likely to report victimizations.

Those reservations notwithstanding, let’s take a look at the demographics of arrests and victimizations.  The specific numbers vary by offense type, but it does seem true that across the five offense types covered by the MPD report, blacks are substantially overrepresented among both victims and arrestees.  For arrests, the numbers range from 79% black for aggravated assault and auto theft to 94% for robbery.  Among victims, the numbers range from 57% for burglary and auto theft to 80% for homicide.  This compares with a citywide 39% black population.  It does seem plausible, then, that overall stop disparities in the city are generally in line with racial disparities in criminal activity.

This reflects a lot of rough estimation, though.  It also does not address such potentially important questions as the relative productivity of stops of minority suspects (recall, for instance, the New York data on the very low weapon recovery rates from blacks and Hispanics who were stoppped) and the relative harshness of the treatment of minority suspects (recall the higher frisk and use-of-force rates against minorities in New York).

One Reply to “Comparing Police Stops of Citizens in New York and Milwaukee, Part II”

  1. The frequency and nature of police stops involving African American and Hispanic men raise questions regarding the true purpose of these stops. Police departments tend to say they are engaged in proactive policing and concentrating their efforts in the proverbial “high-crime neighborhoods.” However, the great majority of the stops do not result in arrests or citations for criminal activity, and many of the stops do not even occur in the purportedly targeted neighborhoods. Might the police – either knowingly or unknowingly – be engaged in a campaign of intimidation and social control?

Comments are closed.