In recent days, I’ve been hearing an abnormal amount of criticism of Donald Trump for “policies on race relations that basically amount to ‘reinstitute stop and frisk.’” As usual, the statements are being made in a tone that makes it obvious that the speaker considers it simply beyond the pale to question whether or not those stop and frisk policies were doing anything right. Compare: “His ideas on improving race relations basically amount to ‘bring back slavery.’” Since I’m hearing most of this come from normal people rather than rabid social justice warriors, I consider it worth addressing. So the point of this post is to quickly hand supporters a little bit of extra ammo in some of those conversations.
Anyone who is paying attention will have been aware for a long time now that the crime rate among the minority — and particularly black — population is in fact higher than the crime rate among the white population. They’ll also have been aware that victim and witness reports make it impossible to dismiss this as the result of biased police, because victims and witnesses actually record a higher black/white disparity  than arrest rates do.
But even knowing this still isn’t enough to refute the possibility that a policy actually is racist. When we’re considering police shootings, for example, it’s an entirely possible hypothetical that blacks could be 2 times more likely to commit crimes than whites, yet 2.5 times more likely to be shot by police, in which case only part of that 2.5x figure would be explained by blacks’ disproportionate commission of crimes, and finding ratios like these would suggest that blacks are still 1.5 times more likely to be shot by police because of racism. So, if you want to argue that there isn’t any institutional racism, you have to show that the first figure is larger than the second one. This is what I  and a few other people like Peter Moskos  have done with the numbers for crime and police shootings. And what people like Moskos and I who have run these figures discover is that once controlling for crime, it is white people who are more likely to be shot per encounter with police — it isn’t blacks who are more likely to be shot in a given encounter at all, at any frequency.
Still, we can’t generalize from findings about one event (shootings) to another (stop-and-frisks). A shooting will be a much rare more event than a stop; the threshold for convincing a cop to make a stop will obviously be much smaller than the threshold for convincing a cop to fire a weapon, considering if nothing else the different cost of mistakenly making a stop versus mistakenly shooting a gun.
This leaves us with a couple different related questions:
- Was stop and frisk enforced in a racist way?
- Was stop and frisk effective at reducing crime?
- Was stop and frisk constitutional?
With regard to the first point, even the critics of stop and frisk acknowledge  that “Marijuana stops are more prevalent in precincts where . . . “high-crime area” justifications are more likely to be reported . . .” Without diverging into too much detail on the point here, I advise any interested readers to see the work of criminologist Frank Zimring  or my essay on the war on drugs  on the point that the war on drugs actually has been an effective tool for police work that has been successful at reducing violent crime — indeed, even including giving police the ability to use marijuana as a pretext for a stop in high-crime areas:
Marijuana was not a priority of the New York City police, yet they had a huge number of public marijuana arrests. Why was that? That was because they were only arresting minority males who looked to them like robbers and burglars and they used as a pretext the less serious crime arrest to find out whether the particular person they were arresting had a warrant out for a felony and was a bad actor. . . . The good news is that drug violence went down tremendously. There are a couple of different ways in which the police department measures the number of killings associated with drug traffic in New York; both of those measures that they use are down more than 90 percent so that the streets themselves have been changed, people can walk there, and the number of dead bodies associated with illegal drug traffic has gone way, way down.
Now to provide a quick rebuttal of the claim that stop and frisk was racially enforced.
We know that stops are more likely to take place in “high-crime areas.” But “high-crime areas” also tend to be minority areas in the first place, whatever the reason is for that being the case (it’s irrelevant to our purposes here).
So what would happen if we had a low-crime minority area? Would stops decrease in frequency?
We actually have a perfect case study for examining this question.
In Brooklyn, New York, the borough of Brownsville is 76.7% black and only 2.6% white . In Buffalo, New York, the borough of Kensington has a similar demographic spread : 82.3% black and 11.5% white. But despite such similar demographics, it turns out that Kensington actually has the lowest murder rate of all boroughs in New York, while Brownsville has the highest .
So, what happens? Does the frequency of stops increase from Brownsville to Kensington because Kensington is 1.07 times as black?
It turns out that the stop and frisk rate  is 2% in Kensington, compared to 29.1 in Brownsville.
The frequency does, in fact, decrease — because Kensington has less crime — regardless of the racial makeup of its residents.
With regard to the second question, was stop and frisk effective at reducing crime?
For a long time, there have simply been no studies that asked this question while adequately controlling for all the other changes that took place over the period of time in which stop and frisk policies were adopted. However, if we dig far enough, we can find some information :
The only study that explicitly poses the question “Does stop and frisk stop crime?” was an unpublished paper by Robert Purtell and Dennis Smith that relied on monthly precinct level data from New York City from 1997 to 2006. After controlling for a large number of variables including the effects of hotspots policing, Purtell and Smith found that stop and frisk helped reduce robbery, burglary, murder, and grand larceny . . . While other researchers have looked into similar questions as the one posed by Purtell and Smith, they do not isolate stop and frisk as a variable and instead combine its effects with other police strategies like 1) firearm reduction, 2) hot spots policing, and 3) order maintenance policing.
So it looks like what is quite literally the one and only study that asks whether or not stop and frisk actually stopped crime strongly suggests that the answer is “Yes.”
Finally, was stop and frisk constitutional?
The Wall Street Journal actually had to fact-check Lester Holt ’s “fact-check” of Trump’s claim that the ruling that stop and frisk was unconstitutional came from a “very against-police judge” who later had the case taken away from her. Their conclusion? “We rate Mr. Trump’s claim true and unfairly second-guessed by a moderator who didn’t give the viewing public all the facts.” Furthermore, the Wall Street Journal’s Law Blog elsewhere notes  that “Mr. Trump, technically, is right that the general tactic of stop-and-frisk was not deemed unconstitutional. The ruling in the civil-rights class-action case Mr. Trump was referring to didn’t order an end to the practice of stopping and searching pedestrians suspected of criminal activity.” And as Rudy Giuliani writes in a Wall Street Journal Op-Ed :
Stop and frisk is based on an 8-1 decision of the Supreme Court, Terry v. Ohio. That ruling hasn’t been overturned or even modified by the court since it was handed down in 1968. Stop and frisk is constitutional and the law of the land. The majority opinion, written by then-Chief Justice Earl Warren, approved the constitutionality of stopping a suspect if the police officer has a reasonable suspicion that a person has committed, or was about to commit, a crime. If the officer also has a reasonable suspicion the person is armed, he can conduct a pat-down — that is, a frisk — of a person’s outer clothing. In many places, this practice is called a “Terry stop,” based on the decision upholding its constitutionality.
As usual, the “whites as oppressors, nonwhites as victims” narratives not only target whites on false premises, they harm nonwhites also due to their own violent ignorance of reality.
At the same time as it is true that blacks are disproportionately responsible for violent crime, so it is also true that blacks are disproportionately the victims of it. As Heather MacDonald notes  , while blacks constituted 78% of shooting suspects in New York City (despite being only 23% of the city’s population), they were also 74% of all shooting victims. So it isn’t whites who are the primary beneficiaries of these “racist” policies in the first place: while whites committed just over 2% of the city’s shootings, they were also under 3% of the city’s shooting victims. Minorities make up nearly 80% of the drop in homicide victims in New York’s record-breaking crime decline  from its highs in the 1960s through to the early 1990s now to current historical lows unseen since our grandparents and great-grandparents were in their prime.
As MacDonald concludes :
To be sure, thousands of innocent New Yorkers have been questioned by the police. Even though such stops may have been justified given the information the officer had at the time, they’re still humiliating and infuriating experiences. But if the trade-off is an increased risk of getting stopped in a high-crime neighborhood versus an increased risk of getting shot there, most people would choose the former.
Well, most sensible people would — but most people aren’t sensible. Instead, we’re at the point where it’s a significant point of condemnation in an election that someone thinks that we shouldn’t be too worried about inconveniencing some blacks if the result is that it means we can also stop a lot more of them from killing other ones. The truly vicious moral condemnation is all too often reserved not for blacks who kill other blacks, but for whites who agree with the idea of doing something to try to help stop it.
For whatever it’s worth, stop and frisk probably isn’t the most significant consideration with respect to whether violent crime will rise or fall. Murders in New York have, in fact, decreased  since the period of time during which the rate of stops was at its highest. But it would be fallacious reasoning to leap from this fact to the conclusion that stop and frisk is irrelevant. Specifically, it would be an instance of post hoc ergo propter hoc —like claiming that, because we have longer lifespans today than we did a few centuries before cigarettes were invented, cigarette smoking must have no impact on a person’s lifespan today. In both cases, the reason for the apparent lack of trend is because variables other than the one you’ve identified have more impact on the outcome in question—not because the variable you’ve identified has no impact on the outcome in question at all. Furthermore, one reason that violent crime tends to fall is that a lot of violent criminals have already been put away—an effort which increased use stop and frisk most certainly contributed to in the past.
The only controlled evidence we have strongly suggests that, all else equal, an increased use of stop & frisk does in fact help to reduce violent crime. So while we shouldn’t exaggerate how important stop and frisk is in the grand scheme of things, we should also be prepared to point out why Trump isn’t the one being ignorant when he suggests that it does—and can continue to do so.