Trump defended the NYPD’s stop-and-frisk policy. The NYPD said Trump got the facts wrong.

During the presidential debate on Monday, Donald Trump claimed that the end of New York City’s stop-and-frisk policy led to a rise in murders, and Hillary Clinton claimed that a judge declared the policy unconstitutional.

The New York City Police Department apparently took issue with both of these claims:

Some of the NYPD’s claims here are questionable. US District Court Judge Shira Scheindlin really did find in 2013 that stop and frisk, as the NYPD enforced it, was unconstitutional. So Clinton’s claim, despite the NYPD’s insistence otherwise, was right.

But it’s true that the end of stop and frisk did not lead to a rise in murders. As the Washington Post noted, there is no correlation between the rise and fall of stop and frisk and the murder or crime rate in New York City. The number of stops associated with the practice rose throughout the 2000s, after New York City’s massive crime drop in the 1990s. And when the number of stops dramatically dropped in the early 2010s, crime and murder rates remained relatively flat — and even fell in the past year, as the NYPD pointed out.

This should have been expected. Previous studies found that stop and frisk had little to no effect on crime in New York City.

Still, the policy did have one effect: It massively increased distrust, particularly among black and brown New Yorkers, toward police.

It’s not hard to see why. According to city data, the great majority of people who were stopped and frisked in New York City were minority residents. Almost all turned out to have no contraband on them — meaning they weren’t guilty of a crime. And 1 percent of black people who were stopped had weapons or contraband on them, while 1.4 percent of white people stopped did — suggesting that black people weren’t more likely to be doing anything wrong despite getting stopped more often.

New York City police at the time claimed that they were simply conducting a majority of stops in high-crime areas, which are disproportionately black and brown. But as criminologist Jeffrey Fagan found, even if you controlled for crime, minority New Yorkers were still disproportionately stopped:

The percent Black population and the percent Hispanic population predict higher numbers of stops, controlling for the local crime rate and the social and economic characteristics of the precinct. … The crime rate is significant as well, so the identification of the race effects suggests that racial composition has a marginal influence on stops, over and above the unique contributions of crime.

As a result of these racial disparities, stop and frisk fostered distrust and resentment toward police. That distrust could have actually led to more crime: The evidence shows that police can better solve and prevent crimes if they have the trust and cooperation of their community.

So Trump was simply wrong — and the NYPD, at least to some extent, disagrees with him, even as he tries to prop up a policy the police department once enforced.