Here's an excerpt from an article by the NYT's mostly pro-cop reporter Joseph Goldstein about the decision against anti-stop and frisk in New York City. I haven't read the 195-page decision, but it sounds kind of incoherent (at least in Goldstein's not hugely sympathetic retelling), even though I have some sympathy for the decision.

Police Dept.’s Focus on Race Is at Core of Ruling Against Stop-and-Frisk Tactic

In the process, Judge Scheindlin coined a term, “indirect racial profiling,” to explain how the department’s reliance on data indicating that black men committed a disproportionate amount of crime led to what she saw as violations of the Constitution.

How the judge came to this decision was revealed over the course of a 195-page opinion, released on Monday.

One witness claimed to have heard Mr. Kelly say that the stop-and-frisk tactics were intended to frighten minority men into leaving any guns they owned at home. A precinct commander described “the right people” to stop, with a reference to young black men. And a statistical analysis of millions of police interactions revealed that few people subjected to stop-and-frisk methods had been engaging in wrongdoing.

These three items of evidence were central to Judge Scheindlin’s conclusion that the Police Department has a policy of conducting stops “in a racially discriminatory manner.”

In particular, the statement attributed to Mr. Kelly, as well as the phrase about the “right people,” served as the lens through which the judge interpreted a mountain of data about 4.4 million police stops. That data showed that stops overwhelmingly involved black and Hispanic people and that the people stopped were rarely found to be engaging in criminal activity.

Judge Scheindlin found that the department operated on the notion that black and Hispanic people were the right people to stop — a theme, she wrote, that was evident in statements made by Mayor Michael R. Bloomberg out of court, and in directives made by commanders during station house roll calls.

The judge repeatedly cited testimony by the onetime highest ranking uniformed member of the department, Joseph J. Esposito, who was asked about the reasons for some stops. At one point, Mr. Esposito explained: “Well, who is doing those shootings? Well, it’s young men of color in their late teens, early 20s.”

It was a similar sentiment that Judge Scheindlin detected in a recording of a deputy inspector exhorting a subordinate to conduct more stops.

That meant stopping “the right people at the right time, the right location,” the deputy inspector, Christopher McCormack, said. Pressed on what he meant, he noted that black teenagers and young men were behind some crimes being committed in his precinct.

At another point, Judge Scheindlin noted that Mr. Kelly had said, according to a state senator who was present for the conversation, that young black and Hispanic men were the focus of the stops because the commissioner “wanted to instill fear in them, every time they leave their home, they could be stopped by the police.”

A key point is that New York City, with its income tax on rich people, can afford to pay cops to get in the face of shifty-looking people far more than most cities can. Without acting in an overtly discriminatory manner, NYC can afford to show its black and Latinos who is boss. If they find that awareness depressing and want to move somewhere in America less affluent, well, the Port Authority Bus Terminal is open 24/7.

Mr. Kelly had filed an affidavit denying that he made such a statement, but the judge credited the account of the state senator, Eric Adams, who testified during the trial.

The city offered explanations for the statistics on millions of stops that so troubled Judge Scheindlin. The low arrest rate of people stopped reflected the fact that officers were proactively stopping criminals before they had a chance to go through with their crimes, according to city lawyers. And minorities were disproportionately stopped because they lived in the same high-crime neighborhoods where the police deployed many of its officers.

But Judge Scheindlin was not swayed by those arguments. She concluded that the low arrest rate meant that the people who were being stopped were rarely the criminals the police sought; all they had in common was the color of their skin.

“The city adopted a policy of indirect racial profiling by targeting racially defined groups for stops based on local crime suspect data,” she wrote.

Perhaps the way to logically unsnarl this is to think hard about the distinction between relative and absolute differences in stops. We're not supposed to think about differences in crime rates by race, so most of the more sophisticated thinking on the subject has been limited to pointing out that blacks are more prone to commit crime, while the unsophisticated gasp in shock.

But, let me try to take this to another level and explain arithmetically why a realist like me has been uncomfortable with Bloomberg's huge intensification of stop and frisk.

To help, let me make up some stylized numbers. Let's say that under Giuliani, the average young black man in New York city gets stopped once every 4 years and the average young white man gets stopped once every 40 years. Moreover, let's assume that that 10 to 1 racial ratio is reasonable considering differences in crime rates and so forth.

Then, under Bloomberg, the mayor and the police chief figure that, heck, New York is rich, so let's quadruple the number of stops. Now the average young black man is stopped every single year and the average young white man every ten years. The racial ratio remains the same ten to one, so how can anybody logically object to what Bloomberg is doing compared to what Giuliani did? The relative proportions remain the same, right?

Okay ... but, think about the change in absolute terms rather than relative terms. Under Bloomberg, the number of stops a young black man can expect to endure per year has increased by 0.75 stops annually (from 0.25 to 1.0). For young white men, however the increase is only 0.075 stops (from 0.025 to 0.100).

While these increases of 0.75 incremental stops for blacks divided by 0.075 incremental stops for white still give the same old relative racial ratio of 10, 0.75 minus 0.075 gives an absolute increase in stops for blacks relatives to whites of .675 stops per year, which might be considered a serious quality of life degradation for young black men.

So, perhaps the question ought to be: rather than focusing on the reasonableness of the relative rates of blacks and whites being stopped and frisked, perhaps the focus should instead by on the absolute overall levels of stop and frisks. In other words, under any reasonable system, blacks are going to be stopped and frisked a lot more often than whites; but given the expectation of diminishing marginal returns, is the absolute number of stops and frisks too high for the pain it inflicts on blacks relative to the gains in reduced crime?

The problem for the judge, however, is that Americans are much more comfortable with judges making decisions about relative matters (e.g., make sure blacks are treated fairly relative to whites) but not absolute matters (e.g., how much stopping and frisking to do).

More fundamentally, it's not all that clear that marginal returns diminish that fast. The NYC homicide rate keeps going down. It could be that, beyond the numbers about guns taken away and the like, stop and frisk may have slammed home a psychological shift away from the criminal energies unleashed during the civil rights era back to a more orderly and authority-fearing culture that preceded the 1960s. But, that's speculative.