The Myth of 'Activist' Judges


In August, a federal district judge, Shira Scheindlin of the Southern District of New York, ruled that the “stop-and-frisk” practices of the New York Police Department (NYPD) are unconstitutional under both the Fourth and Fourteenth Amendments.

Last month, a panel of the Court of Appeals for the Second Circuit removed her from the case on the grounds that she had made inappropriate comments to the press and had invited plaintiffs to mark the case as “related” to a prior stop-and-frisk case she had heard, so that it would be assigned to her.

I argue in front of trial and appellate courts for a living, and I'm mystified by the Second Circuit's ruling.

For one thing, it was sua sponte (meaning the court did it on its own; the City had only requested a stay of Scheindlin's order pending appeal, not reassignment). More importantly, there's no suggestion that there was any error in the trial itself: the briefing on the city's appeal won't even be filed until the spring, and it'll probably be a year before we see a merits ruling, that is, a ruling on the actual constitutional claim.

Reassignment to a different district judge is almost always done as part of the merits ruling on an appeal, and then usually in conjunction with a finding of serious error. After all, reassignment is an assertion by the higher court that leaving the case in a particular judge's hands would undermine the integrity of the judicial system.

So what's going on here?

Jeffrey Toobin of the New Yorker sees old-boy former-prosecutor back-scratching here. (Scheindlin is not of that fraternity.) Perhaps some facts will come to light that we don't yet know. Until then, though, it's a puzzle—and an important puzzle, because there are few more important criminal procedure questions than the legality of aggressive policing techniques like stop-and-frisk.

I can't speculate on what the Second Circuit panel was thinking. But let me offer another observation, one that Toobin didn't make: what I see in this case is the reawakening of the specter of the liberal judge who's instinctively hostile to law enforcement and congenitally soft on the bad guys.

The place you used to see that specter floating around the most was in debates over criminal sentencing around thirty years ago.

In 1984, Congress passed the Sentencing Reform Act (SRA) as a way of preventing sentencing disparity (read: make those hippie judges toughen up). The SRA created mandatory Sentencing Guidelines—a “points system” in which every aspect of the offense conduct was given a certain number of points (yes, really).

Sentencing was a matter of tallying up the points and placing your perp on a chart to translate points into months.

The Sentencing Guidelines Manual would come out every year, getting thicker and more intricate with each new edition, and criminal lawyers spent much of their time mastering its nuances, and arguing with each other (and with judges) about how to classify each new crime. The express purpose of the SRA and the Guidelines was to limit, as much as possible, the ability of judges to use their own judgment about how to sentence criminals. And that purpose was rooted squarely in the assumption that there are some, or many, judges who, left to their own devices, will start getting “activist” and unpredictable.

(I'm ambivalent, by the way, about sentencing discretion. Most criminal lawyers are. Ask one.)

For 20 years after the passage of the SRA, sentencing discretion wasn't much of an issue in federal court, because there wasn't any. But in January 2005, the Supreme Court decided a case called United States v. Booker, which declared that the Guidelines are now “advisory,” not mandatory.

The years since Booker have created a rich opportunity to study judicial behavior, and to see whether judges would seize their newfound sentencing freedom and take it to their favored ideological extremes.

But judicial disparity (meaning different results reached like cases by different judges) is very difficult to define, let alone to study, both because of the judge's role in a case (making it hard to define “results”) and because of the inherently variable fact patterns that come before judges (making it hard to specify “like cases”).

A few years ago, my colleagues and I figured out a new way to measure sentencing disparity precisely, by controlling for variations in offense conduct. We created a database of almost 10,000 cases with the same crime: drug smuggling busts at U.S.-Mexico ports of entry on the Southwest border, and we identified and controlled for the relevant variation (drug type and load size).

You're probably not interested in the methodological details (but if you are, you can read about them here and here!), so I'll jump straight to the findings. We found that previous studies of inter-judge sentencing disparity within courthouses had considerably overstated the amount of variation among judges.

When we were able to control precisely for the details of the offense conduct to create an “expected sentence” measure for each individual case, we found relative consistency among the judges in the district. Admittedly, this is just one study of one courthouse and one type of case. But non-violent drug offenses are precisely the type of case that (in the popular imagination) draw out the “judicial activism” in judges.

The chart below tells the story for the Southern District of California. Ten of the thirteen judges in the district (there are 16, but three hadn't sentenced enough cases for a statistically valid sample) sentenced within 11 percent of the district baseline, and seven of the thirteen were within three percent of it. Thus, a mule doing a pre-run risk/reward calculation can predict that he has a 75% chance of getting a sentence within 10% of the district norm, a 10% chance of getting a sentence around 20% shorter, and a 15% chance of getting a sentence around 25% longer.

The real disparities, it appears, are not among individual judges in a district, but rather among the districts themselves.

For, example, check out the comparison between the light-sentencing Southern District of California and the heavy-sentencing Western District of Texas, for the exact same crime (smuggling drugs through a port of entry).

In the table, I've shown the mean sentences and mean load sizes for marijuana and cocaine, and the “months per kilo” measure derived therefrom. (Months-per-kilo is a very crude approximation of our “expected sentence” measure, but it is intuitively graspable and a good rule of thumb if you're contemplating a smuggling venture.)

That's some pretty serious disparity, wouldn't you say?

Within SDCA, the most Joe Mule has to fear is his small chance of drawing one of the two harsh sentencers and getting a sentence around 25% higher than the baseline. But if he ventures east and tries his luck on the Juarez-El Paso crossing, his baseline for the same crime on the months/kilo metric goes up more than 250 percent!

You see the point? If sentencing disparities are an inter-district phenomenon rather than an inter-judge phenomenon, then they must be rooted in institutional causes such as prosecutorial charging policies or courthouse culture—not a few out-of-step individual judges drunk on discretion.

So I urge readers to be skeptical when you hear talk of wild activist judges imposing their wacko political visions on the world through their cases. There's very little evidence that myth has ever been more than a few exaggerated anecdotes.

Equally importantly, lay readers may tend to overestimate the actual amount of control a trial judge really has over the outcome of a case—even a case, like Floyd, the stop-and-frisk case, in which there's no jury.

Here are three things to remember:

(1) trial judges do not make the law they are applying (legislatures and appellate courts do that);

(2) trial judges do not control the facts presented by the parties (they can limit testimony at the margins, but less so in bench trials than in jury trials, and there has been no suggestion that Judge Scheindlin unfairly limited or excluded evidence);

(3) trial judges do not control the legal claims brought by the parties.

A trial judge (unlike a Supreme Court justice) really is something like an umpire calling balls and strikes.

Judge Scheindlin didn't write the Fourteenth Amendment; she didn't write the numerous cases applying it to police procedures; and she didn't invent the statistical data on NYPD stops. She didn't prevent the City from presenting expert testimony to counter the plaintiffs' experts, or from putting on officers to testify about the 19 stops that were litigated individually.

It's no secret that Judge Scheindlin believes that the NYPD's stop-and-frisk policy was conducted unconstitutionally. She said so in a 180-page decision. So she made some comments to the press. Does that mean she didn't try the case fairly?


As Robert Morgenthau supposedly said of G. Gordon Liddy, “You don't get immunity just because the prosecutor doesn't like you.”

And the “related case” issue is even stranger. It's mandatory that parties identify a new case as “related” to a previous one if the new case “calls for determination of the same or substantially related or similar questions of law or fact.”

Judge Scheindlin said from the bench in a prior stop-and-frisk case that if a new one was filed raising the same legal issue, it would be a related case. That's not “the appearance of partiality”; it's the law. And judges have discretion over whether to accept new cases identified by a party as related. So, trial lawyers reading this who care to comment: look at pages 7-9 of the Nov. 13 Second Circuit opinion and tell me if that colloquy would have raised your hackles in court. It doesn't raise mine.

As readers may recall (and for what it's worth), I'm a former prosecutor. I'm pretty conservative on law-and-order issues, and I have regularly urged Crime Report readers to be cognizant of the difference between policy preferences and constitutional limits when thinking about police practices.

I've read Judge Scheindlin's opinion in the case, and I found it exhaustive, fair and thorough. If there's some glaring irregularity, it's not evident from the documents. The law and the facts here were pretty straightforward.

Here's the law: the Constitution forbids disparate enforcement of the law based on race, and it also requires that police have “reasonable suspicion” before stopping someone for a weapons frisk. Judge Scheindlin did not invent those rules, nor did she invent the rule that litigants can use expert statistical analysis to make their cases.

Here are the discrimination facts in a nutshell: based on the NYPD's own data from more than four million stops, its officers disproportionately target black and Hispanic residents for stops. By “disproportionately,” I mean that however you analyze the data (and by “you” I mean Columbia professor Jeffrey Fagan), and whatever factors you control for, race jumps out.

A few examples: the best predictor for frequency of stops in any given geographic region (however defined) is the racial composition of its residents (and not, say, its crime rate).

Within any given geographic region, blacks and Hispanics were more likely to be stopped than whites; and the likelihood of the stop in fact resulting in any further enforcement action (that is, the likelihood that the stop produced any evidence of crime) was actually lower for blacks than for whites.

Here are the reasonable suspicion facts in a nutshell: bottom line—only 12% of all the stop-and-frisks resulted in any subsequent enforcement action (an arrest or a summons). The opinion has a detailed analysis explaining why the real figure is probably much lower (Fagan was as statistically conservative as possible and gave the NYPD the benefit of the doubt at the margin. Also, many stops generate no report at all, and those by definition did not uncover any crime). Officers found weapons (the ostensible public-safety benefit of the program) in only 1% of stops.

The opinion doesn't just rely on statistics, either. Judge Scheindlin also tried 19 individual stops as part of the case, with testimony from both the cops and the suspects.

OK, cards on the table: the real reason I hope Judge Scheindlin's stop-and-frisk decisions survive is that she cited one of my articles. (See Davis et al. v. City of New York, 902 F.Supp. 2d 405, 412.)

But the serious point, for readers who are trying to figure out what to think about these cases, is that it's not just knee-jerk leftists who think there's something constitutionally suspect about a program that, over the course of a decade, seems drawn to minority residents like a moth to flame, while remaining demonstrably terrible at ferreting out crime.

Caleb Mason is a regular contributor to The Crime Report. He is an attorney at Miller Barondess LLP in Los Angeles, and a former federal prosecutor. He welcomes comments from readers.

Comments are closed.