Science

The FBI Says Hate Crimes Are Soaring. It Actually Has No Idea.

We could have fixed this problem years ago. We didn’t.

The FBI logo with crime scene tape through it and question marks around it.
Photo illustration by Slate. Photo by the FBI.

The FBI released its annual report on Hate Crime Statistics on Tuesday, covering incidents occurring in calendar year 2017, and here’s the gist: It appears to be a pretty hateful, hate-filled time to be American. The latest numbers indicate—or perhaps confirm—that bias-based offenses have been on the rise. In 2017, 7,175 such incidents were reported to the bureau from across the country—the highest number since 2008, continuing what is now a three-year rise. That upward trend also appears to be consistent, broadly speaking, with recent, anxiety-provoking tallies made by local jurisdictions and by nonprofit groups such as the Anti-Defamation League. Indeed, the notion that we’re being pulled beneath a cresting hate crime wave has been prominent in recent headlines.

Is that notion accurate? The truth is that we don’t really know. There’s no doubt the hate crime numbers out Tuesday capture something real about the country’s mood: Gallup polls find that people are increasingly inclined to say they’re afraid of being the victim of a hate crime; with that proportion rising from 16 or 17 percent in the early 2000s to a high of 22 percent in the years since President Donald Trump took office. (For comparison, the number of those who worry over getting murdered has held steady at around 18 percent.) Still, it’s not clear whether this anxiety matches up with any higher real-world risk. It could be that causation works the other way: Hate crime tallies may be biased upward by the fact that we’ve grown ever more attuned to hate—and more likely than before to report bias-related incidents and details to the police. The resulting brume confounds debate and makes it hard to say whether any seeming rise in hate crimes might be genuine.

It’s true the FBI numbers out on Tuesday show we’ve hit a nine-year hate crime high in absolute terms. But we also know these stats are deeply flawed. For one thing, they surely undercount the total by a wide margin. To get a sense of this error’s potential magnitude, one need only consult the National Crime Victimization Survey, a rolling study of teenagers and adults from about 100,000 American households. The results of this research, also run by the Department of Justice, extrapolate to an average of about 250,000 hate crime incidents every year across the country, including those that never get called in to the police. (According to the study, slightly more than half are not reported.) The sources for the numbers from the FBI—as in, which law enforcement agencies are reporting them—also changes year to year, complicating any attempt to look at trends.

That the measurement of hate should be so uncertain even now, as politicians wallow in divisive rhetoric and press reports of hate-fueled violence and intimidation pile up, is altogether maddening. Not just because it complicates efforts to acknowledge (and then, one hopes, reverse) what really does feel like a rising tide of hate-related crimes, but also because we should be better at this by now. The same conundrum turned up 30 years ago, when the federal government faced an apparent resurgence of white supremacist violence. It wasn’t clear, back then, how bad the hate crime problem really was—or whether it was really getting worse—and this uncertainty was used to parry calls for legislation to address it. Eventually, in 1990, Congress passed the Hate Crime Statistics Act to codify the records kept by law enforcement agencies, so that the hate crime problem—whatever its scope—could be fully tracked and understood through yearly FBI reports. We’ve gotten these in every year since 1992, and yet the data remain somewhat unenlightening.

Efforts to improve hate crime accounting have long been driven by the sense we have today that such incidents are in the midst of a rapid and remarkable proliferation. That same dread was pervasive in the late 1970s and early 1980s, and helped inspire the Anti-Defamation League, the Southern Poverty Law Center, and the National Gay and Lesbian (now LGBTQ) Task Force, among other groups, to start counting up incidents of bias crimes via hotlines or by scouring the press. The widespread fear of growing racial and religious violence also led police departments in major cities such as Baltimore, Boston, Los Angeles, and New York to start totaling their hate crime stats around the same time.

In December 1980, Rep. John Conyers Jr. convened a hearing on the growing risks posed by neo-Nazis and the Ku Klux Klan, and by systematic racism in the military and law enforcement. California College of the Arts assistant professor Maxwell Leung has done an excellent job of tracing out the odd, extended story of this hearing—called “Increasing Violence Against Minorities”—and the push for federal hate crimes legislation that it helped inspire. The U.S. Commission on Civil Rights followed with a 1983 report on racial and religious intimidation and violence, which noted that the problem’s scope had been obscured, to some extent, by competing hate crime data sets that were based on different methodologies. Some agencies and groups would draw distinctions between acts of speech and violence, for example, while others didn’t. Certain data sets would also count attacks on poor and homeless people as hate crimes, and there were major differences over whether to include the targeting of women or LGBTQ groups. That helps explain why the commission’s first recommendation was that the government develop an “accurate and comprehensive” way to measure hate crime in America under the direction of the FBI. This idea would be picked up by Congress a few years later, with Conyers leading efforts to standardize and centralize the collection of statistics.

An incident in the Howard Beach section of New York City made the calls for better data more pressing. In December 1986, several black men were beaten by a group of whites outside a Queens pizzeria, and one of the victims—Michael Griffith—was subsequently chased onto a highway, where he was struck by a car and killed. The attack seemed to cause a rapid spike of hate crime in the city: The NYPD reported that its numbers doubled in the year after the attacks. Nationwide, advocates and the media claimed the upward trend in bias-motivated violence extended well beyond New York. “Hate Crimes on Rise in U.S.: Racial attacks, gay bashing burgeon,” the Boston Globe would say in a front-page story a few years later.

Yet there was also widespread disagreement over basic facts. In New York City, officials warned that hate crime numbers might only seem to be on the rise because people were paying more attention. “I think anytime you focus attention on anything like hate crimes, you tend to get an increased level of reporting,” the police commissioner said. The same official also wondered whether crimes were getting tagged with racial motivations even when they would have been committed either way: “Some hoodlums who would snatch your gold chain before will now snatch it and say, ‘Remember Howard Beach,’ ” he said.
Clearly, racial bias among cops could affect reporting, too.

On the national level, the Reagan administration came out against new record-keeping legislation on the grounds that it would be burdensome to law enforcement and probably unnecessary. “There’s no hard evidence to indicate that hate crimes are on the increase,” said a DOJ spokesman at the time. “There have been a few well-publicized cases, but we don’t know about a broader pattern.”

When the Hate Crime Statistics Act was finally signed into law by George H.W. Bush in 1990, it was meant to cut through years of this uncertainty and dithering. By that time, the measure had nearly universal, bipartisan support in Congress. (Sen. Jesse Helms led a small group of dissenters, who claimed that the effort to track hate crimes based on sexual orientation would be a Trojan horse for “the homosexual and lesbian legislative agenda.”) Better data were supposed to help prove that violent racist acts were common in America and—almost as importantly—that they were on the upswing, too.

That latter focus, the “spikes” and “jumps” and “surges” in the hate crime stats (rather than the naked totals), continues to this day, as an end in itself. The feeling of acceleration, the notion that things are always getting worse, offers leverage to anti–hate crime activists and provides them with a way to press for better policies.

It’s also useful simply because it’s hard to get your head around each year’s total on its own: Is 7,175 a lot of hate crimes for 2017, or not that many, or about what you’d expect? How much more should we be doing to control whatever hate crime problem that number represents? Laying out the same statistics over time perhaps allows us to convince ourselves that we’re getting answers to more pressing questions: Is this awful problem getting worse? If it is, then who or what might be to blame?

But the truth is that our stats remain unsound, even decades after all these fights to make them more definitive. The FBI reports required by the 1990 legislation don’t really help us measure long-term trends with any certainty or nuance; instead they function more as broad, symbolic figures of the problem. If you talk to experts in the hate crime field, they’ll tell you that, to some extent, the Hate Crime Statistics Act has fulfilled its promise: Each crime the bureau counts and puts into its data set represents one we know has been reported, investigated, and confirmed. As such, the FBI reports can be a vital baseline measure of the hate crime problem if not the encyclopedic standard for its analysis. When we say, “At least 7,175 hate crimes occurred throughout the U.S. last year,” it means that we have concrete proof of bias-based violence and intimidation as a real, ongoing threat to American life. That is obviously worthwhile.

Yet experts are also quick to say the FBI numbers are deeply flawed. Yes, they show that hate crimes happen, but they grossly underestimate the problem. The lowballing isn’t really systematic, either. Since the reporting of hate crime statistics by law enforcement agencies is optional, and not particularly incentivized, many either choose not to track their hate crime numbers and report them to the FBI, or else they aver—implausibly, impossibly—that no hate crimes of any kind have been committed in their jurisdiction. Sizable American cities including (but not limited to) Honolulu; Miami; Tulsa, Oklahoma; Arlington, Texas; Little Rock, Arkansas; and Newark, New Jersey, have all, at one or many points during the past few years, furnished exactly zero incidents to the national total. According to an Associated Press report from 2016, at least 2,700 police and sheriff’s departments have never filed a single hate crime report, while ProPublica reports that 120 federal agencies—including, yes, the FBI itself—also fail to send in hate crime data for the annual FBI report.

That’s not to say one can’t find some meaning in these gaps, or use them as impetus for action. When a major city claims it had zero hate crimes for the year, for example, activists or officials may then know to treat it as a problem spot, where for whatever reason, police aren’t really working hard to track the problem. These, in turn, may be the jurisdictions where the local government should be prodded or cajoled to greater action.

But on the whole, when it comes to understanding trends in hate crimes over time, these gaps are a disaster. You can’t even expect the data shortfall will hold steady in its scope from year to year, either—it turns out that the percentage of agencies reporting incidents varies widely over time. And to make things more confusing, Congress and DOJ have changed the rules for tracking hate crimes at several points. The original law called for counting any attack showing evidence of bias based on race, religion, sexual orientation, or ethnicity. In 1997, after further legislation, the FBI added categories for incidents showing bias against those with disabilities.
And with the passage of the Matthew Shepard and James Byrd Jr. Hate Crime Prevention Act in 2009, the bureau also started tracking crimes influenced by gender bias and those involving juveniles. Other changes along the way may also have influenced the numbers: Following the 2012 massacre at a Sikh temple in Wisconsin, for example, the Obama administration decided to split out the totals for bias crimes against Arabs, Buddhists, Hindus, Jehovah’s Witnesses, Orthodox Christians, and Sikh Americans.

Yet despite this broadening of categories, and larger pool of potential hate crime victims, the FBI data actually suggest that numbers have drifted downward over time. (The main exception is a sharp, temporary increase driven by anti-Muslim incidents in 2001.) The most recent total out today, of 7,175 for 2017, does represent a nine-year high in absolute terms—but it’s also smaller than almost every single tally from 1995 to 2008. Meanwhile, if you normalize the numbers based on how may law enforcement agencies participate each year, the 2017 figure—which amounts to 0.44 hate crimes per jurisdiction—matches up pretty closely with the average across the past decade’s worth of data. And the victim survey, which (as mentioned above) extrapolates numbers of hate crime incidents based on a repeated sample survey, reports no statistically significant changes in the overall number between 2004 and 2015.

In other words, any trends that one might find in FBI reports are highly suspect, if not spurious. According to American University criminologist and hate crime scholar Janice Iwama, you can get more useful information on how current events and ongoing policies affect hate crime totals when you look in greater detail at smaller units of analysis. For example, one could try to understand how the president’s rhetoric on immigration relates to hate crimes against Latinos across a handful of local communities of varying demographic characters.

Activist groups have also tried to push the government for better tracking of these crimes. One month after the Unite the Right rally and associated murder in Charlottesville, Virginia, dozens of organizations signed a letter to the Department of Justice calling for stronger policies in addressing hate-based incidents, including 10 specific ways that federal data collection might be improved. Law enforcement agencies should be incentivized to report their numbers, the letter argued, and DOJ should undertake a comprehensive study of the reasons why some jurisdictions continue to hold back.

That’s the right idea. After 30 years of sketchy data, it’s time for something better.

The mere feeling that we’re living through a major surge of hate-fueled violence may be useful in that it can be used to justify better anti–hate crime legislation. We don’t need stats to tell us that the racist rhetoric emanating from the White House will continue to cause real and lasting harm. But if the numbers from the latest FBI report reinforce this anxious feeling, it’s only in a vague, symbolic way. We still don’t know the truth about the state of hate and whether it makes sense to think that things are getting much, much worse. That ignorance is irresponsible.