The FBI’s latest numbers showed a 17 percent increase in reported hate crimes in 2017.

But what does this actually say about the actual number of hate crimes occurring in the U.S.? Not much.

The Hate Crime Statistics Act of 1990 mandates that the FBI publish statistics specifically for crimes motivated by bias, and a broad network of state laws require that hate crimes are both tracked and prosecuted. Despite this, a variety of problems plague the implementation of these laws.

A total of 16,149 law enforcement agencies reported 7,175 incidents of hate crimes to the FBI in 2017. That means that 87.4 percent of law enforcement agencies reported zero bias-motivated crimes.

On the surface, this number seems suspiciously low, particularly when considering that the other federal survey of hate crimes, the National Crime Victimization Survey, which looks at how many people say they’ve experienced a hate crime, estimates that “U.S. residents experienced an average of 250,000 hate crime victimizations each year from 2004 to 2015.” There remains an enormous discrepancy between what victims report as a hate crime and what law enforcement agencies do.

The FBI numbers appear even more suspicious in light of the fact that, as the Arab American Institute points out, several high-profile hate crimes are not in the data. For example, Kansas reported no hate crimes for 2017, despite the murder of Indian immigrant Srinivas Kuchibhotla. Adam Purinton reportedly yelled, “Get out of my country,” before shooting and killing Kuchibhotla and wounding two others in an Olathe, Kansas bar. Purinton was caught hours later in Missouri after telling a bartender at an Applebee’s that he’d shot two “Iranians” and needed a place to hide.

Why are hate crimes underreported?

Since 2016, the investigative journalism organization ProPublica has created a national network of news organizations to document hate crimes. The Documenting Hate Project recognizes that “There is simply no reliable national data on hate crimes.”

Journalist Rachel Glickhouse, who helped found Documenting Hate, told me that the FBI data are useful in showing broad patterns around hate crimes occurrences, but there are a variety of reasons why their numbers are likely incomplete.

One problem she noted was that many police departments provide either zero or cursory training around hate crimes. In some cases, state officials were even misinformed about relevant statutes in their states. Different departments also track hate crimes differently, making it challenging to easily find data.

Buzzfeed reporters looked at over 2,400 police incident reports from 10 of the largest police departments that reported no hate crime incidents for 2016. The investigation identified “15 assaults in which the cops’ own narratives suggested that the suspect may have been motivated by bias.”

Lack of trust between law enforcement and communities that have been historically targeted by hate also leads to underreporting, says Arusha Gordon, counsel for the Lawyer’s Committee for Civil Rights’ Stop Hate Project, an organization that trains law enforcement and prosecutors on responding to hate crimes. It’s important to work on “building trust between law enforcement and communities that have been historically targeted by hate.” For example, immigrant and LGBT communities “might not feel comfortable reporting to police because of their citizenship status or because they might have had a previously negative interaction with someone in law enforcement.”

Gordon stressed how important it is to also train law enforcement to recognize “indicators of hate.” For instance, the number 88 is a key symbol in the contemporary white nationalist movement. (The letter “H” is the eighth letter of the alphabet, so 88 stands for “Heil Hitler.”) If a property crime includes an 88 spray-painted on a building, the police officer might not understand the crime’s connection to a hate movement.

Former FBI and and Homeland Security employees have said that law enforcement doesn’t pay enough attention to racist groups or downplays hate crimes incidents to victims. For example, in December, New York City’s deputy human rights commissioner experienced a racist incident, but city police officers discouraged her from reporting it – an incident that’s part of a broader trend of undercounting hate crimes in New York.

Domestic terrorism

Having studied the white supremacist movement in the U.S, I see the problems associated with the FBI’s hate crimes reporting as part of a larger issue related to FBI priorities around fighting terrorism.

Threats by far-right groups aren’t addressed with the same resources, even while they motivate a variety of bias-motivated crimes. In fact, there are no longer any Homeland Security officials dedicated to monitoring right-wing extremism.

In 2009, years before Charlottesville and a string of high-profile white supremacist murders, Daryl Johnson – then an analyst for the Department of Homeland Security – authored a report outlining the threats posted by “Rightwing Extremism.” In hindsight, the analysis predicted much of what has come to pass: a rise in hate crimes and domestic terrorism carried out by white supremacist and far-right groups, many of them former veterans.

However, a broad political backlash from conservative media and veterans groups forced the Department of Homeland Security to rescind the report. Johnson’s team was disbanded the following year.

Former FBI special agent Michael German, who worked undercover to infiltrate right-wing extremist groups before leaving the FBI in 2004, told me that law enforcement might be failing to see hate crimes as possibly part of a broader pattern. For example, if a skinhead group is engaging in organized crime, law enforcement may not see the connections between individual violent acts and never investigate the broader organization.

There are additionally a number of examples of situations when the FBI did not label criminal acts as domestic terrorism, even when the crimes fit the statutory definition. Although far-right violence does not always qualify as terrorism, it does count when it is done in an effort to change the broader political discourse.

Studies show that far-right extremists have caused more deaths than Islamist extremists since 9/11. The most comprehensive database on far-right violence in the U.S. tracked 4,420 violent incidents carried out by far-right actors between 1990 and 2012, including 670 deaths and over 3,000 individuals injured.

In a statement on the 2017 hate crimes report, the FBI said that it is “working with law enforcement partners across the country to encourage reporting of hate crime statistics.” They promise that, in 2019, they will provide training for law enforcement officers in how to identify and report bias-motivated incidents. I feel that this is an important step in working to fix the problems in the U.S. with reporting and prosecuting hate crimes.

Sophie Bjork-James does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.