The FBI is finally taking steps to better track police use of force across the US

The federal government is finally making a serious effort to get an accurate count of how many people are killed and hurt by police each year.

On Thursday, the US Department of Justice announced the FBI will start gathering more data on police shootings and other use of force incidents. Through the new program, the FBI will not only obtain reports of deaths, but also nonlethal police encounters. The scheme, which will start as a pilot initiative, will also provide funding to police departments to help collect a wider range of policing actions, from traffic stops and searches to use of nonlethal force and shootings.

The FBI program will begin in 2017 — initially involving certain local, state, and federal law enforcement agencies, with the potential to expand from there.

Still, there’s one problem: The FBI program will continue to at least partly rely on local police departments voluntarily submitting data on policing actions. That could lead to some of the same big underreporting issues that are present in the database today.

The FBI plans to conduct on-site visits to the agencies to check for underreporting. But it’s unclear how the feds will enforce financial penalties or other incentives to get local agencies to comply with their requirements if they do discover underreporting.

If that doesn’t get resolved, the database could prove just as lacking as the ones that came before it. After all, the current database is estimated to miss more than half of people killed by police. A marginal improvement just won’t cut it when the reporting gap is so big.

The FBI isn’t the only federal agency trying to expand the data that it collects on police. Previously, the Bureau of Justice Statistics (BJS) announced a program separate from the FBI’s that will also try to better track and verify police killing data.

Federal officials wrote in the Federal Register that the BJS database will try to tackle underreporting problems with a “redesigned methodology” involving open sources and local police agencies as well as medical examiner and coroner offices.

BJS said it will enforce the collection of these statistics through existing federal law — which requires police agencies report all deaths under their custody or lose federal funding — and verify the data by itself. It will collect data ranging from the causes of death to demographic data about the victims. The new database will first give annual data for 2016, then move to a quarterly basis afterward.

The FBI and BJS’s new programs will supplant the current databases kept by the two agencies, both of which severely undercount police killings.

A 2015 study by RTI International, which conducted the analysis for the BJS, found that from 2003 to 2009 and 2011, BJS’s Arrest-Related Deaths captured approximately 49 percent of people killed by police, while the FBI’s Supplementary Homicide Reports captured 46 percent. Neither system picked up about 28 percent of law enforcement homicides, meaning more than one-quarter of police-caused deaths weren’t tracked at all by the FBI and BJS’s databases.

Better data would be a big deal for police reforms: Although police killings have been a big part of public and political debates since the Ferguson protests two years ago, the media, criminologists, activists, and even police have had to rely on third-party data to study and evaluate the number of police killings — and that data is very often missing crucial information, such as the victim’s race or the events that led to the killing.

So if the BJS and FBI databases truly manage to be comprehensive, they will help Americans finally know the actual scope of police killings in the US. But it remains to be seen if the federal agencies will actually succeed this time around.