The recent spate of highly publicized killings of black men by police officers has driven many people to wonder how common such police violence is. Incredibly, no one knows—not even FBI Director James Comey. The federal government doesn’t track it, and until this week, no one else did either.

Earlier this week, The Washington Post and Guardian set out to catalogue the number of times people are killed by police—the Post focusing on firearms deaths, while The Guardian looked at all fatalities. They got the data by poring over news accounts, police reports, and other records, trying to get a full picture. Because they use different methods and track slightly different things, they also come up with two different totals—385 so far this year in the Post, 470 for The Guardian.

The federal government could theoretically compel local law enforcement to produce the information, guaranteeing reliable and uniform reporting. So why hasn’t it done so? Tracking and reporting these numbers provides an empirical basis for political debate, and shifts the focus from documenting the problem to proposing solutions.

Senators Barbara Boxer of California and Cory Booker of New Jersey announced a bill Tuesday that tries to solve the problem. The Police Reporting Information, Data, and Evidence Act of 2015 (the name creates a ghastly backronym, the “PRIDE Act”) is a fairly straightforward bill: It creates grants to states and in return requires them to record and report to the Justice Department every case in which an officer shoots or causes serious bodily injury against a civilian, and every case in which a civilian shoots or causes serious bodily injury to an officer. They would also have to report basic demographic data for all victims. The text is short and simple.

The PRIDE Act covers some of the same territory as the Death in Custody Reporting Act, an old law that fell off the books but was renewed in late 2014. The reason the PRIDE Act is still necessary is that its predecessor is widely considered ineffectual. Unfortunately, it’s not entirely clear that the PRIDE Act will solve its problems, as Franklin Zimring, a professor of law at the University of California at Berkeley who has long studied police use of force, told me.

The new bill and the old law share two salient features: They rely on states to provide the information needed, and they use federal grants as the mechanism for compliance. The way they handle the grants is a little different, though. The DCRA threatens to cut existing federal funding to states if they don’t report their numbers. Withholding federal grants, however, hasn’t always been a successful tool to make states comply with laws they don’t like. The Adam Walsh Child Protection and Safety Act, which uses the same mechanism, has only enlisted 17 states. The PRIDE Act instead creates new grants, relying more on carrots than sticks to achieve its aims.

That seems to answer the concerns often raised by local law enforcement agencies. Last fall, James Pasco, executive director of the Fraternal Order of Police, warned The Washington Post’s Wesley Lowery that any proposal to gather data that didn’t provide money for its collection was dead on arrival. “Otherwise it’s an unfunded mandate,” he said. “About 80 percent of police departments have fewer than 10 officers. They don’t have huge data collecting operations. They don’t even have a single person in some of these departments who are dedicated to all the statistical work they have to do now.”

The PRIDE Act might solve that problem (though its cost to taxpayers is unclear), but Pasco’s qualms spotlight the other potential pitfall in the bill.

“The states employ less than 5 percent of the police officers and sheriffs that operate in the state. How are they going to get the information?” Zimring wondered.

It’s a real problem. The Bureau of Justice Statistics tallies more than 12,000 local police departments, and there are thousands of sheriffs, too. As Pasco noted, few of these departments have statistics sitting around, just waiting to be collected by the state. That means the states, in turn, also don’t have the statistics sitting around, just waiting for Congress to pass a law requiring them to send the data to Washington. From a legislative standpoint, it’s much easier to institute a mandate on the states than it is to design—and fund—a program that collects data from every local authority. But it also reduces the chances of harvesting reliable, comprehensive data.

This isn’t the first time that Congress has sought to scrutinize questionable practices in local policing. The last major national outcry over police brutality was in 1992, when riots broke out after Los Angeles police officers were acquitted in the beating of Rodney King, which had been videotaped. Those riots produced a huge amount of media attention, and they also produced calls for reform.

Then-Representative Henry Waxman, a California Democrat, tried to pass a law that would allow the Justice Department to investigate local police departments and sue them for excessive use of force. The bill went nowhere. But in 1994, Waxman managed to get it included in the Violent Crime Control and Law Enforcement Act of 1994, which did pass. The relevant bit, Section 14141, makes it illegal for police to “engage in a pattern or practice of conduct ... that deprives persons of rights, privileges, or immunities secured or protected by the Constitution or laws of the United States.” There’s serious irony here: The act as a whole signaled a major expansion of incarceration and federal involvement in crime. It created many of the tough-on-crime policies that are often blamed for the current backlash against cops, and yet it also includes the tool that Attorney Generals Eric Holder and Loretta Lynch have used to push back against abuses.

“It is the most important regulatory tool that the federal government now has for local police departments,” Zimring says. “That more than anything else is the most substantial legacy of the King case.”

For example, the consent decree that the Justice Department recently announced with Cleveland over a pattern of excessive use of force by officers sprung from an investigation conducted under the auspices of Section 14141. There have been similar agreements with Albuquerque, Oakland, Seattle, and Los Angeles. An investigation that Baltimore requested into its own police department also falls under the statute.

Section 14141 was effective largely because it allows the Justice Department to go directly to local police departments, investigate them, and force them to make concessions. When Section 14141 was passed, though, few expected the small piece of legislation to have a huge impact on a public debate 20 years down the road. If Congress can devise an effective mechanism for gathering data on police-related deaths, perhaps it can produce a similarly outsized impact. In the meantime, that hole in the data seems likely to remain a major obstacle to police reform.