D.C.'s Police Force Shrank. How Did It Get Safer?

A new report details how to do more with less—but police staffing still matters.

Courtesy Heather Diehl/Getty.

Washington, D.C.’s Metropolitan Police Department (MPD) has a staffing problem. Since 2020, the department has lost 17 percent of its sworn force. The current staffing level—3,149 sworn officers—is the lowest recorded figure since 1967.

And yet, remarkably, the District has gotten a lot safer. Last year saw the fewest homicides in a decade, with trends pointing to an even lower rate this year. Carjackings and auto thefts, too, are back down from sky highs in 2023 and 2024. In a 2024 Manhattan Institute report, I argued that MPD was “doing less with less,” keeping crime high even as it declined elsewhere. Now, crime has dropped even as staffing is still down—what gives?

The answer, argues a recent report from my friends at the Niskanen Center’s criminal justice policy shop, is that MPD figured out how to do more with less. They started implementing focused strategies, using proven policing techniques to bring crime down. That’s a welcome finding—but the city still needs to be careful about where its system is going over the long run.

Just a few years ago, D.C. had some major crime problems. As I noted at the time, the city was experiencing surges in homicides as well as in carjacking and auto theft.¹ The homicide rate among young black men nearly doubled between 2014 and 2024. A surge of juvenile carjackers had pushed car theft rates to their highest levels on record.

At the same time, MPD’s staffing had plummeted, the result of a “Defund”-motivated budget cut by the City Council as well as a general malaise among police officers. Police proactivity—measured as a composite of arrests and stops—had fallen by roughly half compared to the pre-Defund period.

Source: Lehman 2024

Two years later, though, crime is much lower, even as staffing hasn’t recovered. What explains the difference?

Obviously, the answer isn’t that a reduced officer headcount resulted in less crime (though some extreme progressives might argue otherwise). Nor, the Niskanen report argues, is it about where police officers are deployed—MPD data indicate that police are active today in roughly the same places they were in 2023.

The difference, rather, comes down to what MPD officers were doing with their time. The number of arrests increased 40 percent between 2023 and 2025, according to the Niskanen report. “Fewer officers were generating substantially more enforcement output,” the report notes.

Those arrests were the result of “proactive, upstream enforcement strategies,” Niskanen argues. MPD began running a series of “focused initiatives” in 2024 that aimed police resources at problematic areas, like the nightlife-heavy H Street corridor or neighborhoods with high levels of gun violence. In these areas, MPD made many more arrests for relatively minor offenses conducive to such strategic enforcement. As a result, narcotics arrests rose 150 percent; traffic stops doubled; and release violations and fugitive charges (arresting people with outstanding warrants) rose 50 percent.

“These are arrests that require officers to go looking for crime before it happens — stopping cars, executing warrants, sweeping for drugs and illegal weapons, pulling people with outstanding violations off the street,” the paper notes. “They signal a deliberate strategy of getting guns, drugs, and dangerous people off the street before violence happens.”

That’s a welcome change of pace. As I noted in my 2024 report, MPD arrests’s had declined steeply as of 2023. This was especially true of the sort of quality-of-life offenses usually used in proactive policing. Narcotics arrests and traffic violations, for example, had fallen 77 percent compared to 2019.

The Niskanen team also investigates the impact of another potential cause of the decline: the deployment of the National Guard to D.C. at the order of President Trump. I’ve argued that the deployment probably helped cut crime. And indeed, the Niskanen team uses data on the timing and location of the deployment to estimate that the guardsmen helped reduce property offending by 24 percent, albeit without any effect on violence.

Source: Niskanen 2026

That’s a real and meaningful reduction. But it’s also not the best bang for the bucks expended on the Guard deployment. The Niskanen report also estimates that spending the same amount on additional MPD hiring could yield 12 times as much in terms of crime prevented; focusing those additional MPD staffers on hot spots would yield a 35-fold increase in returns.

It seems, as previously mentioned, that a resource-constrained MPD figured out how to police smarter to compensate for not having enough manpower to go around. That’s a valuable lesson for other police departments, many of which are just starting to return to pre-2020 staffing levels. Strategic, proactive policing that focuses on where crime is actually happening remains a better approach than the reactive model favored by more lackluster departments.

Washington’s success, moreover, reinforces the importance of using petty crimes to handle big problems. Such “broken windows” policing is sometimes unpopular, because on paper it looks like cops sweating the small stuff for no good reason. But narcotics arrests, traffic stops, and other offenses where probable cause is easy to establish are the backbone of effective, strategic policing. Big cities reduce them at their peril.

At the same time, Washington’s success shouldn’t make us sanguine about the longer-run trend of its declining officer headcount, or those of other big cities. America has an underpolicing problem. As I’ve noted, the ratio of sworn officers per 1,000 Americans has been declining steadily since the Great Recession. Given that policing is one of the few reliable, scalable tools we have for controlling crime, that’s a reason to worry.

It’s possible to compensate for this reduction in staffing levels through smarter policing. But a preferable alternative is to both use officers more intelligently and dedicate more resources to the suppression of crime. As Niskanen’s analysis shows, doing so in D.C. could generate billions of dollars in social benefit. That’s not just money—it’s lives and livelihoods saved.

1. Auto theft is when you steal a car. Carjacking is when you steal a car by force.