A city was handed to the army to be made safe, and a year later it is marginally less safe, cost millions a day, and lost at least one man who could have lived.

"The Pentagon's acknowledgment that the deployment will continue until 2029 confirms what we've said all along. There is no real emergency here."

— Monica Hopkins, executive director of the ACLU of Washington, D.C.

It is now the one-year mark of the military occupation of Washington, D.C., and the ledger has come in. A Reuters investigation released Wednesday tallied what the National Guard has actually done in the capital since it arrived last August, and the result is the quiet, unglamorous collapse of an idea: that soldiers in the street are a synonym for safety.

The troops now outnumber the city's own police force. They cannot make an arrest, because they can only detain and then call an officer to finish the job. Yet in the entire time they have been deployed, they show up in roughly 1.3 percent of the criminal cases filed in the District's Superior Court, the court that handles nearly all local prosecutions. The ones they did touch, charging documents show, were mostly fare evasion and shoplifting in neighborhoods Reuters describes as "wealthier, whiter, and less dangerous than the rest of the capital."

Translation: the army was parked in the parts of town that looked the best in a photograph.

The number that should trouble the most is the one at the center of the case. Trump sent the Guard in to save a city from "crime, bloodshed, bedlam, and squalor." A year on, violent crime has not fallen. Metropolitan Police Department data shows 1,712 violent crimes so far in 2026 against 1,645 at the same point a year earlier, a four percent rise. The presence of soldiers changed nothing about the thing they were sent to stop.

Reuters could find no reference in court records to soldiers doing any kind of law enforcement in the neighborhoods where, over the past five years, about 82 percent of the city's 859 murders occurred. The occupation, in other words, has largely been conducted around the places where the actual harm is happening.

The human cost is not abstract, and it is not what the administration has claimed. In one incident, police arrived to arrest a man who had been stomping on a woman outside a nightclub. The responding officer said there were "approximately 15 National Guard members in the area on routine patrol who did not take any action." In another, soldiers found a man who had been stabbed in an upscale waterfront district. Believing him merely drunk, they left him lying on the ground without performing any emergency measures. When police arrived, they realized he had been "stabbed so severely his intestines were spilling out." He died two hours later in the hospital.

A deterrent, in the administration's framing, is supposed to mean fewer people dying, not soldiers standing in a semicircle around a man who could have lived.

The price is steep enough to have its own math. The nonpartisan Niskanen Center puts a deployed Guard member at about $607 a day, against $384 for an MPD officer. The administration has estimated that keeping the deployment running until the end of Trump's term in 2029 will cost around $1.4 billion. A researcher's earlier estimate, before the force doubled, pegged the daily cost at roughly four times what it would take to house every person living on the city's streets. Applied to today's force, the ratio stretches to nine or ten times.

The Guard, the administration also made clear, was doing real work, but work of a different kind. It helped federal agents dismantle homeless encampments last summer, in what Trump called a war on "vagrancy." So the troops were, quite literally, cleared out of some neighborhoods by the same men who had been posted there to make those neighborhoods safer. The two projects were the same project.

The critics have a simpler word for what is happening. "I can think of many other ways that federal resources could have been invested to make the nation's capital safer," said Clinique Chapman, CEO of the DC Justice Lab, writing in the Washington Post. For what the Guard costs to deploy for six days, taxpayers could have funded a full year of trauma recovery centers for crime survivors in every ward. For 41 days of deployment, the city could have paid living wages and the rent on a one-bedroom apartment for every resident returning from federal prison.

Hopkins is more blunt about the point. The point, she argues, was never to make D.C. safer, because D.C. was never unsafe enough to require an army. "Because the nation's capital is not legally recognized as a state, extremist politicians have used D.C. as a testing ground for anti-democratic orders," she wrote. "We cannot let military deployment on the streets of the United States become normal or acceptable as a part of daily life here or anywhere else."

There is a particular cruelty in the timing of the one-year mark, which lands with the Pentagon having confirmed the soldiers will stay until 2029. The experiment was never going to be measured. A deployment that outlasts the president who ordered it cannot be blamed on the president, and cannot be retired. It simply becomes infrastructure: the thing about the city, the way the flagpoles are about the building, assumed to have always been there.

A year ago the promise was that the army would make the capital safe, and that the proof would be visible in the falling crime figures. The figures rose. The troops stayed. The question was never whether the deployment would work, because no one was willing to define what working would look like or when it would end. The question is why a city already near its lowest crime in decades was treated as a problem only a military could solve, and why, after twelve months of data saying otherwise, the answer the nation received was simply to keep the soldiers there a little longer.