The Setback
R-01 Published
Regulatory Vintage 2025-2026 · Published Sep 9, 2026

Washington, D.C.

The city hit its housing target the year the pipeline ran dry.

Supply Scorecard

How Washington, D.C. measures up

Supply scorecard: sourced metrics
Metric Value Vintage Source
M1 Housing units permitted per 1,000 residents 2.3 2025 FRED
M2 Multifamily units permitted 1,372 2025 DC Policy Center
M3 Decline in total units permitted since 2022 -79% 2022-2025 FRED
M4 Share of residential land where multifamily is allowed 26% 2026 DC Policy Center
M5 Average days added by a zoning appeal 450 2015-2025 DC Policy Center
M6 Year-over-year change in average asking rent -3.8% Aug 2026 RentCafe
M7 Multifamily rental units that are true market rate 21% 2024 DC Policy Center

In July 2024, Mayor Bowser announced that Washington had built 36,216 new homes since 2019. The goal was 36,000. The city had cleared it.

The win was real. It was also already over.

A home takes two to three years to go from permit to move-in day. The homes the mayor celebrated were approved in 2019, 2020 and 2021. Back then the District approved five to eight thousand homes a year.

Then approvals collapsed. In 2022 the city permitted 7,705 homes. In 2025 it permitted 1,591. In January 2026, one apartment building in all of Washington pulled a permit. It had thirty units.

5,945 2019 7,370 2020 4,740 2021 7,705 2022 3,020 2023 1,737 2024 1,591 2025

Homes permitted in the District, by year. Source: U.S. Census Bureau Building Permits Survey.

Whatever opens in 2027 and 2028 is whatever got approved in 2025. That is close to nothing.

The slow part is zoning

Ask people in Washington real estate what went wrong and most say permitting. The record says otherwise. The city’s building department starts reviewing plans within two days almost every time. It finishes the first review within a month 97 percent of the time. That is fast for a big city.

The slow part is zoning, the rules about what can be built where.

Three quarters of the District’s residential land allows only single-family houses. Apartments are allowed on about a quarter of it. Anyone who wants to build more than the map allows has to ask permission. That means hearings before several boards and, in much of the city, a historic review.

None of it runs on a clock. Since 2015, neighbors and other opponents have filed 89 appeals against zoning approvals. They won nine. The losing appeals still did damage. An appeal adds about 450 days to a project whether it wins or loses. Those 89 appeals touched more than 17,000 homes.

So a builder in Washington takes a gamble on every project. They cannot know when approval comes, or if it comes at all.

That is a real problem. It is no longer the main reason nothing gets built.

A normal building loses money

Picture an ordinary apartment building. Forty units, mostly one-bedrooms, built and rented at Washington’s average prices. The kind of building the city says it wants.

It costs about $16.3 million to build. Once it is full, buyers would pay about $11.6 million for it.

The builder loses $4.8 million on day one.

Investors want a building like this to earn about 6.5 percent a year on what it cost. This one earns 3.7 percent. At that rate, the money goes somewhere safer.

That is why one building pulled a permit in January.

Fixing zoning closes a quarter of the gap

Suppose the city passed every reform on the wish list at once. Automatic approval for apartments. Reviews capped at four months. No required parking. Cheaper building-code rules. No affordable-unit requirement.

The building goes from earning 3.7 percent to 4.4 percent. It still needs 6.5.

Earns 3.7% Needs 6.5%
Everything zoning reform can do Still short

Grant every reform at once and the building still does not work.

Most of the shortfall sits where zoning cannot reach.

Faster approvals help least of all. Waiting is cheaper than it sounds, because during approval the builder only owns land. The building does not exist yet. Ten months saved is worth about $160,000 on a $16 million project.

The real cost of delay is risk. Banks charge more, or walk away, when approval might never come. That is the honest argument for fixing the process.

Rents fell, costs rose, jobs left

Average rent in Washington is about $2,450 a month. That is down almost 4 percent in a year. Over the same year, building got about 4 percent more expensive.

The region also lost 122,300 jobs in the year ending February 2026. Federal jobs made up 56,300 of them.

Washington spent forty years building a system to slow growth. Now there is no growth to slow.

The land is worth less than nothing

Remove every rule and the building still cannot pay for its land. For the project to earn what investors need, someone would have to pay the builder about $1.5 million to take the lot.

The market already knows this. Through 2025 and 2026, banks pushed owners to sell development sites, and many went up for sale. One approved site in the region went to a foreclosure auction. Sites that do sell are being replanned smaller.

What would work

Fix zoning anyway. Demand will come back, and the city needs to be ready when it does. A process that adds 450 days no matter the merits is broken on its own terms. Just stop promising it will fix next year’s numbers. If it gets sold as a quick fix, it will be blamed for failing and dropped.

The fastest lever is taxes. Washington charges apartment buildings 85 cents in property tax for every $100 of value. For this building, that is about $98,000 a year, paid out of rent in a market where rent is already too low. A ten-year tax break helps this building about as much as every zoning reform combined. One is a single vote. The other is a rewrite of the zoning code.

Some of this is not a housing problem at all. No zoning code brings back 122,300 jobs.

How this was built →


Want to see how the building’s numbers work? Read the numbers behind this report →