I’ve noticed numerous stories and tweets about a building boom: for example, a recent CNBC story asserts that the number of new apartments is “at a 50-year high.” Various twitterati have used this claim to support their own points of view: some claim that rents are stabilizing because of this new surge in supply, while others argue that the failure of rents to decline shows that new supply doesn’t reduce rents.
This article was written by Michael Lewyn and originally published by Market Urbanism.
But is supply really increasing that rapidly? Federal statistics on housing construction are at a Census housing data webpage. I looked at the “New Housing Units Completed” table and found that about 216,000 housing units in structures with over five units were completed in the first half of 2023.
On the positive side, this is definitely an improvement over the 2010s, when the economy was still recovering from the 2008 recession. For example, in the first half of 2019, just over 169,000 such units were built, and 2018 was pretty similar.
But is construction still up to Reagan-era levels? Not really. In the first half of 1986, almost 258,000 relevant units were completed. And in the first half of 1973, just over 378,000(!) such units were built.
And these levels of construction were in a less populous country. Today the U.S. population is about 335 million, up from about 240 million in 1986 and 212 million in 1973. So if construction had kept up with population, our new unit count would be about 1/3 higher than in 1986, and almost 60 percent higher than in 1973. Instead, construction went down.
To put the facts another way: our half-year multifamily construction rate is about 644 per one million Americans for 2023, down from 1075 per million in 1986 and 1783 per million in 1973. That’s not my idea of a “50-year high.”
