Member-only story
How An Algorithm Is Helping Drive Up Rent
Big landlords are using “YieldStar”, which finds the highest price a market can bear — leaving rents higher, and units emptier
Recently, Annie Lowrey wrote a good piece for The Atlantic noting just how bad rental housing is for Americans right now.
As she wrote, rent is eating an increasingly large share of people’s budgets:
Nationwide, the share of renters who are considered “burdened” — spending more than 30 percent of their income on rent and utilities — has climbed to 47 percent; one in four renters — about 11 million — spend more than half their income on shelter. Renters today spend about 10 more percentage points of their earnings more on housing than they did in the 1970s.
The problem, of course, is that there simply isn’t enough housing. Many American cities haven’t built anywhere near enough in recent decades. The reasons why? Manifold: NIMBYism, local regulations that prohibit dense housing, developers who can make more money on a single giant home than several smaller starter homes, and more. By the estimates of the experts Lowrey consults, the US needs anywhere from 3.8 million to 7 million more units, yowsa.
But recently I stumbled upon another reason why rents may have soared so high:
Many big landlords are using software to optimally price their rent — and discovering that the code can recommend jacking rent high and leaving units empty.
The software in question here is called YieldStar. It’s a product offered by the realty firm RealPage, and what it does is analyze rental markets and predict prices that a realtor could feasibly get for any given property. In essence, it tries to maximize the rent landlords can charge. I learned about YieldStar via this fabulous in-depth ProPublica investigation by Heather Vogell (with data analysis by Haru Coryne and Ryan Little); the material in this blog post is sourced from that story, and you should go read the whole original piece here!
As Realpage tells it (and they may obviously be boasting), the software has high success rates for landlords, who apparently accept the software’s recommendations up to 90% of the time. And when landlords…
