The Algorithm of Housing Anxiety

The use of computer software to set rental rates is a path to societal suffering.

Often we focus on the end results of a fundamentally flawed decision. For example, the media tends to focus on anti-democratic political efforts, and backdoor deals that benefit only the rich and corrupt politicians—not the fact that corporate money (thanks to past Supreme Court decisions) can flow virtually unrestricted into political pockets. The rot, as they say, is at the root. It began with a bad idea.

Along similar lines, we're being told that the nation has a vast and deep housing shortage. Young families can't find homes within their budget, and an increasing number are becoming semi-permanent renters. For example, apartment listing service RentCafe reports that the percentage of rentals to ownership is at the highest level in 55 years. In the biggest 50 metro areas, 101 ZIP codes flipped from homeowning house­holds to rentals over the past 10 years.

Markets ebb and flow, of course, but something else is happening this time. Rents are rising, even in places where intuition might suggest they can't be sustained. But intuition, and human beings for that matter, may have little to do with it. 

According to investigative news site ProPublica, in some cities, as many as 70 percent of rentals run by large compa­nies such as apartment manager Greystar are being priced with a piece of software called YieldStar, produced by property management firm RealPage. What’s chilling, if not surprising, about this landlord-by-proxy technology is that it’s actually promoted as a way to take messy, emotional human beings out of the landlord-renter relationship.

Why does this matter? Because traditionally, renters in many parts of the United States have been able to negotiate for better terms. This, in turn, allows them to save money for a down payment on a home, put a kid through college, and surpass other basic middle-class stepping stones.

One RealPage exec bragged to ProPublica, that as a property manager, “very few of us would be willing to actually raise rents double digits within a single month by doing it manually.”

Note that this caution isn’t out of fear of upsetting tenants. It’s fear of overreaching the market and ending up with empty units. There’s only a certain risk landlords are willing to take.

Not to worry, says the YieldStar computer. The algorithm shows that if you toss out the current tenants and raise rent by 20 percent, you’ll haul in a big return even if the apartment sits empty for three months.

ProPublica reports that one developer for data analysis firm RealPage told them that “leasing agents had ‘too much empathy’ compared to computer-generated pricing.”

Some of you are probably thinking, what’s the problem? Maximizing profits on rentals is the end game, right? Am I a luddite, because the idea of putting human interactions in the hands of a computer makes me uneasy?

There’s a reason, for example, that our air force does not hand over command and control to computer algorithms. What if the algorithm shows that a nuclear first strike has a 60 percent chance of killing 1 billion fewer people than waiting to give a retaliatory response?

For the shrinking middle class, trapped between high home prices and volatile, computer-controlled rents, a genuine desperation is setting in. Forget those retirement aspirations. Back into the workforce you go, this time at the bottom of the ladder to cover your additional living costs.


Rising rents. Nationwide, rents rose dramatically over the past two years. Computerized decision making bears part of the responsibility. SOURCE: Redfin  

Among my renter friends, the cycle goes something like this: shock as you get your new rent notice; panic as you plead with the property owner for some empathy (to no avail); frustration as you seek a new, more-affordable place (there’s nothing); and resignation, as you realize you’re stuck, and getting stucker.

Naturally, YieldStar’s management defends the use of its software, downplays its negative impacts on the rental market, and harps on its continued growth in influence and reach. They note that when it’s first introduced to property managers, the staff tend to resist 10 to 20 percent of the suggestions. Multifamily owners typically like to keep occupancy at 98 percent or so. With computers at the helm, 95 percent is the new comfort level.

Apartment construction is red hot, but few of those units will fall in the affordable category. The system is failing all but the top of the economic ladder. My prediction: You’re going to see a surge of restrictive rental legislation in the next few years at the metro level, not unlike what’s happening where I live, in Portland, Maine. There are rent caps, strict limits on rent increases, and tenant protec­tions that make eviction slow and costly. These are not all good policies, but as housing costs hit existential levels, algorithm proponents may run into something that does not compute: rebellion and rage against the machines.

Ultimately, the attempt to gouge every last housing dollar from the middle class by testing every limit (as computers like to do) may result in a backlash that could set the rental market on its heels for years to come.

Editor's Note: Since writing this piece, warnings about the misuse of algorithms by governments and corporations has proliferated. For example, take a look at this piece from Wired about how the government can ruin lives with algorithms. Another story about Rotterdam shows how use of algorithms to identify the likelihood of people to commit fraud has backfired miserably.

This story is part of a special report presented with generous sponsorship from: ProVia, Whirlpool, Cultured Stone, and Sonos.