Decades ago hospitality pioneers invented revenue management when they decided rate-setting should be data-driven and strategic, not gut-driven and impulsive. Revenue management as a discipline has evolved significantly since then, attracting teams of data scientists using more and more powerful analytics. Eventually it jumped to multifamily with early movers like Yieldstar and LRO, which RealPage owns today.

In multifamily, despite lower technical sophistication of the revenue management tools, less overall adoption of the technology and the relative lack of revenue management professionals, it’s become quite a thing. RealPage and others have been sued by the federal government, several states and by private parties, all basically arguing its revenue management tools facilitated price-fixing by multifamily properties that led to inflated rents “above competitive levels.” Setting aside the fact that you could say the same about potential price fixing in hotels, and there have been zero lawsuits about that, the bigger question is simply, “is it true?” Does algorithmic revenue management inflate rents above competitive norms?

Explosive research released last year says yes: Thanks to algorithmic pricing, rents were $25 per month higher for more than 4 million apartments during 2019, driving $1.5 billion of total rent paid above what normal competition would dictate. But these findings are worth some investigation.¹

The researchers hand-collected data related to management companies’ use of pricing algorithms and mapped that against building-level rents and occupancy from 2005 to 2019. Their findings indicated 10% of all rental housing was using algorithmic pricing, although in buildings with more than 20 units that figure was 30%.

The study broke down into two categories how algorithms help an individual property manage revenue. First, the revenue models purport to provide “responsive pricing,” recommendations that change quickly as market conditions change. The research confirmed this was true. It found “strong evidence” that the software helps set rents that are “more responsive to market conditions.” During downturns, adopters of the technology lower rents faster than non-adopters, and during upswings, “adopters increased rents and reduced occupancy compared to non-adopters.”

Those findings are based on standard regressions and reveal what seems like revenue management working as it should. But the authors weren’t so sure. They questioned whether in certain conditions - notably during downturns - the models suggest prices that don’t maximize the revenue of any individual property but instead optimize pricing for all the properties that are using the software. They called this type of revenue guidance “coordinated prices” - a term that sounded loaded to us.

Why would a software even provide that? The paper presumes any individual building is going to see high demand elasticity as prices change, meaning for any one building, higher rents would lead to significantly fewer people signing leases. But if you imagine demand elasticity for “the group of all” buildings, and revenue manage that group as a portfolio, the picture changes. Renters facing similar prices in any of the available options have fewer alternatives, so they’re more likely to accept the higher rent. That means demand at the group level “would be more inelastic” and all participating buildings would benefit.

OK, but if only 30% of larger apartment buildings use algorithmic software and assuming all those buildings kept prices artificially high, wouldn’t residents just find options in the other 70% of buildings? The paper doesn’t address this question.

In fact it was impossible for the authors to determine if the revenue models were providing “responsive pricing” or “coordinated pricing” from the real-life data they had. So they used the underlying data from 10 cities and created a theoretical demand model that predicted how rents would look if landlords priced for a collective outcome and, alternatively, if they priced for their own individual revenue maximization. The real-life, observed outcomes were consistent with what the theoretical model predicted would happen if the software priced for the collective outcome. From there they estimated the difference in rents that would have been paid without the “coordinated pricing” and got the $1.5 billion figure.

Does creating a theoretical model of renter behavior and comparing results to real life sound like a smoking gun to you? Not to minimize what it suggests, but it didn’t strike us that way. Still, housing gets so much political attention it may not matter: This research has been used in many public contexts to bolster the case that algorithms lead to price fixing, including being cited by The White House (Biden’s, that is).

The bummer in all this is what it means for the multifamily industry. Apartment managers bear the brunt of high investor expectations and low resources, so tools that actually work and make them more operationally efficient (e.g. EliseAI) or effective (e.g. revenue management software) are especially valuable. In the context of a huge apartment supply, most of which doesn’t use revenue management technology, and the wider context of technology guiding both businesses and consumers in their financial decisions, it seems backward to point to revenue management in some multifamily assets and call it a crime.

And of course, we know apartment owners include already-rich people who aren’t hurt by this, but they also include lots of pensioners, individual investors and families looking to real estate for consistent income. The bigger risk here isn’t that algorithms are raising rents—it’s that fear of algorithms is driving the industry backward. In an environment where multifamily margins are tight and operational complexity is rising, we need better pricing tools, not fewer. Penalizing innovation based on theory rather than evidence sends the wrong message—to operators, technologists, and investors alike.

Who: Bedrock

What: Real estate development and investment firm “dedicated to the future of Detroit and Cleveland.”

The Sparkle: Probably the best example of a local development firm that wants to be entrenched locally and profit as the community thrives. Locally focused urban developers are not uncommon, especially in America’s older and larger cities, but there are few examples better than Bedrock of a single firm driving so much value in places without the natural tailwinds of a gateway city. From one-off projects to mixed-use city blocks to entire riverfront districts, Bedrock has gone deep in its two hometowns. Founded by Dan Gilbert of Rocket Mortgage, the firm reflects both his capital and a long-term belief that overlooked cities can reinvent themselves.

Most states have a Springfield or a Washington. Some even have two. But not every state has an Oconomowoc. A what? Exactly. Here are the hardest-to-pronounce town names in every state. And as one of us hails from Upstate NY: Long Live Schenectady!

https://www.mentalfloss.com/geography/maps/hardest-to-pronounce-town-name-in-every-state-map

Editor’s Note: The Real Estate Haystack believes in sharing valuable information. If you enjoyed this week's newsletter, subscribe for regular delivery and forward it to a friend or colleague who might find it useful. It's a quick and easy way to spread the word.