5 ms·

From the blog post: Algorithms that recommend prices to numerous competing landlords threaten to remove renters’ ability to vote with their feet and comparison-

From the blog post: Algorithms that recommend prices to numerous competing landlords threaten to remove renters’ ability to vote with their feet and comparison-shop for the best apartment deal around. From the brief: To participate in the service, landlords must share in “real-time” their “non-public,” “competitively sensitive” data, including actual rents paid, occupancy rates, and records of lease transactions. [Multifamily Compl. ¶¶ 227, 380.] RealPage then feeds “this data into a common algorithm.” [Student Compl. ¶ 5;] [Multifamily Compl. ¶ 380.] The common algorithm uses these common data for a single, common purpose: to generate “forward-looking, unit-specific pricing and supply recommendations” for all participating landlords. [Student Compl. ¶ 5.] To ensure that the landlords abide by these “recommendations,” RealPage puts significant “pressure” on them “to implement RealPage’s prices,” including by requiring clients to submit requests to deviate to the “corporate office” and tracking the “identity of the client’s staff that requested a deviation.” [Multifamily Compl. ¶¶ 17-20, 261-86.] As a result, landlords using RealPage adopt RealPage’s recommendations 80-90% of the time. [Id. ¶ 15.] The complaints allege that RealPage was clear about the purpose of its common pricing scheme: to increase prices above competitive levels through collaboration. Doesn't sound like the natural ebb and flow of the market to me.

scrps 3y ago
That sounds like a dark pool for the rental market.
- kristjansson 3y ago
  The rental market analogue of a dark pool is the rental market. All transactions are secret by default, allowing higher-information parties (landlords) to set prices that are advantageous to them, and prevent transactions from moving the market. These algorithmic pricing systems extend that information advantage by pooling information across landlords (maybe not illegal by itself) and leveraging their position as market-wide price-setter to push rents higher, with the promise that so many units are following their recommendations that renters have ~no alternative (very illegal price fixing)
  - scrps 3y ago
    The rental market is a decentralized network, with companies like these the network gets routed through a central node and the operator of that node has pefect visibility while the connections being routed have no idea what every other connection is doing like a dark pool, it's the centrality that makes it an analogue. I could think of several advantageous moves like artificially inflating rents then favoring a well paying client by keeping their rents lower than the market. Investing in property spread out across shell companies and then have the algo favor them in a non-intuitive way.
    - kristjansson 3y ago
      Dark pools don't provide sellers or buyers an asymmetrical advantage[1] and prices are still set by participants and cleared at market, not set by the operator of the exchange (i.e. dark pool). [1] if they're both participating in the pool, clearly an institution seeing pricing signal from an alternative trading system (even if it's only their own cleared trades) has more information than someone observing only public market transactions.
ummonk 3y ago
Yeah the brief seems to be describing a much more clearcut example of collusion than the intentionally vague and broad blog post is trying to imply.
- chefandy 3y ago
  It does come across as a bit glib, but I empathize with the author. I worked in legal higher-ed for a long time; I have absolutely no legal education, but my job involved reading lots of writing about the law, both for expert and non-expert audiences. Accurately conveying the importance of a topic without getting into the nitty gritty or running into collisions with colloquial and legal word usage is tough. Even that one paragraph I quoted had like 6 citations or something, and that's not just to satisfy technical requirements. There's just so much subtlety behind even very common words like may, must, shall, and will when used in a legal context that it seems like context switching is tough. The word count seems to grow logarithmically with the factual and conceptual detail, and even then you're leaving out a lot of padding beneficial to laypeople. We ran an open database of caselaw and we regularly got questions from well-meaning folks trying to inform themselves about some legal predicament by reading the cases, but it's just not feasible. Lots of folks argue that the whole thing is just based on shitty writing practices and gatekeeping legal knowledge, which might be true, but the difference is still consequential now. I actually think this is one area generative AI will be a huge boon, though I was skeptical until pretty recently. Also, accessibility is a concern. Lawyers doing this sort of work are often very concerned with making their work useful to general audiences. Since subpar access to educational resources and cognitive disabilities are over-represented among low-income people, they're the most heavily affected by this. I wouldn't be surprised if they avoided a lot of nuance assuming curious people would just read the brief. I imagine they'd have written it differently specifically for a largely college-educated professional crowd, but I can definitely see why they didn't.

scrps 3y ago

That sounds like a dark pool for the rental market.

kristjansson 3y ago
The rental market analogue of a dark pool is the rental market. All transactions are secret by default, allowing higher-information parties (landlords) to set prices that are advantageous to them, and prevent transactions from moving the market. These algorithmic pricing systems extend that information advantage by pooling information across landlords (maybe not illegal by itself) and leveraging their position as market-wide price-setter to push rents higher, with the promise that so many units are following their recommendations that renters have ~no alternative (very illegal price fixing)
- scrps 3y ago
  The rental market is a decentralized network, with companies like these the network gets routed through a central node and the operator of that node has pefect visibility while the connections being routed have no idea what every other connection is doing like a dark pool, it's the centrality that makes it an analogue. I could think of several advantageous moves like artificially inflating rents then favoring a well paying client by keeping their rents lower than the market. Investing in property spread out across shell companies and then have the algo favor them in a non-intuitive way.
  - kristjansson 3y ago
    Dark pools don't provide sellers or buyers an asymmetrical advantage[1] and prices are still set by participants and cleared at market, not set by the operator of the exchange (i.e. dark pool). [1] if they're both participating in the pool, clearly an institution seeing pricing signal from an alternative trading system (even if it's only their own cleared trades) has more information than someone observing only public market transactions.

ummonk 3y ago

Yeah the brief seems to be describing a much more clearcut example of collusion than the intentionally vague and broad blog post is trying to imply.

chefandy 3y ago
It does come across as a bit glib, but I empathize with the author. I worked in legal higher-ed for a long time; I have absolutely no legal education, but my job involved reading lots of writing about the law, both for expert and non-expert audiences. Accurately conveying the importance of a topic without getting into the nitty gritty or running into collisions with colloquial and legal word usage is tough. Even that one paragraph I quoted had like 6 citations or something, and that's not just to satisfy technical requirements. There's just so much subtlety behind even very common words like may, must, shall, and will when used in a legal context that it seems like context switching is tough. The word count seems to grow logarithmically with the factual and conceptual detail, and even then you're leaving out a lot of padding beneficial to laypeople. We ran an open database of caselaw and we regularly got questions from well-meaning folks trying to inform themselves about some legal predicament by reading the cases, but it's just not feasible. Lots of folks argue that the whole thing is just based on shitty writing practices and gatekeeping legal knowledge, which might be true, but the difference is still consequential now. I actually think this is one area generative AI will be a huge boon, though I was skeptical until pretty recently. Also, accessibility is a concern. Lawyers doing this sort of work are often very concerned with making their work useful to general audiences. Since subpar access to educational resources and cognitive disabilities are over-represented among low-income people, they're the most heavily affected by this. I wouldn't be surprised if they avoided a lot of nuance assuming curious people would just read the brief. I imagine they'd have written it differently specifically for a largely college-educated professional crowd, but I can definitely see why they didn't.