If you hire your property, there’s an excellent probability your landlord makes use of RealPage to set your month-to-month cost. The corporate describes itself as merely serving to landlords set probably the most worthwhile value. However a sequence of lawsuits says it’s one thing else: an AI-enabled price-fixing conspiracy.
The traditional picture of price-fixing entails the executives of rival firms gathering behind closed doorways and secretly agreeing to cost the identical inflated value for no matter they’re promoting. This kind of collusion is likely one of the gravest sins you may commit in opposition to a free-market economic system; the late Justice Antonin Scalia as soon as referred to as price-fixing the “supreme evil” of antitrust regulation. Agreeing to repair costs is punishable with as much as 10 years in jail and a $100 million high-quality.
However, because the RealPage instance suggests, expertise might supply a workaround. As an alternative of getting collectively together with your rivals and agreeing to not compete on value, you may all independently depend on a 3rd get together to set your costs for you. Property homeowners feed RealPage’s “property administration software program” their information, together with unit costs and emptiness charges, and the algorithm—which additionally is aware of what opponents are charging—spits out a hire suggestion. If sufficient landlords use it, the outcome might look the identical as a conventional price-fixing cartel: lockstep value will increase as a substitute of value competitors, no secret handshake or clandestine assembly wanted.
With out value competitors, companies lose their incentive to innovate and decrease prices, and customers get caught with excessive costs and no alternate options. Algorithmic price-fixing seems to be spreading to increasingly industries. And present legal guidelines might not be geared up to cease it.
In 2017, then–Federal Commerce Fee Chair Maureen Ohlhausen gave a speech to antitrust legal professionals warning concerning the rise of algorithmic collusion. “Is it okay for a man named Bob to gather confidential value technique info from all of the members in a market after which inform everyone how they need to value?” she requested. “If it isn’t okay for a man named Bob to do it, then it most likely isn’t okay for an algorithm to do it both.”
The various lawsuits in opposition to RealPage differ of their particulars, however all make the identical central argument: RealPage is Bob. In line with one estimate, in additional than 40 housing markets throughout the US, 30 to 60 % of multifamily-building models are priced utilizing RealPage. The plaintiffs suing RealPage, together with the Arizona and Washington, D.C., attorneys common, argue that this has enabled a vital mass of landlords to lift rents in live performance, making an present housing-affordability disaster even worse. (In a assertion, RealPage has responded that the share of landlords utilizing its companies is much decrease, about 7 % nationwide. RealPage’s estimate consists of all rental properties, whereas the lawsuits deal with multifamily-building models.)
In line with the lawsuits, RealPage’s purchasers act extra like collaborators than opponents. Landlords hand over extremely confidential info to RealPage, and lots of of them recruit their rivals to make use of the service. “These sorts of behaviors increase a giant pink flag,” Maurice Stucke, a regulation professor on the College of Tennessee and a former antitrust legal professional on the Division of Justice, instructed me. When firms are working in a extremely aggressive market, he stated, they sometimes go to nice lengths to guard any delicate info that would give their rivals an edge.
The lawsuits additionally argue that RealPage pressures landlords to adjust to its pricing ideas—one thing that might make no sense if the corporate had been merely being paid to supply individualized recommendation. In an interview with ProPublica, Jeffrey Roper, who helped develop one among RealPage’s major software program instruments, acknowledged that one of many biggest threats to a landlord’s earnings is when close by properties set costs too low. “You probably have idiots undervaluing, it prices the entire system,” he stated. RealPage thus makes it exhausting for patrons to override its suggestions, based on the lawsuits, allegedly even requiring a written justification and express approval from RealPage employees. Former workers have stated that failure to adjust to the corporate’s suggestions might end in purchasers being kicked off the service. “This, to me, is the most important giveaway,” Lee Hepner, an antitrust lawyer on the American Financial Liberties Undertaking, an anti-monopoly group, instructed me. “Enforced compliance is the hallmark function of any cartel.”
The corporate disputes this description, claiming that it merely affords “bespoke pricing suggestions” and lacks “any energy” to set costs. “RealPage prospects make their very own pricing selections, and acceptance charges of RealPage’s pricing suggestions have been tremendously exaggerated,” the corporate says.
In December, a Tennessee decide rejected RealPage’s movement to have a class-action lawsuit in opposition to it dismissed, permitting the case to go ahead. It could be a mistake, nonetheless, to conclude from that instance that the authorized system has the algorithmic price-fixing downside underneath management. RealPage might nonetheless prevail at trial, and in any case, it isn’t alone. Its major competitor, Yardi, is concerned in an analogous lawsuit. Certainly one of RealPage’s subsidiaries, a service referred to as Rainmaker, faces a number of authorized challenges for allegedly facilitating price-fixing within the lodge business. (Yardi and Rainmaker deny wrongdoing.) Comparable complaints have been introduced in opposition to firms in industries as different as medical health insurance, tire manufacturing, and meat processing. However successful these instances is proving troublesome.
The problem is that this: Beneath present antitrust regulation, exhibiting that firms A and B used algorithm C to lift costs isn’t sufficient; you might want to present that there was some type of settlement between firms A and B, and you might want to allege some particular factual foundation that the settlement existed earlier than you may formally request proof of it. This dynamic can place plaintiffs in a catch-22: Plausibly alleging the existence of a price-fixing settlement is difficult to do with out entry to proof like personal emails, inner paperwork, or the algorithm itself. However they sometimes can’t uncover these sorts of supplies till they’re given the authorized energy to request proof in discovery. “It’s like making an attempt to suit a sq. peg in a spherical gap,” Richard Powers, a former deputy assistant legal professional common within the DOJ antitrust division, instructed me. “It makes the job actually exhausting.”
Within the case of RealPage, the plaintiffs had been in a position to make the peg match. However in Might, a Nevada decide dismissed an analogous case in opposition to a gaggle of Las Vegas motels who used Rainmaker, concluding that there wasn’t sufficient proof of a price-fixing settlement, as a result of the motels concerned hadn’t shared confidential info with each other and weren’t required to simply accept Rainmaker’s suggestions, even when they allegedly did so about 90 % of the time. “The rulings to this point have set the bar very excessive,” Kenneth Racowski, a litigation legal professional at Holland & Knight, instructed me. The RealPage case “was in a position to clear that bar, but it surely would possibly show to be the exception.”
And instances like RealPage and Rainmaker would be the simple ones. In a sequence of papers, Stucke and his fellow antitrust scholar Ariel Ezrachi have outlined methods wherein algorithms might repair costs that might be much more troublesome to forestall or prosecute—together with conditions wherein an algorithm learns to repair costs withouts its creators or customers intending it to. One thing comparable might happen even when firms used completely different third-party algorithms to set costs. They level to a current research of German gasoline stations, which discovered that when one main participant adopted a pricing algorithm, its margins didn’t budge, however when two main gamers adopted completely different pricing algorithms, the margins for each elevated by 38 %. “In conditions like these, the algorithms themselves truly be taught to collude with one another,” Stucke instructed me. “That might make it attainable to repair costs at a scale that we’ve by no means seen.”
Not one of the conditions Stucke and Ezrachi describe contain an express settlement, making them nearly unattainable to prosecute underneath present antitrust legal guidelines. Worth-fixing, in different phrases, has entered the algorithmic age, however the legal guidelines designed to forestall it haven’t saved up. Powers stated he believes present antitrust legal guidelines cowl algorithmic collusion—however he fearful that he could be flawed. “That is the factor that saved me up at evening,” he stated about his tenure on the Division of Justice. “The fear that every one 100-plus years of case regulation on price-fixing might be circumvented by expertise.”
Earlier this 12 months, a handful of Senate Democrats led by Amy Klobuchar launched a invoice that might replace present legal guidelines to mechanically presume a price-fixing settlement at any time when “opponents share competitively delicate info by way of a pricing algorithm to lift costs.” That invoice, like a lot congressional laws, is unlikely to turn out to be regulation anytime quickly. Native governments may need to take the lead. Final week, San Francisco handed a first-of-its-kind ordinance banning “each the sale and use of software program which mixes private competitor information to set, advocate or advise on rents and occupancy ranges.”
Whether or not different jurisdictions observe go well with stays to be seen. Within the meantime, increasingly firms are determining methods to make use of algorithms to set costs. If these actually do allow de facto price-fixing, and handle to flee authorized scrutiny, the outcome might be a type of pricing dystopia wherein competitors to create higher merchandise and decrease costs would get replaced by coordination to maintain costs excessive and earnings flowing. That might imply completely increased prices for customers—like an inflation nightmare that by no means ends. Extra profound, it might undermine the incentives that hold economies rising and dwelling requirements rising. The fundamental premise of free-market capitalism is that costs are set by way of open competitors, not by a central planner. That goes for algorithmic central planners too.
