Nov 22, 2023

Are machines doing the collaborating that competitors maynot? It is an application of artificial intelligence ("AI")that many businesses, agencies, legislators, lawyers, and antitrustlaw enforcers around the world are only beginning to confront. Itis also among the top concerns of in-house counsel acrossindustries. Competitors are increasingly setting prices through theuse of communal, AI-enhanced algorithms that analyze data that areprivate, public, or a mix of both. Allegations in private and public litigation describe"algorithmic price fixing" in which the antitrustviolation occurs when competitors feed and access the same databaseplatform and use the same analytical tools. Then, as some allege,the violations continue when competitors agree to the pricesproduced by the algorithms. Right now, renters and prosecutors areteeing off on the poster child for algorithmic pricing, RealPageInc., and the many landlords and property managers who use it. Private and Public Litigation A Nov. 1, 2023 complaint filed by the Washington, DC, AttorneyGeneral's office described RealPage's offerings this way:"[A] variety of technology-based services to real estateowners and property managers including revenue management productsthat employ statistical models that use data—includingnon-public, competitively sensitive data—to estimate supplyand demand for multifamily housing that is specific to particulargeographic areas and unit types, and then generate a'price' to charge for renting those units that maximizesthe landlord's revenue." The complaint alleges that more than 30% of apartments inmultifamily buildings and 60% of units in large multifamilybuildings nationwide are priced using the RealPage software. In theWashington-Arlington-Alexandria Metropolitan Area that number leapsto more than 90% of units in large buildings. The complaint allegesthat landlords have agreed to set their rates using RealPage. Private actions against RealPage have also been filed in federalcourts across the country and have been centralized inmulti-district litigation in the Middle District ofTennessee(In re: RealPage, Inc., Rental Software AntitrustLitigation[NO. II], Case No. 3:23-md-3071, MDL No. 3071). TheAntitrust Division of the Department of Justice filed a Statement of Interest and a Memorandum in Support in thecase urging the court to deny the defendants' motion todismiss. Even before the MDL, RealPage had attracted the AntitrustDivision's attention when the company acquired its largestcompetitor, Lease Rent Options for $300 million, Axiometrics for$75 million, and On-Site Manager, Inc. for $250 million. The Antitrust Division has been pursuing the use of algorithmsin other industries, including airlines and online retailers. TheDOJ and FTC are both studying the issue and reaching out to expertsto learn more. Journalists and Senators Additionally, three senators urged DOJ toinvestigate RealPage after reporters at ProPublicawrote aninvestigative report in October 2022. The journalists claim thatRealPage's price-setting software "uses nearbycompetitors' nonpublic rent data to feed an algorithm thatsuggests what landlords should charge for available apartments eachday." ProPublica speculated that the algorithm isenabling landlords to coordinate prices and in the process pushrents above competitive levels in violation of the antitrustlaws. Senators Amy Klobuchar (D-MN), Dick Durban (D-IL) and CoryBooker (D-NJ) wrote to the DOJ concerned that the RealPage enables"a cartel to artificially inflate rental rates in multifamilyresidential buildings." Sen. Sherrod Brown (D-OH) also wrote to theFederal Trade Commission with concerns "about collusion inthe rental market," urging the FTC to "review whetherrent setting algorithms that analyze rent prices through the use ofcompetitors' private data ... violate antitrust laws." TheOhio senator specifically mentioned RealPage's YieldStar and AIRevenue Management programs. The Europeans The European Commission has enacted the Artificial Intelligence Act , which includesprovisions on algorithmic pricing, requiring algorithmic pricingsystems be transparent, explainable, and non-discriminatory withregard to consumers. Companies that use algorithmic pricing systemswill be required to implement compliance procedures, includingaudits, data governance, and human oversight. The Legal Conundrum An essential element of any claimed case of price-fixing underthe U.S. antitrust laws is the element of agreement: a plaintiffalleging price-fixing must prove the existence of an agreementbetween two or more competitors who should be setting their pricesindependently but aren't. Consumer harm from collusion occurswhen competitors set prices to achieve their maximum joint profitinstead of setting prices to maximize individual profits. Tocondemn algorithmic pricing as collusion, therefore, requires proofof agreement. It may be difficult for the RealPage plaintiffs to prove thatthe RealPage's users agreed among themselves to adhere to anyparticular price or pricing formula, but not impossible. End usersare likely to argue that RealPage's pricing recommendations aremerely aggregate market signals that RealPage is collecting anddisseminating. The use of the same information service, theirargument will go, does not prove the existence of an agreement forpurposes of Section 1 of the Sherman Act. The parties and courts embroiled in the RealPage litigation areconstrained to live under the law as it presently exists, so thesolution proposed by Michal Gal, Professor and Director of theForum on Law and Markets at the University of Haifa, is out ofreach. In her 2018 paper, "Algorithms as IllegalAgreements," Professor Gal confronts the agreement problemwhen algorithms set prices and concludes that it is time to"rethink our laws and focus on reducing harms to socialwelfare rather than on what constitutes an agreement. "Academics have been critical of the agreement element of Section 1for years, but it is unlikely to change anytime soon, even with theadded inconvenience it poses where competitors rely on a commonvendor of machine-generated pricing recommendations. Nonetheless, there is some evidence that autonomous machines,just like humans, can learn that collusion allows sellers to chargemonopoly prices. In their December 2019 paper, " Artificial Intelligence, Algorithmic Pricing andCollusion ," Emilio Calvano, Giacomo Calzolari, VincenzoDenicolo, and Sergio Pastorello at the Department of Economics atthe University of Bologna showed with computer simulations thatmachines autonomously analyzing prices can develop collusivestrategies "from scratch, engaging in active experimentationand adapting to changing environments. "The authors sayindications from their models "suggest that algorithmiccollusion is more than a remote theoretical possibility." Theyfind that "relatively simple [machine learning] pricingalgorithms systematically learn to play collusive strategies. "The authors claim to be the first to "clearly document theemergence of collusive strategies among autonomous pricingagents." The Agreement Element in the Machine Pricing Case For three main reasons, the element of agreement need not be anobstacle to successfully prosecuting a price-fixing claim againstcompetitors that use a common or similar vendor of algorithmicpricing data and software. First, there is significant precedent for inferring theexistence of an agreement among parties that knowingly participatein a collusive arrangement even if they do not directly interact,sometimes imprecisely referred to as a "rimless wheelhub-and-spoke" conspiracy. For example, in Toys"R" Us, Inc. v. F.T.C., 221 F.3d 928 (9thCir. 2000), the court inferred the necessary concerted action froma series of individual agreements between toy manufacturers andToys "R" Us in which the manufacturers promised not tosell the toys sold to Toys "R" Us and other toy stores tobig box stores in the same packaging. The FTC found that each ofthe manufacturers entered into the restraint on the condition thatthe others also did so. The court found that Toys "R" Ushad engineered a horizontal boycott against a competitor inviolation of Section 1, despite the absence of evidence of any"privity" between the boycotting manufacturers. TheToys "R" Us case relied on the SupremeCourt's decision in Interstate Circuit v. UnitedStates, 306 U.S. 208 (1939), in which movie theater chainssent an identical letter to eight movie studios asking them torestrict secondary runs of certain films. The letter disclosed thateach of the eight were receiving the same letter. The Court heldthat a direct agreement was not a prerequisite for an unlawfulconspiracy. "It was enough that, knowing that concerted actionwas contemplated and invited, the distributors gave their adherenceto the scheme and participated in it." The analogous issue in the algorithmic pricing scenario iswhether the vendor's end users that their competitors are alsoend users. If so, the inquiry can consider the agreement elementsatisfied if the algorithm does, in fact, jointly maximize the endusers' profits. The second factor overcoming the agreement element is related tothe first. Whether software that recommends prices has interactedwith the prices set by competitors to achieve joint profitmaximization—that is, whether the machines have learned tocollude without human intervention—is an empirical question.The same techniques used to uncover machine-learned collusion bysimulation can be used to determine the extent of interdependencein historical price setting. If statistical evidence of collusivepricing is available, it is enough that the end users knowinglyaccepted the offer to set its prices guided by the algorithm. Theeconomics underlying the agreement element in the first place liesin prohibition of joint rather than individual profit maximization,so direct evidence that market participants are jointly profitmaximizing should obviate the need for further evidence ofagreement. A third reason the agreement element need not stymie a Section 1action against defendants engaged in algorithmic pricing is basedon the Supreme Court's decision in American Needle v.NFL, 560 U.S. 183 (2010). In that case the Court made clearthat arrangements that remove independent centers ofdecision-making from the market run afoul of Section 1, if the neteffect of the algorithm is to displace individual decision-makingwith decisions outsourced to a centralized pricing agent, themechanism should be immaterial. The rimless wheel of the so-called hub-and-spoke conspiracy isan inadequate analogy because the wheel in these casesdoes have a rim, i.e., a connection between theconspirators. In the scenarios above in which the courts have foundSection 1 liability i) each of the participants knew that itsrivals were also entering into the same or similar arrangements,ii) the participants devolved pricing authority away fromthemselves down to an algorithmic pricing agent, and iii)historical prices could be shown statistically to have exceeded thecompetitive level in a way consistent with collusive pricing. Theseelements connect the participants in the scheme, supplying the"rim" to the spokes of the wheel. If the plaintiffs inthe RealPage litigation can establish these elements, they willhave met their burden of establishing the requisite element ofagreement in their Section 1 claim. The content of this article is intended to provide a generalguide to the subject matter. Specialist advice should be soughtabout your specific circumstances. AUTHOR(S)