{"id":106668,"date":"2024-12-06T13:24:49","date_gmt":"2024-12-06T18:24:49","guid":{"rendered":"https:\/\/cdt.org\/?post_type=insight&p=106668"},"modified":"2024-12-06T16:42:56","modified_gmt":"2024-12-06T21:42:56","slug":"justice-department-goes-after-algorithm-fueled-price-fixing-in-apartment-rentals","status":"publish","type":"insight","link":"https:\/\/cdt.org\/insights\/justice-department-goes-after-algorithm-fueled-price-fixing-in-apartment-rentals\/","title":{"rendered":"Justice Department Goes After Algorithm-Fueled Price-Fixing in Apartment Rentals"},"content":{"rendered":"\n

On August 23, the Department of Justice, along with eight states, filed an antitrust enforcement action<\/a> against RealPage, charging it with using an algorithm to organize and coordinate a scheme among apartment landlords to inflate rental prices in violation of the Sherman Act. The allegations set forth what appears to be a textbook example of using artificial intelligence to supercharge anticompetitive collusion, a capability that CDT has written about previously<\/a>. <\/p>\n\n\n\n

After requesting and receiving, with the Department\u2019s consent, two extensions, RealPage filed its response on December 3 \u2013 a motion to dismiss for failure to state a claim.[1]<\/strong><\/sup> Essentially, RealPage takes issue with the way the Department has defined the relevant markets for antitrust analysis, and further denies orchestrating or being involved in any pricing coordination among landlords.<\/p>\n\n\n\n

Some might say that the Justice Department has come to the party late. The District of Columbia<\/a> and Arizona<\/a> had already brought enforcement actions under their own laws, and a federal class action was already pending in a federal district court in Tennessee.[2]<\/strong><\/sup> But the Department\u2019s allegations provide far greater detail, benefitting from its greater resources combined with its stronger investigatory authority, which enables it to compel production of evidence during its investigation, before it brings an action, under the Antitrust Civil Process Act<\/a>. <\/p>\n\n\n\n

As charged in the complaint, RealPage has created and advertised to apartment landlords an algorithm-powered system to collect and analyze, on a daily basis, current rental prices and planned future prices, and current availabilities and projected future availabilities for all participating landlords. This information is separately categorized for each individual rental unit, according to size, floor plan, layout, and amenities. RealPage makes explicitly clear to the landlords that it will analyze this information and provide pricing recommendations to each landlord based on this information. This kind of information is competitively sensitive, and in a healthy competitive marketplace it is closely guarded, not shared.<\/p>\n\n\n\n

RealPage\u2019s system has the hallmarks of a classic anticompetitive \u201chub-and-spoke\u201d[3]<\/strong><\/sup> arrangement under which competitors coordinate pricing and output decisions through a central clearinghouse \u201chub.\u201d This kind of arrangement has been found to violate section 1 of the Sherman Act, which prohibits contracts, combinations, or conspiracies in restraint of trade, provided that the evidence sufficiently demonstrates that the competitors along the \u201crim\u201d had a \u201cconscious commitment to a common scheme designed to achieve an unlawful objective.\u201d[4]<\/strong><\/sup> It is not necessary that the competitors along the \u201crim\u201d have direct communication with each other regarding the anticompetitive scheme, because they are communicating effectively through the \u201chub\u201d as \u201cspokes.\u201d<\/p>\n\n\n\n

Here, per the Department\u2019s allegations, RealPage created the \u201chub\u201d and advertised it to landlords, encouraging them to join up. RealPage explained that it would calculate pricing recommendations for them, based on pricing data submitted on a daily basis by every participating landlord in the market area. And that the recommendations would be guided by the highest prices being charged, which would enable each landlord to confidently increase its own rental prices in line with the high end of prices being charged by its competitors. <\/p>\n\n\n\n

So, per those allegations, the landlords were well aware that they would be \u201cspokes,\u201d participating along with their competitors, and that the result would be pricing recommendations that would result in increased prices. Or, as RealPage regularly puts it, would \u201craise all ships.\u201d A RealPage revenue management vice president elaborated that this phrase means that \u201cthere is greater good in everybody succeeding versus essentially trying to compete against one another in a way that actually keeps the industry down.\u201d And even more pointedly, that landlords using RealPage\u2019s software would \u201clikely move in unison versus against each other.\u201d<\/p>\n\n\n\n

Thus, the landlords who joined up were consciously committing themselves to a common scheme not to compete. Their commitment includes paying RealPage a hefty fee in recognition of the value they receive. <\/p>\n\n\n\n

But RealPage has allegedly gone beyond just creating and advertising the hub that enabled and facilitated a conscious commitment to unlawful pricing coordination. It has taken a number of calculated steps to make sure landlords follow through on that commitment. It constantly nudges landlords to follow each other\u2019s price increases. It actively monitors prices charged on literally millions of apartment units \u2013 not only to calculate new pricing recommendations, but also to determine which landlords are complying with its recommendations and which are not. <\/p>\n\n\n\n

Each day, RealPage sends updated pricing recommendations to each landlord. RealPage makes it easy for the landlord to accept the recommendations in bulk \u2013 it can be done with a single keystroke, or even programmed for auto-accept, which RealPage strongly encourages landlords to adopt. Diverging from the recommendation, in contrast, requires the landlord\u2019s property manager to affirmatively give RealPage a \u201cstrong sound business-minded\u201d justification for each divergence, based on something the algorithm is not accounting for, such as local construction or renovations occurring in the building. And whenever RealPage disagrees with the justification, which is usually, the matter is escalated to the property manager\u2019s supervisor, and upward, with increasing aggressiveness.<\/p>\n\n\n\n

Internal training explained that RealPage wanted to \u201cwiden auto accept parameters\u201d by introducing the feature and then \u201ccreating enough trust so that over time we have client[s] that are willing to let auto accept run with very wide parameters\u2026 AKA \u2013 accept all recommendations.\u201d RealPage trains pricing advisors to have an \u201caccountability conversation\u201d or a \u201crefresher on short term vs long term goals\u201d for clients that show less tolerance for increasing auto-accept parameters.<\/p>\n\n\n\n

The result, according to the complaint: more than 85% of final floor plan prices are within 5% of RealPage\u2019s recommendation.<\/p>\n\n\n\n

The Department further charges that RealPage reinforces its algorithm-driven coordinated upward pricing recommendations by discouraging landlords from offering renters discount \u201cconcessions\u201d \u2013 such as a free month\u2019s rent or waived fees \u2013 as landlords in a competitive marketplace would have incentives to offer. In its \u201cbest practices\u201d for landlords, RealPage\u2019s guidance is simple: \u201cEliminate concessions.\u201d A landlord\u2019s agreement to refuse to offer concessions is bolstered by its awareness that competing landlords are receiving the same advice from RealPage.<\/p>\n\n\n\n

Essentially, landlords are encouraged, and then pressured, to turn over rental pricing decisions to RealPage\u2019s algorithm, knowing that it is coordinating pricing among participating landlords, pushing prices higher.<\/p>\n\n\n\n

The Department alleges that RealPage\u2019s pricing algorithm ratchets in only the upward direction. It resists recommending the kind of price decreases<\/em> in response to a decrease in demand that would occur in a marketplace with healthy competition. Instead, the algorithm overrides its normal functioning to coordinate recommended reductions in supply<\/em> \u2013 taking units off the market temporarily \u2013 the classic tactic used by price-fixers to reinforce inflated prices. RealPage refers to this as \u201crevenue protection mode\u201d or \u201csold out mode.\u201d<\/p>\n\n\n\n

Interestingly, the Department has brought the case under both section 1 and section 2 of the Sherman Act. Section 1 prohibits contracts, combinations, and conspiracies in restraint of trade \u2013 often referred to as collusion. It involves a de facto agreement \u2013 a \u201cmeeting of the minds\u201d \u2013 between two or more entities. The prime example of a section 1 violation is price-fixing, essentially what is alleged here. Section 2 prohibits monopolization or attempts to monopolize. It involves having monopoly or market power, and using it to sabotage the competitive efforts of rivals through exclusionary conduct. A section 2 violation can be committed by one entity acting alone.<\/p>\n\n\n\n

In this case, the section 1 collusion claims focus on anticompetitive benefits to landlords, and secondarily on the fees RealPage charges the landlords for participation in the scheme. The section 2 monopolization claims focus on the anticompetitive benefits to RealPage from the massive data it collects from bringing so many landlords into its scheme. The thrust of the claim is that rival apartment rental management software providers cannot compete with RealPage without entering into similar anticompetitive schemes, and even then, RealPage has the overwhelming and entrenched advantage in having signed up so many landlords under its own anticompetitive scheme and amassing all their data.<\/p>\n\n\n\n

The section 1 claim describes long-recognized hub-and-spoke collusion. The section 2 claim is not<\/em> long-recognized. The factual allegations describe a market share RealPage has achieved, and the barriers to entry resulting from the vast data RealPage has amassed that potential entrants could not get similar access to. There are not, however, allegations of actions that RealPage has taken to maintain a monopoly by sabotaging the competitive efforts of potential rivals \u2013 at least not in a way familiar in past antitrust cases. Instead, the alleged monopolization is more of a by-product of the collusive scheme. Indeed, RealPage has moved to dismiss the section 2 claim on the basis that it has not engaged in any exclusionary conduct \u2013 that amassing the data is not exclusionary.<\/p>\n\n\n\n

The kernel of the Department\u2019s section 2 monopolization claim is that, by creating and managing the hub-and-spoke collusive scheme, RealPage has intentionally made it impossible for a rival to offer the same benefits to landlords without engaging in a similar collusive scheme \u2013 in other words, impossible for them to compete on the merits lawfully \u2013 and which would likely be futile even if they did attempt it. According to the Department\u2019s information, RealPage controls at least 80 percent of the market for apartment rental management software. No other revenue management company could hope to match RealPage\u2019s access to landlords\u2019 nonpublic, competitively sensitive rental data.<\/p>\n\n\n\n

One potential advantage to the Department of bringing a section 2 claim is that the potentially available remedies go beyond ordering the cessation of the unlawful conduct, and include structural relief \u2013 that is, in this case, requiring RealPage to divest some of the parts of its operation that collectively enable the unlawful coordination among landlords. Structural relief is not generally available to remedy a violation of section 1. The Department may believe that the anticompetitive conduct alleged here cannot effectively be remedied without breaking apart the arrangement.\u00a0<\/p>\n\n\n\n

It is also noteworthy that the Department has brought a civil enforcement action, not a criminal prosecution. Ordinarily, when the alleged conduct is clearly price-fixing among competitors, a criminal case is warranted. The courts have long considered price-fixing to be a per se violation, and do not accept any mitigating justifications.[5]<\/strong><\/sup> Three considerations likely influenced the Department\u2019s decision. <\/p>\n\n\n\n

First, although the caselaw is ostensibly clear that price-fixing is per se unlawful, the courts have in practice recognized exceptions, when there are novel factual circumstances that courts have not previously considered. The Department may have decided that the factual allegations in this case were potentially too complex to rely on criminal prosecution and the requirement to prove the violation beyond a reasonable doubt. Indeed, the district court in the federal class action in Tennessee ruled that that case could not be brought as a per se violation. <\/p>\n\n\n\n

Second, although section 2 of the Sherman Act explicitly provides for criminal penalties equal to those in section 1, section 2 violations had not, until very recently, been criminally prosecuted for decades. And of the 168 criminal section 2 cases \u2013 brought between 1903 and 1977 \u2013 all but 20 were multiple-conspirator cases that could have been brought under section 1, and many were. Of the 20 single-defendant cases, only 12 resulted in guilty findings, most as a result of a nolo contendere plea and a fine. Only three cases resulted in prison; two of those involved crimes of violence, and in the other, the individual served only a single month in prison.[6]<\/strong><\/sup><\/p>\n\n\n\n

The Biden Antitrust Division revived section 2 criminal prosecutions for the first time in almost 50 years. It has brought three, none of which has reached a verdict or resulted in prison. All were against individuals \u2013 not companies \u2013 whose conduct clearly met long-established standards for guilt. Two of them involved multiple conspirators and were also brought under section 1.[7]<\/strong><\/sup><\/p>\n\n\n\n

Third, if the case were pursued criminally, the Department would have had to go it alone; the states would not have been able to join. Notably, the enforcement actions the Department is pursuing against Google and Apple for monopolization under section 2, each joined by several states, have been civil, not criminal. <\/p>\n\n\n\n

As the case moves forward, RealPage will have ample opportunity to justify its marketing and use of the pricing algorithm.<\/p>\n\n\n\n

More broadly, the issue will not be whether the services RealPage provides include some that are lawful and even procompetitive. It will be whether the algorithm-directed pricing system specifically is permissible under the antitrust laws.<\/p>\n\n\n\n

And the mere fact that an algorithm is driving the coordinated pricing will not excuse it. As then-Acting FTC Chair Maureen Olhausen remarked, antitrust enforcers evaluating the use of an algorithm in commerce will follow the \u201cguy named Bob\u201d rule \u2013 \u201cEverywhere the word \u2018algorithm\u2019 appears, please just insert the words \u2018a guy named Bob\u2019 \u2026 If it isn\u2019t okay for a guynamed Bob to do it, then it probably isn\u2019t ok for an algorithm to do it either.\u201d[8]<\/strong><\/sup><\/p>\n\n\n\n

\n\n\n\n

[1]<\/sup> Available at https:\/\/ecf.ncmd.uscourts.gov\/doc1\/13314442744<\/a>.<\/p>\n\n\n\n

[2]<\/sup> RealPage, Inc., Rental Software Antitrust Litig. (No. II), 2023 U.S. Dist. LEXIS 230200 (M.D. Tenn. Dec. 28, 2023).<\/p>\n\n\n\n

[3]<\/sup> E.g.<\/em>, United States v. Apple, Inc., 791 F.3d 290, 314 (2d Cir. 2015).<\/p>\n\n\n\n

[4]<\/sup> E.g.<\/em>, id.<\/em> at 315. See<\/em> Interstate Circuit v. United States, 306 U.S. 208, 227 (1939) (\u201cAcceptance by competitors, without previous agreement, of an invitation to participate in a plan, the necessary consequence of which, if carried out, is restraint of interstate commerce, is sufficient to establish an unlawful conspiracy under the Sherman Act.\u201d)<\/p>\n\n\n\n

[5]<\/sup> United States v. Socony-Vacuum Oil Co., 310 U.S. 150, 218 (1940).<\/p>\n\n\n\n

[6]<\/sup> See <\/em>Daniel A. Crane, Criminal Enforcement of Section 2 of the Sherman Act: An Empirical Assessment, 84 Antitrust L.J. 753 (2022).<\/p>\n\n\n\n

[7]<\/sup> Two of those cases were resolved with a guilty plea. Zito pled guilty in September 2022 to attempted monopolization, with an invitation to divide markets that would have given him a monopoly in highway crack-sealing services in Wyoming and Montana. He was sentenced to six months\u2019 home confinement followed by three years\u2019 probation. Tomlinson pled guilty in April 2024 to conspiring to rig bids for forest fire-fighting fuel truck services in Idaho and Nevada and to de-prioritize bids of others. He is apparently still awaiting sentencing. The case charged violations of both section 1 and section 2. https:\/\/www.justice.gov\/atr\/case\/us-v-ike-tomlinson-and-kris-bird. The third case is still pending. Martinez and his co-conspirators are charged with violations of both section 1 and section 2 \u2013 price-fixing of transmigrante forwarding agency services for shipping vehicles from the United States through Mexico to Central America, and conspiring to monopolize by threatening potential competitors with violence. https:\/\/www.justice.gov\/atr\/case\/us-v-carlos-favian-martinez-et-al.<\/p>\n\n\n\n

[8]<\/sup> Should We Fear the Things That Go Beep in the Night? Some Initial Thoughts on the Intersection of Antitrust Law and Algorithmic Pricing,\u201d Maureen K. Ohlhausen, Acting Chairman, U.S. Federal Trade Commission, May 23, 2017, p. 10, https:\/\/www.ftc.gov\/system\/files\/documents\/public_statements\/1220893\/ohlhausen_-_concurrences_5-23-17.pdf<\/a>.<\/p>\n","protected":false},"featured_media":86099,"template":"","content_type":[7251],"area-of-focus":[5660,7258],"class_list":["post-106668","insight","type-insight","status-publish","has-post-thumbnail","hentry","content_type-blog","area-of-focus-competition","area-of-focus-open-internet"],"acf":[],"_links":{"self":[{"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/insight\/106668","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/insight"}],"about":[{"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/types\/insight"}],"version-history":[{"count":10,"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/insight\/106668\/revisions"}],"predecessor-version":[{"id":106678,"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/insight\/106668\/revisions\/106678"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/media\/86099"}],"wp:attachment":[{"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/media?parent=106668"}],"wp:term":[{"taxonomy":"content_type","embeddable":true,"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/content_type?post=106668"},{"taxonomy":"area-of-focus","embeddable":true,"href":"https:\/\/cdt.org\/wp-json\/wp\/v2\/area-of-focus?post=106668"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}