In its proposed settlement (the Settlement) with RealPage, Inc. (RealPage), the Department of Justice (DOJ) aims to reset how algorithmic pricing tools operate in the multifamily housing context. Framed as part of the DOJ’s broader prohibition against algorithmic coordination, the Settlement targets core inputs, features, and governance of RealPage’s revenue management products. The Settlement imposes restrictions on data use, requires redesigns of pricing features, subjects RealPage to a court‑approved monitor, and it requires RealPage’s cooperation in the government’s ongoing litigation against property managers that used the software. The Settlement demonstrates the broader lesson that competing landlords must make independent pricing decisions—and the rise of AI does not change this requirement. 

The DOJ’s allegations

In its Complaint, the government alleged that RealPage’s software drew on nonpublic, competitively sensitive information from landlords to calibrate rental prices, included features designed to restrain price decreases or steer pricing alignment, and hosted meetings where competitors shared sensitive insights. Though fueled by new technology, this theory is textbook in antitrust: access to current or future looking, granular, competitively sensitive data, combined with common pricing logic, can suppress competition and increase prices, in violation of the antitrust laws. 

Settlement terms

First, under the terms of the Settlement, RealPage is required to stop using competitors’ nonpublic data in its  revenue management product and is barred from training models on active lease or forward‑looking data from unaffiliated properties.  Model training is limited to historic, backward‑looking nonpublic data that is at least 12 months old. The Settlement order also bans runtime use of unaffiliated property data and prohibits sharing or making that data accessible within the product—even in aggregated or anonymized form—with narrow, time‑limited exceptions to supplement a property’s own transactional history using out‑of‑market data when no same‑owner surrogate is available. RealPage must retrain models to comply and may not identify geographic effects narrower than statewide for specified models, eliminating neighborhood‑ or submarket‑level effects learned from nonpublic cross‑owner data. 

Second, the software must be redesigned to remove mechanisms that prop up prices or encourage  competitors toward common pricing ranges. More specifically, auto‑accept functions must require user‑set parameters, “governor” guardrails must be symmetric with user‑adjustable bounds, features cannot reduce target lease counts to juice price recommendations, and “sold‑out” guardrails must rely solely on a property’s own information. Crucially, users must be able to set parameters permitting recommendations to dip below previously defined “floors” to the same extent they can exceed “ceilings,” addressing a common critique that algorithmic tools embed one‑way ratchets against price decreases. RealPage may not require  acceptance of recommended prices, and it may not deploy any feature that uses unaffiliated nonpublic data in ways inconsistent with the Settlement. 

Third, RealPage is prohibited from conducting market surveys used to gather nonpublic competitive intelligence for pricing and from discussing nonpublic data‑based analyses or pricing strategies at RealPage‑hosted meetings attended by competing property managers or owners. Pricing advisors are likewise barred from disseminating unaffiliated nonpublic data when advising on prices or occupancy.

Fourth, the Settlement also requires the acceptance of a court‑approved monitor—selected by DOJ and approved by the court—with broad access to review RealPage’s code, model training documentation, and runtime logic to ensure compliance with the terms of the Settlement. The monitor will issue periodic reports and can hire experts at RealPage’s expense. 

Fifth, RealPage must also create and implement an antitrust compliance program, designate an antitrust compliance officer, ensure annual training, conduct periodic audits of feature compliance and data sourcing, and file annual certifications from its general counsel and compliance officer. If prohibited topics surface in RealPage meetings, the company must promptly report detailed information to DOJ and the monitor. RealPage must also cooperate fully in the DOJ’s litigation against property management companies that used its software, including by producing documents and making personnel available as needed for testimony.

Timelines, and the Tunney Act 

Within 180 days of the court’s stipulation and order of the Settlement, RealPage must cease runtime use of unaffiliated nonpublic data, retrain models on compliant datasets, and implement the required feature redesigns. Within 60 days, pricing advisors must cease disseminating unaffiliated nonpublic data, and within 30 days, the company must submit its antitrust compliance policy and designate a compliance officer.  The proposed Settlement would remain in effect for seven years, subject to early termination after four years if the DOJ deems continued oversight unnecessary. 

As required by the Tunney Act, the Settlement and competitive impact statement will be published in the Federal Register, triggering a 60‑day public comment period. After considering comments, the Middle District of North Carolina may enter the final judgment upon a public‑interest finding.

Practical implications for owners, managers, and tech providers

The Settlement contains clear guidance that property owners, managers and technology providers can utilize to minimize antitrust liability.  Property owners and managers should avoid any and all reliance on competitors’ nonpublic,  pricing, occupancy, or demand information and take all measures to ensure independent price setting.   

For pricing‑tech providers, the Settlement sketches a compliance blueprint for AI in concentrated or sensitive markets: train on aged, backward‑looking data that excludes active leases, constrain geographic effects to broad regions, eliminate one‑way pricing guardrails, prohibit cross‑client data sharing in runtime, and audit both code and conduct.