"Premature Antitrust Standards in Algorithmic Pricing" is an article by Jay Ezrielev, founder of Elevecon LLC, published in Fall 2025 in Antitrust (American Bar Association), Vol. 40, No. 1. It argues that courts are establishing antitrust standards for common data algorithms (CDAs) without an adequate understanding of their economic effects, that the standards rest on faulty economic reasoning, and that the emerging legal framework could cause long-term damage.
Summary of argument
The article defines common data algorithms (CDAs) as pricing software that collects data across multiple competing firms, applies algorithms to a common dataset, and calculates individualized pricing recommendations that each firm may accept or reject. CDAs have been litigated in rental housing (RealPage, Yardi), hotels (Cendyn/Gibson), and insurance.
Ezrielev's central claim is that courts are condemning CDAs prematurely. He argues that legal standards are being set before economic research has established whether these tools harm or benefit competition, that the distinction courts have drawn between nonpublic and public data is a poor basis for liability, and that restricting CDA access to nonpublic data could reduce pricing efficiency, increase costs, and harm consumers. In his account, the genuine collusion risk arises from data exchange, which can enable firms to detect deviations from a collusive arrangement, rather than from data use as an algorithm input.
Court decisions discussed
The article surveys four decisions, distinguishing the rental cases that survived from the hotel cases that were dismissed.
| Case | Industry | Outcome | Standard |
|---|---|---|---|
| RealPage (M.D. Tenn.) | Multifamily rental | Allowed to proceed | Rule of reason |
| Duffy v. Yardi (W.D. Wash.) | Multifamily rental | Allowed to proceed | Per se (stricter) |
| RealPage (student housing) | Student rental | Dismissed | — |
| Gibson v. Cendyn (D. Nev. + 9th Cir.) | Hotels (Las Vegas) | Dismissed with prejudice | — |
Ezrielev identifies the dividing line drawn by courts as the nonpublic/public data distinction: the rental cases survived because plaintiffs alleged pooling of nonpublic, proprietary data, while the hotel cases were dismissed because no comparable data-sharing was alleged.
Key claims
Ezrielev advances several economic and legal critiques of the courts' reasoning.
On the assumption that data-sharing harms renters, he argues that courts treat pricing as zero-sum, inferring that if lessors benefit from sharing data, renters must be harmed. He counters that pricing is not zero-sum and that more accurate, market-reflective pricing can benefit both sides through more efficient resource allocation, for example by expanding occupancy rates.
On the evidence the RealPage court accepted, he challenges the inference drawn from a shift from negative to positive vacancy-rent correlation over 2011–2022, which plaintiffs characterized as a strategy change from prioritizing occupancy to prioritizing rents. Ezrielev attributes the shift to macroeconomic cost shocks, including interest rate fluctuations, the pandemic, and inflation, rather than to algorithmic collusion. He notes that with demand shocks rents and vacancies move in opposite directions, while with cost shocks they move together.
On lessors' incentives, he argues the alleged conspiracy is self-undermining. If CDA recommendations produced above-market rents, individual lessors would experience declining occupancy, contrary to their self-interest, and lessors could in any case obtain supracompetitive rents without paying for the software if such a conspiracy actually worked.
On the nonpublic data distinction, he argues that using nonpublic data in an algorithm is not the same as exchanging nonpublic data through an algorithm. Using nonpublic data as an algorithm input does not let competitors detect deviations from collusion, the mechanism that sustains tacit collusion, whereas exchanging data does enable such detection, which he identifies as the real risk that courts conflate with mere use. He adds that "nonpublic" is ambiguous, since nominally private data may be inferrable from public sources, and that banning nonpublic data use may incentivize firms to make data public, potentially increasing collusion risk.
On the state of the economic research, he argues it is insufficient to support firm legal standards. Few empirical papers exist on CDA effects. Peng and Shin find that CDAs lead to more competitive hotel pricing. Ezrielev describes Calder-Wang and Kim's RealPage study as having significant limitations. He notes that theoretical models of hub-and-spoke collusion, such as Harrington's, assume mandatory acceptance of recommendations, which is inconsistent with actual CDA fact patterns where users can reject recommendations, and that algorithmic game-theory models of learned collusion are separate from hub-and-spoke frameworks and do not generalize to realistic markets.
The article's policy conclusions follow from these critiques. Ezrielev argues that courts should wait for economic research to catch up before establishing legal standards; that the nonpublic/public data distinction is a poor basis for antitrust liability; that restricting CDA access to nonpublic data could reduce pricing efficiency, increase costs, and harm consumers; that the real collusion risk comes from data exchange rather than data use; and that premature standards could discourage beneficial algorithmic pricing innovation.
Relation to AI governance
The article addresses competition law applied to AI systems, a dimension of AI regulation distinct from the safety, geopolitical, and workforce concerns that recur across the source base. It illustrates that AI regulation is not monolithic, in that antitrust, safety, and export-control frameworks apply different logic and can conflict. Ezrielev's "premature standards" argument parallels Amodei's call for surgical, evidence-based regulation in The Adolescence of Technology, in that both warn against regulation outpacing understanding. Algorithmic pricing is an early example of the AI Diffusion challenge, in which AI systems embedded in markets create competitive dynamics that existing legal frameworks struggle to address, and the debate over nonpublic data echoes broader data governance questions in AI policy.
Provenance
The article appeared in Antitrust (American Bar Association), Fall 2025, Vol. 40, No. 1, authored by Jay Ezrielev, founder of Elevecon LLC.