This browser is not actively supported anymore. For the best passle experience, we strongly recommend you upgrade your browser.
Welcome to Reed Smith's viewpoints — timely commentary from our lawyers on topics relevant to your business and wider industry. Browse to see the latest news and subscribe to receive updates on topics that matter to you, directly to your mailbox.
| 3 minutes read

Unpacking DOJ's Theory of Per-Se Price Fixing for "Recommended Price" Algorithms

As early iterations of AI-powered tools enter the workplace, only time will tell how these technologies reshape business operations and decision making. In the interim, the Antitrust Division of the Department of Justice ("DOJ") recently filed a statement of interest brief that suggests competitors who adopt pricing algorithms may face per se liability for price fixing. Is DOJ carving out a no-go zone for current (and future) business intelligence technologies? Probably not. Rather, the coming wave of AI solutions will be judged by (i) whether they supplant independent price setting and (ii) the degree of non-public data used to generate outputs.

First, some background. A series of civil class actions have challenged the use of property management software by landlords for the purpose of generating real-time, property-specific pricing recommendations.  According to plaintiffs, landlords provided non-public data to the software (e.g., actual rents paid, occupancy rates) and then received recommendations on forward-looking rents based on an algorithmic analysis of all available data. On November 15, 2023, DOJ filed a statement of interest and supporting memorandum in these actions. That memorandum supports the sufficiency of plaintiffs' allegations and opposes defendants' motion to dismiss. DOJ offers the following summary of their position:

Although not every use of an algorithm to set price qualifies as a per se violation of Section 1 of the Sherman Act, it is per se unlawful when, as alleged here, competitors knowingly combine their sensitive, nonpublic pricing and supply information in an algorithm that they rely upon in making pricing decisions, with the knowledge and expectation that other competitors will do the same.

Second, a caveat. DOJ's statement of interest assumes plaintiffs' allegations to be true for purposes of assessing whether they state a claim under Section 1 of the Sherman Act. The actual evidence may fail to support some or all of those allegations. I take the same approach here and ask the following general question: Is DOJ categorically opposed to price algorithms that fit the summary above, or is more required? Answering that question may prove critical for new AI platforms that function, by definition, on the combination and collective analysis of large data sets.

A close read of DOJ's memorandum suggests that the legality of AI output -- even in regard to pricing – will not hang in the balance of circumstantial evidence regarding competitors' “knowledge and expectation[s].” Rather, DOJ also discussed two additional elements of the algorithm that appear material to their theory of per se liability.

  1. The presence of enforcement mechanisms that drive adoption of a specific recommended price.  DOJ appears to give material weight to allegations that the software enforced adoption of its recommended prices by requiring additional administrative steps to deviate from the recommendation and/or monitored the identify of personnel that made those deviation requests. In short, any effort to promote or enforce the recommended price may appear inconsistent with non-committal pricing data that a user was free to adopt, modify, or reject. One can imagine various alternative outputs to a specific recommended price, such as a range of prevailing prices that could be filtered on certain data points, which may provide a similar level of pricing intelligence while preserving the autonomy of each user's decision making.
  2. The use of non-public data to generate forward-looking prices.  A core theme of DOJ's memorandum is that the use of technology does not materially change existing precedents on potential anticompetitive conduct. For example, competitors could not divide a market through the use of some common mapping technology, just as they could not do the same via an in-person meeting and a paper map. Thus, as a general matter, exchanges of non-public data between competitors are highly choreographed affairs that involve third parties who aggregate sufficiently historical data for retrospective analysis only. If an algorithmic pricing solution is going to utilize non-public data, participants must ensure that the inputs are manipulated in a way that removes the potential for data outputs to chill future price competition. 

A forthcoming AI revolution will certainly lead to new interpretations and guidance from DOJ and related antitrust enforcers. For now, early adopters are best advised to proceed with caution and to vet any pricing technologies with experienced antitrust counsel.



antitrust, ai, supply chain, emerging technologies