Although antitrust scrutiny of “big tech” companies has increased drastically over the past decade, much of the national debate has concerned issues of monopolization and the Sherman Act—the dominant federal antitrust statute. But with rapid developments in artificial intelligence and machine learning, algorithmic price fixing has become an increasingly pressing threat that the Sherman Act is ill-equipped to tackle. Under the current framework, the challenge of establishing the existence of an agreement between competitors in cases in which the algorithms have evolved beyond their programmers’ intentions presents difficulties to regulators.

This Note underscores that section 5 of the Federal Trade Commission (FTC) Act offers a broader mandate to antitrust regulators and argues that it is the best vehicle for bringing price-fixing claims against companies that use algorithms to collude with one another. Unlike the Sherman Act, the FTC Act bans unilateral conduct and invitations to collude. It can reach collusive conduct without the showing of an agreement. While circuit courts have curbed the scope of the FTC Act by finding that tacit collusion is lawful, the FTC still retains the authority to bring standalone actions for section 5 violations. This Note presents a framework for regulators to consider when bringing a standalone section 5 action for algorithmic collusion. It argues that algorithms function as public announcements and should be scrutinized more carefully—first, for certain market characteristics indicating that collusion would benefit market players, and second, for the existence of plus factors that demon-strate the likelihood of actual collusion.

The full text of this Note can be found by clicking the PDF link to the left.


It started with flies. At its peak, The Making of a Fly, a biology textbook for fly researchers, was priced at almost $24 million on Amazon. 1 Olivia Solon, How a Book About Flies Came to Be Priced $24 Million on Amazon, WIRED (Apr. 27, 2011), [https://​​]. The book, originally printed in 1992, was out of print by 2011, but Amazon had listed seventeen copies for sale. 2 Id. While the fifteen used copies started at $35, the two new copies—sold by two different sellers—started at well over $1 million. 3 Michael Eisen, Amazon’s $23,698,655.93 Book About Flies, It Is Not Junk: A Blog About Genomes, DNA, Evolution, Open Science, Baseball and Other Important Things (Apr. 22, 2011), []. Every day, the prices of both copies increased by set multiples; the sellers had pegged their prices to each other’s, causing a feedback loop, which only stopped once someone noticed the outlandish amount. 4 Id. (“Once a day profnath set their price to be 0.9983 times bordeebook’s price. The prices would remain close for several hours, until bordeebook ‘noticed’ profnath’s change and elevated their price to 1.270589 times profnath’s higher price.”). Eventually, the price peaked at $23,698,655.93. The next day, the price dropped to $106.23. 5 Id.

While this particular incident might be explained away by poor over­sight on the part of the sellers, the use of algorithmic pricing on Amazon, as well as on other online platforms, has become increasingly common, 6 See Le Chen, Alan Mislove & Christo Wilson, An Empirical Analysis of Algorithmic Pricing on Amazon Marketplace, in Proceedings of the 25th International Conference on World Wide Web 1339, 1345 (Ass’n for Computing Mach. 2016) (estimating that more than a third of all Amazon sellers who changed prices frequently used algorithms). This fraction has likely increased since then. See, e.g., David Grossman, Left to Their Own Devices, Pricing Algorithms Resort to Collusion, Popular Mechanics (Feb. 12, 2019), https://​www.​​technology/​robots/a26309827/left-to-their-own-devices-pricing-algorithms-resort-to-collusion [] (“A study from 2015 showed that a third of all items on Amazon had prices set by an algorithm, and chances are that percentage has only risen.”). sometimes causing incredibly high—or low—prices. 7 See David Z. Morris, What Causes Crazy-High Prices on Wayfair and Amazon?, Fortune (July 14, 2020), []; Rupert Neate, Amazon Sellers Hit by Nightmare Before Christmas as Glitch Cuts Prices to 1p, Guardian (Dec. 14, 2014), []. In particular, the prevalence of algorithmic pricing has led to increased concerns about ef­fective regulation of these algorithms to prevent price fixing between com­petitors, whether inadvertent or not.

Thus, this Note proposes a framework for the federal enforcement of antitrust prohibitions against price collusion between algorithms. As more companies adopt algorithms to determine prices, recent legal scholarship has noted that competitors may be able to collude with one another, re­sulting in price fixing, without triggering scrutiny under the federal anti­trust statutes as they are currently being enforced. 8 See, e.g., Ariel Ezrachi & Maurice E. Stucke, Artificial Intelligence and Collusion: When Computers Inhibit Competition, 2017 U. Ill. L. Rev. 1775, 1794, 1796 [hereinafter Ezrachi & Stucke, When Computers Inhibit Competition] (describing the challenges of reg­ulating certain types of algorithms); Ariel Ezrachi & Maurice E. Stucke, Sustainable and Unchallenged Algorithmic Tacit Collusion, 17 Nw. J. Tech. & Intell. Prop. 217, 255–58 (2020) [hereinafter Ezrachi & Stucke, Algorithmic Tacit Collusion] (noting the need to up­date “current antitrust policies” to resolve issues of algorithmic collusion). And as these pricing algorithms often use fast-developing technologies like machine learning 9 See, e.g., How Do I Turn Smart Pricing On or Off?, Airbnb Help Ctr., (on file with the Columbia Law Review) (last visited Jan. 12, 2021) (providing an opt-in algorith­mic pricing feature for hosts). or artificial intelligence, 10 See, e.g., Uber AI, Uber,​?_ga=2.​145952564.​1449153791.1600984909-1051261321.1600984909 [] (last visited Jan. 12, 2021) (de­scribing how artificial intelligence “powers many of the technolo­gies and services underpin­ning Uber’s platform”). companies may—even unwittingly—coordinate with other competitors to arrive at a supracompetitive price. Well-meaning companies may not intend to collude, but the “black box” nature of algo­rithms can often result in unintentional price coordination. 11 See infra section II.A.2.

Price collusion claims can be litigated under either section 1 of the Sherman Act or section 5 of the FTC Act. 12 See The Antitrust Laws, FTC: Protecting Am.’s Consumers, https://​www.​​tips-advice/competition-guidance/guide-antitrust-laws/antitrust-laws [https://​perma.​cc/Q4QJ-RN7L] (last visited Jan. 12, 2021) (listing both the Sherman Act and FTC Act as mechanisms to enforce price-fixing prohibitions). But most price-fixing cases are brought under the Sherman Act, as for years the FTC expressed reluctance to challenge practices on a standalone section 5 basis when the Sherman Act could sufficiently address the uncompetitive practice. 13 FTC, Statement of Enforcement Principles Regarding “Unfair Methods of Competition” Under Section 5 of the FTC Act (Aug. 13, 2015), https://www.​ftc.​gov/​system/​files/documents/public_statements/735201/150813section5enforcement.pdf [https://​perma.​​cc/BFC7-K2N6] [hereinafter FTC, Statement of Enforcement Principles] (“[T]he Commission is less likely to challenge an act or practice as an unfair method of competition on a standalone basis if enforcement of the Sherman or Clayton Act is sufficient to address the competitive harm arising from the act or practice.”). This policy statement was rescinded in 2021, but it has not yet been replaced with further guidance. FTC, Statement of the Commission: On the Withdrawal of the Statement of Enforcement Principles Regarding “Unfair Methods of Competition” Under Section 5 of the FTC Act (July 9, 2021), [] [hereinafter FTC, Statement on Withdrawal]. But while price col­lusion can be per se illegal under section 1 of the Sherman Act, the statutory case law requires evidence, whether direct or circumstantial, of an agreement between competitors to fix prices to prove a statutory vi­ola­tion. 14 See, e.g., Bell Atl. Corp. v. Twombly, 550 U.S. 544, 553 (2007) (“[T]he crucial question is whether the challenged anticompetitive conduct stem[s] from independent de­cision or from an agreement, tacit or express.” (second alteration in original) (internal quo­tation marks omitted) (quoting Theatre Enters. v. Paramount Film Dist. Corp., 346 U.S. 537, 540 (1954))). On the other hand, section 5 of the FTC Act does not require an explicit showing of an agreement for antitrust claims. 15 See, e.g., In re Musical Instruments & Equip. Antitrust Litig., 798 F.3d 1186, 1196 (9th Cir. 2015) (“[U]nlike § 1 of the Sherman Act, a violation of § 5 of the FTC Act does not require allegation and proof of a contract, combination, or conspiracy.”). Section 5 is thus more expansive than the Sherman Act, as it also bars unilateral conduct such as invitations to collude. 16 Barbara Blank & Eric Sprague, A Compliance Check for Collaborators Who Also Compete, FTC: Protecting Am.’s Consumers (Aug. 16, 2016),​news-events/blogs/competition-matters/2016/08/compliance-check-collaborators-who-also-compete [] (“Invitations to collude, which are solicitations by one competitor to another to coordinate on price, output, or other important terms of competition, can cause competitive harm, and therefore violate Section 5 of the FTC Act.”). Consequently, it is easier to file suit under the FTC Act against individuals or companies that rely on more complex algorithms to fix prices without directly communicating with one another. But after a series of “adverse appellate rulings in the 1980s involving Commission attempts to expand Section 5 beyond the Sherman Act,” the FTC generally abandoned bringing standalone section 5 price-fixing claims in court, a policy that was formalized in 2015. 17 James C. Cooper, The Perils of Excessive Discretion: The Elusive Meaning of Unfairness in Section 5 of the FTC Act, 3 J. Antitrust Enf’t 87, 88 (2015); see also FTC, Statement of Enforcement Principles, supra note 13.

Several academics have noted the Sherman Act’s deficiencies when it comes to vig­orously enforcing against anticompetitive algorithmic con­duct. 18 See, e.g., Salil K. Mehra, Antitrust and the Robo-Seller: Competition in the Time of Algorithms, 100 Minn. L. Rev. 1323, 1328 (2016) (“The Sherman Act contains a gap in its coverage under which oligopolists that can achieve price coordination interdependently, without communication or facilitating practices generally escape antitrust enforcement, even when their actions yield supracompetitive pricing that harms consumers.” (footnote omitted)). Others have pointed out that a potential solution for algorithmic collusion may lie in the FTC Act instead. 19 See, e.g., Ezrachi & Stucke, Algorithmic Tacit Collusion, supra note 8, at 233 (“One way is for the US Federal Trade Commission to attack practices that facilitate tacit collusion under its broader powers under Section 5 of the FTC Act, which it hasn’t actively pursued in the past few decades.”). But no attempt has yet been made to map out an outline for bringing a standalone section 5 claim in algorithmic price-fixing cases. Thus, this Note aims to provide an overview of how section 5 can be used effectively to regulate different types of algo­rithmic collusion.

Specifically, this Note argues that the FTC Act, given its broader pow­ers, remains the best enforcement mechanism to challenge algorithmic price fixing, and prior cases that limited the scope of section 5 should be revisited in light of rapid technological developments. Part I describes the current statutory regime of federal antitrust law with a focus on the Sherman and FTC Acts. While price fixing is a per se violation of the Sherman Act, a collusion claim requires evidence of an agreement. In con­trast, the FTC Act contains a broader prohibition on all unfair methods of competition and does not require an agreement among competitors.

Part II describes four broad categories of pricing algorithms and the challenges they pose to enforcement. While some algorithms are built with a clear intention to collude, others are more opaque. A conventional ap­plication of the antitrust framework is sufficient to regulate the simpler algorithms. But with the advent of artificial intelligence and other technol­ogies, the more sophisticated algorithms are capable of evolving beyond what their programmers had initially built them for. 20 See Joe Devanesan, Google Has Found a Way for Machine Learning Algorithms to Evolve Themselves, TechWire Asia (May 21, 2020),​2020/​05/​google-has-found-a-way-for-machine-learning-algorithms-to-evolve-themselves [https://​perma.​​cc/​7QWD-DY8W] (“[A]lgorithms can be tested against standard AI problems for their abil­ity to solve new ones. . . . And crucially, those machine learning applications will be free from human input.”). It is these algorithms that will present major challenges to antitrust enforcers.

Consequently, Part III suggests that the best way to tackle these algo­rithms is under the FTC Act: by treating collusive algorithms as announce­ments and therefore as invitations to collude. Research has indicated that algorithms can evolve to “broadcast” their pricing intentions to other cho­sen algorithms while also masking these communications from third par­ties. 21 See infra notes 191–194 and accompanying text. In doing so, they function as announcements. And when a market satisfies certain structural characteristics that can facilitate collusion, courts should more closely examine any evidence of specific plus factors 22 Plus factors are “economic actions and outcomes, above and beyond parallel con­duct by oligopolistic firms, that are largely inconsistent with unilateral conduct but largely consistent with explicitly coordinated action.” William E. Kovacic, Robert C. Marshall, Leslie M. Marx & Halbert L. White, Plus Factors and Agreement in Antitrust Law, 110 Mich. L. Rev. 393, 393 (2011); see also infra notes 177–180 and accompanying text. that might be more prominent in an algorithmic setting. If these structural market characteristics are accompanied by specific plus factors, this should weigh in favor of finding evidence of an invitation to collude.