On Simple Competition Policy

The Network Law Review is pleased to present you with a special issue curated by the Dynamic Competition Initiative (“DCI”). Co-sponsored by UC Berkeley and the EUI, the DCI seeks to develop and advance innovation-based dynamic competition theories, tools, and policy processes adapted to the nature and pace of innovation in the 21st century. This special issue brings together contributions from speakers and panelists who participated in DCI’s second annual conference in October 2024. This article is authored by Keith N. Hylton, William Fairfield Warren Professor, Boston University; Professor of Law, Boston University School of Law.

***

These remarks address the topic of “simple competition policy for a complex world”.[1] There is much to be said in favor of simplicity in legal rules. Complex legal rules are often difficult for the individuals who must comply with the rules to understand what they mean. Complex rules are often confusing, and take a long time to understand, when in the real business world decisions must be made quickly – at the pace at which business occurs. Businessmen do not have the luxury that academics possess to read, discuss, and think about rules for long periods of time. Markets often force them to act before having the time to fully understand a complex environment.

Complex rules, because they often create clouds of confusion, permit regulators to act with inordinate discretion in interpreting the rules. Discretion opens the door for the private preferences of regulators to determine how rules will be enforced.[2] The result of this process may be a system directed by individual preferences rather than laws or rules.

The Roman emperor Caligula posted his edicts so high that ordinary Roman citizens could not read them.[3] This is equivalent to having complex rules.

1. Simple Per Se Rules as a Fantasy

Some commentators have suggested that the establishment of more per se prohibitions in competition law would improve society’s welfare by making antitrust enforcement more predictable and certain.[4] Of course, certainty could also be achieved by per se legality rules, but because such rules would weaken antitrust enforcement, few commentators offer them as solutions to the complexity problem.

The notion that per se prohibitions would make competition law enforcement more predictable is a fantasy. Any statutory system will involve some complexity. Jeremy Bentham gave the example of a statute that says (substituting my own informal terms), “You can’t lay hands on a priest”.[5] He goes on to spend a page or two describing the different interpretations such a simple prohibition could generate – ranging from the literal interpretation that an ordinary person cannot lay their hands on a priest to the more technical interpretation that a police officer cannot arrest a priest. Between these endpoints there are numerous combinations and permutations of the endpoint interpretations to consider. The point of this example is to show that even simple per se prohibitions can generate complexity pretty quickly.

We should abandon the notion that there is some world consisting of simple per se antitrust prohibitions that operates beautifully in comparison to our current situation. Complexity in the enforcement of law is unavoidable.

2. Optimal Competition Law and Policy

I will offer some broad remarks on what I think is an optimal design for competition law enforcement and try to link my proposed design with some current discussions of competition policy.

The first question is one of institutional design. How do you design a competition law system that is relatively simple and effective at regulation? I think there is a lot to be said for a simplified version of the common law system: a system of adversaries going against each other in a court or some other judicial form and a neutral judge in between them. The judge is by nature going to be weighing off or comparing the different interests of the opposing parties and trying to reach some sort of balance between their arguments – in a sense, balancing the welfare effects of alternative decisions.

That is a very simple model of the common law system, and you can operate with a simple set of meta-norms, such as a requirement to act reasonably or to minimize total costs.[6] You could describe the meta-norms at work in several different ways, and through this simple system you can generate relatively specific and complex rules that apply to specific fact-settings.

For the most part, that is the way American antitrust has evolved. It started off with simple rules under the Sherman Act.[7] The Sherman Act consists essentially of two paragraphs. You can boil it down to just two sentences: (1) “Do not fix prices,” and (2) “Do not monopolize”. After enacting these two sentences, the whole matter was kicked over by the legislature to judges to make decisions in a common law fashion that apply to specific fact-settings.

There are several noticeably beneficial and effective aspects of the way the American system developed antitrust law. I would not favor an approach to competition law that involves a lot of statutes or a lot of statutory rules, because I think it is a fantasy to believe that such an approach would simplify competition policy and make enforcement easier.

Also, statutory systems generate rent seeking; people appealing to the legislature and paying off politicians for getting the rules that they want. A mostly or completely statutory system of competition law is bound to be a bad system. I wouldn’t favor a system with many regulators, or one of regulators supposedly enforcing “ex ante rules”. Such a system is likely to be heavily taxing on businesses. I have been struck by the number of new regulations being proposed in the EU for generative AI.[8] These regulations are like barnacles on a ship that prevent it from moving forward. At this stage of the innovation process it is probably better for government to stay on the sidelines and wait to see how things develop in the AI space. Unless there is an example of an extremely important harm that is impending, regulators should just wait until a party comes before them who has suffered a real harm.[9]

The other part of this “common law system of competition law” consists of rules on evidence. The common law system operates largely with a “preponderance rule” though sometimes the rule is biased (e.g., “reasonable doubt” rule) in favor of one party or the other depending on the reasons for doing so.[10] For example, if the state is going against an individual the state may be required to prove guilt beyond a reasonable doubt. There are good arguments for making the burden of proof or the standard of proof higher for the state – the main one to prevent the state from being corrupted. The standard of proof decides how you determine whether someone has proven something, and the burden of proof determines who has the burden of proving something. All of these rules are part of a simple common law system.

It has become common now to talk about competition policy as involving a tradeoff between “static costs” and “dynamic costs”.[11] Static costs are associated with the traditional consumer welfare effects of monopolization. Dynamic costs are associated with the innovation suppression that can occur because of antitrust enforcement. One could refer to a balancing of the profits, or I would like to think of it as normal rents from control of assets, against the consumer surplus or consumer welfare effects of a decision – where the profits or rents represent (or are proportional to) “dynamic costs” and the consumer surplus loss represents “static costs”.

Competition policy tries to strike a balance between these two costs.[12] A system of no intervention of any sort says, in effect, that you can make whatever profits you can, and the state won’t intervene. That is also a system that maximizes innovation because no one ever needs to worry about any kind of antitrust penalty under a system where there is no threat of antitrust regulation. Then you have the opposite system, which is full of per se prohibitions, that says you can’t do anything, or that anything you do could be an antitrust violation – any action that might maximize or increase your profits and possibly work against the interest of consumers is a violation of the law. Such a system minimizes static costs, and, on the other hand, it also minimizes the dynamic benefit from investment.

Thus, we have a tradeoff in antitrust – and you see the same tradeoff in a lot of fields. You see the tradeoff between static and dynamic costs in intellectual property law as well. Ron Cass and I wrote a book on the intellectual property laws, and throughout our study, we examined the balance of dynamic and static costs implicit in key rules.[13] Intellectual property is a classic illustration of this kind of balancing done through the common law process. There are certain doctrines that have been developed in intellectual property that can be described as the result of balancing static and dynamic costs. For example, the abstraction principle in patent law is clearly a result of balancing the static and the dynamic costs of patent protection.[14]

The distinction between static and dynamic costs can also be framed as “competition within the market” versus “competition for the market”.[15] Many discussions of antitrust focus on the static problem alone – that is, competition within the market. For example, many commentators focus on efficiencies that are passed on to the consumer in competition regulation. Focusing solely on efficiencies likely to be passed on to consumers exemplifies an excessively static approach. The better approach to efficiencies, in my view, would be to look for “replicable efficiencies”. If the efficiencies are replicable, then even if the firm does not pass the efficiencies on to the consumer right away, another rival can enter and replicate the efficiencies, and competition between the new firms subject to a lower cost base would eventually benefit the consumer. The focus on “passing on” implicitly assumes that the probability of entry is zero. There was also a mention of “capabilities” as an alternative to the standard efficiencies examination in competition law.[16] An examination of capabilities would certainly be a move in the right direction. My concern here is defining capabilities. I see three components of the term: (1) animal spirits, (2) technical abilities, and (3) objective conditions that would justify the belief that entry is plausible. In a rational competition law system, I would put more weight on the objective conditions – specifically, what are the objective profit incentives for an actor to take one course versus another. The other features are subject to “gaming” in the regulatory process. Animal spirits, for example, looks to whether the actors in a certain industry are aggressive and eager to seize opportunities. These characteristics certainly exist and vary across industries and societies, but they are also difficult to measure. The same can be said of technical abilities. However, objective conditions often are measurable. Objective evidence on profit conditions and costs can be quantified to some extent, offering a reliable measure of capabilities. This approach can be applied immediately to merger regulation.

Applying the foregoing to the DOJ 2023 Merger Guidelines,[17] I will confine myself to two parts: the new reporting burden and the discretion-maximizing nature of the revisions. The new reporting burden increases the cost of reporting by a factor of four – and that is by the agencies’ own admission. The costly new burden is likely to weigh more heavily on smaller firms than on those involved in larger mergers. The reporting burden may be costly enough to strangle some of the smaller mergers in their infancy – which can harm innovation. The discretion-maximizing nature of the new rules will also increase the burden of regulation. The discretion-enhancing rules throw up a cloud of uncertainty over mergers. This is a paradox, given that the merger guidelines were initiated in 1968 to provide clarity and predictability to firms contemplating mergers.[18]

3. Conclusion

Simple competition policy, for a complex world, starts with institutional design. The Sherman Act appears to offer an institutional framework that practically approximates the ideal that I have proposed here. It is a framework that is flexible and responsive, and at the same time capable of regulating down to a fine level of detail. However, what precisely is the regulatory goal and how it should be realized in each case or application is something that requires time for courts, regulators, and commentators to discover.

***

Citation: Keith N. Hylton, On Simple Competition Policy, Network Law Review, Spring 2025.

* I thank Nicolas Petit and Thibault Schrepel for helpful suggestions on this paper.

References:

• [1] This is the title of the roundtable panel at the Dynamic Competition Initiative’s 2d Annual Conference, held in Florence, September 21, 2024, https://www.dynamiccompetition.com/event/dci-workshops-2024/. For a recording of the panel, see https://www.youtube.com/watch?v=e42h0jxCo2M.
• [2] The idea that discretion in the interpretation of public laws may generate arbitrary and abusive enforcement has been around for a long time. See, e.g., William Blackstone, Vol. 1, Commentaries on the Laws of England, at 67 (“[T]he freedom of our constitution will not permit, that, in criminal cases, a power should be lodged in any judge, to construe the law otherwise than according to the letter. This caution, while it admirably protects the public liberty, can never bear hard upon individuals. A man cannot suffer more punishment than the law assigns, but he may suffer less. The laws cannot be restrained by partiality to inflict a penalty beyond what the letter will warrant…”).
• [3] Id. at 46.
• [4] See, e.g., the first speaker in the DCI roundtable panel recording on YouTube, supra note 1. The call for abolishing or weakening the consumer welfare standard would inevitably, if widely followed, lead to the adoption of more per se rules. For one example of an arguments for weakening the consumer welfare test, see Marshall Steinbaum and Maurice E. Stucke, The Effective Competition Standard, A New Standard for Antitrust, Roosevelt Institute, 29 (2018), available at http://rooseveltinstitute.org/wp-content/uploads/2018/09/The-Effective-Competition-Standard-FINAL.pdf. Admittedly, one could draw a distinction between simplicity and the per se nature of a rule. For example, one could argue that simplicity might be achieved without per se rules. To some degree, I am taking this position here. However, there is a common viewpoint that per se rules are simpler and would therefore make enforcement more predictable. I am questioning the validity of that common viewpoint.
• [5] Jeremy Bentham, A Comment on The Commentaries: A Criticism of William Blackstone’s Commentaries on the Laws of England 15-18 (Charles W. Everett ed., 1928).
• [6] See, e.g., Keith N. Hylton, Law and Economics versus Economic Analysis of Law, 48 European Journal of Law and Economics 77 (2019).
• [7] One could argue that the Sherman Act is “short” but not “simple”. According to this view, while the injunctions in the Sherman Act can be stated briefly, they embody complex ideas. For example, monopolization is a term that represents a complex idea. Because this is true, my use of the term “simple” is arguably misleading. Still, I would argue that simplicity consists in the presentation of a straightforward, concise injunction that most people would understand. For example, the injunction issued by a parent to a son to “be a good boy” is also short and at the same time complex. But most children would understand what the injunction means in most settings.
• [8] For discussion, see Thibault Schrepel, Decoding the AI Act: A Critical Guide for Competition Experts (July 22, 2024). Amsterdam Law & Technology Institute Working Paper Series 2024, Available at SSRN: https://ssrn.com/abstract=4609947; Marco Almada, The EU AI Act in a Global Perspective (January 05, 2025). Handbook on the Global Governance of AI (Furendal & Lundgren, eds, Edward Elgar 2025 forthcoming), Available at SSRN: https://ssrn.com/abstract=5083993 or http://dx.doi.org/10.2139/ssrn.5083993; Behrang Kianzad, Fairness, Data and Competition Law: Reconciling Fairness Norms in DMA, Data Act and AI Act With EU Competition Law (November 15, 2024). Available at SSRN: https://ssrn.com/abstract=5030204 or http://dx.doi.org/10.2139/ssrn.5030204.
• [9] The traditional theory that AI will lead to anticompetitive harm is based on the notion that AI use will enhance market concentration, as few firms will be able to afford the costs of using AI. See Anton Korinek and Jai Vipra, Concentrating Intelligence: Scaling and Market Structure in Artificial Intelligence (October 02, 2024). Institute for New Economic Thinking Working Paper Series No. 228, Available at SSRN: https://ssrn.com/abstract=5044496. However, the recent launch of the AI distillation DeepSeek suggests that even this traditional argument may turn out to be wrong. New technological applications can easily and quickly overturn early-accepted views of the likely harms to competition resulting from a developing technology.
• [10] Keith N. Hylton and Vic Khanna, A Public Choice Theory of Criminal Procedure,15 Supreme Court Economic Review 61-118 (2007).
• [11] See J. Gregory Sidak and David J. Teece, Dynamic Competition in Antitrust Law, 5 J. Competition Law & Economics, pp. 581–631 (2009).
• [12] See, e.g., Keith N. Hylton and Haizhen Lin, Innovation and Optimal Punishment, with Antitrust Applications, 10 Journal of Competition Law and Economics 1-25 (2014).
• [13] Ronald A. Cass and Keith N. Hylton, Laws of Creation: Property Rights in the World of Ideas (Harvard Univ. Press, 2013). See also, Keith N. Hylton, Property Rules, Liability Rules, and Immunity: An Application to Cyberspace, 87 Boston University Law Review 1-39 (2007); Keith N. Hylton and Mengxi Zhang, Optimal Remedies for Patent Infringement, 52 International Review of Law and Economics 44-57 (2017).
• [14] Id. at 59.
• [15] The phrase is originally due to Joseph Schumpeter. See Michael L. Katz, Big Tech Mergers: Innovation, Competition for the Market, and the Acquisition of Emerging Competitors, 54 Information Economics and Policy, Issue C (2021).
• [16] Petit, Nicolas and Teece, David J., Capabilities: The Next Step for The Economic Construction of Competition Law (December 09, 2024), available at SSRN: https://ssrn.com/abstract=5060004 or http://dx.doi.org/10.2139/ssrn.5060004; David J. Teece, Gary Pisano, Amy Shuen, Dynamic capabilities and strategic management, 18 Strategic Management Journal 509–533 (1997).
• [17] U.S. Dep’t of Justice & Fed. Trade Comm’n, Merger Guidelines (2023), available at: https://www.justice.gov/atr/2023-merger-guidelines. For a recent review and critique, see Keith N. Hylton, Getting Merger Guidelines Right, 65 Review of Industrial Organization 213 (2024).
• [18] Id. at 214.