How the Dynamic Competition Paradigm Challenges Antitrust Law's Static Framework in the Digital Economy

April 1, 2026

Executive Summary

At the 2026 David E. Nelson Memorial Lecture at Berkeley Law, economists David Teece and Nicolas Petit argued that antitrust law’s structural, efficiency-based paradigm is systematically misreading competition in digital markets — missing real gatekeepers, gerrymandering market definitions, and ignoring the innovation capabilities that determine competitive outcomes. The single most important takeaway: enforcement frameworks that cannot account for dynamic competition and technological capability risk harming both innovation and consumer welfare alike.

Instructor(s)

Prof. David Teece, Professor, Haas School of Business, UC Berkeley
Prof. Nicolas Petit, Professor, European University Institute
Dr. Bowman Heiden, Founding Faculty Director, Haas School of Business, UC Berkeley
Daralyn Durie, Partner, Morrison Foerster

Keywords

dynamic competition antitrust law digital economy • static vs. dynamic competition paradigm • United States v. Google, Section 2 Sherman Act default search monopoly • FTC v. Meta market definition gerrymandering • Digital Markets Act DMA consumer harm Europe • dynamic capabilities innovation ecosystems antitrust • Neo-Brandeisian antitrust enforcement critique • algorithmic collusion antitrust analysis • Nvidia Grok capability acquisition antitrust review • “how should antitrust law account for innovation ecosystems” • “dynamic competition standard for digital platforms 2025 2026” • Schumpeterian competition technological change competition policy

Legal Analysis

United States v. Google and the Missing Gatekeeper: How Static Market Analysis Distorts Digital Antitrust Cases

The structural deficiencies of contemporary antitrust analysis are nowhere more visible than in United States v. Google LLC, in which Judge Amit Mehta of the United States District Court for the District of Columbia held in 2025 that Google had unlawfully maintained its monopoly in general search under Section 2 of the Sherman Act, 15 U.S.C. § 2, through a roughly twenty-billion-dollar revenue-sharing agreement that secured Google’s placement as the default search engine across Apple’s Safari browser, Siri, and Spotlight. Nicolas Petit observed that the government’s theory rested on a framing that legal doctrine reinforces but business reality complicates: “How can a monopolist be the one ending up paying Apple $20 billion in rent just to appear on the screen? Who’s actually the gatekeeper taxing the other here?” That question — pointing to Apple’s enormous distributional leverage over mobile query traffic — received no meaningful treatment within the liability framework because antitrust’s structural lens fixes attention on the alleged monopolist rather than on the systemic architecture of competitive bargaining. David Teece sharpened the indictment by noting that the judicial record also failed to incorporate evidence of Microsoft’s strategic failures with Bing — specifically, the sustained underinvestment and operational mismanagement under Steve Ballmer — as a causally independent explanation for the competitive outcome. “Everybody knows it’s important,” Teece said. “People in the street know it’s important. People in Silicon Valley know it’s important. It’s not talked about in the courtroom.” The explanatory gap is, in Teece’s view, paradigmatic: both government and defense economists remained anchored to static micro-economic tools that have no vocabulary for organizational capabilities. The remedy phase further exposed the problem when Judge Mehta himself acknowledged that no party had raised the competitive significance of generative artificial intelligence at the liability trial, despite that technology’s already demonstrable presence in the market at that time. Petit characterized this sequence as an indictment of the process itself: the question is whether the billions spent in litigation and enforcement are an efficient route to conclusions that industrial observers could have identified at the outset.

Market Definition as Gerrymandering: FTC v. Meta, the Digital Markets Act, and Enforcement That Harms Consumers

The Federal Trade Commission’s ongoing action against Meta Platforms illustrates a second mode of analytical failure — the construction of artificially narrow relevant markets that systematically exclude the most significant competitive pressures. In its complaint, the FTC proposed a market in which Facebook and Instagram competed solely with Snapchat and MiWi, an application that, as Petit noted, “no one knows.” The proposition that Meta does not compete with TikTok or YouTube for user attention is, Petit argued, empirically untenable: video content on Instagram Reels, Facebook Reels, TikTok, and YouTube Shorts is functionally interchangeable from a user-experience standpoint, and when the market is defined to include those platforms, Meta’s share falls to approximately thirty percent — well below conventional monopoly thresholds. Judge Boasberg’s appellate review recognized the definitional defect, yet the Commission appealed, producing what Petit described as a “logical dead end”: under the FTC’s own market theory, a merger between Meta and TikTok would be presumptively lawful, an outcome that flatly contradicts any coherent account of competition in social media. The European experience under Regulation (EU) 2022/1925, the Digital Markets Act (DMA), provides an additional and cautionary data point. In implementing the DMA’s self-referencing prohibition, Google removed interactive map functionality from European search results — rendering the embedded map, in Petit’s description, “a dead map, just a picture of a map.” Meta, responding to the DMA’s data-unbundling requirements, initially presented European users with a binary choice between consenting to targeted advertising or paying a monthly subscription of approximately ten US dollars — a reform that added friction and out-of-pocket cost in the name of competition. The European Commission subsequently required Meta to abandon the subscription model. In both cases, the regulatory intervention produced immediate, observable consumer harms while the competitive benefits remained speculative. Petit summarized the structural lesson: “By focusing on making sure that the gatekeepers are not too big, the agencies in Europe have forgotten about the consumer.”

Dynamic Capabilities, Schumpeterian Competition, and the Paradigm Shift Antitrust Economics Has Yet to Complete

The deeper problem identified by Teece and Petit is not doctrinal but epistemological: antitrust economics imported the wrong framework when the Chicago School displaced the Harvard structuralist tradition — exchanging Joe Bain’s structure-conduct-performance model for static price theory, while leaving innovation outside the analytical core. Teece traced this deficit to his 1990 address at a Berkeley centennial symposium on the Sherman Act, where he argued that “it’s fantastic that we have now joined law and economics and competition policy with the Chicago School coming in — the problem is we brought in the wrong economics. We brought in static microeconomics and innovation is absent.” The central organizing principle of Teece’s dynamic competition framework, drawn explicitly from Joseph Schumpeter and developed over three subsequent decades, holds that “the primary driver of competition is innovation, and all the static stuff is relatively unimportant.” This reorientation has profound implications for every major analytical step in antitrust methodology — market definition, harm assessment, and remedy design. If competitive position is determined not by current market share but by the accumulation and deployment of organizational capabilities — the repeated, describable processes of knowledge creation that reside within firms — then concentration ratios and static price-effect models are measuring the wrong variables. Teece noted that the term “dynamic competition” did not appear in the economic literature until he introduced it in 1992, and that the concept remains absent from economics department syllabi: “If you go to an index of an economics textbook and look under capabilities, you will find zip.” Petit situated this knowledge gap within a broader institutional pathology: competition agencies are evaluated on case volume, not welfare improvement, producing what he called “policy-affirming evidence making rather than evidence-based policymaking.” The iRobot example — in which European Commission intervention blocked Amazon’s proposed acquisition of the robotics firm, which subsequently entered insolvency and was sold to a Chinese acquirer — illustrated Petit’s contention that enforcement decisions made without a rigorous account of dynamic capabilities can destroy technological progress rather than preserve it. Both speakers drew on the framework of Thomas Kuhn’s The Structure of Scientific Revolutions — itself written at Berkeley — to argue that the anomalies accumulating in antitrust practice are now sufficient to force a paradigm transition, provided that law schools, business schools, and economics departments can overcome the institutional incentives that sustain the static model. Petit predicted that the path forward requires building a “fine-grained study of dynamic adoption and dynamic capabilities” into the evidentiary standards of antitrust proceedings, with capability combination analysis in merger review supplementing, and ultimately displacing, fragile merger-simulation price models.

Generated by AI from the Interview/Transcript below.

Key Takeaways

Static economics misframes digital competition. Antitrust analysis built on static microeconomics cannot capture the innovation dynamics that actually determine competitive outcomes in digital markets; as Teece argued, “if you drive for efficiency, you’re going to drive away from innovation.”
Missing the gatekeeper in Google. The United States v. Google LLC liability framework failed to interrogate Apple’s distributional leverage, asking who is the monopolist without asking, as Petit put it, “who’s actually the gatekeeper taxing the other here?”
Market definition as gerrymandering in Meta. The FTC’s proposed market excluding TikTok and YouTube from Meta’s competitive environment is internally incoherent; under that definition, Petit noted, a Meta–TikTok merger would be presumptively lawful.
DMA enforcement produced consumer harm. European Digital Markets Act implementation resulted in degraded product functionality (Google Maps) and new subscription fees (Meta), illustrating that structural regulation without capability analysis can harm the very consumers it purports to protect.
Capabilities absent from antitrust vocabulary. The economics field has no operative language for organizational capabilities, leaving courts without evidentiary tools to assess why one firm outcompetes another; Teece observed, “you will find zip” in any economics textbook index under “capabilities.”
Institutional incentives entrench the old paradigm. Competition agencies are measured by case volume rather than welfare outcomes, generating what Petit called “policy-affirming evidence making” rather than evidence-based policy — a structural moral hazard that perpetuates outdated enforcement theories.
Innovation ecosystem health is the right enforcement metric. Teece argued that the appropriate evaluative lens for any contested business conduct is “what does it do to the robustness of the innovation ecosystem,” not whether it increases or reduces static efficiency.
Algorithmic collusion requires an innovation-welfare test. When asked about algorithmic coordination among AI systems, Teece reframed the inquiry: “Does algorithmic collusion help innovation or hurt? If it hurts it, we should block it” — rejecting the threshold question of whether coordination technically constitutes collusion as analytically insufficient.
Capability analysis can extend to non-merger transactions. Petit argued that antitrust law’s historical flexibility — demonstrated by the Sherman Act’s adaptation to trust structures in 1890 — is sufficient to assert jurisdiction over hybrid transactions such as the Nvidia–Grok licensing-and-hiring arrangement, and that capability-alignment analysis can assess net innovation welfare even without the predictive precision of merger-simulation models.
Paradigm shift requires institutional convergence. Teece called for the intellectual — not merely organizational — integration of law schools, business schools, and economics departments around a dynamic competition framework, predicting that “10 years from now, no one will be talking about innovation efficiency” as the awkward hybrid concept gives way to a fully articulated new paradigm.

B-CLE Recording (CLE: Free) | Youtube Recording | Resource(s) | Agenda

Download the transcript and slides here!

Interview/Transcript

This interview/transcript was based on a conversation about the 2026 David E. Nelson Memorial Lecture: The Disconnect between Business and Law Schools on the Digital Economy, hosted by Berkeley Center for Law & Technology, UC Berkeley School of Law and Haas School of Business, UC Berkeley. The conversation was presented by David Teece, Professor, Haas School of Business, UC Berkeley, and Nicolas Petit, Professor, European University Institute and occurred on April 1, 2026.

Daralyn Durie 00:25

Good afternoon everyone and welcome to the 2026 David E. Nelson –sorry, I got it — and welcome to the 2026 David E. Nelson Memorial Lecture at the Berkeley School of Law. It is my privilege to represent Morrison Foerster here today, and I’d like to start by telling you just a little bit about Dave Nelson and why his partners, firm, family and friends wanted to recognize him by funding this annual lecture series. Dave Nelson’s legal education took place here at Berkeley. He graduated from the school in 1959 — can you all hear me? Great — he then joined Morrison Foerster in 1960 and in what may sound to some of you like an anachronism, he practiced with our firm for his entire career, which spanned over three decades. He was a tech lawyer ahead of his time, representing companies like Memorex and Fujitsu. Dave was also very active in public service, and he was particularly committed to civil rights, to education, and to improving the legal profession. He served on the Berkeley Board of Education, the Board of Directors of the Oakland Museum, and was chairman of the San Francisco Committee for Civil Rights Under Law. The David Nelson Memorial keynote address is intended to honor his contributions to the profession, and this is an appropriate place to do so, given his long standing relationship with Berkeley Law School, his interest in education, and his work at the intersection of technology and law, so I’d like to express the appreciation of Morrison Foerster and the Nelson family to the Berkeley Center for Law and Technology for its assistance in making this possible. Today’s program, Two Schools, One Economy: The Disconnect Between the Business and Law Schools on the Digital Economy, explores a central tension of our time, the contrast between how business and law approach innovation, market power and competition in the digital age. And we are extremely fortunate to have two distinguished scholars who bring complementary perspectives to this conversation. First, it’s my pleasure to introduce Professor David Teece of the Haas School of Business here at UC Berkeley. Professor Teece is a leading economist and a global authority on innovation, technological change and competition policy. His work has shaped how we understand dynamic competition and the strategic behavior of firms, not that law firms would ever engage in strategic behavior, and he has advised businesses and governments around the world, while also contributing extensively to academic research. Joining him is Professor Nicolas Petit of the European University Institute. Professor Petit is a prominent scholar of competition law, known for his work at the intersection of law and economics, particularly in the context of digital markets. His research examines how antitrust frameworks adapt to technological change and innovation driven industries, and he has been influential in advancing more economically grounded approaches to competition policy together, Professors Teece and Petit will help us think more deeply about how legal and business perspectives can, or perhaps should better align in shaping the future of the digital economy. So with that, please join me in welcoming Professor Nicolas Petit.

Nicolas Petit 04:24

Ladies and gentlemen, it’s great to be here at Berkeley. I’m going to start with a little bit of a provocation. All right, so we want to start talking about the disconnect between law and business, law schools and business schools, and I want to open with a picture. This is Leland Stanford. That’s why I’m saying that provokes. And this is the namesake of a neighbor of this university in a management science seminar, that person is a case study in entrepreneurial success. He is a builder. He mastered the idea of network effects before we even invented the concept. He built a moat, very large market share, exceptional profits. Now to a lawyer, Leland Stanford, is exactly the opposite. If the reason for the Interstate Commerce Act of 1887 is the trigger for the Sherman Act of 1890, if the railroad monopoly octopus a cautionary tale of power that needs to be constrained. So this brings us to the central paradoxes of our digital economy. Are people like me — I’m a professor of law — confusing power with success and monopoly with dynamic competition? David and I, we believe that this disconnect is deep and is harming our ability to enact and practice policy in the digital economy. So I’m going to talk about that first, and then David will come in, and then I’ll come back in with final remarks. So let me start with exhibit number one, US v Google. So in 2025 Judge Mehta here in the US District Court declared Google guilty of unlawful monopoly maintenance under Section two of the Sherman Act. The violation centers on a $20 billion agreement where Google receipts a massive share of advertising revenue back to Apple in exchange for that, Google is the default search engine on safari, Siri spotlight and other access points on the iPhone. Okay. Now you can look at this through two lenses — the low crowd, the low community, sees a giant paying for exclusivity to kill rivals, with Apple being just a nearby standard of very bad violation of antitrust law. But from a business perspective, one is tempted to ask, how can a monopolist be the one ending up ending up paying Apple $20 billion in rent just to appear on the screen? Who’s actually the gatekeeper taxing the other here? That is the hard question. So the Google case may be a story of missing the real gatekeeper. All right, now let’s move to exhibit number two, the FTC, ongoing complaints, I would say even crusade against Meta. So this is not about missing the gatekeeper. It’s probably about missing the market. So to win a section two case, I won’t lecture alone in section two, but I just need to recall this, to win a section two case, the complainant, the plaintiff, needs to prove monopoly power in a properly defined relevant market, we say, in the digital economy, market definition looks like a gerrymandering exercise. So in its complaints against Meta, the FTC proposed the markets where Facebook and Instagram only competed with two apps, Snapchat and another app that no one knows called Miwi. Now look at the screen. These are four identical videos of a big dog. One is an Instagram reel, the other one is a Facebook reel. One is a TikTok video, and the other one is a YouTube short. Does anyone in this room believe Meta doesn’t compete with YouTube and TikTok? So in the real world, in the business world, Meta is in a competitive race for users attention against Tiktok, YouTube and maybe even Netflix and others. And when you define the market this way, Meta share is about 30% far from the monopoly level on appeals. The judge on appeals, Judge Boasberg, saw right through that gerrymandering and said that the case could not stand on that market definition. Yet the FTC continues to insist that Meta is not competing with YouTube and TikTok, and has appealed the judgment of Judge Boasberg. So here’s the logical dead end for you to consider. If the FTC is right, a merger between Meta and TikTok would be presumptively lawful under US anti trust law, right? So this may be, this disconnect, may be the cost of a field that has drifted away a little bit from business reality and reality seen ads by users. All right now, let me move to to exhibit three. So I talked about missing the gatekeeper, missing the markets. Let me talk a little bit about missing the remedy or harming consumers. Let’s go to Europe. I didn’t want to be too hard on the United States, so we can cross the Atlantic and go to my continent. So since 2022 in Europe, we have a sector specific law for large digital platforms, gatekeepers, which is called the DMA. The Digital Markets Act, probably heard about this. The DMA bans a lot of things, but two things it bans very effectively is one practice called self referencing. And the second thing is it forces platforms to unbundle data unless they have explicit consent from users. All right. So following these rules, you can look on the left hand side. Following these rules, Google removed in Europe the ability to click through Google maps from search results in the EU so the map is still there in Europe, but you cannot click on it. It’s dead. It’s a dead map. It’s just a picture of a map. There is not nothing active behind them. On the right hand side, you can see what Meta did in response to these requirements to unbundle data unless you have user consents. The choice that Europeans had immediately after the DMA entered into force was either to keep using Facebook for free and give consent to targeted advertisements, or to pay a 9.99 US dollars a month for a subscription to a Meta or a Facebook version without bundling up data, without communication between personal data on the platform of Meta. So this is interesting, because by focusing on making sure that the gatekeepers are not too big, the agencies in Europe have forgotten about the consumer in the name of competition, they’ve added friction and prices, right? So just notes, because I don’t want to be, you know, criticized for telling a story that’s not true, Meta had to abandon that practice. So the EC required Meta to move to ditch the subscription. So now Meta has to offer the same service without data bundling, for free in Europe, and it’s been fighting that in court. All right, I finally go to exhibit four, and this is where the disconnect just doesn’t produce problem with the remedy, a problem with the market or a problem with identifying the gatekeeper, but may even be failing society overall. So as we all know, we face a lot of questions about privacy, hate speech, untrustworthy information online, and they’re important questions. Yet for many lawyers in Brussels and in the US, the answer is simple: concentration, economic concentration, produces adverse below verse, which means, if we add competition to markets, we’re going to have more privacy, we’re going to have better content, we’re going to have better speech, and we’re going to have better information, right? So this is the sort of the theory, the sort of implicit theory behind the hate against economic concentration is that concentration comes with adverse spillovers on information, democracy, content and civility online. Now we can think about this in economic terms and think about more this more deeply. So suppose for a minute that we break up Google and Meta to achieve a perfectly competitive structure in advertisements. Now all these firms, they are going to be fighting for advertising. Advertisers, right? They are selling ads. So they’re going to fight for advertisers. Advertisers pay them, so they’re going to fight for that. What are they going to do? Well, they’re going to do what competitors do. They are going to lower the prices of ads, and they’re going to increase the quantity of ads, and they are going to increase the targeting quality of ads. That’s what’s going to happen. So the bottom line is that if you introduce function in these markets, what you’re going to end up with is not less data extraction, but probably more data extraction. Now I want to finish with before I end over to David with one word. I think it would be dishonest to say that the legal system is completely blind to these issues. And at least the two cases that I mentioned, US v. Google and FTC vs. Meta, the courts eventually acknowledged that technological change needed a more relaxed approach to antitrust enforcement or to antitrust remedies. So in US v. Google the judgment acknowledged that a year and a half after the liability judgment, Generative AI had come into the space and required a more soft handed approach to resolving the antitrust violation. The question is, is this trial and error approach worth it? So is it efficient to spend billions of dollars and euros in regulation and litigation for a judge or an agency to realize what we knew on day one? So when Judge Mehta says in the remedy trial and judgments that no parties had mentioned Gen AI in the liability trial that was a year and a half before, but Gen AI was already there, if you study industry 18 months before, is this correct? The parties probably remain silent, because the rules of antitrust procedure do not allow them to raise these arguments. So the question that we have today is why isn’t anti trust law and economic regulation updating to that economic reality and making space for these arguments to be heard and acted upon? And why, in particular, as anti trust, instead of updating in that direction, actually decided to revert to old school structuralist arguments made in the 1920s and recasted today has a new and fashionable enforcement trend under the name of neo-brandeisians, right? Why that? Right? And so I want to hand over to David, because he has very strong views about why we did an update.

David Teece 18:06

Well, thank you, Nicolas, and it’s great to be here. And to honor David Nelson, what I learned is that he understood Silicon Valley. And if you’re going to be anti trust lawyer or an anti trust economist, you better understand dynamic competition. We have an enormous benefit here at Berkeley and privilege, because we’re on the edge of Silicon Valley, sort of in it, but not in it. So we can actually don’t get consumed by it, but we can look at it and be part of it. And Nicolas highlights that while law schools and business schools are sometimes different in the way they look at competition. And let me tell you that inside business schools, different groups, or at least between business and economics, there’s different views about competition. Why is that? Because the world is so complicated we’ve got to simplify it. So what do we do? We create paradigms, which is great, so long as they’re good and so long you don’t get locked into them. But I’ve been part of a journey that began here at Berkeley, trying to break the world out of what I call structures, or the static competition gridlock. And I managed to enter this field of, you know, antitrust, economics, at a time when economics was getting into antitrust, which was the Chicago revolution. And my research agenda was early and young, but it was 1990 in the centennial of the Sherman Act. So right here at Berkeley, we held a conference with all the top names from around the country and around the world, at antitrust law and antitrust economics, at the local reader and Turner and Baxter Easter Burke and Posner and Williamson, and little fellas like Richard Gilbert and David. East were there as well, and I gave a talk. And my background, you should know, is, I’m an innovation scholar, first. Believe it or not, it’s very hard to study innovation in an economics department, because economists will know that innovation is important, but they hate it, but it messes up formalization. They can’t formalize it’s too complicated, so they ignore it. And I was bold enough, as a young scholar, I had tenure so I could speak my truth my mind, to say that it’s fantastic that we have now joined law and economics and competition policy, with the Chicago School coming in, the problem is we brought in the wrong economics. We brought in static microeconomics and innovation is absent. And if I can put it this way, it’s not too distant from the truth, a 40 year journey, mainly a soul journey until Nicolas and a few other colleagues have joined in the last five to 10 years to actually bust the old paradigm. And I want to tell you that I think we’re busting it. And the static competition paradigm is giving away, giving way slowly, to the dynamic competition paradigm. But we need to understand that this is not people. Sometimes it’s people with vested interests, but it’s actually people who don’t have the mindset to challenge the mindset that they inherited from the textbooks. And yes, Nicolas is helping out by putting up the cover of a book that was written here for a big week as one of the most influential books of all time, written by Clinton The Structure of Scientific Revolutions. What he points out, thinking so much of the social sciences, you’re thinking about physics and chemistry and so forth. But it’s also true the social sciences is that the way knowledge evolves, the way we advance, is we establish a paradigm and it becomes the mainstream. We use this term. This is a mainstream or textbook view, but every now and again, you got to rewrite the textbooks. And you know, in physics and disciplines like that, when you start seeing the data, it doesn’t fit with the theory. Your theory isn’t going to last all that long. But in economics, it’s not that way. If you’ve got another bunch of theorists that like your theory, you’re in good shape. So evidence is sort of brushed to one side, particularly amongst the theorists and the high priesthood. And this is true in antitrust economics, not just, you know, economic theory, more generally, the high priesthood maintains their position and their righteousness and sort of treats everything else at the back of the hand, but because there’s so much new complexity out there. Now the old framework is breaking down, but the old framework actually also began here at Berkeley, so Berkeley has set a stage in all of this. If you go back to the 30s. The structuralist school began here at Berkeley and at Harvard. There was a fellow called Ed Mason at Harvard and a fellow called Joe Bain here at Berkeley. And Bainian economics is still alive. And in fact, Nicolas is talking about the brown shoe case in Google. It’s old Bainian economics. It came from this very campus here, and was bread and butter for anybody studying and just the field of industrial organization. I remember when I first came to Berkeley, you go to the library, there was like 10 feet — this is when you went to the library for books — it was 10 feet of books of Joe Bain industrial organization because that was a standard textbook. But innovation is left out, and it’s been a battle to get it in. And the structuralist school basically held that market structure drove performance. So that’s why everything starts with looking at market definition, and all of that structure conduct performance now that that’s the dominant theory, but a number of us have said, Well, we think the line of causation may be completely wrong, and said performance actually is what drives structure, and it’s innovation that drives structure anyway. Economists have spent quite a bit of time looking at the linkages between market structure and innovation, but they haven’t looked at all at innovation and its impact on market structure. There’s an endogeneity. The world is complex and it’s a system but Systems Thinking exist. In the engineering school over there, but it doesn’t exist in the economics department over there or in the business school. And this was keenly aware to me, somehow or other, I had the good luck to read Kuhn’s book, and that gave me the confidence to stand alone against the whole damn field for 30 years. I actually — you will not find the term dynamic competition in the literature until 1992 when I put it in there, and the mainstream folks at the Federal Trade Commission, the Department of Justice, and also, quite frankly, in the defense bar, because it will take classes from, well, it’s the same school, you know, treat this stuff with the back of the hand until they no longer can. So we’re at the moment where you can no longer treat it with the back of your hand, in which case the question is, what do we do? And when you dig, and I’m talking now as an economist, because we have been so influential in competition policy, but we have an enormous Achilles heel in economics which no one admit to you, and that is, we do not have a meaningful, robust or realistic theory of the business enterprise, there have been four Nobel laureates gone out for work done on the theory of the firm. Our colleague here, my dear friend, Oliver Williamson, now passed away, for transactions cost economics. Ronald Coates from the lawyer actually Lauren economics, Chicago School on early work on transactions cost economics, and then pot at Harvard, and there’s one other and the OH ation Agion just got theory the firm. But despite this, it is pathetic, utterly pathetic how much we know in economics about firms and firm behavior. Yes, there’s stuff on game theory and so forth, but people don’t stick their nose very close to the ground to look very closely at what goes on in Silicon Valley. And anyway, it’s now time to create this new paradigm. And of course, we’re not going to find the answers just in the field of economics, but we can find the answers in business schools. In business schools, you have both economists, but you also have Organizational Behavior people. You have people that are seriously interested in studying technological innovation, and they’re not held to the same formal standards. They can’t prove a theorem, they’re not going to die professionally, whereas in economics, if you can’t prove a theorem, you’re going to die professionally, but you can also live by doing good, solid empirical work. So we’re now in this lovely position where there’s an open mind. And as Thomas Cohn pointed out, what happens is, when the anomalies get so great, the mainstream starts to break down. But it’s a sociological phenomena. It’s not just a pure linear phenomenon. So you know, you can think of going back to the Copernican revolution. We used to believe that most people believed that the sun rolled around the Earth, and it took a while for that to get overturned. And there’s a lot of gnashing of teeth and people dying on their sword every time you have these major revolutions, but now we’re into dynamic competition. So the good news is that people are talking the language of dynamic competition. The bad news is no one really understands it, because you’ve still got to dig deep, as David Nelson did, and live the life of competition in Silicon Valley, and also you’ve got to come up with a richer story and understanding of what the business enterprise is all about. Here’s another liability we have in economics. We have no language or theories in economics about the capabilities of business organizations. If you go to an index of an economics textbook and look under capabilities, you will find zip, except if you’re looking at economic development in third world countries. So capabilities coming back to the Google case, you know, how did Google Beat Bing? I mean, the government is in there, essentially saying, Google, you did all these bad things, and poor little Microsoft wasn’t able to compete because of the horrible things you did. Did evidence get into the court about what Microsoft was doing and not doing? Did any evidence get into the court? Steve Ballmer was really kind of out to lunch for decade and a half. He made strategic mistake after strategic mistake after strategic mistake. Did not invest enough in Bing. You know this is the stuff that’s important. Everybody knows it’s important. People in the street know it’s important. People in Silicon Valley know it’s important. It’s not talked about in the courtroom. Why not? Because neither the government’s economist nor even Google’s economists had the smarts to figure this out, because the Google folks are also locked to the old paradigm. In fact, Hal Varian, the chief economist, is sort of a static economist, so you know, you think Google is going to launch its best defense? No, it’s not. So the judge makes errors. Google makes errors. The government makes errors, because they’re all floundering around somewhere between the old paradigm and the new but the new paradigm is not yet fully in place. That’s why Berkeley has joined with the European University Institute to bring robustness more empirical work demonstrate how to use this in cases where the rubber meets the road. So, let me sort of back up just a little bit to say that, yes, we don’t have capabilities language, but the other thing that we have done, and once again, Berkeley has been center for all this, is introduce information, at least the study of information, into competition policy. And Hal Varian, of course, who was Google’s chief economist, was also Dean here at Berkeley of the School of Information Sciences. And my colleague, Al Shapiro, wrote this book, which became sort of the new mainstream textbook around digital competition. It’s called Information Rules. And I’ve just got a little provocative paper out there saying information really doesn’t rule capabilities do that, we ask information matters. Don’t get me wrong. I’m not saying what they said. It’s wrong, but it’s all about network effects. It’s all about switching costs. It’s all about standards. These things mattered. But if you don’t have an understanding of capabilities, and you’re going in trying to understand the behavior of enterprises and competitive outcomes, if you don’t have the language of capabilities, I guarantee you you’re not going to get very far. So the challenge is, okay, we now have to develop a theory of capabilities. Well, economics dropped the ball on this 120 years ago to the disastrous impact on just about everything. In my view, the reason why we’re in such a pickle in global competition is because economists all think that the Wealth of Nations comes from trading, which it does, or from Exchange, which part of it does, but the other half of it comes from From production, from capabilities. We don’t have anything on that in our field, so we’ve been misleading Washington and being very wonderful in the way we’ve put forward the benefits of open trading system and the value from Exchange, but without a theory of production. Guess what? We’ve got denuded corporations, we’ve got unemployed people, and we’ve got an industrial enterprise that’s failing, and China is sort of jumping in. Doesn’t believe doesn’t have the same economics, does not have the same textbooks as we have. Thank God it didn’t have our textbooks. China is lucky, so it’s able to gap fill where we’re missing the boat because we’re not looking at the things that really matter. So anyway, what we’re done is, and I can say this is where business schools, I can say so this is the most cited paper in all of Business and Economics. Now it’s my paper on dynamic capability. So people are latching on to this thing. This matters, but now we’re into the awful business of, how do you talk meaningfully about it? And there’s a lot we can talk about, and I’m not going to take time to do that now, other than to say that there are quite a number of books and very large literature in business schools. It hasn’t made its way into the economics departments yet. You know, there’s an allergic reaction when you hit the Economics Department capabilities. Please don’t bring it in, because settled life. And this is paradigm. This has come back at Thomas Kuhn. Yeah, you know, science evolves for the comfort of those within it. To some degree, it’s not just about pure intellectual inquiry. It’s about the comfort of the incumbents, and it’s worth a reread, particularly since it was written here at Berkeley, and our culture here at Berkeley, this is a wonderful thing about Berkeley, is we don’t honor the status quo, and as a consequence of that, we’ve been able to launch this new paradigm, dynamic competition. So here’s sort of a summary, if you go back a little bit about some of the differences. And you know, the primary difference is this one. Anyone that’s familiar with the Chicago School and modern antitrust economics, it’s all about efficient conduct. What is that what innovators are thinking about? No, I’m thinking about how to do, how to innovate, and often and usually, if you drive for efficiency, you’re going to drive away from innovation. If you drive for innovation, you’re going to drive away from efficiency. So these two things are at war with each other a good deal of the time. So you’ve got this rich literature now — that I’ve contributed to early on in my career about the efficiency arguments, and we still use efficiency arguments, and Department of Justice and the FJC have struggled to deal with it — now they talk about innovation efficiency, they can’t drop efficiency. They don’t know that actually innovation and efficiency are at war with each other, but they try to capture the overview by talking about innovation efficiency. It’s nonsense. But this is what happens when one paradigm is giving away to another, you get these awkward looking hybrids floating around that will eventually die. I guarantee you, 10 years from now, no one will be talking about efficient innovation. They won’t, but they have to, for the meantime, because we’re trying to bridge between the old and the new. And I can make jokes about it, because I’ve lived on the other side for 25 years, and how have I been able to get there is in part by actually being engaged with litigation like folks over here and by looking at documents and data and talking to executives and actually believing them rather than, you know, most economists, they have their textbook view, and they’ll listen selectively. When they hear something from somebody in industry that fits the textbooks, they’ll latch on to it. The rest of it are just babbled to them because they have no framework with which to incorporate it. So if David Nelson was here, it’d be great to have a rich conversation with him, because I’m sure that he’s the kind of person we could have all learned from, and probably some of you did learn from. So anyway, this is the main distinction. It’s between innovation and efficiency, but there’s many, many other elements as well. And another key one is time horizon. And it’s just pathetic that in anti trust, you know, people are looking for effects in a one or two year time frame. You know, innovation actually in the digital world, some of it does take place very quickly, but frequently it’s much longer than that. And you know, the guidelines tests and all of those things have this all wrong as an indication that the world the paradigm is changing. The antitrust law journal just did two volumes. They’ve never done two volumes before on dynamic competition. Might say that I have some role in stimulating that, and I’ve got my own article in there, which any of you are interested is a good, quick way to get in. It’s written for lawyers, not for economists, so it’s a fairly quick way to get into the dynamic competition paradigm. So that’s just giving you a little flavor, maybe we should stop talking for the meantime, there’s a lot more we could say. Actually, there’s a conclusion somewhere, implications, yeah, so it’s, it’s back to Schumpeter. I mean, the other thing that drives me nuts is that nobody reads anymore, and so if you ask an economist, what’s Schumpeter about, they’ll say, oh, he said that big firms are necessary for more innovation. Yeah, he said that. He said the opposite to that in a different book. But the key thing he said that matters, and which is the foundation stone of dynamic competition, is that the primary driver of competition is innovation, and all the static stuff is relatively unimportant. So if you take Schumpeter seriously, you know he was throwing water over 100 years ago on the mainstream, and then I just found, much to my delight the other day, Mason, who’s a structureless friend of Bain, looking at Schumpeter, and Mason says Schumpeter is clearly correct, but we don’t know how to do it. So this is the old story. You know, the man is looking for his keys under the street line and say, Well, you know, you didn’t drop your keys over here, but why are you looking there? Was because where the light is, that’s what we’ve been doing for close to 100 years in competition economics, and we’ve been, you know, some of this has had big impacts on policy, but it’s been intellectually, I don’t want to say dishonest, but it’s intellectual laziness by the profession for a long period of time, and a false sense in economics that we suffer from physics envy. We want to have things buttoned up, nice and neat and have formal theorems and be able to prove results. Because of that, physics envy, our field is reluctant to grapple with reality as it is. So my message here is we have to learn to deal with the reality of this is and Nicolas was putting up these, who’s a monopolist? You know, Apple was paying money. Google is paying money to Apple. Google is being accused of monopoly. There’s something not right there, right? You have to know that this is a really weird case. If the folks had the monopoly, it’s sort of paying money to hold on to things and and then you dig deeper and say, Why is it, why is it that Bing didn’t succeed, or hasn’t succeed? And when you look at it closely, you know that it’s a huge, multi billion dollar investment job every year. And if Microsoft is not willing to do that, the chances that they can succeed alone. So you have to deal with these tensions or paradoxes. To use Nicolas, when you run into these paradoxes, they shouldn’t just amuse you and intrigue you. You should say, is there something wrong with my underlying structure of knowledge? And in a place like Berkeley, that’s our job. And you know, we’ve let the professional world down to some extent by not doing our job well enough. But we’re now on to it, and with the EU together, we’re going to make up for the sins of the past. Thanks, David.

Nicolas Petit 43:15

Okay, so I am going to talk a little more. It’s always difficult to speak after David, because I want to address the next question, which is really, why didn’t we update until now? So why didn’t law schools, why didn’t lawmakers? Why didn’t law enforcement agencies didn’t update their views based on the experience and evidence that’s that’s kicking in. The paradox is that David mentioned the anomalies. The anomalies reveal, when they accumulate, they reveal, as David very eloquently put it, flaws in the structure of knowledge that’s underlying the application and and that’s actually not my question. It’s not David’s question. It’s a question that was raised recently by very famous economist, Jason Fairman. Was a Harvard macroeconomist, and he learned just like us that iRobot, a company doing very sophisticated vacuum cleaners, was going bankrupt couple of months ago, a year ago, two years ago, it had tried to sell to Amazon. Amazon wanted to buy iRobot, and the European Commission in Brussels decided that this was not a good a good acquisition, and the parties withdrew their proposed transaction. A year and a half later, iRobot is basically selling to a Chinese company for scraps. And there’s probably lots of technological progress that has been lost in this very unfortunate intervention. And so. I just like — maybe to advance, it’s important to understand why the field has not updated, if we want to be effective at trying to improve it, right? So that’s this is the reason and I think that there are probably reasons that have to do with a form of cultural problem, phenomenon in the field of interest and economic regulation, there is a cultural thing, and its drivers are sociological, institutional and probably ideological or political. Now I just, I just want to say, before I go deeper into that, that I’m not a conspiracy theorist. I believe in the power of governments, in policy making. I believe in collective decision making that can improve welfare. I do certainly believe in that, but I also believe in fact and science.

Speaker 45:51

So you have to say you’re not a conspiracy theorist. That’s true.

Nicolas Petit 45:59

Yeah, I’m not all right. So some facts. So in 2024, DG competition. DG comp is the European competition agency. DG comp publishes a paper which says that trends, economic trends, about rising economic concentration, rising markups in the economy, rising profits have moved in Europe in the same direction as in other regions of the world. So the US and Europe have almost experienced the same increase in market power. Firms are making more money. They are bigger. They’re making more profits. And yet, at the same time, the EU has spent 20 years telling us that we had the most effective and strong competition enforcement program. So when you put these two point data points together, it probably implies that so if this, if the market forces that shape firm size and economic profits are the same across the Atlantic, it implies that the current antitrust program of Europe is a very expensive way to achieve very little. Right? That’s the first thing. Now comes in the Draghi report. You might have heard about the Draghi reports, former Italian Prime Minister, governor of the central the European Central Bank, very famous, very you know, esteemed person across the policy world. The Draghi report diagnoses that Europe is suffering from a massive productivity gap driven by its lag in digital innovation and Draghi notes, I quote, a close link between the size of companies in Europe and technology adoption. The evidence is that European businesses are too small, and size enables technology creation and technology diffusion. When 30% of large EU businesses in Europe in 2023 have adopted AI, only 7% of small firms have adopted AI, there is a gap between firms of different sizes. So size enables adoption because it’s very obvious. It’s an old economic problem. You can spread the high fixed cost on a larger number of units, and that allows you to scale technology now that evidence, unfortunately, it’s like, you know, uncontrived evidence has fallen of this on different years. And I go here to maybe the first point, the sociological point, because it feels sometimes in competition policy that we’re facing a form of secular religion defended by a band of adepts of the cults who really are strongly resistant to any challenge to the mysteries of their beliefs. And so in that, in that superstition, size is a taboo. Size is bad. Big is bad. So that’s the first thing. The second thing is, there is, and it’s probably fair to say that there is indeed a specific institutional architecture that supports the superstition. So we have, and it’s a good thing, a sprawling network of hundreds of commission agencies in the world, experts that meet every year, like in DC last week, at the antitrust the ABA interest spring meeting, a great meeting in an environment. They meet in an environment that rewards activity above everything else. The KPI of comption agencies, how many cases they’ve done? It’s not where their commission has improved, and these agencies for a good reason, they want to do cases. This is the metric of success, and that creates a very profound moral hazard. So we have sometimes a legal community that has become dependent on these agencies. They have to give away some cases to make the system work. Their goal is not to fix the system, it’s to maintain access to the system. So we found ourselves in a standard sociological problem, the emergence of a very strong and deep epistemological community, a world where the enforcers and the practitioners speak to each other and maintain the system. You can you can think about the contrast with central banking economics. Central banks have no practitioners. They invest millions in independent academic research because they know that being wrong about the economy has immediate catastrophic consequences. In antitrust, that investment in research is missing. What we see is investments in what I call policy affirming evidence making, rather than evidence based policy making and the process is a process in which the desired policy act outcome dictates the evidence that you’re going to produce. That’s what happened with that report from BJI competition. All right, so the second driver that I want to talk about, that maybe is beyond this sort of sociological institutional points, is a latency loop. So I like to say to my students, I teach something called law and tech. It’s abstract level, and I talk about AI, I talk about nuclear power, I talk about digital platforms. Of course, I often tell them that we often complain that law is very slow in a context of technological change. But the good news, well, the bad news is that it’s much faster than economics. So law is slower than business and technology, but it’s comparably much faster than economic theory. And this creates a fundamental misalignment, because by the time a formal economic theory reaches the courtroom, the business model that it seeks to, that it could help address, has already been regulated or addressed by the legal structure. So because there is no real time economic theory to guide new legislation, we see a fairly dangerous pivot. We cannot wait for empirical verification. We cannot wait for experience, and we affirm legal principles without the benefits of economic evidence, we double down on what David was mentioning, all the economic intuitions that we cannot be sure are robust to the change in technology, in the technological structure, right, and and that faulty process. I want to say that because this is a topic that’s very close to my heart, and David has been mentioning that, that faulty knowledge transmission process has been aggravated by a 20 century trends that few people know of, and that trend is the divorce, the divorce between law schools and economics and business schools. So by separating these disciplines, we have limited the vital spillovers that could enable independent economic research to inform evidence based policymaking. So of course, today, commission agencies and regulate regulators in the world have their own in house economists, but but the incentives have changed. Unlike in an economics or Business and Management Science departments, we are no longer necessarily in the realm of objective inquiry with without a goal, non goal ended inquiry. We’re in the realm of a structure in which the economics are used to build cases and advance a policy agenda, right? So it’s not the I won’t. I don’t want to bore you with history, but the process that we’re seeing today is the same as what happened with the Sherman acts, right? Industrial Revolution, massive changes in transportation, mass production, mass finance, emergence of large agglomerations of power, in oil, steel and banking. Sherman Act adopted 1891 judgments of the Supreme Court trans Missouri decide that all agreements, all contracts between companies, are prohibited as unlawful, a very radical, restrictive, aggressive enforcement of the law, rigid reading of the text, the word every contract in the Sherman Act interpreted as outlawing all transactions between competitors. All it took a time for the law to move back to something more sensible, the Chicagoan revolution in the 1970s and post Chicago revolution in the 1980s helped improve the legal infrastructure through the consumer welfare standard. Now one more point, and that’s the point about politics, or maybe ideology. So all these ideas in the courts and and. In agency. In the agency world, they also feed on on a fascination with power structures that we see in law schools. So in the circles these days, a movement has become very popular and fashionable. The law, any political economy movement tells a very simple story: profits. And those will benefit from profits construct legal institutions that entrench power, increase inequality and social injustice. And the Neo-Brandeisian movement, the Lina Khan and Tim Wu kind of literature, is the anti trust version of that movement. It tells a huge make believe theory that profits concentration, which in turn eliminates small businesses and democracy. If you study the data on entry by small businesses in the European and the American economy these past 20 years, you’re going to see a huge increase, not a crushing of the small business economy. So you see this in the works of people like Tim Wu, the FTC and Lina Khan. If you look for economic evidence in these works, you find zippo, nothing zero. You find anecdotal reference to pieces in the New York Times and other other press outlets, you find a highly funky version of economic history that forgets about decades of policy making, the term Arnold tenure at the anti trust agencies, and and so it tells a very strange story about power that I think we should fight because the prescriptions from this diagnosis are radically different to what we should, I think, be doing. They want to return us back to the early days of the Sherman Acts, where every large business entity is prohibited from entering into contracts with others, structural presumptions against M and A transactions that allow small businesses to make a killing and incentivize people to create more and new businesses. So we want to, so economics has not been very effective at doing that. So what we want to do, and I’ll close here, what is we believe is missing, is, is a fine grained study of dynamic option and dynamic capabilities. I’m going to give two definitions here, because I think we’ve left you without maybe a good definition of these concepts. So what we call dynamic competition is basically competition through technology, right? And competition through technology determines business selection. Capabilities are the set of processes of knowledge, processes that reside within the business organization, that can be repeated and that can be described these capabilities, these dynamic capabilities for innovation, are the enablers of competitive advantage and survival. Both concepts occupy minimal space in antitrust policy, very little. And so with David, we believe that we need to enter into this process of epistemology — sorry, this is a big word. It’s very difficult. You heard. I’m not a native speaker — epistemological inquiry to build up on the foundations of price theory, transaction cost economist and game theory and modernize the tools. I also want to say something because maybe that ambition would be mistaken. This is not a zero enforcement program. It is not about saying anti trust enforcement and economic regulation should disappear and see the way to the free market. It is a program that seeks to study under which conditions it is appropriate to enforce anti trust regulation in the context of technological change. The focus is not on business size, and it is not on the market. It is on technological change, value creation and value capture. And we want to, we want to think about technological creation and technological diffusion as the first step in antitrust enforcement. All right, so I’ll close with just a reference to Foucault. I couldn’t come to Berkeley and not talk about Michel Foucault, especially because he’s French, just like me. And I want to connect what I said more broadly to something that also is slightly, I think, forgotten. So if you believe that political economy is a study of power structures emerging from capitalism, I think I advise you to re read Foucault, the Burke of bio politics. Political Economy, according to Foucault, is the study of Public Law. Political Economy emerged as a branch of law that used economics to evaluate the activities of the ruler of the sovereign and add checks upon public power. Before that, public law had mostly been about trying to legitimize the activities of the ruler. So given what we’re seeing today in US politics and outside as well, also in European politics, sometimes I think the advancement of executive power and encroachment and basic individual freedom and collective freedoms justifies a program just like ours that thinks a little bit about how to keep government in check, especially when the dangers and the adverse spillovers might limit innovation that drives prosperity grows and removes inequality in society in the long term, that’s what I wanted to say. I want to thank you for your attention.

Speaker 1:01:04

All right, well, we can take some questions if you have time. That’s good, yes, press releases, conclusionary slide for a second. Yeah. The bottom the last item the robustness of the innovation system around a digital platform. So most of the criticism that you somewhat disparagingly made of the Tim Wu School of anti trust, which is feels to me, more like a political exercise than strictly an economic exercise. It’s about the actions of the large platforms which no one can deny are dominating the economy, and he uses particularly Amazon as a good example, and the way in which those who were selling on the Amazon platform have been squeezed badly by Amazon using its power over that platform to extract more of the economics. So I was interested in this last comment, and how do you feel? Do you feel that the platforms themselves should be the focus, or you’re saying that, basically, innovation around the platforms may be of greater value than restricting the platforms themselves.

David Teece 1:02:24

Let’s look at this last point here. I mean, first of all, I’ve not thrown away — I’m not throwing away the consumer welfare criteria. Just said, make it long term. Because if you take the short term view, you pay no attention to innovation. So if you have the long term view, then at least you get innovation in but the proxy that I’m seeking is the lens that I think to evaluate this stuff without saying what the answer is, is, when you look at business conduct, such as what you put your finger on, what does it do to the robustness of the innovation ecosystem, because in the long run that came out that’s very hard to figure out. But that doesn’t mean we don’t stop trying to figure it out. I mean, really, what’s going on for decades? It goes back to that statement of Ed Mason, this stuff is hard to figure out. Well, that doesn’t mean you don’t try and you go to something, you know, it’s wrong. It means you start to try and figure it out. And and the significant work on innovation ecosystems in business schools now, we’re just starting to sort of come up with metrics for you know, what’s the health of an innovation ecosystem? But it’s the right question to ask, and it’s just shocking that we’ve gone for century, or at least half a century, without asking, in my view, the right questions. We’ve done it for expediency and because, yes, the law needs predictability, but you also have to change when you know it’s wrong, and you know, people still say this today. Well, I still want to use concentration ratios, because I know how to measure well, actually you don’t, if you don’t know where the boundaries of the market are. And one of the things that Nicolas and I find ourselves saying that I have the belief that I have because I also operate as a small time venture capitalist, and I built companies to billion dollars or so and more. As a CEO and founder, you worried not about people in the market that you can see. It’s the unseen competition that scares the living daylights out of you, and that’s potential competition. So one of the things in this anti trust law journal, paper of mine say, look, potential competition is more important than actual. But as Nicolas says, that doesn’t cut for giving a license to the incumbent at all. You know, it can cut both ways. A lot of people shied away from this in the agencies because they think, Oh, you’re just trying to create more freedom for big tech firms. But actually, my boldman, what really got me fired up about this was I worked on an antitrust case, the cool case way back in the 80s. And it was a company with revenues like a million, 100 million to 200 and then we get an antitrust cases thrown at them left and right, and you had anti trust economists jumping on, and this is where sale up comes up with his raising rivals cost, because Gore hired up a whole bunch of chemical engineers. Gore made Gore Tex, of course. They had this substance called PTAB, which was proprietary. They made their own proprietary resins, and they wouldn’t sell the residents to somebody else. Why should they? But you know, Steve Salove came up with this, well, you’re raising rivals cost because you’re not letting other people have your proprietary technology. This is absolute nonsensical stuff that can only come out of someone with a static view that doesn’t think about the innovation ecosystem. Because if you say, you know, it’s like taking away your patents. Worse than that, it’s taking away your trade secrets and your patents, because you’ve got to sell this somebody else, you know. How can you come up with that? It’s only if you don’t see innovation as the primary driver of competition. So it was the small company’s predicament, as much as it is the big company’s predicament. It’s the innovator’s predicament that we’re trying to protect, not the guys that sit on their laurels and don’t do anything, which is sort of what mainstream competition policy is indirectly ended up supporting. And in Europe, of course, you have the living dead, you have all these companies that are going nowhere because there’s no entrepreneurship and there’s no innovation. So we don’t want to end up where Europe is. We got to be where China is, and China is where we are, doing a better job at it than we are. They’re not — they don’t have a static view. They don’t have this efficiency nonsense. You know, not very much about innovation ecosystems and policies to support robust innovation ecosystems. So we don’t have to look very far to a real world case where it’s being done quite well. And antitrust sort of, it’s just another tool to achieve this bigger goal. And, you know, we look at it and say, Oh, it’s not very rigorous, it’s supple and agile, because it’s got this bigger goal of supporting national competitiveness, which is what our antitrust policy should be doing as well. In fact, right here in this Sherman Act conference, the headline of a reader’s paper was, antitrust is America’s industrial policy. Pretty pathetic one that it is indirectly. It does shape industrial outcomes. Problem is it’s not focused on what matters for great power arrival rates. So this is, this is not child play stuff anymore. This is the future of democracy, just as Tim Wu sees it as being the platforms, as being the problem. I actually see it is anything we, if we, if we don’t have the most robust innovation ecosystems, we lose, we lose everything. And so this is not just idle stuff that anti trust people worry about. This is something society has to be concerned about, and, and, and that involves rolling our sleeves up and understanding innovation and entrepreneurship in a way doesn’t exist as well as it must and law and in economics.

Audience 1:08:54

So how do we get people to come together? Just a short walk to the business school? That’s too far.

David Teece 1:09:01

Well, look, we do have law and economics programs, but they’re not intellectually linked. And they may be organizationally linked, but they need to be intellectually linked. And this is where we go. If we get dynamic competition, we can link it to technology policy and industrial policy. We do not have coherent policies in the United States, and if we have an innovation first approach that brings technology policy, industrial policy and competition policy together, instead of that, right now, they’re fighting each other, and this is nonsense. You can’t run a great country and support a great economy while policies are inconsistent. Why they’re inconsistent? I’ve been in meeting with prophetic antitrust economists. Don’t ever mention the word industrial policy, it’s like anathema, because why are they saying that? Because they don’t want their comfortable little silo of antitrust economics to be disruptive, because that hurt them professionally. They’re not saying it because it’s correct as a matter of science, but it’s exactly what Thomas Kuhn was saying. Is that we take these professional positions for our own comfort, and that’s why you need robust intellectuals such as what Berkeley produces. This is our greatest — we are the greatest place on the planet for intellectual diversity, and it’s time that we harnessed it in the law school, in the business school, in the economics department, to help with national competitive industrial policy and technology policy, all the great issues of our day that determine the future of our children. And that’s why this is important. That’s why I could work on this for 30 years with having no followers, because I know it’s what’s critical for democracy.

Audience 1:10:48

Thank you for the lecture. I really appreciate it. And thank you for mentioning focus theory. And I have a question it’s like right now for not only the AGI itself, but also like the you know, the algorithmic collusion itself is invisible. Like, not only the market itself is dynamic, but the product, like the AGI itself is dynamic. So here is a problem that, how can we redefine the agreement when we use a traditional anti monopoly law to refine or to polish? Like, what do you mean by algorithmic collusion?

David Teece 1:11:22

I haven’t thought about that one, but what you — I think what you’re saying is, when you have algorithms that are connected and communicating, does that constitute collusion? And you know, traditional contract law and competition policy view, is there an offer and there is an acceptance. And of course, some extent in these algorithms, it happens in real time, right? I don’t know. I haven’t — I’ve been focused, actually, my early work was on collusion and cooperation. And the other thing that annoyed me back then, we fixed this one. We did actually have some impact, is if you looked in the economics textbook for the word cooperation, you only got collusion. There was no room for beneficial cooperation amongst competitors. But of course, in the technology space, that happens all the time, you know, Microsoft has to do deals with Google and so forth. So I don’t know is the answer and but the answer, the way I’d look at the problem, the way I’d frame it, is to say, does that form of collusion hurt innovation or help it? Does algorithmic collusion help innovation or hurt? If it hurts it, and I would say, we should block it if it helps it. Then yeah, so I frame the I don’t think, you know, framing it in terms of, is it collusion or not? Is a wrong? It’s only half the question. The real question is okay, whether it is or whether it isn’t, what’s the impact on innovation, beneficial or negative?

Speaker 1:13:02

And if you use the word collusion, you already sound bad, right? So then you have to say, what is it really?

Audience 1:13:07

I have a question on the same note, I work in the semiconductor industry. So business innovation, the pace of business innovation, is a lot faster than anti trust actions. So the most current example that I have question on this. For example, Nvidia. Nvidia, a few years ago, they tried to buy arm, but it was not approved by by the government, because of anti trust concern. But recently they they acquire. They didn’t acquire, but they got license from their one of their competitor from inference, AI inference, from Grok. They buy their licensing, and then they, they recruited top 20 of their employees, and they now work for Nvidia. And they basically going back to your definition of whether this helps innovation or hurts innovation, it helps innovation for Nvidia itself, but it doesn’t help the innovation for people outside of Nvidia. So in [unable to transcribe].

Audience 1:14:19

So my question is whether, what is the definition of antitrust in this cas? What is the definition of it? Is it? Is it considered antitrust action by bypassing the current rule?

David Teece 1:14:39

Well, of course, there’s been lift out hires going on forever, often as trade secret issues, as much as there are antitrust issues involved in that. But when you tighten up on antitrust, on M and A activity, it’s sort of a natural consequence. And. So I suppose, look, I come back to the same metric. You have to look at the bout for world, if those employees stayed where they were, is competition more robust than if they move across to or doesn’t it make any difference? I mean, if it’s a small group, it probably doesn’t make a difference, and it’s probably not something that warrant trust inquired. But if it’s a big group, and really puts a company on a different trajectory, then it’s an anti trust issue, which way it cuts. I don’t know. You’d have to look at it. This is where you have to be forward looking. Mergers is the only place where you have to be forward looking, because you are somewhat predictive, you know, looking at competitive effects on a go forward basis.

Audience 1:15:45

How do we find faults in their actions? If they’re not colluding, they’re doing everything within the rule. But let’s say the enhances Nvidia’s competition.

David Teece 1:15:57

But –uy I’m not sure I can answer your question directly, but I’ll try and answer it indirectly, because people sort of bring up this question, how can the authorities, how can the agencies figure out these forward looking effects, and because sometimes executives and commentary and venture capitalists have trouble figuring it, well, actually the government has an advantage. The government can actually do civil investigations and get information from other parties in the ecosystem, which private parties can’t do. So I no longer give an excuse to people in government when they say we can’t do this. Well, other people are doing it every day for a living. That’s true, but you have a special advantage, because you can do civil investigations and get data and information that other people don’t have. So, you know, that’s sort of my general approach, process wise, is dig deeper and learn something. I mean, Nicholas did the great favor to me, because I didn’t believe agencies could be this dumb. But remember the Giphy Facebook merger and the UK authorities blocked it because they said little Giphy could be a competitor at Microsoft. I can tell you, even down to the bottom quartile in my business school class, they’re going to say this is ridiculous, but it made it through the competition authorities, and Nicolas brought the senior guy into the classroom, and he was bold enough to do it, and I got a chance to ask him. I said, you know, when there’s no business model that’s viable for Giphy. It’s run out of money, doesn’t have any revenues. How is this company going to succeed? To which his answer with a straight face, well, it’s an established company. It’s been around for eight years. Yeah, it’s an established company without any of the key ingredients that are needed to compete, and it’s about to go out of business and Facebook is sort of doing a rescue buy out here. But he said this with a straight face, which to me, says we have failed in the academic world to educate people in the agencies about the very basic ABCs of business. I mean, you know, this is where the economics department is not talking to the business school, because any business school student would have at least raised their eyebrows about this, but the fact that, with a straight face, a well trained master’s level, graduate levels, very nice person from the UK competition authority could say it was an established company, therefore it had a chance, based purely on time, it’s about to be what you know, Nicolas calls the living dead. Well, it’s probably actually the dead dead. But the notion that it can compete with anybody, let alone a major platform, is just completely nuts. But this shows you that there’s something wrong with the paradigm. It’s back to Nicolas’s starting point when you see these anomalies, and this is what Thomas Kuhn pointed out when these anomalies you know show up in what appears to be paradoxes between what you think you know and what the paradigm is telling you is the case and what you observe when you start that littering the field, it’s time to question the paradigm.

Nicolas Petit 1:19:18

Yeah, I think so. I have a take on your question. So if your question is, if Nvidia create this transaction with Grok, and it’s this sort of weird, M and A, it’s like, we’re going to take some people in, then license some technology, and it’s not a merger, right? And is that a problem, right? And we have two there are, I think there are two sub questions, and there are two sub answers to your question. The first one is, is the law today — can the law that we have claim jurisdiction over these transactions, these sort of funky M and Acircular deals that we see in the in the AI space. And I think the answer has to be an affirmative yes, that anti trust law has been over history, able to catch transaction that were slightly off what the statute said. And the best example is the Sherman Acts. At the time of the shermanite in 1890 what happened is that these railroads companies, the steel company, the aluminum companies, were entering into all these sort of weird transactions that didn’t make sense. Anti trust, using the Trust has the form of organization for the industry and so and so the law has proven very flexible across time to actually catch these things, right? So, in a merger case, you try to establish control, right? And the the agencies would look into this and see whether there’s an element of control. So that’s the first sub question and sub answer. So I think that is not such a problem for interest law to claim review of these issues, right? And they’ve done so. I mean, the in the UK, there have been several investigations over these AI partnerships. The more difficult question is the economy question, which is this transaction improving welfare by the question we would ask is creating a net innovation benefits, right? And that’s where the power of what David wrote, and others like [unable to transcribe] other colleagues is so good. Their work on capabilities tell you that when, when Firm A and Firm B merge, this is a capability combination that might have more power than if Firm B and firm C merge, right? And so you can study the alignment in capability and the net innovation benefit that is going to come out of one capability combination versus the other, and it’s not as predictive as what commission economies do when they try to study the price effects that come out of a merger, or the output effect that come out of a merger with a model. But these models are very predictive, but they are very fragile as well, right? So, so the standard of the standard of admission of the towards capability analysis, and we’re trying to build this right, needs to be a little tolerant. To some form of uncertainty about the degree of prediction you’re going to have, just like we do with merger simulation models, right? So, but so we think we have some grip on that question, and there’s grip in the capabilities literature to study these kinds of effects. I hope that answer your question.

Prof. Nicolas Petit
European University Institute

Nicolas Petit is Professor of Competition Law at the Department of Law of the European University Institute (EUI). His research in recent years has focused on EU and US competition law, competition and innovation, law and economics, and law in a context of technological change. Nicolas Petit uses mixed approaches including law and economics, doctrinal and empirical legal methods.Nicolas Petit’s research focuses on a range of topics, including the application of economic analysis to competition law, the intersection between competition law and intellectual property, and the regulation of digital markets. He is known for his contributions to the development of the “more economic approach” to competition law and policy, which emphasizes the use of economics and empirical evidence in antitrust analysis.

Quicklinks

Support

Quicklinks

For Students

How the Dynamic Competition Paradigm Challenges Antitrust Law’s Static Framework in the Digital Economy

Legal Analysis

United States v. Google and the Missing Gatekeeper: How Static Market Analysis Distorts Digital Antitrust Cases

Market Definition as Gerrymandering: FTC v. Meta, the Digital Markets Act, and Enforcement That Harms Consumers

Dynamic Capabilities, Schumpeterian Competition, and the Paradigm Shift Antitrust Economics Has Yet to Complete

Key Takeaways

Interview/Transcript

Prof. Nicolas Petit
European University Institute

Quicklinks

Support

Quicklinks

For Students

Legal Analysis

United States v. Google and the Missing Gatekeeper: How Static Market Analysis Distorts Digital Antitrust Cases

Market Definition as Gerrymandering: FTC v. Meta, the Digital Markets Act, and Enforcement That Harms Consumers

Dynamic Capabilities, Schumpeterian Competition, and the Paradigm Shift Antitrust Economics Has Yet to Complete

Key Takeaways

Interview/Transcript

Prof. Nicolas Petit European University Institute

Prof. Nicolas Petit
European University Institute