Over the past four years, antitrust scrutiny has increasingly focused on large technology firms. Ginger Zhe Jin and Liad Wagman discuss the complexities of antitrust enforcement and policy in the digital age, highlighting the challenges of promoting innovation while fostering competition, and areas where consumer protection and antitrust are colliding or are set to collide. To that end, the authors identify several key questions that the next administration of the United States should address to better delineate between legal and illegal competitive practices in the digital age, with implications for the broader economy.
Editor’s note: This article is part of a symposium that asks how the next presidential administration and its antitrust agencies should reorient competition policy for the next four years. Contributions from Herbert Hovenkamp, John Kwoka, Steven Salop, and Ginger Zhe Jin and Liad Wagman can be read here as they are published in the coming days.
During the Biden administration, a number of antitrust cases and policies emerged, either targeting large technology firms or broadly relating to the information and communication technologies (ICT) sector. We identify four crucial questions that the next administration should address to draw a clear distinction between legal and illegal competitive practices. While our primary focus is on the ICT (henceforth, tech) sector, the questions we raise hold broad significance for antitrust enforcement and competition policy, as tech firms develop, provide, and apply technological advancements throughout the economy.
Recent US Antitrust Actions in Tech
Since the Biden administration, and to some extent earlier under the Trump administration, the American antitrust agencies have intensified their scrutiny of business models and merger and acquisition (M&A) activities within tech, particularly when they involve major digital platforms such as Google/Alphabet, Apple, Facebook/Meta, Amazon, and Microsoft (collectively also known as GAFAM).
For example, a 2021 Federal Trade Commission (FTC) study analyzed the acquisitions that GAFAM firms conducted between 2010 and 2019 and were exempted from reporting to the antitrust agencies under the Hart-Scott-Rodino (HSR) Act. The FTC subsequently attempted to block Meta’s acquisition of Within Unlimited and Microsoft’s acquisition of Activision Blizzard. Although the FTC’s attempts were unsuccessful—both mergers were consummated after judges ruled against the FTC—a number of other M&A deals in tech, including Nvidia’s attempt to acquire Arm, Amazon’s attempt to acquire iRobot, and IQVIA Holding’s attempt to acquire Propel Media, were abandoned by the merging parties after antitrust agencies raised competition concerns.
On the policy front, the United States Department of Justice (DOJ) and the FTC finalized the 2023 Merger Guidelines, with at least two guidelines having significant overlap with tech (guidelines #4 on nascent competition and #8 on serial acquisitions), and proposed sweeping changes to the HSR reporting requirements for all sectors.
Beyond merger review and policy, the DOJ sued Google in 2020 for monopolizing the search engine and search advertising markets, and in 2023 for monopolizing digital advertising technologies. The FTC sued Facebook/Meta in 2020 for its acquisitions of Instagram in 2012 and WhatsApp in 2014, alleging that these acquisitions left consumers with significantly fewer choices for personal social networking, and deprived advertisers of the benefits of competition. In 2023, the FTC sued Amazon for illegally maintaining monopoly power with alleged harms of inflated prices, degraded quality, and stifled innovation for consumers and businesses.
The DOJ and FTC under the Biden administration did much to highlight the tension between large digital platforms and smaller entities that use, compete, partner, or complement them. In doing so, they motivated legislators, the courts, and the general public to reflect on antitrust practices in the digital age. However, they have also left open several important questions regarding tech regulation that the next administration should clarify. We highlight four of these key questions.
Question 1: When is big bad?
Digital and computing technologies enable considerable economies of scale and scope. Take online search engines as an example. According to the judge’s ruling in the DOJ’s recent search engine case against Google, constructing a general search engine is an extremely capital- and human-resource intensive endeavor, as developing the technical infrastructure alone requires billions of dollars. In the case, it was highlighted that Google incurred $8.4 billion in 2020 to operate its search engine and another $11.1 billion to run search ads; Microsoft invested nearly $100 billion in search over the past two decades; and the capital expenditure required for Apple to reproduce Google’s technical infrastructure for search was estimated at $20 billion. Building upon its investments in search engine and search advertising technologies, Google has subsequently expanded into cloud computing, video streaming, mobile operating systems and browsers, mapping technologies, hardware, AI and machine learning tools, and self-driving technologies, among others.
Economies of scale and scope break down market barriers, leading to unprecedented increases in the size of top firms and in the concentration of sales in a fixed geography, language, or set of products. However, as summarized by Carl Shapiro and Ali Yurukoglu (2024), these increases in size and concentration do not necessarily imply a reduction in market competition. For example, two nationwide retailers competing in each metropolitan statistical area (MSA) could be more competitive than a scenario where each MSA has a separate local monopolist, even though the number of retailers across the entire U.S. may decline from 300+ to two and the nationwide sales share of each retailer increases dramatically. Consequently, traditional proxies such as firm size and sector-wide concentration can be misleading indicators of future antitrust enforcement and policy.
In light of this, a key question is how to distinguish whether increased scale and concentration is due to anticompetitive behavior or economic efficiency. This is a perennial debate in the American antitrust community between “big is bad” and “big is efficient,” with extensive legal and economic arguments for both. To reach some consensus in the context of today’s technologies, antitrust agencies must acknowledge the tech-enabled economies of scale and scope in the first place, and recognize that their nature and extent may go beyond what the courts have seen in the “traditional” economy.
That being said, technology-enabled economies of scale and scope have the potential to enable new business strategies that harm market competition. As the judge ruled in the Google Search case, it is possible that “big is bad” because “big is efficient.” Smaller search engines may not be able to persuade downstream original equipment manufacturers (OEMs) and internet service providers (ISPs) to adopt the same default setup and exclusive contracting as Google did, precisely because their search engines may not be as popular as Google is among billions of users. An ongoing challenge is how to deter the anticompetitive actions that build upon bigness in tech, while still allowing the market to reap the efficiencies from the considerable economies of scale and scope associated such technologies.
A related challenge is whether and how to recognize efficiencies across markets. Many multi-sided platforms derive network efficiency from interactions among their different sides. If the relevant market in an antitrust analysis does not include all sides, potential remedies to anticompetitive behaviors risk reducing the economic efficiency outside the market under consideration. All of this further depends on the market definition that is recognized by the courts, as well as how the courts and the antitrust agencies acknowledge the connections among the related markets.
Question 2: When and how does competition promote innovation?
In traditional antitrust cases, consumer harm is evaluated by the extent to which a challenged practice has led to higher prices, reduced quality, or both, relative to a hypothetical scenario without the practice. However, due to technological advancements, many products and services offered by digital platforms (such as general search, social media posting, and social media consumption) are provided to end-users at no monetary cost. Consequently, antitrust agencies often frame anticompetitive harm in tech cases in terms of potential future innovation: perhaps the current difficulty of competing with an established monopolist has deterred future innovators from entering the relevant market, or perhaps a more competitive market would lead to superior products and lower prices. While these speculations are intuitive and well-intended, they are not always consistent with the academic literature. A large economic literature has shown that innovation does not always increase with market competition. For instance, Edward Chamberlin (1962) points out that profits earned by successful innovators in imperfectly competitive markets can provide incentives to engage in R&D, implying that innovation may decrease with competition if competition dissipates the expected returns from innovation.
However, the market may under-innovate because inventors may not appropriate all of the social returns from their inventions due to competition or regulation, or over-innovate if highly innovative firms can reap large market shares from competing firms. To the extent that large firms in a concentrated market can better appropriate the value of their innovation, they may have more incentives to innovate; but they may also be reluctant to innovate out of the concerns that new products may cannibalize or replace their existing products. Because of these countervailing forces, empirical researchers have documented an inverted U-shaped relationship between product market competition and innovation.
The correlation between firm size and innovation is also complicated. Ufuk Akcigit and William Kerr (2018) show that total R&D expenditures and patent counts tend to increase with firm size, but R&D expenditure or patent per employee is higher in smaller innovative firms. Conditional on being innovative, the researchers find that smaller firms are more likely to explore new areas with major innovations rather than exploit existing areas. Recent research by Jane Olmstead-Rumsey (2022) confirms that, on average, smaller innovative firms have higher relative quality of patents than market leaders, but their innovations have become more incremental since 2000.
In short, the total volume and quality of innovation are not always greater, faster, or better if they come from small startups as opposed to large incumbents, nor does innovation automatically increase when a market consists of a large number of small firms rather than few large firms. After all, significant innovations often require considerable R&D expenditures in order to reap economies of scale and scope. In a 2020 report by the Boston Consulting Group, Apple, Alphabet, Amazon, Microsoft, and Meta ranked #1, #2, #3, #4 and #10 among the 50 most innovative firms worldwide.
Even if small startups possess greater agility and innovation capacities, their ability to execute such innovation is contingent upon adequate funding and/or the anticipation of sufficient returns. Early investments from large incumbents can help to bolster startup funding. More importantly, an ecosystem dynamic exists where innovation thrives within smaller firms because larger firms perceive it to be more advantageous to acquire innovations rather than pursue them internally. This, in turn, incentivizes smaller firms to innovate with the hope of an exit, including the option of being acquired. Shifting to an alternative environment that diminishes the market share of “big” firms would necessitate substantial changes in the ecosystem of startup investment.
These complexities suggest that promoting increased competition within a market does not necessarily guarantee greater or improved future innovation. If reduced innovation constitutes a significant aspect of the alleged anticompetitive harm, it becomes imperative to construct a counterfactual scenario where the challenged practice and the proposed remedy affect not only the likelihood of future market entry but also the prospects of future innovation. This is considerably more challenging than simply asserting that the challenged practice has discouraged entry.
Consider the Google Search case as an example. Suppose Microsoft could have achieved a larger market share in general online search had Google not secured agreements with Apple, Android OEMs, and ISPs to use Google as the default search engine. This may have incentivized Microsoft to invest more heavily in traditional search technologies (collecting, indexing, and ranking websites) and less in disruptive technologies like Large Language Models (LLMs). This, potentially, could have led to slower progress in OpenAI and similar startups, resulting in less competition among LLM providers than we currently observe among Google, OpenAI, Microsoft, Amazon, and Meta. Would this counterfactual scenario have resulted in more or less innovation compared to the current state? More broadly, how can antitrust agencies and the courts ensure that the remedied counterfactual world yields more and superior innovation than the status quo?
Question 3: Is the consumer always right?
The academic literature demonstrates many behavioral biases in human decision-making. For example, many consumers may not bother to consider alternative choices to the default (e.g., a browser’s default search engine), even if those alternatives are only one click away; most consumers readily click on “I consent” without reading privacy notices, even if those same consumers express significant privacy concerns in surveys; some social media users are more reactive to sensational postings even if they know such postings are more likely to exaggerate or mislead; and business managers may not bother to do their own extensive market analysis if a black box algorithm is readily available. As has been evident in a number of consumer protection cases, these behavioral biases may enable new business models to emerge, which leverage users’ limited attention and bounded rationality for profit.
Not only do these considerations blur the boundary of antitrust and consumer protection, but they also challenge a core principle of competition policy: Should antitrust authorities take consumers’ behavioral biases as part of consumer preferences and only consider a firm’s behavior anticompetitive if it generates harm in this bias-driven world? Or should antitrust authorities encourage or even mandate firm behavior that mitigates or fixes such biases? In other words, should the counterfactual world that has sufficient market competition be (i) a world where firms compete for or despite human biases, or (ii) a world where firms are required to address these biases, or possibly compete to address them?
This question is of immediate relevance in the potential remedy for the Google Search case. Under option (i), the remedy can be a ban on the exclusive terms of the default contract, but it still allows (or even encourages) firms to contract on the default setting. Then the question is whether Google is still allowed to bid, and, if the answer is no, whether that will harm market competition for bidding for default status. Under option (ii), the remedy may mandate consumers to make an explicit choice of a search engine, even if consumers with default biases and limited attention would rather not make such choices at all. Whether such a framework hurts some consumers, at least in the short run, is a relevant question.
The blurry line between consumer protection and antitrust is further manifested in the FTC’s recent activities relating to “Dark Patterns.” The examples that the 2022 FTC staff report mentions as “dark patterns” include disguising ads to look like independent content, making it difficult for consumers to cancel subscriptions or charges, burying key terms or junk fees, and tricking consumers into sharing their data . The use of negative options (e.g., automatic renewal unless the user opts out) could be another example, depending on how it is marketed to consumers. Recent research by Liran Einav, Benjamin Klopack, and Neale Mahoney (2023) finds that many consumers do not cancel their unwanted auto-renewed subscriptions due to inattention, and, relative to a counterfactual in which consumers are fully attentive, inattention could raise seller revenues by a significant percentage, between 14% and upwards of 200%. Because of such revenue benefits, it is conceivable that market competition would exacerbate the use of dark patterns, which raises the question of whether antitrust enforcement should accommodate or mitigate consumer behavioral biases.
Arguably, an even more complicated trade off arises at the intersection of privacy and competition. For example, in the name of increasing user privacy (where some of the current news coverage lies), Apple and Alphabet are creating situations where their own apps have access to users’ private data but other apps may face barriers to access the same data. For example, Apple Health may have fewer barriers to access health data than Peloton, and Apple Cash, iMessage and Apple TV may have fewer barriers to access users’ contacts data than PayPal/Venmo, WhatsApp and Netflix. This is leading to a collision between antitrust and consumer protection, and any proposed remedy depends on whether consumers’ behavioral biases are viewed as part of consumer preferences that market competition should aim to satisfy, or as a problem that market competition should help address. Under either scenario, it is problematic that one competitor sets market rules that disadvantage other competitors.
Question 4: What is the Consumer Welfare Standard, exactly?
There is a pressing need to clarify the welfare standard for antitrust enforcement and policy: precisely who constitutes the “consumers” within the “consumer welfare standard”? Does this encompass upstream or downstream firms, as well as employees or contractors providing labor to these entities? In a multi-sided market, should the welfare standard prioritize competition on a particular side, or consider the welfare of all market participants? How to weigh and balance the welfare of various parties within the same “market,” especially if a business practice advantages consumers on one side but disadvantages suppliers on the other, or benefits certain users but harms others within the same side? While the Supreme Court’s decision in Ohio v. American Express touched on these issues, it is far from providing a comprehensive resolution.
The current ambiguity creates inconsistencies in how antitrust cases are evaluated and decided. Different courts and agencies may apply varying interpretations of the consumer welfare standard, leading to unpredictable outcomes and legal uncertainties for businesses. Moreover, this lack of clarity can be exploited by firms engaging in anticompetitive behavior. By selectively focusing on benefits to a narrow group of consumers, or benefits to users on a particular market side, or benefits to users in one quality dimension like privacy, they may attempt to justify practices that harm other market participants or competition overall.
A Call for More Economic Analysis
To answer the above four questions, economic analysis should assume a more prominent role in antitrust cases and policy formulation even though the views of certain antitrust policymakers and the implications of the 2023 Merger Guidelines have suggested the opposite. The call for economic analysis is built upon two key reasons:
First, the increasing complexity of platform business models, often involving multiple sides and intertwining with consumer behavioral biases, renders intuition and traditional business practices from decades ago out of date. As the digital economy increasingly enables data-driven decisions, antitrust agencies must modernize their ability to collect and analyze data, and understand the economic forces and market dynamics that underpin the incentives and biases of firms and users in the marketplace. This calls for a systematic development of economic theories and data analyses—not shying away from them.
Second, to construct a counterfactual world that has sufficient competition, one must incorporate contemporary technological advancements rather than rely on decades-old technologies. Such counterfactual analyses inherently demand rigorous economic scrutiny, and its relegation to the sidelines is bound to lead to suboptimal decisions.
Authors’ Disclosures: Some content of this article was presented at the 2024 Technology Policy Institute (TPI) Aspen Forum, when Jin participated in the panel “Antitrust: The Next Four Years.” Jin and Wagman worked full-time at the U.S. Federal Trade Commission in 2015-2017 and 2020-2022, respectively. Wagman is an academic affiliate at the International Center for Law & Economics (ICLE). Jin was on academic leave to work full-time at Amazon from 2019 to 2020. Both Jin and Wagman have provided consulting services to a few companies covered by this article. All opinions and errors are their own.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
