By Joe Kennedy
Progressive Policy Institute
Innovation Frontier Project

INTRODUCTION

On October 6, 2020 the majority staff of the Subcommittee on Antitrust, Commercial and Administrative Law of the U.S. House of Representatives Committee on the Judiciary issued a 400-page report on competition in digital markets.[1] It was the culmination of a 16-month investigation involving seven hearings, hundreds of interviews, and literally millions of pages of evidence. The long-awaited report received extensive media coverage, and criticism, after its release, but much of it was superficial and based on initial impressions. Since the subcommittee report is widely seen as a precursor to legislation in the 117^th Congress, this review offers policy makers a comprehensive and detailed analysis of its strengths and weaknesses.

The subcommittee had an agenda, and its report should be seen as part of a broader political push to use the nation’s antitrust laws to curb the power of large technology companies. Critics of “Big Tech” run the gamut from right-wing populists, who see social media platforms as hostile to their worldview, to the progressive left, which claims these companies represent a danger to competition, innovation, local communities, and even democracy.[2] The report itself was produced by staff for the Democratic majority on the Committee, and no Republicans signed onto it. Ken Buck, a Republican member from Colorado, issued his own “Third Way” report that agreed with a few of the recommendations in the majority staff report, but did not endorse it in total.

Animated by such allegations, antitrust enforcement has recently become more aggressive, especially targeting large internet companies. Earlier this year the U.S. Department of Justice filed an antitrust suit against Google alleging the company had illegally maintained its monopoly power by signing agreements with distribution platforms, such as web browsers and operating systems, to make Google the default search provider. This was followed by a similar suit by most of the nation’s state attorneys general. The Federal Trade Commission and numerous attorneys general filed complaints against Facebook alleging the company had substantially reduced competition in the social networking market by acquiring Instagram and WhatsApp and that it had harmed consumer privacy with its data practices.

The House subcommittee report likewise focuses exclusively on Amazon, Apple, Facebook and Google. It alleges that the four companies have engaged in a number of anticompetitive actions, which have harmed consumers, competitors, and innovation. These conclusions came as little surprise, since they echoed previous statements and writings by both the subcommittee’s chairman David Cicilline (D-RI) and prominent subcommittee staffers.

On the positive side, the report gives readers a detailed if one-sided view of how the Big Tech firms operate and it offers several constructive suggestions for updating U.S. antitrust laws. It also makes a strong case for more frequent use of retrospective merger analysis. A great deal of merger enforcement requires agencies to make assumptions about the future. Comparing these assumptions with actual outcomes would likely improve future decisions.

The report also calls for more Congressional oversight of tech competition and regulatory transparency, including a requirement that antitrust regulators at the Department of Justice and the Federal Trade Commission offer written justifications for their actions. Importantly, the authors also favor providing both the Department of Justice and the Federal Trade Commission with additional resources to perform these duties.

Unfortunately, however, the subcommittee report suffers from three major flaws that make it an unreliable guide to legislation or regulation.

First, the subcommittee report is marred by many factual errors and inaccurate assumptions, which are detailed below.

Second, the report is largely silent on why these platforms are so popular with U.S. and indeed global consumers. Clearly, the palpable benefits they provide need to be weighed against the alleged harms from market power or anticompetitive actions. To the extent that the report does acknowledge such benefits as lower prices or better services, it complains that they create unfair competitive barriers to new companies trying to enter the market.

Third, it makes an unconvincing case that its proposed reforms would produce better outcomes for society. For example, its call to ban all mergers by dominant companies would likely deter venture capital investment and thus slow innovation, while prohibiting platform owners from competing with third parties would needlessly deprive consumers of choice.

Ultimately, the report fails to achieve its avowed goal of providing “a comprehensive understanding of the state of competition in the online marketplace” (p. 10). Instead, the subcommittee has produced a tendentious and empirically thin account that does not offer the targeted companies a chance to respond to its assertions of abuse.

THE NEED FOR STRONG ANTITRUST ENFORCEMENT

Large tech companies certainly are not immune to the temptations of anticompetitive behavior. In 2010, the DOJ’s Antitrust Division filed a complaint against several tech companies, alleging that they had mutually agreed to refrain from “cold calling” each other’s employees, thereby reducing competition and wages. These allegations also led to a class action civil suit. As a result of these actions the companies agreed to refrain from similar actions in the future and entered into settlements totaling $435 million.

In 2012, the Justice Department filed a case against Apple and five book publishing companies alleging that they conspired to raise the price of e-books. The publishers complained that Amazon was selling e-books for its Kindle reader at very low prices. The DoJ alleged that they struck an agreement with Apple to illegally collude to charge higher prices for e-books sold to users of Apple’s new iPad as well as Amazon’s Kindle. The publishers settled with the government. Apple challenged the government’s case in court, eventually losing and paying $450 million to settle the case.

The outcome of both these cases suggests that more vigorous enforcement of current antitrust laws, rather than a slew of new laws and rules, could be sufficient to handle clear cases of abuse. Some critics of the current antitrust regime have argued that the government should have fined the companies more and that the civil settlements should have been larger. But these objections go to the application of current laws rather than the need for new ones.

The subcommittee report is not wrong in arguing for diligent antitrust enforcement. Even experts who don’t think new antitrust laws are necessary agree that the laws we have could be better enforced.[3] The report describes several practices that, if verified, could justify enforcement action. These include agreements between companies to coordinate on limiting competition and responding to lawsuits, manipulating platform results to favor their own products over those of competitors in a way that degrades the experience of users, and providing misleading information to government agencies.

DUBIOUS PREMISES AND FACTUAL ERRORS

While the report included some valid claims about the need for more vigorous antitrust enforcement, its core argument is much more extreme:

Although these four corporations differ in important ways, studying their business practices has revealed common problems. First, each platform now serves as a gatekeeper over a key channel of distribution. By controlling access to markets, these giants can pick winners and losers throughout our economy. They not only wield tremendous power, but they also abuse it by charging exorbitant fees, imposing oppressive contract terms, and extracting valuable data from the people and businesses that rely on them. Second, each platform uses its gatekeeper position to maintain its market power. By controlling the infrastructure of the digital age, they have surveilled other businesses to identify potential rivals, and have ultimately bought out, copied, or cut off their competitive threats. And, finally, these firms have abused their role as intermediaries to further entrench and expand their dominance. Whether through self-preferencing, predatory pricing, or exclusionary conduct, the dominant platforms have exploited their power in order to become even more dominant. (p. 6)

Given the size and complexity of each firm, as well as the differences between them, one would expect these conclusions to rest on a strong body of factual evidence. Indeed, the reliance on empirical evidence showing harm to consumers traditionally has been the bedrock of antitrust policy. That is not the case here.

One conceptual problem with the report is that it narrowly defines the markets these companies compete in. These are large companies and if the market they do business in is defined narrowly in terms of product and geography, one would expect to see them raise prices and reduce output in order to increase profits. As it happens, however, they do not. Instead, what is happening in many markets for digital services is a rapid decline in prices matched with a fierce attempt to attract more users, the opposite of what we expect from oligopolies. On the other hand, if their markets are defined more broadly – and accurately – it becomes clear that tech companies face greater competition and have less scope for behavior that harms consumers.

Of course, defining markets correctly isn’t always easy. For example, consider the market for apps. iPhone users have few options to download apps other than from its App Store; they cannot take advantage of Android apps. From this, the subcommittee leaps to the conclusion that competition for apps is lacking. That ignores the fact that purchasers of new phones undoubtedly take the availability and security of specific apps they want into consideration when deciding between Apple and Android phones. Similarly, many consumers who want to purchase a book and have it delivered to their home the next day probably go to Amazon. However, for most purchases, consumers also consider other choices, including alternative websites and brick-and-mortar stores. The latest ecommerce figures show that brick-and-mortar retail still accounts for 86% of total sales — even during the pandemic.[4]

Or take the case of advertising markets. Google and Facebook don’t compete directly in search or social networking, but they do compete for online advertising dollars. Although advertisers increasingly feel a need to be online and smaller companies have fewer alternatives, the largest companies also purchase advertising in other media, including television, radio, and print. They use high-priced consultants and sophisticated software to measure the return from different media and negotiate the best prices. And while Google and Facebook offer users very different services, they do compete fiercely for users’ time.

The report is also riddled with factual errors and displays a sloppy approach to sourcing. For instance, in emphasizing the dangers of Big Tech’s market power, the authors warn that: “Just a decade into the future 30% of the world’s gross economic output may lie with these firms, and just a handful of others” (p. 11). This misstates a McKinsey report that made this prediction: “If digital distribution (combining B2B and B2C commerce) represents about one-half of the nonproduction portion of the global economy by that time, the revenues that could theoretically, be redistributed across traditional sectoral borders in 2025 would exceed $60 trillion – about 30 percent of the world revenue pools that year.”[5] The McKinsey report is clearly referring here to ecommerce within the entire global economy, not just the four tech companies. To put the report’s erroneous forecast in perspective, last year the combined annual revenue of the four Big Tech firms was about half a percent of global economic output.[6]

More generally, the report contains 20 references to data in Statista. Like Wikipedia, Statista is a useful aggregator of statistics from a variety of sources. But the ultimate source of its data is available only to subscribers, not to those using its free service. It’s important to know where Statista is getting its data (and it would be reassuring to know the authors of the report do too).

Other assertions seem calculated to exaggerate the companies’ market dominance. After acknowledging that publicly available third-party estimates show Amazon controls about 40 percent of online retail sales, the report simply asserts, without citations, that “this market share is likely understated, and estimates of about 50% or higher are more credible.” (p. 15). It later notes, “The company is consistently one of the highest-priced stocks on Wall Street, which is a clear indication investors expect Amazon to maintain and expand its market power” (p. 252). But share price has nothing to do with market power. A company can double its stock price simply by doing a reverse stock split that exchanges every two existing shares of stock for one new share. A better gauge of market power is a company’s price/earnings ratio, which itself can be difficult to calculate.

Four Mistaken Assumptions

An implicit assumption running through the report is that important questions such as privacy and the preservation of a free press raise antitrust issues. The rapid rise of large internet companies with access to huge amounts of data that compete with press sources by providing content and selling advertisements clearly has had large effects and there is a strong argument for a policy response. But it’s not self-evident that applying antitrust doctrine to these new problems is the right response.

For example, the rise of the credit reporting agency and the ubiquity of credit cards have engendered deep concerns about data security, accuracy, and protection against fraud. Although both industries are highly concentrated, antitrust enforcement has not been the main enforcement tool for these problems. Instead, specific laws such as the Fair Credit Reporting Act spell out clear market rules that firms must play by. Similarly, tech markets would benefit from a national privacy law that avoids the heavy-handed regulation embraced by the EU’s General Data Protection Regulation or California’s Consumer Privacy Act. And if Congress believes some or all news publishers deserve help in order to strengthen society and democracy, it can pass legislation such as the Local Journalism Sustainability Act, which would give taxpayers certain tax credits for supporting local newspapers and media.

A second flawed assumption involves the value of data and the degree to which it conveys market power. The report assigns superpowers to data:

Data allows companies to target advertising with scalpel-like precision, improve services and products through a better understanding of user engagement and preferences, and more quickly identify and exploit new business opportunities. Much like a network effect, data-rich accumulation is self-reinforcing. (p. 42)

Each of the four companies collects massive amounts of data as part of its normal function. Much of this data is never used. What is used obviously gives the firms a degree of market power, partly because it improves the value of the products they offer. For instance, Google’s knowledge of what search results users click on helps it move the most popular choices up to the front, reducing time spent searching.

Although data is extremely important, it does not convey overwhelming market power. Unlike oil and most other physical goods, the same data can be used repeatedly by many people and, as we have seen, most of it is relatively cheap. Data can also become obsolete and its marginal value can decline quickly. More data, even a lot more, does not always convey a greater competitive advantage. For example, the accuracy of internet search did not decline when companies significantly lowered the amount of time past searches were stored for.[7] Economists Anja Lambrecht and Catherine Tucker have argued that data itself seldom provides a company with a competitive advantage, especially in the face of a superior product offering.[8] The algorithms, business models, and, especially, the products themselves are what really matter.

A third assumption involves the staying power of these companies. There is no doubt that each of the four firms is a powerful presence in its core markets. But is this because they consistently offer users the best options or because of anticompetitive tactics? Even as it repeatedly accuses the firms of inhibiting competition, the report also acknowledges that their products are the best. Google search, for example, is by any objective measure far better than its main rival, Bing. Amazon’s breadth of selection and quick delivery are unmatched. Facebook offers more coverage and services than any rival. The staff merely asserts without much evidence that these services would be even better if the dominant companies were more stringently regulated to prevent them from subverting competition.

Nor does the subcommittee seem to appreciate the long list of companies once considered dominant that have fallen from their high perch.[9] Google was preceded by Yahoo! and AltaVista. Blackberry and Nokia were the dominant phone companies until the iPhone came along. A&P, Sears, and Walmart dominated retail sales long before Amazon. Finally, Myspace cratered shortly after its sale to News Corporation for $580 million.

Each market continues to experience rapid technological change and an inflow of new venture capital funding. This results in constant improvements in their offerings. Companies that fail to continuously offer the best value are unlikely to last. For example, although the report points out that the ongoing pandemic has dramatically increased Amazon’s revenue, it fails to mention that the firm’s market share has fallen due to shipping delays and out-of-stock items.[10]

Fourth, the report discounts the benefits of vertical expansion into adjacent markets. In fact, it views such activity exclusively as a ploy by companies to leverage market power in one industry to create it in others. Yet, consumers often benefit from having one company solve multiple problems at once. One of the enduring advantages of Apple products is their unified design in which every piece of hardware and software is built around the others. Similarly, Tesla’s electric vehicles are an integrated package — the company develops the software, batteries, and many of the vehicle’s parts in-house. Entry into new markets creates more competition and usually helps consumers.

A Jaundiced View of Acquisitions

Similarly, the report overstates the anti-competitive danger of mergers. Most criticism of the failures of past merger policy in Big Tech centers on Facebook’s acquisitions of Instagram and WhatsApp. This accounts for only two of the over 600 acquisitions by the four companies mentioned in the report.

Both proved to be huge successes. But neither looked like sure bets at the time they were announced. In fact, four years after it was completed, Facebook’s purchase of Instagram was ranked 15^th on a list of worst tech deals of all time.[11] In any event, these mergers were carefully scrutinized by the relevant regulators. The Federal Trade Commission (FTC) and the European Commission examined the WhatsApp deal and approved it, while the FTC and the United Kingdom cleared the Instagram purchase.

No one knows whether Instagram or WhatsApp would have become such large platforms if Facebook had not purchased them, especially once Facebook introduced its own competing products. Retrospective studies repeatedly show that historically a high portion of mergers and acquisitions fail to earn back their cost of capital.[12] In light of deals such as News Corporation and Myspace, Time Warner and America Online, and eBay’s purchase of Skype, there is little reason to think the tech community is immune from making poor business decisions.

The report depends on two widely cited studies to support its position that mergers represent a growing threat to innovation. Last year, economists Sai Krishna Kamepalli, Raghuram Rajan, and Luigi Zingales published a paper showing that the prospect of an acquisition by an incumbent platform can undermine early adoption by users and create a “kill zone” where new ventures fail to get funding.[13] And indeed, the report alleges that funding for venture capital has fallen significantly (p. 47). Another study by Colleen Cunningham, Florian Ederer, and Song Ma showed that technology in acquired companies was less likely to be developed when it overlaps with the acquirer’s existing products, especially when the acquirer faces weak competition.[14]

The Kamepalli study is based on a small sample of nine acquisitions, seven by Google and two by Facebook. Even worse, according to Mark Jamison, the acquisitions that are included in the sample don’t meet the assumptions the authors chose for their model and should be tossed out.[15] The Cunningham paper looks exclusively at the pharmaceutical industry, in which trade secrets, patent protection, and heavy product regulation play a much larger role than in the tech industry. As a result, the study may have limited applicability to other sectors of the economy.

The report claims erroneously that venture capital investment in the United States has not been growing. Although it leveled off in 2019, tech funding was still 54 percent above the 2017 level.[16] The number of angel and seed deals rose by almost six-fold between 2006 and 2019, peaking in 2015. The number of early deals rose by 2.4 times. The report claims that “The rates of entrepreneurship and job creation have also declined over this period” (p. 47). But the data for this statement end a decade ago in 2011.

There’s no doubt that acquisitions can present antitrust problems and merit close scrutiny by regulators. But experience also shows that acquisitions put talent and technology in the hands of companies that can deploy it more quickly and over a wider market. Mergers also encourage investment in new firms by giving venture capitalists another exit strategy beyond an IPO.

Incorrect Claims About the State of Competition Beyond Big Tech

The report misjudges the context in which the broader antitrust debate is playing out. Concentration is rising across the economy, but in most industries it remains well below levels that have traditionally caused concern. Although the Big Tech firms represent a rising share of the S&P 500, the share of the stock market accounted for by the top five firms remains well below where it was in the 1960s, when antitrust enforcement was much more active.[17] Nor are companies making significantly greater profits, especially when we limit ourselves to domestic nonfinancial firms.[18]

The existence of high markups on products is also not necessarily a serious concern.[19] In markets with large fixed costs and economies of scale, a company can charge large markups in price over what it costs to produce each additional unit and still suffer losses because margins are not high enough to recover large fixed costs. In general, there are also reasons to think that estimates of marginal costs for companies across the economy are not adequately measuring increased investments in hard-to-measure intangible assets. The rise of superstar firms across a wide variety of sectors reflects their success in combining investments in information technology with a diffuse set of intangible assets including organizational transformation, software production, worker training, and brand equity, not in quashing competition.[20]

Running through the report is an implicit assumption that big companies have an obligation to help their competitors succeed, or at least go to great lengths to avoid harming their prospects for success. This assumption collides with a 40-year consensus that antitrust law should be largely indifferent to whether particular actions harm competitors and instead focus on their effect on consumers.

To what extent are large firms obliged to help their rivals? Normally a firm could refuse to deal with a strong competitor if it believed cooperating would disproportionately help the other firm. It could also cut off existing suppliers and customers without a clear reason. And it would never have to worry whether the introduction of a new improvement would harm competitors. The report strongly implies that at some point, large firms lose the right to behave like any other company. But it does not spell out what actions Big Tech should be permitted to take to improve its market share.

For instance, Google and Facebook constantly make changes to the algorithms behind Search and News Feed. Because of their dominance, these changes can have large effects on other companies. “Due to their outsized role as digital gateways to news, a change to one of these firm’s algorithm can significantly affect the online referrals to news publishers, directly affecting their advertising revenue,” the subcommittee notes. (p. 63). Similarly, while making it clear that leading firms should seldom be allowed to acquire new firms, it seems uncertain about whether the firm should be allowed to develop a competing product that might put a rival out of business. It would be reassuring if the report had clearly stated that even the largest firms have the right to improve their products, even if doing so has a negative effect on the profitability of competitors.

 

WEIGHING RISKS AND BENEFITS

Any objective evaluation of competition in technology markets must examine both the benefits and problems created by technology. Textbook economic theory assumes that every company has a small share of the market. In real life, many markets consolidate around a few large producers. While this gives companies the ability to influence volume and prices, larger firms often succeed because they are more productive and therefore are able to offer more variety and lower prices. That’s why U.S. antitrust law has traditionally focused on weighing the threat to competition against the promise of greater efficiency. Each of the four companies examined by the subcommittee delivers significant economic and social benefits that consumers obviously value highly. The report, however, not only discounts the benefits tech companies provide, but views them in a sinister light as threats to competition.

Is Big Always Bad?

In many industries, size brings strong advantages. One involves economies of scale: As companies produce more, the marginal price of producing each additional unit of a good or service falls. That’s why car manufacturers, for instance, need to achieve a certain size before they can compete effectively with incumbents. One reason why Google’s search engine is the most popular is that doubling the number of users does not double its costs and its sheer size provides the company with more data.

A second advantage involves network effects, where the value of a product to any user increases with either the number of other similar users (direct effects) or the number of users on the other side of the market (indirect effects). Facebook becomes more valuable to you every time one of your friends joins it. It wouldn’t be very convenient if your friends were spread out over 20 social networking sites, even if those sites competed fiercely on prices and services. As the number of users of any platform grows, so does its value to advertisers (an indirect network effect).

Markets can benefit from economies of scope as well as scale. It is sometimes more efficient to have one company produce all of a product than to divide the work among many firms, each producing a single component. Google’s dominance of various submarkets of internet advertising may raise anticompetitive concerns. But it likely makes the integrated system more efficient and cheaper than if it was subdivided into many parts. Yet the subcommittee report views large company size as inherently suspect and seldom acknowledges the manifest economic and consumer benefits it brings.

How Valuable Is Data?

Even as it acknowledges that many services offered by tech companies are either free or inexpensive, the report minimizes the value proposition for users. It argues that “not withstanding claims that services such as Google’s Search or Maps products or Facebook are ‘free’ or have immeasurable economic value to consumers, the social data gathered through these services may exceed their economic value to consumers” (p. 46). In other words, the report claims that consumers are trading their personal data for too little in return.

Yet the subcommittee offers little evidence that consumers attach great value to the personal data the companies collect. Certainly, they value privacy, and few would care to have their medical prescriptions or credit cards made public. But companies that don’t protect that information already pay a heavy price. But do you really care if some algorithm sends you a coupon for the new shoes you were shopping online for or if Google Maps knows where you are driving if the aggregated data helps you avoid traffic jams? The experience of the Internet era, here and abroad, is that most consumers are willing to trade some degree of privacy for the ability to communicate instantly and cheaply with friends and family across the globe, to have access to information about anything, and to purchase products more conveniently and at a lower price.

Your personal data has very little market value. An article in the Financial Times examined the market price charged by data brokers. General information including gender, age and location was worth only $0.005 per person. Information about someone shopping for a car cost $0.0021 per person. Finally, knowledge that a woman is in her second trimester of pregnancy and therefor likely to be a significant purchaser of baby products, commands $0.11 per woman.[21]

But people do place a huge value on the services they get in return. A recent online choice experiment involving users found that individuals would have demanded $17,530 to give up search engines for a year.[22] The equivalent values for email and maps were $8,414 and $3,648, respectively. Considering just the top five functions, the average person attached a value of $31,607 to services that they essentially get free or at very low cost. For comparison, the report notes that Facebook reported an average revenue per user of only $36.49 in the United States and Canada in July 2020 (p. 171). Rather than face these realities, the report mostly ignores the vast disparity between the costs and benefits of today’s tech platforms, while asserting they would be greater if the market power of the providers was reigned in.

Do Big Tech Companies Stifle Innovation?

Another key charge in the report’s sweeping indictment of tech platforms is that they inhibit economic innovation. In reality, a raft of studies shows they are the source of much of the innovation the U.S. economy has enjoyed in this century. In a 2018 survey of the top 1,000 global companies, Amazon and Alphabet (Google’s parent) take the top two positions, investing $22.6 billion and $16.2 billion, respectively. Apple was eighth ($11.6 billion) and Facebook was 16^th ($7.8 billion). For comparison, the big manufacturers Ford and Merck U.S. spent $8.0 billion and $10.2 billion respectively and ranked 12^th and 15^th respectively.[23]

Tech research and development, moreover, is not confined to existing markets. These firms are leading investors in frontier technologies including artificial intelligence, autonomous vehicles, quantum computing, robotics, and cloud computing. In each case, they face strong competition, often from established companies with deep expertise in the field. Would tech firms keep pushing the envelope if their profitability was substantially damaged? The report doesn’t address the question. Yet U.S. leadership in these technologies has huge implications for both future productivity and national security, given fierce technology competition with China.

The tech companies innovate constantly to make their core products better. Yet the report routinely misrepresents innovation as a threat to competition. It notes, for example, that a “challenge facing upstart search engines is the growing number of features and services that a general search provider must offer to be competitive with Google” (p. 83). True, but competition of course is what impels companies of all kinds to grow large, so they can enjoy economies of scale. All large companies benefit from economies of scale. In the same paragraph the report lists some of these “mandatory high-quality search features:” maps, local business answers, news, images, videos, definitions, and “quick answers” (p. 83). All are bundled in the service at no cost to users.

When Are Low Prices Bad?

Many popular services such as Google Search and Facebook, as well as many apps, are priced at zero. In fact, prices for digital services of all kinds are falling. A 2019 report by the Progressive Policy Institute noted that the cost of internet advertising had declined by 40 percent since 2010, while other forms of advertising had not gotten cheaper.[24] Price declines have had an even broader effect. Economist Thomas Philippon estimated the rise of ecommerce has amounted to a permanent increase in consumption of one percent.[25] Meanwhile, The Economist reported that inflation of online prices is running one percentage point below general inflation, saving consumers millions of dollars.[26]

The subcommittee report complains that “Amazon’s below-cost prices on products and services tend to lock customers into Amazon’s full marketplace ecosystem” (p. 297). Antitrust theory recognizes the danger of “predatory pricing,” in which a company lowers prices to drive out competition and then raises them later to boost profits. But the report adopts an absurdly expansive definition of the concept to include “any situation where a dominant firm prices a good or service below cost in a way that is harmful to competition” (p. 297). It’s true that when any company, large or small, lowers prices, it will undercut competitors. On the other hand, it may sharpen price competition, benefitting consumers. A better approach would be to focus on whether consumers are harmed in either the short- or long-term. Although the report mentions many instances of lower prices, it provides little proof of harm to consumers. If lower prices result from greater efficiency, they are a blessing.

It’s worth noting that the subcommittee’s dark view of low prices is not shared by most antitrust courts and practitioners. Consider the example of diapers.com, owned by Quidsi, also an Amazon third-party seller of diapers. Amazon dramatically lowered the cost of diapers on its platform, causing large losses to both companies. Eventually, Amazon purchased Quidsi, thereby eliminating a competitor from the market. To the subcommittee, this is a clear example of Amazon using its size to drive out a competitor so that it could later charge above-market prices.

But that’s not what happened. For one thing, it was actually Quidsi that first started selling diapers below cost to gain market share. As Quidsi’s founders explained:

[W]e started with selling the loss leader product to basically build a relationship with mom. And once they had the passion for the brand and they were shopping with us on a weekly or a monthly basis that they’d start to fall in love with that brand. We were losing money on every box of diapers that we sold.[27]

Amazon’s actions created a dilemma for Quidsi. It was not big enough to sustain the losses needed to match Amazon’s low prices. But it could not raise prices without losing market share to Amazon and other sellers. Eventually it accepted a $545 million offer from Amazon.

The story doesn’t end there. Even with Quidsi, Amazon was part of a highly competitive market for diapers that includes strong brands such as Pampers and Huggies, as well as large retailers including Walmart and Target. And while Amazon had acquired 43 percent of the online baby supply market by 2016, 80 percent of all sales occurred offline. Unable to make money, Amazon eventually wrote off Quidsi at a loss.[28] Meanwhile Quidsi’s former owners used their money to start a new online retail company, Jet.com, which they eventually sold to Walmart for $3.3 billion.[29] What the subcommittee staff see as a clear-cut case of predatory pricing was actually a costly strategic error by Amazon management.

Confusion about Winners and Losers from Competition

The authors of the Cicilline report take the view that pro-consumer behavior from Big Tech should be viewed as anticompetitive behavior simply because it raises the bar for competitors. For example, looking at the impact of Google Maps, it states, “Whereas market leaders TomTom and Garmin sold navigation services through subscriptions, Google was offering its service for free–a fact widely seen as disfavoring the incumbents, whose stock prices fell upon Google’s announcement” (p. 232). Similarly, it notes “Other retailers are unable to match Amazon on its ability to provide free and fast delivery for such a large volume and inventory of products” (p. 260). Offering consumers more services at no additional cost is a textbook case of increasing the level of competition in a market.

In discussing Facebook’s development of Open Graph, which lets third-party apps interact with data on Facebook, the subcommittee alleges that it gave Facebook the ability to prioritize access to its social graph (a list people who are “friends” or who follow each other in a social network) “effectively picking winners and losers online” (p. 149). The report recognizes that “Facebook’s Open Graph provided other companies with the ability to scale through its user base by interconnecting with Facebook’s platform. Some companies benefited immensely from this relationship, experiencing significant user growth from Open Graph and in-app signups…” (p. 149). There is no attempt to measure the large benefit conferred on companies that used Open Graph against the losses to those that Facebook allegedly discriminated against. Nor is there any discussion of why the subcommittee thinks it should be illegal for a company to introduce a product that benefits some developers but not all.

In contrast, the report does acknowledge some of the benefits of app stores, which “provide mobile device users with a sense of trust and security that the apps they install from an app store have been reviewed, will not harm the user’s mobile device, will function as intended, and will not violate user privacy…. By reducing the costs of app developers, app stores help make software applications more affordable for consumers” (p. 94).

Nonetheless, elsewhere the report criticizes both Apple and Google for using their app stores to disadvantage potential competitors. Even if we assume such favoritism would be anticompetitive, it is not clear how regulators would distinguish it from legitimate efforts to protect the integrity of each platform’s app store. The problems posed by each case would have to be measured against the benefits, which would be a regulatory nightmare.

Finally, the report acknowledges that, “For many sellers, there is no viable alternative to Amazon, and a significant number of sellers rely on its marketplace for their entire livelihood” (p. 274) and “Due to a lack of alternatives, third-party sellers have no choice but to purchase fulfillment services from Amazon” (p. 287), it adds. But 56 percent of Amazon merchants also sell on eBay and 47 percent have a personal site.[30] Meanwhile, Shopify is growing rapidly, partly because it makes it easy for sellers to create their own websites. For those sellers limited to Amazon, the report does not explore how much worse off they would be if Amazon did not allow third-party sellers. Nor does the report ask whether third-party sellers are happy or unhappy with the arrangement. The subcommittee simply asserts that stricter antitrust enforcement would provide a superior outcome.

COUNTERPRODUCTIVE REMEDIES

Having misjudged the nature of digital markets and the main source of competitive advantage, the report then proceeds to recommend legislative and regulatory solutions that will be difficult to implement and unlikely to increase consumer welfare.

Banning Acquisitions

The subcommittee proposes to limit a large firm’s ability to acquire future companies. This supposedly would ensure that new companies could grow to the point where their innovations would be available to all companies, not just the acquirer, and possibly even become a serious challenger to the dominant firm. The report also calls for eliminating current reporting exemptions for dominant firms (defined as having a market share of over 30 percent) so that even the smallest transactions are subject to antitrust review. It would also reverse the burden of proof. Instead of requiring the government to show an acquisition would harm consumer welfare by raising prices or slowing innovation, the company would have to show that consumers would benefit and that these benefits could not be achieved in other ways (p. 387). The standard would change such that acquisitions by dominant firms would be forbidden even if they resulted in efficiencies and even if there was no reason to think that the acquired firm would be a successful challenger (p. 393).

This ban on acquisitions would likely have a chilling effect on venture capital firms, which often fund new companies in hopes they will be acquired by larger ones. It also fails to recognize that firms differ in their core strengths. Smaller firms are often more able to concentrate on developing a small number of important innovations while larger firms have a better capacity to integrate innovations into an existing product and scale them up for a national or even global audience. Left on their own, small firms may not have either the resources or the skills to take on an established firm.

Breaking Up Big Tech

The report also advocates structural reforms that would break up Big Tech firms either by forcing them to sell off major parts of their business or prohibiting them from offering their own products and services on their platforms that compete with third-party sellers. This reflects the subcommittee’s suspicion that dominant firms have an irresistible incentive to discriminate against third party products in favor of their own brands.

Physically breaking up Big Tech would be a Herculean task. Forcing Facebook to sell Instagram or separating Amazon Marketplace from Amazon Web Services would require the courts or regulators to make complex decisions about organizational structure and market prices. It tends to assume that management structures are not heavily intertwined. The government has not made a major effort to break up a monopolist since the case against Microsoft at the turn of the century (and even then, an appeals court reversed the structural breakup and imposed a less severe behavioral remedy). A review of past cases concluded that, with the exception of the AT&T case, in which the same results could have been achieved with a regulatory rule, breakups failed to increase competition, raise industry output, or lower prices.[31]

Forbidding companies from using their own platforms even as they host third parties would ban a common business practice and harm consumers. Chairman Cicilline has compared such a ban to the Glass-Steagall statute, which prohibits commercial banks from owning investment banks or other businesses. But that statute was passed in order to prevent banks from extending the benefits of federal deposit insurance to other activities (not to prevent banks from competing with third parties). Instead, it may have decreased the stability of the financial system by preventing financial institutions from diversifying their risk.[32]

Some third-party products are protected by copyright or patent law. But where this is not the case, there is nothing wrong with one company trying to copy or improve on another’s product. In fact, patent holders are required to disclose their innovations precisely to allow others to build on them (with the appropriate licensing agreements). Large retailers including Walmart and Costco and large grocery chains routinely offer their own branded products next to those of major third parties. These offerings give consumers more choice and lower prices. Eliminating them would kill jobs and harm consumers. The subcommittee staff tries to distinguish Amazon from brick-and-mortar stores by arguing that the latter have much less detailed information about the competing products they offer and far less information about buyers’ shopping habits and preferences than does Amazon (p. 282). However, the report does not show that this additional data conveys a substantive advantage. Nor does it acknowledge that the brick-and-mortar retailers are rapidly trying to catch up.[33] For example, Walmart is the third biggest spender on IT in the US, trailing only Amazon and Google.[34]

Forcing companies to choose between offering their own products and opening up their platforms to third-parties may also be counter-productive. Although Amazon is unlikely to abandon third parties, other retailers, such as Macy’s, may become less likely to open theirs up, thus reducing the alternatives to Amazon for third-party sellers. It is also not clear that private labels are always a threat to third-party offerings. (Your friendly neighborhood grocery store does this all the time). Although selling its own brand may result in more revenue, it may lower margins, because of the extra costs incurred from manufacturing and because the revenue is offset by the loss of commissions on third-party sales. In fact, platforms that charge higher commissions to third-party sellers have less of an incentive to offer their own labels. Finally, platforms may get around this ban on owning the platform and competing on it by purchasing third-party products before offering them to consumers. This would not increase the bargaining power of outside suppliers. Neither would it protect suppliers from suddenly being terminated in favor of another product.

The subcommittee advocates invoking the essential facilities doctrine, which would force Big Tech firms to deal with all competitors on equal and fair terms like utilities. Such solicitude to competitors is not required of ordinary companies. In fact, the doctrine, which does not explicitly exist in statutes and has never been recognized by the Supreme Court, is rarely invoked. Doing so would involve either the courts or regulators in numerous disputes about whether specific contract terms were in fact fair. The Supreme Court explained the dangers in a recent case: “Enforced sharing also requires antitrust courts to act as central planners, identifying the proper price, quantity, and other terms of dealing–a role for which they are ill-suited. Moreover, compelling negotiation between competitors may facilitate the supreme evil of antitrust: collusion.”[35]

CONCLUSION

The market power of Big Tech raises many serious policy issues related to consumer prices, innovation, privacy, data security, misinformation, radicalization, and inequality. Most of these don’t involve antitrust. Those that do, such as consumer prices and innovation, require a careful weighing of the benefits against the threat to competition.

Unfortunately, the subcommittee has mostly missed the opportunity to grapple with these real problems arising from the tech sector. It has instead attempted to apply the static lessons of historical battles against monopoly to the fluid, fast-changing realities of digital innovation and competition today. Its report is mainly a parable about the perils of bigness. Big companies are indeed powerful and merit strong public oversight. But large size is often a marker of legitimate business success. The main reasons the Big Tech companies are so successful and big is that they provide the best products and services for billions of consumers in the United States and around the world. The report offers scant evidence that these companies are actually suppressing competition or innovation in our very dynamic digital markets.

Getting this right matters. Internet platforms may be the business model of the new century. Business models that combine massive data with sophisticated artificial intelligence and bring together different parts of a market may be the key to making large improvements in the quality and affordability of major sectors including education, health care, construction, energy transmission and government services, all of which continue to lag behind the digital sector in boosting productivity and good job growth. [36]

The report has pointed out legitimate issues involving use of data, arbitrary treatment of counter parties, and providing false information to regulators. But existing antitrust laws are adequate to handle these issues without invoking the drastic remedies discussed in the report. The answer, as it has been for over four decades, lies in an unbiased study of each market focused on how different actions affect consumer welfare. The report offers no compelling case for why that standard should be changed, and it fails to show how consumers have been harmed by the leading technology companies.

ABOUT JOE KENNEDY

Joe Kennedy is currently Senior Principal Economist for MITRE, Inc. He has spent 30 years working at the intersection of public policy and technology. Previous positions include Senior Fellow at the Information Technology and Innovation Foundation, Chief Economist at the Department of Commerce, and General Counsel of the Senate Permanent Subcommittee on Investigations.

He is also President of Kennedy Research, LLC. The views expressed in this paper are his and are not necessarily shared by the Progressive Policy Institute (PPI) or any other organization. Mr. Kennedy is grateful to Alec Stapp of PPI for helpful editorial and substantive suggestions. All errors remain his responsibility.

 

Download the paper here: