When should we think of Amazon’s or Facebook’s incremental changes to fee structures and terms-of-service agreements as “taxes and regulatory policy, set by private monopolies”? When should our antitrust regulation “rely more on presumptions and bright-line rules that outright ban certain business practices by dominant firms”? When I want to ask such questions, I pose them to Lina Khan. This present conversation focuses on Khan’s contribution to the recent Investigation of Competition in Digital Markets report from the US House Judiciary Committee’s Subcommittee on Antitrust, Commercial, and Administrative Law (for which Khan served as counsel). Khan is an associate professor at Columbia Law School, where she teaches and writes about antitrust law, the antimonopoly tradition, and law and political economy. The New York Times has described Khan’s work as having “reframed decades of monopoly law,” and Politico has called her “a leader of a new school of antitrust thought.” Her article “Amazon’s Antitrust Paradox” was awarded a 2018 Antitrust Writing Award, and her article “The Separation of Platforms and Commerce” won the 2019 Jerry S. Cohen Memorial Fund’s Best Antitrust Article on Remedies. Khan has served as legal advisor to Commissioner Rohit Chopra at the Federal Trade Commission, and as legal director at the Open Markets Institute.
A related conversation on the Investigation of Competition in Digital Markets report, with Representative David N. Cicilline, Chairman of the Subcommittee, can be found here.
ANDY FITCH: First, why should we see the core business models (and longstanding business practices) of Google, Amazon, Facebook, and Apple not just as sometimes operating to the detriment of individual consumers, but as systemically harming a much broader range of workers, entrepreneurs, independent businesses, open markets, and public spheres? In what ways do the stakes here extend to the foundational health of our economy and of our democracy?
LINA KHAN: Google, Amazon, Facebook, and Apple control the infrastructure on which digital commerce and communications take place. They function as gatekeepers. They’ve used their gatekeeper power both to extort and to exploit the individuals and entities that rely on their technologies. They’ve maintained and extended their power through serial acquisitions and through coercive and predatory tactics. Meanwhile, the targeted ad-based business models of Facebook and Google incentivize maximal surveillance and invasive data collection. Each of these dynamics imperils the health of our economy and democracy. A few facets in particular stand out.
First, the fact that these firms control the infrastructure of the digital age means their decisions have far-reaching effects. A single fee hike or wage cut by Amazon can bring about a wealth transfer from millions of businesses or workers to Amazon. A single change by Google can mean that millions of school kids locked into Google’s education technologies must surrender even more personal data. So the sweeping ramifications of Amazon’s or Google’s decisions, for example, coupled with the lack of any public checks, means that the power they wield can function as a form of private governance. These are taxes and regulatory policy, set by private monopolies.
Second, these firms are directly shaping which information and ideas people access, and which information and ideas get produced in the first place. Control over online video, online search, and email (for Google), and over social networking (for Facebook) directly shapes what news people see or don’t see. These firms’ duopoly over digital advertising has directly contributed to the decimation of news publishers, and to the subsequent rise of “news deserts” across the country, where swaths of counties no longer have a source of local-news coverage. Meanwhile, Amazon’s power in the book market, and the industry-wide ramifications of that power, shape which books get published. Concentrated control, coupled with a business model that incentivizes the dissemination of disinformation and inflammatory content (in the case of Facebook and Google), or that strategically treats books as a loss-leader in order to sell a much broader panoply of consumer goods and Prime memberships (in the case of Amazon), has proven a highly damaging combination. Our recent election cycle, with public officials begging Mark Zuckerberg to ensure that Americans weren’t flooded with false claims about who won, underscored both the dangers and the democratic stakes.
Third, concentration of economic power concentrates political power. The dominant platforms are among the top spenders on lobbying. They have a small army of staffers and consultants and trade groups who spend hundreds if not thousands of hours seeking to influence laws, regulations, and policy outcomes in both formal and informal ways. And fourth, the surveillance technologies these firms have developed and deployed to track users and workers (amid few real limits on how the firms can use this data) threaten human privacy and dignity, while creating and exacerbating deep asymmetries of power. In particular, the dominance of Facebook and Google, coupled with their business models, has meant that the price of accessing digital infrastructure is surrendering to surveillance.
Along those lines, with dominant platforms offering so much “for free” (and sometimes experiencing extraordinary growth before generating profits), why does a late-20th-century antitrust focus on consumer-pricing metrics seem inadequate? How might we need to redefine the personal and the collective “price” paid for such services and products? And/or which economic measurements might better help us to assess monopoly dominance?
Relying exclusively on price-centric models blinds us to the many coercive and predatory ways that dominant firms use their economic power — ways that, in some instances, existing antitrust laws already prohibit. Recent lawsuits filed against Facebook and Google note that these firms have abused their monopoly power in ways that harm user privacy. These lawsuits offer a small but important step forward for antitrust enforcers. More broadly, antitrust law should rely more on presumptions and bright-line rules that outright ban certain business practices by dominant firms. The current approach (which, in many cases, requires proving the “anticompetitive effects” of a business practice, and sometimes even requires weighing these effects against potential “benefits”) has created a much more permissive regime. Lastly, we need to broaden the range of disciplines and methodologies that carry weight in antitrust analysis. We need to incorporate learning from financial analysts, accountants, technologists, and business historians.
In recent years, the set of economists studying monopoly power and concentration has broadened beyond industrial-organization economists to include labor economists, macroeconomists, and public-finance economists, among others. This already has helped to spotlight key dimensions of monopoly power that IO economists were missing. But economics, like all disciplines, suffers from blindspots. So we need to go further.
Now for individual firms, could we start with Facebook’s acute dominance within social-networking spheres — with this platform today mostly just “competing” against its own adjacent corporate holdings? How might Facebook’s history of purchasing potential rivals, its tacit establishment of innovation kill zones, its impeding of American entrepreneurship, epitomize the need for presumptive prohibition on digital mergers and acquisitions? And why should Facebook’s near-perfect market knowledge (far beyond that of regulators) in various domains call forth a proactive incipiency standard protecting nascent competitors, and preventing vertical consolidation?
Facebook offers a case study in permissive merger enforcement. As noted in both the House report and the recent complaints filed by 48 state attorneys general and the Federal Trade Commission, Facebook maintained its monopoly through serially acquiring rivals. One business that it purchased, Onavo, even enabled Facebook to identify and closely monitor rival apps diverting attention from Facebook — positioning it to swoop in and buy up a competitor before others (including antitrust enforcers) fully understood what was going on. Collectively, Amazon, Apple, Facebook, and Google have purchased over 500 companies, and not a single one of these acquisitions was blocked by antitrust enforcers.
Antitrust enforcers can begin to remedy this multi-decade institutional failure by revising merger guidelines, and by taking a much more assertive and forward-looking approach. Lawmakers should consider a presumptive ban on acquisitions by these dominant firms. Antitrust law reflects a preference for growth through internal expansion and investment, rather than through acquisition. Legislating a presumptive ban would reassert this preference. It could be especially impactful amid the COVID-19 recovery, given that the dominant platforms have only grown richer during the crisis, and are sitting on huge sums of cash that they could use to go on a buying spree.
Then for Google, how does a convergence of structural forces (such as the high ongoing costs of maintaining search infrastructure, the contractually determined placement of Google products on a vast array of browsers and devices, the interlocking hold of Android and Cloud and Chrome and Gmail and Maps and any number of applications and services, the invaluable caching of multifaceted click and query data, the self-advantaging establishment of both formal and informal industry standards, and so on) contribute to the perpetuation of a monopolistic marketplace — with, for example, even search-engine rivals, or specialized vertical-search alternatives, dependent on Google as their primary gateway to users?
Google has benefited from the natural-monopoly features of web crawling, and from the self-reinforcing advantages of click and query data — which let Google constantly refine its search algorithm. But as spelled out in both the House report and the antitrust complaints filed recently, Google has maintained and extended this monopoly through a series of anticompetitive business practices. This has included demoting and scraping content from specialized search engines, which Google viewed as a threat, given that more specialized providers could have made Google less relevant for certain valuable search queries. This business conduct has also included coercive contractual agreements with mobile manufacturers, required by Google to pre-install Google Search and a set of Google apps, as a condition of using the Android operating system and its Play Store. Google also pays Apple up to 12 billion dollars a year to be the default search engine on Apple devices. These combined strategies have let Google maintain its search monopoly even as internet usage has shifted from desktop to mobile.
More generally, Google has engaged in an extensive set of tying arrangements (conditioning access to one Google service on the use of another) across its ecosystem, which has allowed it to monopolize display advertising, digital mapping, browsers, along with a host of other services — all of which again feed valuable data back to Google, further strengthening its dominance. And now, as search migrates to voice (including smart speakers, cars, and the “internet of things” ecosystem), Google is trying to use this same playbook to further reinforce its monopoly.
Amazon provides its own striking case of a firm first acquiring gatekeeper status (here in the realm of e-commerce), and then exploiting this position to the longer-term detriment of suppliers, workers, corporate clients, and individual consumers. So how does Amazon’s current business model exemplify the need for structural separations between dominant platforms and adjacent lines of business?
Structural separations and line of business restrictions have been a mainstay tool applied to network monopolies and dominant intermediaries — including railroads, telecommunications firms, and banks. As one of my articles describes, policymakers in earlier eras recognized that allowing a dominant intermediary (on which swaths of businesses depend) to integrate into adjacent lines of commerce would create a core conflict of interest, incentivizing discrimination and self-preferencing. Coupled with forms of common-carriage rules, structural separations were foremost designed to prevent such conflicts of interest and discrimination by dominant intermediaries. A host of other goals (including promoting media diversity, promoting economic stability and administrability, and preventing concentrations of power) also motivated these structural separations.
Amazon has at least three distinct lines of business that can be viewed as infrastructure: its online retail platform, its fulfillment and logistics network, and its cloud-computing provider (Amazon Web Services). The House investigation found that serving as a dominant intermediary across these markets, while also entering adjacent lines of business, has enabled Amazon to engage in a series of harmful business practices. These include: misappropriating the data of dependent third parties (in order to compete against them), using dominance in one market as leverage in negotiations in a distinct line of business, using integration to tie together its products and services (such as conditioning access to the Buy Box on use of Amazon’s fulfillment services), and using monopoly profits from one line of business to subsidize entry into an unrelated area — where Amazon can bleed millions of dollars over years to drive out competitors. Structural separations would help address this conduct. The House report also recommends a host of related tools that have been applied to dominant intermediaries, including nondiscrimination rules, interoperability, and a heightened set of obligations and prohibitions, as well as deep reforms to the antitrust laws.
Portrait of Lina Khan courtesy of Shah Ali.