Thank you, Chairman Cicilline, Ranking Member Sensenbrenner, and Members of the Subcommittee for the opportunity to testify today.
My name is K. Sabeel Rahman and I am President of Demos. Demos is a dynamic “think-and-do” tank that powers the movement for a just, inclusive, multiracial democracy. Our name—meaning “the people”—is the root word of democracy, and it reminds us that the promise of a truly inclusive democracy demands that we ensure that “we the people” can exercise real power over our political and economic futures—and that we must dismantle those systemic forms of racism that exclude Black and brown communities from that shared future.
I want to thank the Committee for its leadership in convening these hearings on the vital issue of the need to strengthen antitrust laws and promote competition in the online economy. In this moment of deepening economic inequality and escalating crises of disinformation, the long-term vitality of our democracy and our economy depend on a reinvigorated approach to anti-monopoly policy, particularly in context of dominant technology platforms like Amazon, Alphabet, Facebook, and Apple.
As these hearings have highlighted, these dominant tech firms now possess a concentrated economic power that threatens economic well-being and innovation, and undermines democratic ideals.
We have seen how Amazon has, for example, leveraged its dominance over online retail transactions to undercut competitors and engage in predatory pricing.1
That same market dominance has enabled Amazon to pressure state and local governments for more favorable regulatory treatment and subsidies, even as it undermines enforcement of workplace safety laws that would protect Amazon workers—particularly Black and Latinx “essential” workers, who face astronomically high injury rates more than double the industry average,2
and that have become hotspots for COVID-19 transmission.3
In this era of continued quarantine, we see Google’s growing dominance over public education in an era when COVID-19 has driven a shift to remote learning.4
And with the election rapidly approaching, we have also seen the dangers of the ad-based and data-mining business model of online information platforms like Facebook. Facebook’s algorithms—designed to maximize user attention in order to sell targeted ads—fuel the rampant spread of misinformation in ways that alter the dynamics of the 2020 elections,5
exacerbate the dangers of voter suppression,6
while also accelerating the spread of extremism, racial violence, and hate speech.7
Here too it is often Black and brown communities that frequently bear the brunt of harassment, hate speech, and voter suppression that Facebook’s corporate policies enable to flourish on its platform.8
The question now is how Congress and our federal government must respond to these various challenges.
In this testimony this afternoon, I will make the case that these many different problems share a common root: the problem of unchecked private control over essential social, economic, and political infrastructure. Tech platforms like Facebook, Google, Amazon, and Apple represent essential infrastructure just like the railroads, bridges, and telegraph lines of a century ago. This poses unique challenges for public policy.
I will also argue today that we have a robust and historically-effective policy toolkit to address the problem of private control over infrastructure. Limiting this problematic form of private power requires rediscovering familiar but forgotten tools, including antitrust law, public utility-style regulation, and a willingness to consider cases where public control of key infrastructure would benefit the public rather than private provision. Reviving and deploying these policy interventions will be critical to secure an equitable economy and an inclusive democracy in the years ahead.
The Policy Problem: Unchecked Power Over Essential Infrastructure
We are used to thinking of infrastructure in physical terms: roads, bridges, railroads, power lines, sewer systems. But infrastructure can also be economic—think for example about systems of financing and credit essential to businesses and households, or to how the COVID-19 pandemic has reminded us how critical child care and healthcare infrastructure is for supporting families and workers and businesses alike. Infrastructure can also be digital. The conduits of commerce are increasingly online through retail platforms like Amazon. The infrastructure of information now depends on web services like Amazon’s AWS or online media platforms like Facebook or YouTube.
We can think of infrastructure as those goods and services that have three key characteristics. First, these are goods and services that have economies of scale: there are efficiencies to be gained by consolidation and unification; a digital or telecom network that covers the whole country is more valuable than one that is limited to just one neighborhood. Their social value hinges on these goods and services being available at scale to as many users as possible.
Second, these are goods and services that open up a wide range of downstream uses and capabilities.9
Think of the economic and social activity made possible by railroads, the telegraph, and now, online media platforms. Infrastructural goods and services are prerequisites for a wide range of uses and activities.
Finally, because of both the scale and necessity of these goods and services, infrastructure also creates a risk of vulnerability. Whenever a good or service is necessary and irreplaceable, everyone who uses it relies on its provider and is vulnerable to provider decisions that affect access to or quality of the essential good or service. As a result, a firm that controls infrastructure holds arbitrary power over everyone who depends on that infrastructure. This control can have tremendous impacts on equity, inclusion, and democracy. Overly restrictive or extractive control over infrastructure can transform it from an empowering foundation to a “bottleneck”, constraining who can use these vital goods and services, widening inequality and disparities in well-being and inclusion.10
Today’s online giants represent a digital infrastructure upon which our economy and our democracy depend—and which creates specific power imbalances that public policy must remedy. There are three specific types of infrastructural power that any policy agenda must attend to.11
Transmission power: This type of infrastructural power stems from private control over the transmission of goods, services, or information. Consider how Amazon’s command of a shipping and logistics system enables Amazon to manipulate the flow of goods, and to mine troves of consumer data. That manipulation and data can then be used to choke out competitors, alter prices, and influence search results in ways that maximize Amazon’s own profits, leaving entire economic sectors, from book publishing to apparel manufacture, at the whim of Amazon’s decisions.
Gateway power: Another kind of infrastructural power arises when firms control gateways to information or other critical goods and services. For example, access to the internet is increasingly mediated through the gateway of Google Search. By controlling the point of entry, Google heavily influences the types of information and commerce that people are able to access. As a result, even small and hidden changes to the algorithms of Google Search can make or break news media, entertainment outlets, and other content producers. At the same time, people using Google Search navigate the landscape constructed by these algorithms, which aim to maximize company profits, not to provide the most accurate information or news.
Scoring power: A third type of infrastructural power is scoring power, exerted by ratings systems, indices, and ranking databases.12
These scoring systems appear objective and neutral but are grounded in data and analytics that reproduce existing patterns of racial, gender, and economic bias. For example, private credit reporting agencies like Experian, TransUnion and Equifax produce credit reports and scores used for lending, insurance, and employment decisions, yet these scores rank and categorize consumers on the basis of borrowing and payment behavior that is shaped by immense racial wealth disparities, which are themselves the products of centuries of discriminatory public policies.13
As a result, evaluations of credit history make Black and brown consumers appear less worthy of affordable credit, insurance, and employment opportunities, reproducing historic discrimination.14
Similarly the scoring systems and algorithms that shape the flow of information, ads, and commerce on Facebook, YouTube or Amazon magnify these racial disparities and can induce highly problematic forms of targeting and filtering of information flows.
These three forms of infrastructural power create an urgent problem for public policy. Conventionally, we fear that government regulation might “interfere” with otherwise free markets—but that viewpoint misunderstands the reality we are in. In truth, we live in an economy that is already heavily governed and regulated by private actors—tech giants like Facebook, Alphabet, and Amazon (as well as dominant “offline” firms as well in sectors like pharmaceuticals, finance, and food production). Yet unlike governmental regulators, the decisions made by private firms with infrastructural power are not subject to mechanisms for democratic representation, participation, or public accountability.15
Absent a government check on this tremendous concentration of private power, vital infrastructure will increasingly be subverted to private profit rather than meeting public needs.
This is why we need the kind of revived antitrust policies that this Committee is considering.
Reviving and Adapting a Policy Toolkit: Breakups, Public Obligations, and Public Options
While today’s tech giants are a 21st century phenomenon, the policy problems they pose are long-standing ones, familiar to our history of economic regulation. A century ago, the rise of industrial monopolies in railroads, telecommunications, finance, and other sectors sparked a wave of policy innovation leading to vital legislative and regulatory interventions like the Sherman Antitrust Act and the rise of public utility commissions at the state and federal level. These innovations were designed to address the problem of private control over infrastructure—the same kinds of problems that today’s tech giants pose. The proposals you have heard over the course of these hearings, while adapted to our modern context, are in fact a revived form of long-standing traditions in American law and public policy.16
In this last part of my remarks, I would like to outline a policy framework for legislative and regulatory action in response to the problems of tech firms, monopoly power, and infrastructural power.
There are three policy strategies in particular that Congress and regulators at the FTC, FCC, and elsewhere should consider.
First, we must limit the dangers of infrastructural power by breaking up dominant firms, imposing firewalls and structural limits on the power of these firms to control essential infrastructure. This means developing policies that include separation by size (“breaking up” market dominant firms); separation by function (splitting platforms from commerce, for example);17
laws requiring interoperability to mitigate against undue consolidation and merger activity; and laws prohibiting tying contracts or predatory pricing. These limits can be legislated, and enforced by federal regulators.
The impact of these policies would be to break up the private control over essential online infrastructure, and reduce the incentives for self-dealing. Imagine, for example, if Amazon could not hold both the online retail portal and the production and selling of its own branded goods on that portal: there would be far less likelihood of Amazon leveraging its platform dominance to give its own products a leg up in competition. Or consider how a structural limit on ad-based revenue would change the incentives for Facebook, removing the profit motive that currently fuels its preference for attention-maximizing algorithms that accelerate the spread of disinformation18
Divestiture or breakup could be a required remedy under these policies, which could in turn spur greater innovation and economic creativity in the future. Indeed, structural separations and breakups have in recent decades been disparaged as overly costly and economically harmful, but these critiques are not borne out by the historical evidence. Looking back at key cases of breakup and structural separation from AT&T to the separation of investment and commercial banking to major antitrust cases of the early twentieth century, the record indicates that breakup—and the threat of breakup—have been essential to enabling the very innovation that eventually gave rise to today’s dominant tech companies.19
Second, we should through legislation and regulatory enforcement impose public obligations and basic standards of nondiscrimination, fair dealing, fair pricing, and accountability over these infrastructural firms. Over a century ago, common carriage requirements were critical to preventing discrimination on railroads, and ensuring that all comers could access new transportation infrastructure to engage in commerce and travel. Historically, public obligations have also encompassed requirements for basic health and safety—for example, assuring that goods are not toxic or harmful to consumers. It was the rise of these kinds of public obligations that helped drive the development of our modern forms of labor, consumer, and business regulations.20
Similar public obligations were at the heart of the net neutrality debates in previous years: requirements of common carriage and anti-throttling obligations were meant to ensure that internet service providers did not leverage their control over access to the internet to favor paying information providers or business allies over other content providers and businesses.
In context of today’s tech giants, these types of measures today will be critical policy tool that complements the structural separations, firewalls, and breakups noted above. For example, we might require by legislation and/or regulation “rules of the road” for platform firms to treat all businesses fairly (by not skewing search results, for example), or requiring fiduciary obligations for how tech firms treat personal user data.21
Nondiscrimination requirements could protect businesses from being squeezed out of Amazon or Google search. Price regulations could prevent predatory pricing on online platforms. Portability requirements and interoperability standards could help ensure equal access and ability to exit market dominant platforms and closed tech ecosystems.
Third, we should consider the degree to which some of these essential infrastructures can be provided not by private, profit-seeking firms, but by public providers, either on an exclusive basis or as “public options” that compete alongside private alternatives.22
In some markets, a public alternative could help remedy the problems of infrastructural power, especially if the public option operates on a non-profit basis, with statutory requirements for nondiscrimination, fair pricing, and the like. These public options could provide a ‘plain vanilla’, non-exclusionary alternative—which in turn would impose competitive pressures on private firms to match these socially-beneficial terms of service. In the internet service debate for example, the attempts to create municipal broadband networks represents a “public option” response to the infrastructural power of internet service providers like Comcast or Spectrum.
In the tech platform domain, there have been proposals for example for a “public” digital infrastructure to offset the monopoly power of today’s private infrastructure firms.23
Public options are also a common intervention in “offline” policy debates. In the healthcare space, debates over Medicare for All are about public provision of healthcare services. Similarly, proposals for establishing a publicly-run credit registry to displace the discriminatory and extractive oligopoly of private credit bureaus like Experian, Equifax, and Transunion who manage consumer credit information and the private investment ratings agencies like Moody’s Standard and Poor, and Fitch—all of which leverage their private control of this essential infrastructure to generate profit in ways that are extractive, racially-discriminatory, and prone to self-dealing and longer-term systemic risk.24
Conclusion: Building an Inclusive Economy and Democracy in a Moment of Crisis
From Amazon’s increasing stranglehold on our economy to the dominance of Alphabet over the flow of information on the internet to the closed ecosystem and market dominance of Apple to the proliferation of extremism and disinformation on Facebook, today’s technology giants pose immediate challenges for our economy, our democracy, and the ideal of an equitable, inclusive society. These various challenges share a common root, arising from the fact that these online firms increasingly operate as digital infrastructure for our shared economic, social, and political life, and yet these firms leverage this dominant position to advance their private interests in ways that harm the public welfare. While the technology is new, this kind of infrastructural power is a familiar problem that previous generations of American policymakers have successfully tackled, deploying a range of tools like breakups and firewalls, structural separations and breakups; public obligations and regulatory standard-setting; and direct public provision and/or the creation of public options. Adapting these policy strategies today will require new legislation from Congress, and new creativity from federal regulators at the FTC, FCC, and elsewhere.
The hearings that this Committee has hosted over these last few months have helped document the scale of the problems raised by tech giants, and the kinds of policy solutions we need. In this moment where Americans across the country are demanding dramatic action to address the worst economic collapse since the Great Depression, the pervasive forms of racialized violence against Black and brown communities, and the ongoing attacks on our democratic system itself, a robust anti-monopoly and antitrust agenda will be critical to improving competition, advancing economic and racial inclusion, and rebalancing our democracy.
Thank you for your time and I look forward to your questions.
1See e.g. Lina Khan, Amazon’s Antitrust Paradox, Yale Law Journal 126:3 (2017).
12See e.g. Danielle K. Citron & Frank A. Pasquale, The Scored Society: Due Process for Automated Predictions, Washington Law Review 89 (2014); Pasquale, Black Box Society, Harvard University Press, 2015.