Press Association via AP Images
This article appears in the Fall 2018 issue of The American Prospect magazine. Subscribe here.
In recent years, a new consensus has emerged that something has to be done about tech platforms. The market power of giant platform companies such as Amazon, Facebook, and Google puts them in a position to undercut businesses that rely on the platform, to dominate and rig markets, and to vacuum up user data. It also means they have an outsized ability to influence government, particularly by hiring legions of lobbyists.
In response, some critics have suggested they should be broken up; others that they should be regulated as if they are public utilities. But as even Facebook's Mark Zuckerberg has acknowledged, the debate on whether to regulate is now over. The question is “How do you do it?”
Commentators and experts have outlined a variety of proposals for regulating tech—ranging from a data tax to privacy rules—but there seem to be fewer ideas on how to address the problems that arise from the basic structure of tech platforms. The answer to Zuckerberg's question, however, is less complicated than one might imagine. As innovative as tech companies are, traditional principles of antitrust and regulation provide a roadmap for action.
Think of platforms as having two elements: First is the platform itself. A platform can serve as a marketplace or exchange (like Amazon's Marketplace) or as basic infrastructure for third parties to customize (like a developer platform), and it usually exhibits network effects—it becomes more valuable the more that people use it (like Google Search). Tech platforms with large networks also capture huge quantities of data, which makes them extremely valuable. As a result, tech platform mergers have become common. Facebook, for example, owns Instagram and WhatsApp. Google has a variety of tentacles in the online advertising space, including DoubleClick (recently retired).
The second feature of tech platforms is that the platform is distinct from the users who are on the platform, even as many platforms also have business lines or products that operate on the platform. To put it another way, the platform owns some of the users—which in turn compete with outside users. For example, Amazon Basics produces and sells goods on Amazon's marketplace—alongside other companies' goods. Google gathers reviews of restaurants and it shows that data on Search—alongside companies like Yelp, which do the same thing.
Both of these aspects of tech platforms are problematic. When one tech platform buys another, it doesn't just mean a larger entity. Tech mergers mean that a single company gets data from multiple different platforms—and that a rising platform in one sector can't compete with a platform in another. For example, Instagram might have emerged as a competitor social network to Facebook—but Facebook bought it. This preserves Facebook's dominance in the social network arena and gives Facebook even more data. When platforms buy up-and-comers before they really gain traction, the result is less competition in the marketplace.
The answer here is antitrust enforcement. The Department of Justice and Federal Trade Commission have considerable powers to block these mergers or unwind them. But they have not exercised their authorities. Part of the reason is ideology. Since the 1970s, antitrust scholars, policymakers, and regulators have been enthralled by a neoliberal approach to antitrust that focuses on consumer prices, and is deliberately blind to the broader harms of concentrated markets. Part of the reason is institutional. The courts have been captured by this ideology as well, and they frequently narrow the scope of antitrust enforcement. But the final reason is a matter of political will. The antitrust agencies have been far too timid for far too long. Legislators and the public should hold them accountable—and if the agencies won't act, Congress should pass new legislation to guarantee a more vigorous enforcement regime.
A second set of problems—conflicts of interest—emerges when a company both owns the platform and has business lines that operate on the platform. Google does better when its review product is used more than Yelp's. Amazon does better when Amazon Basics sells more product than John Q. Public's small business. The platforms, therefore, have a motive to favor their products over their competitors'. They also have the means. The platforms collect data on everything taking place on the platform, which means they know what products are doing well and which ones aren't. The platforms also control placement on the platform, which means they control whether or not consumers see different products. This means that Amazon will know if John Q. Public's winter scarves are selling like hotcakes. If it wants, it can have Basics make scarves that are featured as a top search result—while simultaneously demoting John Q. Public's scarves to page 10 of the search results. This problem isn't theoretical. Some companies have complained that Amazon and Google have done exactly this to them. And in 2007, the European Union fined Google $2.7 billion because it gave special preference to its shopping comparison product over similar competitors.
Other problems emerge from this structure as well. A platform that owns a business can engage in predatory pricing, cutting the price of its own product in order to run a competitor into the ground. The company can then keep prices low or jack up their own prices, but the key is that the competitor is out of business. Platforms can also engage in “tying”—linking their products together. This is a competition problem for two reasons. First, products aren't on a level playing field when the platform, for example, can include native apps on its platform pre-installed and ready for use. In addition, competitors will have to offer every single product in order to compete. Imagine trying to compete with the integrated Google ecosystem—you'd have to offer search, mail, documents, maps, and everything else Google offers. This is partly why our tech behemoths keep getting bigger: To compete with one another, they each have to offer everything.
MORE THAN A CENTURY ago, policymakers confronted similar problems when dealing with the railroads, telephone, and telegraph, and other early network industries. Over time and across industries, three simple principles emerged for how to address this kind of challenge: quarantine, nondiscrimination, and regulation of rates.
For example, consider the law of innkeepers and common carriers, which is sometimes called the law of public accommodations. The idea is that some businesses are essential for society and commercial activity to function properly. As a result, they have a duty to accept any customers in a nondiscriminatory fashion. Innkeepers, as an example, were not allowed to discriminate against customers. If you think back to the early republic, this makes sense. Travel between cities was grueling and took weeks. Without a place to stay along the way, commerce and travel would come to a standstill. Innkeepers therefore had a chokehold on commerce—they could charge high rates and there was little travelers could do about it. The rule of nondiscriminatory access addressed both the power innkeepers had over consumers and the importance of their services.
As another example, think about the telephone industry more than a century ago. Phones are only useful if they are part of a network; multiple people need to have them. But putting in phone lines is expensive, so competition doesn't make much sense—at least it didn't when there were only landlines. If there were different proprietary networks, you wouldn't be able to call everyone, and each company would face huge costs in building phone lines to individual homes. The answer is to allow one telephone company to become a monopoly—to link everyone in one network. But monopolies are problematic because consumers are captive, and the monopolist can raise rates and lower the quality of service.
Public utilities regulation provided an answer. Instead of government provision of telephone service, the government would regulate. It would quarantine the monopoly, separating it from owning any other business lines. This is important because it prevents the monopolist from using its power to exploit businesses or individuals. In addition, the government would regulate the terms and rates of service to ensure they were fair and nondiscriminatory. This prevents the monopolist from exploiting its consumers. In the telephone industry, there was a third element: protection for the monopoly. The utility was given an explicit monopoly, so competitors couldn't steal the most valuable consumers. In return, the utility had to provide access to everyone.
Of course, this system didn't work perfectly. In the early 20th century, federal policy gradually required the Bell telephone system to divest non-telephone properties, required that it not discriminate against independent systems, and then recognized it as a monopoly subject to regulation of rates and terms of service. But technological advances and policy decisions eventually undermined this system. Still, the enduring principles of public utility regulation can still be applied to protect both consumers and competitors if the political will is there.
THESE BASIC PRINCIPLES—quarantine, nondiscrimination, and rate regulation—can be readily applied to tech platforms. Platforms should be separated completely from all of their subsidiary business lines that operate on the platform. This is a form of “breaking up” the platforms. Amazon shouldn’t be able to run both the marketplace and the Basics brand that sells on the marketplace. Mergers like Amazon-Whole Foods should have been blocked and should now be unwound. This would ensure that platforms can’t preference their own product lines, and it would improve free and fair competition.
Platforms should also be required to treat all users with fair, neutral, and nondiscriminatory terms. This would guard against platforms cutting special deals with some entities over others. Rate regulation might not be necessary at first (or at all). But if platforms, over time, adopt a business model that is premised on charging rates and if competitor platforms don't enter the fray, their monopoly status might ultimately require rate regulation to prevent them from exploiting captive users.
The Federal Trade Commission could act to implement this agenda. It can both act on a case-by-case basis with these principles in mind, and it has significant power to make rules to prevent unfair competitive practices. If the FTCis unwilling or unable to act, Congress could also pass a statute enacting these principles into law. A statute regulating tech companies could require the FTCto designate certain companies as “platforms” based on a multifactor test. The FTCwould have to consider whether the platform is an exchange or marketplace, the extent to which third parties use it and offer custom services or products, whether the platform benefits from network effects, and if it could deny access or discriminate against users. Designated platforms would then be required to divest all their other business lines and would be subject to a nondiscrimination duty. In addition, Congress should empower state attorneys general to designate platforms, given the limited resources of the FTC.
Tech platforms might seem like entirely new creatures that defy regulation. But in reality, there are a longstanding set of principles designed for firms with this structure. All that is needed now is action.
