Update (10/29/2020): Our discussion of interoperability measures was altered to acknowledge that, while this is contemplated in the leaked document, the mentions are lacking the specificity of other measures under consideration.
At the end of September, multiple press outlets published leaked set of antimonopoly enforcement proposals proposed for the a new EU Digital Market Act, which EU officials say they will finalize this year.
The proposals confront the stark fact that the Internet has been thoroughly dominated by a handful of giant, U.S.-based firms, which compete on a global stage with a few giant Chinese counterparts and a handful of companies from Russia and elsewhere. The early promise of a vibrant, dynamic Internet where giants were routinely toppled by upstarts helmed by outsiders seems to have died, strangled by a monopolistic moment in which the Internet has decayed into "a group of five websites, each consisting of screenshots of text from the other four."
Anti-Monopoly Laws Have Been Under-Enforced
The tech sector is not exceptional in this regard: from professional wrestling to eyeglasses to movies to beer to beef and poultry, global markets have collapsed into oligarchies, with each sector dominated by a handful of companies (or just one).
Fatalistic explanations for the unchecked rise of today's monopolized markets—things like network effects and first-mover advantage—are not the whole story. If these factors completely accounted for tech's concentration, then how do we explain wrestling's concentration? Does professional wrestling enjoy network effects too?
A simpler, more parsimonious explanation for the rise of monopolies across the whole economy can be found in the enforcement of anti-monopoly law, or rather, the lack thereof, especially in the U.S. For about forty years, the U.S. and many other governments have embraced a Reagan-era theory of anti-monopoly called "the consumer welfare standard." This ideology, associated with Chicago School economic theorists, counsels governments to permit monopolistic behavior—mergers between large companies, "predatory acquisitions" of small companies that could pose future threats, and the creation of vertically integrated companies that control large parts of their supply chain—so long as there is no proof that this will lead to price-rises in the immediate aftermath of these actions.
For four decades, successive U.S. administrations from both parties, and many of their liberal and conservative counterparts around the world, have embraced this ideology and have sat by as firms have grown not by selling more products than their competitors, or by making better products than their competitors, but rather by ceasing to compete altogether by merging with one another to create a "kill zone" of products and services that no one can compete with.
After generations in ascendancy, the consumer welfare doctrine is finally facing a serious challenge, and not a moment too soon. In the U.S., both houses of Congress held sweeping hearings on tech companies anticompetitive conduct, and the House's bold report on its lengthy, deep investigation into tech monopolism signaled a political establishment ready to go beyond consumer welfare and return to a more muscular, pre-Reagan form of competition enforcement anchored in the idea that monopolies are bad for society, and that we should prevent them because they hurt workers and consumers, and because they distort politics and smother innovation—and not merely because they sometimes make prices go up.
A New Set of Anti-Monopoly Tools for the European Union
These new EU leaks are part of this trend, and in them, we find a made-in-Europe suite of antimonopoly enforcement proposals that are, by and large, very welcome indeed. The EU defines a new, highly regulated sub-industry within tech called a "gatekeeper platform"—a platform that exercises "market power" within its niche (the precise definition of this term is hotly contested). For these gatekeepers, the EU proposes a long list of prohibitions:
- A ban on platforms' use of customer transaction data unless that data is also made available to the companies on the platform (so Amazon would have to share the bookselling data it uses in its own publishing efforts with the publishers that sell through its platform, or stop using that data altogether)
- Platforms will have to obtain users' consent before combining data about their use of the platform with other data from third parties
- A ban on "preferential ranking" of platforms' own offerings in their search results: if you search for an address, Google will have to show you the best map preview for that address, even if that's not Google Maps
- Platforms like iOS and Android can't just pre-load their devices exclusively with their own apps, nor could Google require Android manufacturers to preinstall Google's preferred apps, and not other apps, on Android devices
- A ban on devices that use "technical measures" (that's what lawyers call DRM—any technology that stops you from doing what you want with your stuff) to prevent you from removing pre-installed apps.
- A ban on contracts that force businesses to offer their wares everywhere on the same terms as the platform demands—for example, if platforms require monthly subscriptions, a business could offer the same product for a one-time payment somewhere else.
- A ban on contracts that punish businesses on platforms from telling their customers about ways to use their products without using the platform (so a mobile game could inform you that you can buy cheaper power-ups if you use the company's website instead of the app)
- A ban on systems that don't let you install unapproved apps (AKA "side-loading")
- A ban on gag-clauses in contracts that prohibit companies for complaining about the way the platform runs its business
- A ban on requiring that you use a specific email provider to use a platform (think of the way that Android requires a Gmail address)
- A requirement that users be able to opt out of signing into services operated by the platform they're using—so you could sign into YouTube without being signed into Gmail
On top of those rules, there's a bunch of "compliance" systems to make sure they're not being broken:
- Ad platforms will have to submit to annual audits that will help advertisers understand who saw their ads and in what context
- Ad platforms will have to submit to annual audits disclosing their "cross-service tracking" of users and explaining how this complies with the GDPR, the EU’s privacy rules
- Gatekeepers will have to produce documents on demand from regulators to demonstrate their compliance with rules
- Gatekeepers will have to notify regulators of any planned mergers, acquisitions or partnerships
- Gatekeepers will have to pay employees to act as compliance officers, watchdogging their internal operations
In addition to all this, the leak reveals a "greylist" of activities that regulators will intervene to stop:
- Any actions that prevents sellers on a platform from acquiring "essential information" that the platform collects on their customers
- Collecting more data than is needed to operate a platform
- Preventing sellers on a platform from using the data that the platform collects on their customers
- Anything that creates barriers preventing businesses on a platform or their customers from migrating to a rival's platform
- Keeping an ad platform's click and search data secret—platforms will have to sell this data on a "fair, reasonable and non-discriminatory" basis
- Any steps that stop users from accessing a rival's products or services on a platform
- App store policies that ban third-party sellers from replicating an operating system vendor's own apps
- Locking users into a platform's own identity service
- Platforms that degrade quality of service for competitors using the platform
- Locking platform sellers into using the platform's payment-processor, delivery service or insurance
- Platforms that offer discounts on their services to some businesses but not others
- Platforms that block interoperability for delivery, payment and analytics
- Platforms that degrade connections to rivals' services
- Platforms that "mislead" users into switching from a third-party's services to the platform's own
- Platforms that practice "tying"—forcing users to access unrelated third-party apps or services (think of an operating system vendor that requires you to get a subscription to a partner's antivirus tools).
One worryingly underspecified element of this list: interoperability rules for dominant companies. The walled gardens with which dominant platforms imprison their users are a serious barrier to new competitors. Forcing them to install gateways—ways for users of new services to communicate with the friends and services they left behind when they switched—will go a long way to reducing the power of the dominant companies. The "greylist"'s mention of data portability is a good start (that would let users leave one company for a rival), as is nondiscrimination in how APIs are accessed, but we'd welcome more specificity than "forms of interoperability" as the full extent of the proposal. Interoperability is a more durable remedy than passing rules to force those dominant actors to use their power wisely.
That said, there's plenty to like about these proposals, but the devil is in the details.
In particular, we're concerned that all the rules in the world do no good if they are not enforced. Determining whether a company has "degraded service" to a rival is hard to determine from the outside—can we be certain that service problems are a deliberate act of sabotage? What about companies' claims that these are just normal technical issues arising from providing service to a third party whose servers and network connections are out of its control?
Harder still is telling whether a search-result unduly preferences a platform's products over rivals: the platforms will say (they do say) that they link to their own services ahead of others because they rank their results by quality and their weather reports, stores, maps, or videos are simply better than everyone else's. Creating an objective metric of the "right" way to present search results is certain to be contentious, even among people of goodwill who agree that the platform's own services aren't best.
What to do then? Well, as economists like to say, "incentives matter." Companies preference their own offerings in search, retail, pre-loading, and tying because they have those offerings. A platform that competes with its customers has an incentive to cheat on any rules of conduct in order to preference its products over the competing products offered by third parties.
Traditional antimonopoly law recognized this obvious economic truth, and responded to it with a policy called "structural separation": this was an industry-by-industry ban on certain kinds of vertical integration. For example, rail companies were banned from operating freight companies that competed with the freighters who used the rails; banks were banned from owning businesses that competed with the businesses they loaned money to. The theory of structural separation is that in some cases, dominant companies simply can't be trusted not to cheat on behalf of their subsidiaries, and catching them cheating is really hard, so we just remove the temptation by banning them from operating subsidiaries that benefit from cheating.
A structural separation regime for tech—say, one that prevented store-owners from competing with the businesses that sold things in their store, or one that prevented search companies from running ad-companies that would incentivize them to distort their search results—would take the pressure off of many of the EU's most urgent (and hardest-to-enforce) rules. Not only would companies who broke those rules fail to profit by doing so, but detecting their cheating would be a lot easier.
Imposing structural separation is not an easy task. Given the degree of vertical integration in the tech sector today, structural separation would mean unwinding hundreds of mergers, spinning off independent companies, or requiring independent management and control of subsidiaries. The companies will fight this tooth-and-nail.
But despite this, there is political will for separation. The Dutch and French governments have both signaled their displeasure with the leaked proposal, insisting that it doesn't go far enough, signing a (non-public) position paper that calls for structural separation, with breakups "on the table."
Whatever happens with these proposals, the direction of travel is clear. Monopolies are once again being recognized as a problem in and of themselves, regardless of their impact on short-term prices. It's a welcome, long-overdue change.