As always when I talk politics, I remind folks that this blog is my personal blog, which I had well before I joined my current employer Public Knowledge. I’ve been commenting on Presidential campaigns since well before I joined PK, and I don’t run any of this stuff in front of my employer before I publish it.
Friday March 8, the Presidential campaign of Elizabeth Warren, not to be confused with the actual office of Senator Elizabeth Warren (D-MA), announced Warren’s plan for addressing the tech giants. Warren has been drawing attention to massive concentration in industry generally and tech specifically since well before it was cool, so the fat that she is out of the gate with a major proposal on this early in the 2020 campaign is no surprise. Nor is it a surprise that her proposed plan would end up breaking up, in some significant ways, the largest tech platforms.
What makes Warren’s contribution a potential game changer is that she goes well beyond the standard “break ’em up” rhetoric that has dominated most of the conversation to date. Warrens proposal addresses numerous key weaknesses I have previously pointed out in relying exclusively on antitrust and is the first significant effort to propose a plan for permanent, sustainable sector specific regulation. As my boss at public knowledge Gene Kimmelman has observed here, (and I’ve spent many 10s of thousands of words explaining) antitrust alone won’t handle the problem of digital platforms and how they impact our lives. For that we need sector specific regulation.
Warren is the first major Presidential candidate to advance a real proposal that goes beyond antitrust. As Warren herself observes, this proposal is just a first step to tackle on of the most serious problems that has emerged in the digital platform space, the control that a handful of giant platforms exercises over digital commerce. But Warren’s proposal is already smart in a number of important ways that have the potential to trigger the debate we need to have if we hope to develop smart regulation that will actually work to promote competition and curb consumer abuses.
I break these out below . . . .
In addition to promising to appoint vigorous antitrust enforcers who will reexamine the major tech mergers of the last 10+ years to see if any of them can/should be unwound, Warren proposes designating certain entities as “platform utilities.” (As you can guess, I’m not wild about calling these utilities, but let that pass.) Warren proposes that companies that offer an online marketplace, online exchange, or platform for connecting third parties, and have annual global revenue of $25 billion would be designated “platform utilities.” These platform utilities would not be allowed to both operate the online platform and compete on the platform at the same time. Most people call this “structural separation,” although in telecom law we’ve sometimes called this sort of arrangement “cross-ownership prohibitions.” (For example, not being allowed to own both a daily newspaper and hold a broadcast license in the same market is the “newspaper/broadcast cross-ownership limit.”) In addition to structural separation, these digital utilities would:
- Have a requirement to deal with others in a fair, reasonable and non-discriminatory manner;
- Not be allowed to share or transfer data to third parties.
Smaller companies, with annual global revenue between $90 million and $25 billion in annual global revenue, would be “platform utilities,” subject to the requirement for fair, reasonable and non-discriminatory dealing and the prohibition on sharing or transferring data to 3rd parties, but could still participate on their own electronic platforms.
The two examples Warren sites are Amazon, Google Adwords and Google Search. Warren says these three would be full fledged “platform utilities” (Amazon and Adwords are markets. Search connects third parties.) So Amazon would need to spin off its own retailer, Amazon Basics. Google would need to separate both Google Adwords and Google Search, both from each other and from all other Google businesses. Warren would allow enforcement by federal regulators and state attorneys general. Warren would also create a private right of action for injured parties. Penalties could run as high as 5% of annual revenue.
I’m not endorsing this proposal as the end-all-and-be-all of platform regulation. As I’ve previously observed, regulating the broader “digital platform” space is insanely complicated and will take a long time to get right. Indeed, Warren herself notes this is just the beginning of the discussion on how to ensure both competition and consumer protection in what has become an increasingly important part of our lives. Furthermore, as I noted back in 2017 with regard to privacy , we can expect this to be an iterative process. Even if we try to pass a comprehensive piece of legislation (similar to the Communications Act) for digital platforms, we can expect we will need to tweak and fine tune it over the first few years. And, of course, while this proposal is strong for a plain word proposal, there is always the little matter of translating this to actual legislative language.
But none of that takes away from the importance of what Warren has done here. I want to make very clear that Elizabeth Warren’s proposal does a lot of very smart things; it provides an excellent place for us to start the debate we need to have on regulating digital platforms. Many of these are things I (or others) have proposed either with regard to digital platforms, or in related contexts (such as privacy), are part of this proposal. But I want to focus specifically on what makes this so important and so smart structurally.
The Proposal Actually Defines Something That Can Be Regulated.
As I’ve argued before, one of the big differences between sector specific regulation (like the Communications Act) and generally applicable law (like antitrust) is that sector specific regulation requires you to define the sector. This has a lot of advantages over generally applicable law, since it lets you target the issues of the sector with precision and allows you to create an enforcement agency with specific expertise. In addition, generally applicable laws may not adequately capture the nature of the harms we want to address. This is particularly true for antitrust. Even setting aside the increasingly narrow way federal courts have interpreted the role of antitrust, it is not the end-all-and-be-all of regulation. There are a lot of things it doesn’t do, or would do much less well than a targeted sector specific regulator would. Most importantly, even if we could use existing antitrust to break up all the platforms people think have gotten too big, it doesn’t stop the next big platform from coming around. Sure, “more vigorous enforcement” will help. But if there is a systemic problem, such as an economic feature that drives this particular segment to concentration quickly and in ways that elude traditional antitrust detection, then we’re going to have a problem sometime in the future.
Rather than focus on specific companies, the Warren proposal makes a stab at defining something that could actually get translated into legislative language as a clearly definable market. Sure, it’s still somewhat vague (how do we define an “online marketplace?” Is it different from just the online extension of a traditional retailer? Should it be?). But the entire point is to have an actual debate to inform policy, which means starting somewhere. Every sector specific statute has to start with a general definition somewhere, then get refined to the point where you can actually put it in legislative language. Even then, we know there will be edge cases that require judgments by a regulator seeking to enforce the law and ultimately affirmation or rejection by a reviewing court. That’s how law works.
But the discussion has to start somewhere. I’ve made my pitch for a somewhat broader definition of digital platform, as have others, But Warren is the first actual sitting member of Congress (and Presidential candidate) to shift the debate from “these specific companies are too big” to “here is a clearly definable sector of the economy that has systemic problems, is a major part of our lives, and needs a systemic set of solutions rather than just going after some specific companies.” This will, I hope, get us past the “Google, we hates it Precious” mentality and start us on the more productive discussion of “what sort of things require general, sector specific oversight to work properly?”
The Proposal Gives Clear Lines of Demarcation For Its Proposed Structural Separation, And Imposes Rules To Keep Them Separate.
In the debate over whether or not to break up various online platform companies, I have occasionally mentioned something I call “the starfish problem.” Starfish have amazing powers of regeneration. Tear up a starfish and not only does the original regrow, but a sufficiently large piece may grow into a whole new starfish. Suppose you unwind Facebook, WhatsApp and Instagram. Rather than three competitors, you may well have three dominant firms in their own silos (social blogging, messaging and picture sharing) rather than one dominant firm — not much of an improvement. Or, given the nature of how online competition seems to work, we might have fierce competition between the three until one company once again combines all the features of the current Facebook and assumes dominance.
Any good break up works to keep that from happening. That’s why they take so long in antitrust (in addition to being generally disfavored under current antitrust law as a “means of last resort”). The final break up of AT&T was about a decade in the making, and took a great deal of time, effort and expense to properly implement. It’s success is an important milestone in telecommunications and an important lesson in the value of breaking up a company that controls a vital sector of the economy. But it is also a caution that it takes a lot of work to do it right. We have seen what happened after the 1996 Act took the guardrails off and actively encouraged incumbent monopolists (telcos v. cable) to compete with each other. The result was not the robust competition hoped for (unless you live in the exciting fantasyland of the current FCC), but a winner-take-all fight where cable emerged as the winner and now dominates broadband, video and voice as an unregulated oligopoly. (And I, for one, welcome our new starfish overlords . . .)
Warren’s proposal highlights a fairly clear line of demarcation between offering the online marketplace and participating in the online marketplace. Again, I don’t want to understate the difficulty in applying the distinction here in practice. The history of structural separation in telecom shows us that incumbents are very inventive in how they try to insert loopholes and circumvent the safeguards imposed by structural separation. Indeed, the fact that Warren had to clarify that Apple’s App store would qualify as a “platform utility” under her proposed definition shows that even this fairly straightforward and clear line of demarcation will require debate and work to translate into legislative language and actual practice.
At the same time, I don’t want to minimize the importance of this accomplishment either. Warren is the first person of significance in Policyland to go beyond “break ’em up” to “here is how we ought to break ’em up and keep them from coming back together.” That is potentially a huge step forward in moving this debate from general complaints about size to an actual plan of action. It provides a focus and hopefully a deeper investigation into what lines of division make sense to resolve systemic problems in the sector, and how to solve these problems on a systemic, sustainable basis.
Warren’s Proposal Recognizes That Dominance Matters In Sector Specific Regulation.
One of the major arguments against sector specific regulation is that it invariably favors large companies because only large companies can meet the compliance cost. Generally, I find this argument way overdone, but it is not wholly without merit. Fortunately, it is fairly easy to solve, as Warren’s proposal demonstrates. Take into account the size/dominance of a particular firm when applying the regulation. Warren requires “full” structural separation for the largest platforms – those with over $25 billion in revenue. Like AT&T when broken up, you either get to be the long-distance company or the regulated local phone company, but we can’t let you be both because it was simply impossible to effectively police a combined incumbent dominant in long distance and holding a monopoly on the local network through which all long distance companies must flow to reach customers.
But Warren recognizes that if we want to have platforms compete with each other, we need to make it relatively easier for new entrants. Indeed, we should recognize that a new entrant in the platform marketplace might have to “seed” the platform with retail to attract customers, as potential retailers on the platform might not see the value in colonizing the new platform which starts with zero customers (this is one of the big problems in industries with network effects, bit I digress). Warren proposes to exempt platforms with low enough global income entirely, and to require a softer form of structural separation (operating through a separate affiliate) for intermediate sized platforms. This provides a balance between creating what amounts to a regulated “platform utility” monopoly (similar to the local telephone network/Baby Bells in the AT&T break up) and allowing platforms to compete against third parties on the platform they control.
Again, I can’t say at this stage whether Warren has gotten the exact dollar figures right, or even if global revenue is the right measure of dominance. (I have proposed an entirely different measure for dominance — Cost of Exclusion (CoE) — which moves away from revenue figures and market share entirely.) But Warren’s proposal smartly recognizes that size and dominance matter, and that sector specific regulation can make these distinctions in ways that make sense and are broadly applicable.
Warren’s Proposal Recognizes the Importance Of Multiple Enforcement Remedies.
Finally, and of considerable relevance to a number of other debates, Warren recognizes that the policy should be enforceable not simply by federal regulators, but by states and by private rights of action. I will point out that statutes such as the antitrust laws and the Communications Act have a similar scheme of permitting consistent state regulation and enforcement, and allowing for private rights of action so that individuals who believe that they are harmed by a violation of the law can control their own fate and take individual action to protect themselves and remediate the harm done.
Given what we have seen of the privacy debate, where industry and numerous legislators have argued for a single national policy and a single national enforcement agency (often the FTC, which has enough trouble enforcing the authority it already has), it is both refreshing and exceedingly important to see a powerful voice such as Elizabeth Warren defend state level enforcement and private rights of action. An honest debate should no frame the question of preemption as a requirement to justify more than one tool of enforcement. It should carefully consider the question of whether to preempt, and how much to preempt, on the merits.
Where Do We Go From Here?
As Warren herself recognizes in her blog post, this only starts the debate rolling. I will add that, characteristic of Warren, she has started the debate with strong medicine (structural separation). I expect that to receive a fair amount of pushback — not just on the details but whether structural separation is really the right call for this sector.
We should welcome this debate, not dismiss it. Structural separation, while a powerful and effective tool when properly applied (if you are using the Internet, you are enjoying the fruits of structural separation as mandated by the FCC’s Computer Proceedings), is not undertaken lightly. Congress certainly has the power to create structural separation, but doing so will require a strong record that not only explores the proper lines of demarcation and appropriate mechanisms for implementation, but justifies why such a strong remedy is needed. Those who feel the case is obvious should not bristle at being asked to explain it and defend it from the industry that will resist it. It is this debate that sharpens and refines legislation and allows us to reach consensus as a society as a whole. That doesn’t mean conceding to industry or seeking some sort of “compromise” that industry finds acceptable — although that sort of back and forth is always part of how the sausage ultimately gets made. Rather, I mean persuading enough people generally that this is the right approach.
The enormous contribution of Elizabeth Warren’s proposal is therefore not so much in the substance (who knows if this is what we settle on as the right answer), but for smartly laying out a proposal that addresses the right questions. Until now, we have engaged in a tail chasing exercise over defining the problem. Warren’s proposal puts critical questions on the table with some proposed solutions to provide substance to the debate. We no longer need to remain in academic/philosophy mode identifying the inherent problems with sector specific regulation. We can start reacting to the specifics that Warren has proposed, and demand that counter proposals address the same questions of line drawing and policy trade offs.
Going forward, I would like to see legislators (and other potential Presidential candidates, of course) either endorse Warren’s approach or put forward their own solutions so that the debate we need can finally happen. Even those utterly opposed to any new regulation should be required to make more than a general case with platitudes about “first do no harm” and “regulatory humility” and other phrases that have now become the catechism of the Church of the Gods of the Marketplace. Warren has laid out a case for regulation and a proposed solution that provides answers to the basic questions around specifics. Time for everyone who is serious about this question in Policyland to up their game and respond in kind.
In short, we still don’t have the right answers, but we at last have a framework proposed by a serious policy wonk with the power to create real legislation (even as a Senator, let alone if she becomes President). After years of handwringing on one side and broad statements about breaking up companies on the other, we actually have a vehicle for a very real policy debate. It won’t get solved this Congress (the Telecommunications Act of 1996 started life as the Telecommunications Act of 1990), but we may finally be past the debate on “should we even have a debate on regulating digital platforms” to debating “how should we regulate digital platforms.”
Stay tuned . . .