Governments around the world seem to be on a path to turn the leading tech firms into public utilities. Interoperability is likely to be one of the relevant tools for doing that, but history has shown that it would be incredibly difficult and costly.
The House Judiciary Antitrust Subcommittee is holding hearings on antitrust and Big Tech to move towards new legislation that would regulate online platforms. One of the key issues raised during the first hearing, held on February 25, was interoperability. To put it simply: When does a platform have to make available all or parts of its platform to another firm that wants to build its business around the platform? Rep. Joe Neguse (D-CO), who tweeted on this after the hearing, believes that there is bipartisan support to move forward to pass new laws requiring gatekeepers to create more interoperability with their platforms.
The above question is a well-known issue in antitrust and regulated industries, and history makes clear that forced interoperability is almost never easy to do. Governments around the world seem to want to turn the leading tech firms into public utilities and seem to assume that they can do so with minimal cost. That seems unlikely.
When it comes to how interoperability might be applied to the major online platforms, the draft Digital Markets Act of the European Union, released on December 15, 2020, would first create new obligations for what it terms gatekeepers. One of those obligations would apply to gatekeeper online search engines—Google isn’t namechecked, but is the obvious target—and would then require the gatekeeper search engine to “provide to any third party providers of online search engines, upon their request, with access on fair, reasonable and non-discriminatory terms to ranking, query, click and view data.” In other words, Google would be forced to sell its data to competitors at a price regulated by the government—forced wholesaling.
The United States tried that approach for local telephone networks under the 1996 Telecommunications Act. AT&T agreed to be broken up in 1982 into a competitive firm based on long-distance service and—so AT&T hoped—the convergence of communications and computers. Local telephone service, still thought to be a natural monopoly, was separated off into seven regulated companies—the “Baby Bells,” or RBOCs (regional Bell operating companies).
By 1996, Congress wanted to bring competition to local telecommunications, and the core idea was to do a virtual breakup of the local grid. The plan was to break the grid into pieces (unbundled network elements or UNEs) and let new competitors mix and match those pieces with whatever they wanted to bring to the table to create new local telephone competition. Of course, it wasn’t completely clear what the UNEs should look like—the FCC was to issue regulations—and then there was the question of price. The statute (helpfully?) provided that prices should include a reasonable profit, be nondiscriminatory, and be based “on the cost (determined without reference to a rate-of-return or other rate-based proceeding).” Again, the FCC was to interpret that and came up with something known to the trade as TELRIC (total element long-run incremental cost). A forward-looking cost, not a backward-looking cost.
The industry fought about this for the next decade. Cases went to the US Supreme Court twice (with decisions issued in 1999 and 2002) where the fights were over the scope of pieces that had to be made available and the prices for those. The FCC issued four sets of rules before the rules finally passed muster with the courts in 2006, and by then, of course, the technology had largely moved on to wireless and the landline phones at the heart of the 1996 regulations were a dying industry.
“Regulators and legislators should carefully consider the costs and benefits associated with greater interoperability obligations gatekeeper by gatekeeper and should not quickly assume that these obligations can be created in an easy, effective way.”
I can easily imagine having been in favor of the interoperability rules of the 1996 Telecommunications Act. From the outside, they look like smart, light regulations where there are real opportunities for new competition. I don’t think that the historical record matches that hope. The lesson for today is that regulators and legislators should carefully consider the costs and benefits associated with greater interoperability obligations gatekeeper by gatekeeper and should not quickly assume that these obligations can be created in an easy, effective way.
Today, governments around the world seem to be on a path to turn the leading tech firms into public utilities. Interoperability is likely to be one of the relevant tools for doing that. The ongoing dispute over whether app developers can avoid paying Apple a 30 percent fee for in-app purchases could easily be framed as an interoperability issue. Apple could easily make it possible for app developers to complete their in-app transactions with a third-party payment system like Visa or Mastercard, but Apple doesn’t do that because in-app sales are the point at which Apple gets app developers to pay for all of the benefits they get from distributing their apps through the App Store.
There are pending antitrust cases related to this issue around the world, and the European Commission and the UK Competition and Markets Authority are looking into that. To what extent Facebook needs to make its social graph available to third parties also could be addressed through interoperability regulations. The Cambridge Analytica scandal could easily be seen as interoperability gone bad. In a bid to become a platform, in 2007, Facebook opened up the social graph so that others could build on that. That is what interoperability looks like: Cambridge Analytica exploited the social graph to export customer data in ways that were regarded as problematic. Interoperability creates attack surfaces for outsiders to take advantage of.
As those examples suggest, interoperability regulations will operate differently from gatekeeper to gatekeeper, but there are likely to be common issues over the scope of the forced wholesaling that will arise, as well as disputes over pricing. And that is without taking into account how gatekeepers might reset their businesses in the face of these regulations. As I have written about here before, Spotify seems to think that it should be able to get free distribution on the Apple app store. It doesn’t pay for downloads for advertising-supported apps and now it hopes to pay credit-card style fees (2-4 percent) for in-app sales of its premium service. That means it would pay nothing for distribution through Apple or for access to the customer base that Apple has created. Apple might very well reset pricing on the app store to include charges for free app downloads (above some number of downloads) to make up for losses if the 30 percent fee is cut back.
All of that is to say that not only will interoperability regulations confront fierce fights over the scope of the obligations and pricing, the actual effects of the regulations have to be assessed against how business practices and technology will continue to evolve. Given the stakes, enormous resources will be devoted to the fights over these rules and we should be reasonably confident that the benefits of this path exceed the costs. Sitting here right now, that is far from obvious.
