For the past year, IGP research has been exploring the interaction of competition policy and cybersecurity/privacy concerns, particularly as it pertains to the large platforms. We have shown how privacy and security have become product differentiators in the digital economy. As large platforms such as Apple and Google compete by finding ways to better protect their users’ data, however, some businesses complain that this increases the dominant platforms’ market power or harms competitors. And indeed, in some cases it may – raising tough policy choices: more competition, or more security?
The European Commission’s Digital Markets Act (DMA) has intensified policy debate over the competition policy – cybersecurity tradeoff. The DMA contains a new requirement that messaging services, like Meta’s WhatsApp and Facebook Messenger or Apple’s iMessage, must interoperate with smaller developers. The interoperability requirement for messaging services has generated serious concerns from the industry about whether it would undermine the end-to-end encryption (E2EE) that leading messaging services such as WhatsApp have implemented, as well as other security and anti-spam measures.
The technical impact of interoperability on E2EE is being actively debated by the Global Encryption Coalition, which will probably issue a statement on the topic. What’s clear from their deliberations is that forced interoperability could undermine E2EE in various ways and, while it may be technically possible to retain E2EE after an interoperability mandate, it would take a lot of costly and time-consuming work at the protocol level to make it work.
Messaging and Competition
We’ll leave the technical debate aside for the moment, and focus on the competition policy issues. Two important points have not been widely discussed or debated.
- The competition policy case for mandated interoperability in messaging services is a very weak one;
- Once the first point is understood, one can appreciate how the absence of interoperability can be pro-competitive. It allows service providers who improve the security and privacy of their messaging systems to gain exclusive benefits from attracting more users.
Both points are based on a dynamic understanding of competition, however, not the static, neoclassical “perfect competition” models that the European Commission seems to be relying on.
Are network effects a problem in messaging?
An interoperability mandate is supposed to overcome network effects. Network effects occur when the value of a product depends on how many other users use the same service/good. A crude understanding of the network effect sees more users as always better. Not quite true. A larger user base certainly increases the probability that someone you want to communicate with will be available, and makes it easier to coordinate joint consumption. But the other users have to be people that you are actually interested in communicating with. In some cases, adding users can decrease value; e.g., by generating too much traffic or by adding spammers, fraudsters or just plain idiots to your universe of joint users.
Network effects create barriers to entry primarily when there are demand-side economies of scope. A demand-side economy of scope occurs when users have to pay something to participate in multiple, competing networks. Think of a mobile phone subscription. You’re probably paying around $40 a month for it. If mobile networks didn’t interoperate, you’d have to pay another $30-40 a month to get access to anyone not on your existing network. If there were three competing, unconnected networks your costs for universal access would triple. Interconnection and interoperability generate significant demand-side economies of scope by enlarging your access universe at a small (or no) extra cost.
High switching costs can also create barriers to competition. Switching costs are sunk costs incurred by any investments you made in your original service, plus any additional costs required to move. Number portability was an effective way to reduce switching costs in mobile phone service. If there was no number portability, changing your network operator means changing your number. Think of all the costs you would have to incur to move from one provider to another: updated contact information, possibly new business cards or stationery, altering records with your number at banks, utilities, work, and phone applications, informing your contacts, etc. Switching costs also affect any attempt to move from Facebook to some other social media platform. It’s difficult if not impossible to transfer all the postings and links. In those cases, interoperability can eliminate those extra costs.
So what about messaging? Here interoperability generates almost no demand-side economies of scope. This is because the cost of adding a new messaging app to your phone is effectively zero. Messaging apps are generally free to download and use. The only switching cost is the small amount of time it takes to download and register for another app.
Switching costs for messaging services are also minimal; indeed, due to the absence of subscription costs one doesn’t really need to “switch” from one exclusive provider to another, one simply adds a new one. It’s easy to inform your main communication partners, or indeed any partner of interest, that you are using X messaging app and they should join you to benefit from mutual use. The “network effects” argument for messaging apps is extremely weak at best, possibly non-existent. The only real barrier to offering a new messaging service is that any start up needs to get public attention through marketing and an initial critical mass of users. Not easy, but not a special problem requiring regulatory action. If the new service actually has something to offer, it can attract users. And that leads to our second point.
Models of competition
Forced interoperability would make it unnecessary for a start-up to attract new users on their merits. They would instantly have access to anyone and everyone in the world using WhatsApp or iMessage. Is this a good thing?
In some ways, exclusivity of the user base is good for privacy. Under current conditions – without interoperability – service providers compete to attract users by offering features (e.g., better encryption or privacy, spam blocking) that benefit users. They can also implement unique policies regarding such things as security, content moderation and the like. If those features are exclusive to their platform, they benefit from attracting users who like those policies, and their rivals don’t. Compulsory interoperability may actually undermine those incentives. Assuming that it is even possible to retain E2EE under conditions of compulsory interop, the most relevant privacy feature would be homogenized; i.e. there would be a lowest common denominator level of E2EE for all services. Competition in that dimension would cease. Methods of spam control and content moderation might also be affected.
The DMA includes a threshold that exempts smaller messaging services, but that exemption misses the point: WhatsApp and iMessage already contain a huge chunk of the world’s internet users. Any newcomer of any size could obtain access to them merely by requesting it. What happens to the incentive to add features that meet user needs, such as new privacy or security benefits?
Making messaging services compete to add users and/or avoid losing them, in other words, is a good thing for the consumer and for competition. A supposedly pro-competition regulatory measure that eliminates the need for this is not doing the world’s internet users any favors. Further, the forced standardization of the protocols required to retain E2EE among competing apps would institutionalize and probably stagnate key infrastructures of the digital economy when it is still emergent and growing. Historically, compulsory, state-mandated standards tend to eliminate dynamic, Schumpeterian forms of competition and lead to stabilization on a few large, regulated incumbents. The most positive forms of interoperability on the internet have been achieved voluntarily and in a dynamic way, due to the mutual benefits of compatibility.
Big Tech Derangement Syndrome?
The competition-security tradeoff will be an important policy problem going forward. Think of app store gatekeeping and how limiting that power will also affect the intermediaries’ ability to weed out abusive or harmful apps. That is a debate we need to have.
But the DMA’s interoperability mandate is an easy one to reject. The case for consumer benefits from mandated interop are, as explained above, almost nonexistent. At the same time, the DMA would weaken systems competition among platforms and possibly weaken E2EE.
We are not advocating this, at least not yet, but if the EC is truly concerned about the market power Meta allegedly obtains from WhatsApp, it makes a lot more sense to consider a structural remedy that divests WhatsApp from from Meta and makes it (once again) a free-standing service. This is a much cleaner solution that serves their stated competition policy aims much better. Competition for exclusive users would continue but one messaging service would no longer be privileged by Meta’s deep pockets. Separating iMessage from Apple would be far more complex, and the rationale for it more questionable, but probably not as thorny as trying to force the entire digital economy into new standards and protocols for encrypted messaging. Mandating interoperability in a space that has no major demand-side economies of scope makes no sense, unless the real goal is just to “get” big tech.