What if Internet centralization is good?

Mean US marketshare across data centers, DNS servers, web hosts, reverse proxies, and certificate authorities. (The big jump in 2018 comes from consolidation among certificate authorities).

The US dominates services core to the Internet’s functioning. Companies in US jurisdiction provision more than 50% of the marketshare in these services—a power the US can and does flex. Effectively, a single jurisdiction dominates Internet governance.

The puzzle then becomes: why is the Internet as open as it is?

Sure, “something something US norms and values.” But is that a satisfying (or even a believable) explanation?

Before we go any further, remember: a common assumption among folks who work on the Internet is that centralization is bad. It causes fragility. That makes the US’s dominance over the Internet dangerous. That’s certainly an assumption I’ve inherited.

What if that assumption is wrong? What if it’s the other way around: that the Internet is relatively open because a single jurisdiction dominates its governance?

Hegemonic stability theory

Hegemonic stability theory, or HST, postulates that international systems are more likely to remains open1 when one nation-state, the hegemon, dominates—as long as that country is “benign” (meaning other countries understand what it plans to do).

Here’s the logic behind HST: Collective goods are under-provided in diffuse ownership situations because their costs are internalized, and their benefits are spread out.2 In other words, when everyone has to chip in separately, maintaining collective goods doesn’t pay.

One solution is to have a central actor—a hegemon—that’s big enough to provide the collective good cheaply relative to the excess benefits it gains from providing it. In other words, a hegemon is big enough that everyone else can free-ride on it, and the hegemon still benefits more than it pays.

Academics have applied HST most rigorously to global trade. The paradox there is similar to ours: the US (and before it, Britain, and before that, the Netherlands) dominated global trade systems—yet those systems remained relatively open. Why?

HST’s argument: a hegemon provisioning an open system is best for everyone. As long as the hegemon remains “benign,” people gain from free-riding on the open system. Meanwhile, the hegemon gains from its dominant position in that system. A win-win.

HST has its critiques: it may overfit to an Anglo-American regime of global trade, and there’s no telling if a Chinese hegemon (for example) would be different. But it’s a theory with some empirical support.

Does the Internet benefit from a hegemon?

The question then becomes: is Internet governance a collective good that is conceptually similar to global trade?

There’s little doubt (at least in my mind) that the US has a hegemonic grip over the Internet. As widespread as the assumption is that a decentralized Internet is a healthy Internet, the Internet has remained relatively open throughout the US’s dominance over it. Is this openness despite its centralization, or because of it? If Internet governance is conceptually similar to global trade, the answer may be the latter.

Boy, am I going to get some angry emails from this one.

Remember, I’m not making a claim one way or the other here. My goal is to understand what makes an open, robust Internet. It’s a fundamental question, and we think we know the answer: decentralizing ownership, spreading it around across actors and jurisdictions, removes fragility. But can we prove that?

I’m also not questioning the fundamental design of the Internet (the Internet Society has nicely laid out a few critical properties of the Internet’s design). My question is how best to achieve these critical properties in practice. Again, does the Internet (as we understand it) more or less embody these critical properties in spite of or because of a hegemon’s dominance over its core infrastructure?

Nor—and this should go without saying—am I claiming that US hegemony is “good” in any normative sense. Even if Internet governance is a collective good similar to global trade, the fact that the US is the hegemon is the function of historical circumstance. It may change soon if it isn’t changing already.

Which brings me to the reason this question matters at all. If the Internet benefits from a hegemon, then the increasing threats we see to the Internet’s interoperability—think Russia disconnecting from the global Internet—could be a consequence of the US’s dominance breaking up a bit rather than a reaction to it increasing.

That would be a reversal in the presumed direction of causation. The dominant narrative of Internet fragmentation is that countries are opportunistically trying to exert local control as they become stymied by an increasingly US-led Internet.3 What if they’re instead projecting global power as the US’s grip loosens? That would certainly shift the way actors, particularly the US, think about what to do over the next hundred years of probable conflict.

How do you test this theory? That’s another post.

Hegemonic decline

In international trade, Britain was the hegemon until the 1880s. Then the torch passed to the US. Today, the Trump administration made it difficult to trust that the US would be a benign hegemon—it questioned the US’s ability to commit to staying benign, as the trade regime could flip on its head every two to four years.

The trouble comes when a hegemon declines. We’ve seen hegemons decline in global trade before. But the US has always been the hegemon of Internet governance. What happens when an Internet hegemon declines?

That’s a question for another day.

Thanks to Steve Weber for the feedback & conversations.


The term “stability” in HST is an ambitious and poorly-defined concept. The theory is best established in open trade; it’s more of a hegemonic “openness” theory.


Olson, Mancur, The Logic of Collective Action (Cambridge: Harvard University Press, 1965).