If (as Walt seems sometimes to be suggesting in this post), realism is nothing more than the claim that national interests predominate in explaining international outcomes, then realism is theoretically very nearly vacuous. Moreover, the candidate ‘rival paradigm’ explanations are, under this broad definition, actually realist too. They have quite as much to say about state interests, and perhaps more to say about power relations than Walt does. If Walt has an actual realist explanation of what is driving European states apart – one that would presumably be rooted in the security dilemma or some other systemic phenomenon – it would be very nice to know what it is. He certainly doesn’t tell us about it in the post as it stands.
Indeed, an institutionalism without interests makes no sense. Why else create institutions?
But Farrell missed one strand I'd like to emphasize. If Walt really wants to use the EU as a test of rival paradigms, he's going to lose. In fact, he's already lost. Realists (generally) claim that institutions are merely extensions of state power, and thus do not constrain the behavior of (powerful) states in any significant way.* And yet it's clear that membership in the EU has caused states to behave differently than they otherwise would. Rather than devalue and/or default, Greece and Ireland are engaged in austerity. Rather than let them devalue and/or default, Germany and France are bailing them out in the short-run, and are considering a range of Euro-wide fiscal policies that would bring policy convergence on a number of issues. None of these actions would be taken if the EU didn't exist. So the institution is something more than an extension of Franco-German power, even if Franco-German power is reflected in the institution. Saying that the EU will not do anything that France and Germany don't want it to do is not the same as saying that France and Germany would behave the same way if the EU did not exist.
Note that at least some of the old-school institutionalists that Walt seems to be referring to wouldn't like an argument like the previous sentence. But I don't know of anybody who still hews to a that functionalist of a line. Certainly not Keohane. So I'm not sure who Walt sees as his foil. As Farrell notes, he references Andrew Moravcsik and Barry Eichengreen but neither is really an institutionalist the way Walt seems to be describing.
(Moravcsik's book on the EU has "State Power" in the title, and in this paper and this paper Moravcsik attacks neofunctionalist approaches to the EU. Moravcsik is definitely a liberal, but not the sort Walt seems to be referring to.)
But all of this ignores the real problem, which is that neither dogmatic realism nor dogmatic institutionalism helps us understand what's going on. Of course power and national interests are important, and of course the institution is important too. The maintenance of the institution itself is of national interest to member states. Neither is it a question of whether deeper integration will occur; it already has occurred. Never before had anything like the EFSF -- fiscal transfers from some EU states to others -- existed. Never before had the ECB intervened in bond auctions. The question now is whether this new integration becomes legally codified, or whether the cost of it becomes unbearable, and it is abandoned. Neither realism nor institutionalism helps us understand which is more likely.
So it's not just that Walt has the wrong answer, although he does, it's also that he's asking the wrong question.
*It's probably better to say that realists generally think of institutions as intervening variables, but not important independent variables. Like much else in realism this begs a question -- why would rational states waste time creating institutions if that's all the were -- but there it is.