The Myth of Systemically Risky Institutions

I had the pleasure of sitting next to my friend Peter Wallison at dinner last night and he reminded me of a great AEI Outlook he published last month on “The Error at the Heart of the Dodd-Frank Act.”  The question he poses is brilliant for its simplicity and seems to be exactly the right question.  So let me set it up.

The key idea of the bank bailouts and then Dodd-Frank is that there are some financial institutions that are “systemically risky.”  As I understand the theory, the basic idea is that some large institutions are so “intertwined” with others that the failure of one will lead to the failure of others as a result of a domino-type theory.  The failure of Lehman Brothers and the supposed subsequent market response is cited as the example that presumably demonstrates this.

For current purposes I will accept arguendo the conventional wisdom about the failure of Lehman Brothers spooking the markets, even though I find John Taylor’s critique of that argument to be highly persuasive.  But that’s not my focus here.

So here’s the testable hypothesis Peter frames: “Name one major institution that failed because Lehman Brothers failed.”  Or more precisely–name one institution that failed (or would have failed) because it was so intertwined with Lehman Brothers that it was unable to sustain Lehman brothers collapse.  Peter doesn’t see any; nor do I.

So what really happened, as Peter notes, was not a problem of intertwining, but rather a common shock that hit all of the banks at once, the souring of the mortgage market.  But that is not a problem of intertwining. That is a problem of a common shock, exacerbated by regulatory rules (such as Basel II) that encouraged homogeneity among banks’ asset-holdings and mark-to-market accounting rules that magnified the effect of the common shock.  But this has nothing to do with “too big to fail” or “systemic risk”–the common shock would have the same effect whether there were 10, 100, or 1000 institutions if they all have the same balance sheet structure and mark-to-market accounting rules that magnify the shock.

The reality, Peter notes, is that large financial institutions are so diversified that it is absurd to think that the failure of one (such as Lehman) could possibly take down others through their “intertwining.”

By contrast, if the problem is a common shock problem, this leads to an entirely different set of policy responses, such as eliminating regulations that encourage the uniformity in bank balance sheets.  Or as one person recently said it to me more succinctly, once rules are set up (like Basel II), they create uniform arbitrage opportunities as a result of mispricing (such as the mispricing of AAA-rated bonds under Basel) and so bank balance sheets tend to herd in the same direction which works great when things are going great but creates a huge amount of correlated risk when things turn bad.  Second, of course, we want to be wary of rules such as mark-to-market that then amplify the impact of those highly-correlated risks.

I don’t know that Peter has settled the question definitively, but it seems to me that he has asked the right question.