Peter Wallison has an important Op-Ed in the Wall Street Journal last week (AEI link) titled "Dodd-Frank and the Myth of Interconnectedness"
The chain of bankruptcies is one of the central myths of the financial crisis. A owes money to B, B owes money to C, C owes money to D. If A fails, it wil result in a chain of bankruptcies where B, C, and D fail too.
As Peter points out, it simply did not happen. We had a run, not a chain of bankruptcies.
A (L actually!) failed, investors noticed that B, C and D were invested in many of the same things, and stopped lending to B, C and D; B, C and D also stopped lending to each other. Everyone tried to buy Treasuries. Since the banks were funding themselves by overnight debt, they were supremely exposed to such a run.
Getting it wrong matters. To produce a sensible regulation of the financial system, it helps to have a vaguely coherent idea of what happened -- and what did not happen. The Dodd-Frank/Fed approach seems to be "we don't know what happened, really, so we'll just regulate everything that moves."
On the chain of bankruptcies theory, and as directed by Dodd-Frank, the Fed is getting ready to monitor and regulate all links between "systemically important" institutions, and implement regulatory limits on their cross-exposures. I reviewed this in an earlier WSJ oped, and pointed out how fairly hopeless the effort is.
The central idea of Dodd Frank, which the Fed is now endorsing and implementing, comes down to this: no "large," "systemically important," "interconnected," or whatever (nobody knows what these words mean) will be allowed to fail. The Fed will be looking over their shoulders the whole time.
That approach will necessarily mean protecting them from competition. How else do you make sure they're "strong," and will never get in trouble? And it's only a matter of time that "policy goals" get enmeshed with "regulation." See last week's agreement whereby banks agreed to lower principal amounts on one group of homeowners, as "settlement" against their paperwork problems with another completely unrelated group. The Fed is pushing hard for banks to "do more" for housing... you can see where this will go fast.
If, instead, we had a run, as I and Peter believe, that sends you thinking in a completely opposite direction. In my view, it means we need to get rid of the institutional focus -- protecting specific "systemically important" institutions -- and instead focus on the run-prone assets. Peter seems to lean more to the "common shock" problem. But either line of thought is a long way from what's going on in the dungeons of the Fed.