Streetwise Professor

Barney Frank and Timmy! Geithner have reached an agreement on a bill intended to reduce systemic risk.  I emphasize “intended to,” because it is by no means clear that it will have that effect.

The WSJ reports that one element of the proposal “would require financial firms with more than $10 billion of assets to pay for the unwinding of a collapsed competitor.”  The Journal continues:

During the recent financial crisis, the government bailed out several large firms at taxpayer expense, fearing their collapse could sink the economy or financial markets. Policy makers hope the new rules will discourage companies from growing too large and would cushion the blow if they collapse.

A future flashpoint: Which companies will have to pay for the collapse of one of their competitors? Initially, the costs of safely unwinding a failing firm would be borne by the Federal Deposit Insurance Corp., based on borrowing from the Treasury Department. But under the proposal, the FDIC would seek to recoup the money from banks with more than $10 billion of assets — a category that includes roughly 120 U.S. banks. The FDIC would likely be able to also recoup money from broker dealers, insurers and possibly even hedge funds.

So let me see if I get this.  Banks that know that taxpayers will bail them out grow too large and take on too much risk.  But banks will be more cautious, won’t grow so large, and won’t take on so much risk if they are assured of being bailed out . . . by other banks?  Huh?  From the perspective of the incentives of the managers of a given financial institution, does it matter where the money to bail them out comes from?  Not bloody likely.

WIth 120 banks (and perhaps broker dealers, insurers, and hedge funds as well) on the hook for the failure of one (or more) of their number, this loss sharing rule will provide little additional incentive for banks to monitor the riskiness or size of other banks, or take actions that would reduce the growth or risk taking of other banks.  Why should a bank incur the costs of taking actions that penalize the risk taking of another bank when (a) it internalizes the entire cost of those actions, and (b) any benefit accruing to such actions is shared by 118 other banks?

Put differently, if there is a too big to fail problem, it is because benefits are privatized and losses are socialized.  Under the present framework, if the ad hoc, seat of the pants approach deserves such a name, losses are socialized broadly among the taxpaying public.  Under the proposed alternative, they are socialized more narrowly, among the claimants of the banks (and other entities) that are subject to being dunned by the FDIC.  Socialization occurs regardless, and as a result, the proposal does nothing to affect the incentives of financial institutions to gamble on somebody else’s dime–because they can still gamble on somebody else’s dime.  Only the owner of the dime changes. Hence my earlier skepticism that the bill will do little to reduce systemic risk.

By itself, this provision will affect the distribution of losses from the failure of a large financial institution, but it is unlikely to affect the frequency or magnitude of those failures because losses are still socialized.  It does not affect incentives in a meaningful way; it affects how the consequences of those incentives are shared.

The relevant question is whether the share-among-120 banks (and others) rule spreads risk more efficiently than the share-it-among-the-taxpayers rule.  I have no strong initial reactions to this choice, though it does seem problematic to adopt a rule that would impair the ability of financial institutions to lend during periods of market stress, as this rule would.  A likely outcome is that the Fed (or the Treasury) would feel compelled to provide assistance to these financial institutions if the additional cost of paying for the failure of other institutions impairs their ability to lend.  Relatedly, it seems odd to fight the alleged effects of financial contagion by creating a new channel whereby the effects of the failure of one institution are imposed on others.

In sum, the first priority of an effort to reduce systemic risk should be to attack TBTF by reducing the socialization of risk.  This piece of the proposal doesn’t do that.

Rather than a serious effort to address systemic risk, this seems to be populist boob bait, a response to popular outrage against taxpayer banking bailouts, rather than a serious attempt to address TBTF.  Not a surprise, and quite understandable, but not a major improvement of incentives.