Martin Wolf clearly worries about moral hazard:
It is desirable that institutions are prevented from exploiting explicit and implicit guarantees in order to make speculative investments of little economic benefit. The spectacle of businesses prospering from activities from whose consequences they have had to be rescued and from whose impact the public is still suffering is distressing.
the moral hazard explanation for what went wrong doesn’t really hold much water. In order to believe that the banks engaged in reckless behavior because they assumed that if they got into trouble, the government would bail them out, you have to believe not only that financial institutions thought it would be fine if their share prices were driven down to near-zero as long as they were rescued in the end. You also have to believe that the banks knew that what they were doing was reckless, and that there was a meaningful chance that it would wreck their companies, but decided that it was still worth doing because if everything went south, the government would step in.
Wolf concedes that some of Obama’s ideas are not workable:
Would it really be possible to draw and, more importantly, police a line between legitimate activities of banks and activities “unrelated to serving their customers”? If institutions were encouraged to lend to customers, would they be allowed to securitise and sell those loans? Would they be allowed to hedge the risks of lending? If not, why not? If yes, when does this become speculation?
And Economics of Contempt can easily find reasons that capping bank size is a bad idea:
Now imagine that we cap bank size at, say, $100bn in assets. What happens if a bank with $99bn in assets fails? The way the FDIC resolves failed banks smoothly is through P&As, but the only banks big enough to buy the $99bn failed bank would surely be over the $100bn cap if they agreed to the purchase. So the choice is effectively between (1) a disorderly liquidation by the FDIC, which would pose exactly the kind of systemic risks that proponents of capping bank size are trying to avoid; or (2) granting an acquiring bank (or banks) a waiver from the $100bn cap and proceeding with a P&A.
And, as my recent post on Shadow Banking has (I hope) demonstrated, a shadow banking system using bonds or other vehicles can achieve many of the same things as banking – but may fall beyond the realm of ordinary regulation. Does this matter? If such entities can always fail, then perhaps not. Not enough has been said on this topic. I personally would rather see less attention aimed at the features of individual institutions, and more at the system as a whole.