Raghuram Rajan offers a thoughtful comment on regulating big banks (hat tip Mark Thoma). The upshot is that bigness, interconnectedness, and proprietary trading are difficult to define, and hard caps are likely to be gamed and evaded. Rajan advocates more subtle strategies “such as prohibiting mergers of large banks or encouraging the break-up of large banks that seem to have a propensity for getting into trouble.” That strikes me as weak tea. I’d prefer that regulators explicitly target a diffuse market structure under which any bank can be let to fail, and require them to impose graduated-to-prohibitive taxes on a very wide variety of markers on bigness and badness in order to meet that target and eliminate incipient too-big-to-fail funding advantage. Successful banks should grow by division rather than accumulation.
Like Ezra Klein, I was especially intrigued with Rajan’s concluding paragraph:
In reality, proposing limits on size and activity is just an attempt to diminish the deleterious effects of another previous and now anachronistic intervention — deposit insurance. When households did not have access to safe deposits, deposit insurance made sense. With the advent of money-market funds, households gained access to near riskless deposits. Money-market runs can be eliminated by marking them to market daily; they do not need deposit insurance. To encourage community-based banks, deposit insurance may still make sense because small banks are poorly diversified and subject to bank runs. But for large, well-diversified banks, deposit insurance merely contributes to excess. We will bail out these banks anyway in a time of general panic. Why encourage the poorly managed ones to grow without market scrutiny by giving them deposit insurance along the way? Why not phase out deposit insurance as domestic deposits grow beyond a certain size? That would be far more effective in reducing risk than size or activity limits, and far easier to implement.
I’ve a bunch of nits to pick with this: money-market funds are not perfect substitutes for state-guaranteed assets, mark-to-market doesn’t limit aggregate risk or prevent runs (it just makes the race to the exit a bit fairer, and the collapse a bit sharper, as everyone learns when to get nervous at the same time). If large numbers of ordinary households, swayed by yield or convenience, hold funds at an uninsured Chase (JPM), Citi (C), or BoA (BAC) (cue the television commercials claiming “large, well-diversified banks” to be solid as Gibraltar), that fact would make each of those banks individually too big to fail even in the absence of a “general panic”. No system that expects sales clerks and schoolteachers to monitor financial firms is reasonable or politically sustainable. (If ordinary households can’t be persuaded to do without FDIC insurance, Rajan’s proposal would amount to an end of large depository institutions, which wouldn’t be a bad thing. Depositors would migrate to smaller banks, and large institutions would fund themselves as pure-play investment banks, unless regulated by other means.)
Of course I have a better idea. Rather than insuring banks, the government should insure depositors individually for losses they suffer on deposits at any FDIC-approved bank, up to a pretty high limit (say $1 million, indexed to inflation). Ordinary households would be unaffected by this change, as most families hold balances far less than the insurance cap. They could continue to deposit funds at the FDIC-approved bank of their choice without fear. Affluent households would no longer be able to play the wasteful game of multiplying FDIC insurance by splitting funds among different types of accounts and institutions. The affluent would be expected to monitor and help discipline the firms in which they invest. This is both fair and credible politically. It’s fair, because pushing wealth forward in time requires hard information work, and those who wish to push a lot of wealth forward (and earn interest on top!) ought to contribute to the effort. It’s credible, because ex post facto bailouts for underinsured depositors would be a hard sell when the underinsured include only wealthier depositors, who would not be reduced to penury but, at worst, to a level of affluence most households never achieve, simply by maxing out their government insurance.
Insuring depositors rather than banks wouldn’t, and doesn’t purport to, resolve the too-big/bad/sexy-to-fail problem. It would be a modest change that would eliminate some of the gaming that permits affluent investors to shirk their duty of discipline, and that would improve incentives to monitor by reducing the likelihood notionally uninsured depositors will be bailed out. But all of this matters very little in a world where even junior, unsecured creditors of some banks enjoy an implicit state guarantee.