Four months ago, at the Jackson Hole conference, Kashyap, Rajan, and Stein proposed something of a financial catastrophe bond scheme for financial companies, in which they be required to meet a more stringent capital reserve requirement conditional on a financial catastrophe; the expectation was that this requirement might be met with an insurance mechanism of some kind that would cause the financial company to receive an equity infusion in that situation. They suggested a kind of secured insurance, but it seemed to me that a financial catastrophe bond (which they mention late in their paper) was both simpler and otherwise superior; the company would have a liability that, in the event of crisis, would disappear, converting to stockholders' equity.
The usual problem with insurance is moral hazard, and there's some of that here. There is some hope that, to a reasonable approximation, the financial company would not be able to instigate a financial crisis, but it does seem as though some behavior can be expected to reap worse whirlwind in case of financial crisis than otherwise, and that this sort of private behavior on a large scale intensifies systemic fragility. Historically financial crises are frequently preceded by bursts of poor loan underwriting; this is the sort of macroeconomic misalignment that finds itself in need of wrenching correction when the mania abates. A bank that finds itself choosing between loans that are relatively uncorrelated with the likelihood of disaster and loans that are feeding a mania (and are likely to go bad when it ends) will find the latter partially insured by the KRS scheme, and may well find them more relatively attractive than they should.
What seems perhaps safer to me is, rather than to have the liability disappear altogether, to have the liability convert to preferred equity. Concerns about an effective bank run or even an unwillingness of banks to lend should be alleviated just as well in this case as in theirs, insofar as the assets and liabilities pari passu and junior to senior unsecured debt work out the same. The common equity, however, is not shielded from the losses; the catastrophe bond claim remains senior to it. If equity drops dangerously low, the large subordinated class allows for a straightforward nondisruptive prepackaged bankruptcy, wiping away the liability only by handing the ownership of the bank over to the holders of what were originally the catastrophe bonds. It might, in fact, be worthwhile to give that preferred stock voting rights even before control of the company might pass to it; in any event where the liability is converted, it is likely that a good portion of the marginal dollar in assets belongs to that class rather than common equity. To avoid a bank's being permanently stuck with a capital structure that it may not like, it might make sense that such a preferred stock also be callable — it would be expected that calling the preferred stock would not be allowed by regulators unless the equity position of the company were comfortable.
Such a catastrophe-bond-converting-to-preferred-stock would surely be more expensive than ordinary unsecured debt, with which it would be pari passu if the institution fails idiosyncratically, but would likewise be less expensive than preferred equity. Which it falls closer to would provide a hint as to market perceptions of the correlation between the bank's risk and financial sector disaster risk. It should be somewhat cheaper than the original financial catastrophe bond variant. It also, insofar as management can be trusted to exercise its fiduciary duty to the shareholders, also largely eliminates the moral hazard problem mentioned in the second paragraph. It's a little bit more complicated (though not more complicated than the original secured insurance suggestion that KRS made), and it's more complicated in such a way that I expect there are complications I've missed. (One I've not missed so much as glided past is the disposition of dividends on that preferred stock.) Still, I think this is a step in a positive direction from what I've read before.