To reduce anticipated bank failures to a tolerable level, we need either to increase capital or reduce risk.
It seems ironic that there have been very high-level, very detailed efforts to set rules that require much more capital: Basel III. Seeing this, one might imagine that prior tightening of capital requirements, Basel II, had produced providential or at least positive results. But it did not. We had the worst banking crisis since the 1930s — despite implementation of Basel II. Why? Could it be that this happened because banks took progressively greater risks, outstripping the expanding protection provided by increased capital? Is it possible that they could do this again, with Basel III? Does more capital invite more risks? If so, risk-based capital, which requires more capital for some types of assets, can't provide a solution.
The contrast with money market funds is stark. They have no capital at all. Their balance sheet and off-balance-sheet activities are tightly restricted. Yet there was only one significant failure among them: Reserve Fund. This failure resulted directly from the bankruptcy of Lehman Brothers. Lehman, as a systemically important bank holding company, probably would not have been allowed to fail in today's regulatory environment.
We might well think harder about appropriate limitations on money fund assets, but "more capital" is unlikely to be the right answer here.
The converse of this is that if banks could engage in any activities they chose, then no reasonable level of capital could possibly be sufficient to prevent substantial failures. Their range of activity obviously must be and has been constrained, though until now generally in ways that relate to permissible assets, rather than to limitation of diverse contractual obligations such as derivatives.
We need to give much more attention to reducing the riskiness of banking by tightening these restrictions rather than by inflating the capital cushion.
Should banks be able to own or even originate mortgages with 100%-plus loan-to-value? If not, some reasonable limitations need to be set — as in the past. Should banks be able to extend unsecured credit to consumers with a credit score of 500, or with an effective APR of more than 200%? If not, an appropriate line will need to be drawn.
It will at best take years to implement, probably unevenly, some version of Basel III. Limitations on bank assets and activities can be applied much faster than that. The potential benefit from and justification for such limitations does not require international universality.
For better or worse (and I have yet to hear anyone else say it's worse) riskiness of banks is no longer just a function of the liquidity and the fluctuations in value of the assets they have on their balance sheets. Banks also enter into contracts, such as derivative contracts, which not only engender counterparty risks, but also multiple exposures situated outside the balance sheet. They engage in activities, such as high-volume asset sales for securitizations, which create new potentials for liability — such as put-back risks. An industry which previously accepted 100% of the risk on mortgages it originated now seeks to evade keeping even 10% of that risk. Servicing of assets for others likewise entails potentially disproportionate risks.
All these exposures should be limited both qualitatively and quantitatively. Could we accommodate the multiple loss potentials revealed since 2007 just by assessing capital requirements that vary with asset class or even off-balance-sheet items?
The concept of risk-based capital requirements is impeccably sound in its simplest forms. But, who will administer the application of increasingly complex definitions? How is gaming against them to be limited? Will the capital calculations become as complex as the Internal Revenue Code — and can we place much reliance on such a kluge?
Current experience also shows how concentration risks can accumulate both at the level of the individual institution and for the financial system as a whole. This happened with various forms of commercial real estate lending, for example. It highlights the obvious fact that if more equity were required from developers and owners, as a result of limitations on bank loan structures — then less capital would be needed by the banks themselves. Surely this would be preferable.
The same principle can be applied to many other types of bank lending: Instead of pouring more capital into banks, substitute the equity capital of would-be borrowers and counterparties for some of the debt capital recently provided to them by banks. We can call that risk retention, or alignment of interests, or avoidance of moral hazard. Leaving aside assumption of credit risk by the government for social purposes, if residential total loan-to-value had been limited to 80%, we would be hearing much less about underwater homes — and resulting strategic defaults.
