Almost everybody pontificating about financial regulation seems to be recommending increased capital ratios, increasing the ratio of firm’s capital to assets. It is also true that financial regulation around the world includes minimum capital ratios. The reasoning seems to be that if you increase a financial institution’s capital, it is less likely to fail, but that is actually not true.

The problem with regulating capital ratios is that the capital ratio is not a sufficient measure of the firm’s risk. The firm’s risk is a function of both the capital ratio and the riskiness of the assets it holds, and that poses serious regulatory challenges.

A firm has a preferred risk profile. If you limit the capital ratio, the firm can achieve its target risk profile by increasing the riskiness of its assets. This means that if you want to regulate the riskiness of a firm using a capital ratio, you must also control the riskiness of its assets. No problem there.

Well, actually, there is a problem. Controlling the riskiness of a firm’s assets is impossible.

If regulation is to have any impact, the firm has a big financial interest in having riskier assets than the regulators would prefer—an interest that is only made larger by free (too-big-to-fail) or under-priced (FDIC) insurance—and the regulators have only a limited interest. Therefore, the firm will succeed in working around any constraints, and they will always be at least a step ahead of the regulators. The regulators are just no match for a determined firm with a big financial incentive.

Even if regulators could control the risk of firms’ assets, they still could not control the risk of the firm, because risk is dynamic, changing over time. In my banking days, we had a saying: In good times you don’t need capital; in bad times you will not have enough capital.

The problem is that in bad times, everything goes bad. We say the covariance goes to one. It may seem farfetched that a New York grocer, a Los Angeles realtor, and a Chicago lawyer would have economic prospects that were related, and that is true in most states of the world. This is the argument for geographic diversification. However, in a big event like we saw in September 2008 the grocer, the realtor, and the lawyer were all in the same boat. The probability of each of them failing to meet obligations has gone up, but more importantly, the probability of all of them defaulting has gone way up.

There must be a better way to regulate financial companies.