Mark Thoma and
John Cochrane applaud
Alan Greenspan for his call for higher bank capital requirements. While Mark does not agree with Greenspan on the assertion that higher bank capital obviates the need for Dodd-Frank style regulation, John applauds that portion of Greenspan’s latest as well. This did not stop John for having a little fun as the inelegant way Greenspan talked about the expected return to equity versus the expected return to assets:
Competition for equity capital should drive the risk adjusted rate of return for bank equity to be the same as for other businesses. If banks issue more capital, the raw rate of return to equity should decline. So should the variability (beta, risk) of that return. (Other things held constant, which may well be why the historical record is muddy.) In fact, Alan seems precisely to be making the banks' argument. They claim that the return on equity capital is independent of leverage. They have to pay (say) 10% to shareholders, but only 1% to debt holders, so debt is a cheaper source of financing. Banks claim that forcing them to issue more expensive capital will force them to raise loan rates and strangle lending. Which, curiously, Alan seems to be endorsing.
Admati, DeMarzo, Hellwig, and Pfleider made the same argument over four years ago:
We examine the pervasive view that “equity is expensive,” which leads to claims that high capital requirements are costly and would affect credit markets adversely. We find that arguments made to support this view are either fallacious, irrelevant, or very weak. For example, the return on equity contains a risk premium that must go down if banks have more equity. It is thus incorrect to assume that the required return on equity remains fixed as capital requirements increase.
After all – this is nothing more than the
Modigliani-Miller proposition. Leverage affects the expected return to equity but not the expected return to assets. But I still have a serious problem with claims that banks and other businesses would have the same return even if defined in terms of a return to assets. Banks likely have less operational risks than other businesses. I read all of this very early this morning before my first cup of coffee and fired off this at Mark’s place:
CAPM types would see this as an issue of what the unlevered beta is (see R.S. Hamada's 1969 and 1972 papers). For a lot of manufacturers, this beta is near 0.8 so the risk premia is 4% (see my Econospeak post on Cochrane's other ramblings). Banks are not the same sector and their unlevered betas are closer to 0.2 which is why their expected return to assets is generally only 1% more than their cost of debt. Banks and other businesses are fundamentally different businesses. If Greenspan thinks otherwise - why should we listen to him at all?
Then I remembered that one can find estimates of unlevered betas by sectors from
Aswath Damadoran who notes that the unlevered betas for the banking sector are half that of a typical business. I would hope that the former chairman of the Federal Reserve understands all of this and was just having a bad writing day.
You seem to be assuming that the CAPM is valid, so the required return on equity depends on beta. My impression is that the CAPM doesn't work at all and that average returns on equity are more related to own variance than to beta.
ReplyDeleteIn any case, I think you can cut the CAPM middleman out of the argument about the return on banks equity for current leverage. How about looking at the average return over the past few decades (I guess it will be much higher than the return on banks' debt).