Stanford B-school prof Anat Admati has a piece in DealBook arguing that banks should have much higher equity cushions and that you should ignore banks' complaints that those cushions will lower ROEs.
We love this article only because she says something that should be engraved on the foreheads of anyone who wants to comment on bank equity:
Bankers use confusing language that presents equity, or “capital,” as a pile of money that banks must “hold in reserve” or “set aside” passively. This confuses capital requirements, which concern funding only, with liquidity or reserve requirements, which concern how funds are invested. It is you, investors, who hold banks’ equity or capital, not the banks that issued it.
In general the column is a summary and popularization of a longer paper comprehensively examining and rejecting various arguments against increasing bank equity requirements. Admati makes the basic Modigliani-Miller point that considering return on equity while ignoring capital-structure risk is a mistake:
Since investors must be compensated for bearing risk, higher leverage increases the required, or expected, return on equity. To judge whether a manager has created value, one cannot simply look at the return on equity; one must adjust for risk. A bank manager can attempt to reach a “target return on equity” by taking on more risk and by using more leverage, but this, in and of itself, does not create value. It does, however, increase fragility and systemic risk.
But, of course, equity holders like the increased leverage because it gives them option value, as they get all the upside while creditors and/or the government bear the downside. That said, equity holders too might prefer somewhat reduced risk of being written down to zero or almost zero a la Lehman and Bear.
We ran a toy model using 10 big banks of various flavors (JPM, BAC, C, WFC, USB, PNC, RF, STI, GS, MS) and got this pile of nothing:
So, erm, yes. Lots of noise here but this suggests that maybe even shareholders have concerns other than ROE.