I just finished reading Alan Greenspan’s paper for the spring Brookings economics confab, The Crisis, and then a bunch of reactions around the econo-blogosphere. The paper is well worth reading – it’s time to get beyond the blame game and the mea culpas and mea non-culpas, in order to get to longer term regulatory reform. Of the blog reactions, the most interesting, I thought, was Greg Mankiw, who was a respondent on the paper at Brookings:
Alan proposes raising capital requirements and reducing leverage, but he suggests that there are limits to how much we can do so. If we reduce leverage too much, he argues, financial intermediaries will be not be sufficiently profitable to remain viable. He offers some back-of-the-envelope calculations that purport to show how much leverage the financial system needs to stay afloat.
When I read this part of the paper, my first thought was: What about the Modigliani-Miller Theorem? Recall that this famous theorem says that a firm’s value as a business enterprise is independent of how it is financed. The debt-equity ratio determines how the risky cash flow from operations is divided among creditors and owners. But it does not affect whether the firm is fundamentally viable as an on-going concern. It seems to me that, as least as first approximation, the logic of this theorem should apply to financial intermediaries as well as other types of business. If not, we need some explanation as to why.
Note that this is a different objection to imposing higher/firmer/objective/fixed minimum/what have you capital versus leverage requirements than is sometimes made – viz., that no one, least of all regulators, is in any good position to be able to determine the proper level, and that, therefore, the problem ought to be to ensure that risk falls so as to ensure that those that should care, do care. Greenspan is suggesting, rather, that without some minimum level of leverage that might well turn out to be crisis-inducing risky, financial firms will not have a sufficient level of profitability to remain in business. I might have misunderstood that reading the paper, so if (and only if) you have read the paper, feel free to correct me in the comments.
If my understanding of the paper is correct, I think my reaction would be … as compared to what alternative on a risk adjusted basis? It seems to me that the problem identified here is a “gotta get up and dance” issue – if on a short term basis, all your competitors are engaged in a certain level of leverage, and are hitting certain rates of return while, in fact, taking inefficiently high risks considered over a longer run, then, sure, you might not remain in business. If, on the other hand, leverage for all market players (at least in the taxpayer guaranteed sector) is constrained, sure, rates of return will be lower. But if the effect, on a long term basis, is to force return to take into account risk, and properly price it for all players, then capital will flee the sector and threaten, perhaps, to put financial firms out of business only insofar as capital wants to take greater risks for greater returns in other asset classes and investment opportunities.
But of course I might have misunderstood something fundamental, particularly since Greg Mankiw zooms in on something quite different, Miller and Modigliani. He asks what M&M would have to say about Greenspan’s argument – and maybe he is saying something similar to what I suggest above, although I have focused on short term versus longer term risk versus return:
I have a hunch as to where, from the Modigliani-Miller perspective, Alan’s calculations go awry. Alan assumes that the rate of return on equity must be at least 5 percent. But this number should be endogenous to the degree of leverage. If a bank is less levered, its equity will be safer. (It will be like a combination of today’s equity and bonds.) As a result, the required rate of return should fall.
Thus, Mankiw goes on, a less levered – indeed, wholly unlevered – bank should do just fine with a rate of return that reflects the decreased risk. Investors who want that kind of safety as part of their portfolio will gravitate to that bank. The problem, as I suggest above, is when risk and return in the capital market for all firms is skewed so as to favor getting up to dance in the short term. But then Mankiw raises the general question of the applicability of M&M to this case:
To put the point most broadly: The Modigliani-Miller theorem says leverage and capital structure are irrelevant, while undoubtedly many bankers would claim they are central to the process of financial intermediation. A compelling question on the research agenda is to figure out who is right, and why.
Actually, this seems to me to put M&M in a highly specific context – rather than being the classic question, does capital structure matter to the value of the firm? – this question is quite exact – do leverage and capital structure matter to the process of financial intermediation? If the question is financial intermediation alone – so-called narrow banking – I’d hazard that it still matters. Mankiw proposes a thought experiment with a wholly unleveraged bank – could it supply financial intermediation? Answer, presumably yes, at least if the playing field for capital is level, so that return reflects risk.
But isn’t the more difficult question, if one is following the symmetry of M&M, to ask, can a bank perform the functions of financial intermediation with something close to total leverage? Wouldn’t M&M suggest it should be able to do that as well? And hasn’t the meltdown suggested that, for some reason, it doesn’t work that way – rather than being symmetric as pure M&M in a pure world would suggest, in our world, firms, leverage and assets are asymmetric? To be sure, the financial intermediation part might have worked just fine, which, true, confirms Mankiw’s point. But the firm itself seems not to have worked – meaning, those arguing that in our world, capital structure and degree of leverage matter, and even matter with respect to a firm conducting financial intermediation, given that if the firm goes kaput on account of overleverage, the intermediation collapses with it? Which is another way to say, Mankiw suggests through the unlevered thought experiment that financial intermediation can be “unbundled” from the rest of a financial services conglomerate, and that seems right. But it seems equally right – and not consistent with “pure” M&M in a “pure” world – that you can’t successfully “bundle” them, at least not to the leveraged limit.
If that is so, how does one account for it within the broad framework of M&M? With some (lots of) trepidation, let me sound very much like a commercial lawyer here and note something that is always uppermost in my mind when I teach M&M. It is that what is called “debt” for these purposes does not look entirely and purely like debt, at least not when one looks at the actual contractual pieces of paper. Mankiw gives a concise explanation of M&M when he says:
The debt-equity ratio determines how the risky cash flow from operations is divided among creditors and owners. But it does not affect whether the firm is fundamentally viable as an on-going concern.
“On-going concern”? When that term is raised, at least to the lawyers reading the contracts, I think (and think we lawyers think), what is called “debt” involves, at the granular level, a huge number of financial options built into and around the instruments. It might be better thought of (and valued at the risk margins) as bundles of options ultimately enforceable, or not, in various ways through bankruptcy. Mostly they are exercisable at the option of the creditor, but sometime they are implicit rights exercisable by a borrower-owner that has been, e.g., insufficiently constrained by lending covenants, including the ability to leverage up in direct and indirect ways. Some of them are really exercisable in the discretion of the bankruptcy courts or creditors’ committees.
Collateral and margin calls, the spill on effects of mark to market rules, lots of things have the consequence of meaning that, whether the business is viable as an on-going firm, really amounts to the question (in reverse, so to speak) of whether the firm can be pushed into bankruptcy or insolvency (or close enough) by economic actors holding explicit but embedded, or implicit, options within credit instruments. Debt is an extraordinarily lumpy bunch of contractual thingies seen at the granular level, especially if you have to take into account what other debt instruments outstanding mean for this debt instrument. If the question of M&M is at the margins, and in particular a question specifically of the on-going enterprise – not just its “value” but the fundamental solvency-insolvency-bankruptcy law question of its “viability,” then those implicit and embedded options presumably rise in value, but in ways that are highly contingent and uncertain.
Add to that the related observation that financial institutions tend to ramp up by asset, acquiring each new security or bunch of securities, on a semi-smooth and semi-continuous curve of assets and returns and risks and leverage. But collapse – because of the way in which the law governing insolvency, creditors’ rights, bankruptcy, and so on operate, collapse is not continuous and by asset. Like a series of brick floors smashing into the next one below in an unsupported building in an earthquake. House of cards, quite exactly – built card by card, collapsing as a house. It is not simply an unwinding of its assets and leverage in the way that it flowed up the curve, but instead one (or more) institutional collapses. Rise by asset, fall by institution.
If one is operating on the margins of those risks, and if those risks are risks not of debt per se, but of options that are closely tied in but not quite visible, including sometimes the option of pushing the enterprise all at once, and not just some particular asset, over the cliff – then it seems to me that M&M implies something quite different. The valuation will look at lot more at volatility as we get closer to the margins, for one thing, and it seems unlikely that it will look very much like what classic M&M says for debt and equity, even if M&M is the starting model, which I think it is.
Meaning: If debt and equity are seen instead as bunches and bundles of implicit options, running sometimes in favor of the owners and sometimes in favor of the creditors, and if rise and fall are asymmetric as described above, then … M&M seems to me to be still applicable, but to mean something quite different when it is about financing by means of multiple options rather than classic debt and equity as the only two possibilities. And isn’t this really what bankers mean, or at least part of what they mean, when they say that leverage and capital structure are far from irrelevant? Lying behind the owner-creditor divide are different ways, sometimes highly contingent and indeed inconsistent, of allocating control, where “control” can mean anything from operational control to the ability to push the firm itself over the edge?
But look, I feel sheepish daring to comment on Greg Mankiw and Alan Greenspan, so feel free to correct me. It would help, however, if you have read the Greenspan paper. I’d be particularly interested in what people with experience in bankruptcy and creditors’ rights – including our own Todd Z and others – might say if they were commenting on Mankiw. Have I understood Mankiw and Greenspan or, for that matter, M&M all wrong at least in this context?