Monday, September 28, 2009

What Maynard Keynes, James Dean, and Now Richard Posner All Have in Common

By Bob Hockett

The G-20 group of industrialized and industrializing countries have just met in Pittsburgh to consider coordinated next steps we might take to restore and maintain global financial stability. Meanwhile, the Angelides Commission – the U.S.’s latterday “Pecora Commission” – finally has begun meeting in Washington to ascertain “what went wrong” in financial markets these past several years or more. More or less simultaneously, House and Senate Committees chaired by Congressman Frank and Senator Dodd, respectively, have at last begun meeting to consider possible improvements to – perhaps even a complete “overhaul” of – our U.S. “system” of financial regulation.

In the lead-up to these interesting developments, a lively debate has emerged in the academy, the press, the “blogosphere,” and even the more popular media, not only about “what went wrong,” but also and relatedly about more fundamental questions. Many distinguished names have figured into the discussion thus far, two of the more recent being those of Paul Krugman at Princeton and John Cochrane at Chicago. (See recent posts on Leiter, for example: Lawyers in particular will be intrigued to know that even Judge Posner has now got into the act, with a brief book, several articles, and media interviews on the subject already to his name. (See this one, for example.)  He has also announced a recent conversion of sorts to a species of Keynesianism – though some who might wish to be cool in an inconsequential sort of way will be tempted to point out that what he now champions is not what the cognoscenti call “Keynesianism” – which Keynes himself disavowed because of its inattention to uncertainty – but “post-Keynesianism.” (The story of “Keynesianism’s” many prefixes over the past seven decades, from what Joan Robinson called “bastard-Keynesianism” on down through “new,” “neo-,” and “post-Keynesianism,” makes for a fascinating, if bemusing, tale better told elsewhere.)

Now as I say, the questions that have emerged in the current discussion are not only about what went wrong, but also more fundamental in character. They implicate not only regulatory policy, but also financial theory, monetary theory, macro theory more generally, and micro theory as well. “Maybe markets are not efficient after all,” we hear some now saying. “Maybe market actors are not rational either,” some also say. “Maybe it’s all just psychology,” or “irrational exuberance,” or Tobin-style “liquidity preference as behavior towards risk,” or some other such thing, we hear more and more often these days. Well, maybe indeed. But I’ve come to think that at least some of the discussion we’re hearing these days rests on a misconception or two, and that it would be helpful to highlight them even for those who are skeptical of market efficiency or trader rationality. Doing so should enable us better to recognize which proposed improvements to our financial system are apt to be most effective, and which of them less so. And relatedly, it should also facilitate the identification of reforms that believers and skeptics alike, where market efficiency and rationality are concerned, can join energetically in supporting.

So what are the misconceptions I think I detect? These: My impression is that there is a tendency among some participants in the current discussion – as well as among some policy-makers over the past decade or so – to equivocate, first, between two senses of "efficiency," and second, between two senses of "rationality." I suspect, moreover, that these equivocations might well be partly responsible for the pass in which we now find ourselves. Here's what I mean.

First, on "efficiency," the so-called "efficient capital markets hypothesis" (also “ECMH,” sometimes "EMH") familiar to finance is a pretty well corroborated conjecture concerning the speed with which the capital markets aggregate "information." But the information in question, it bears emphasizing, is not restricted to facts actually bearing upon firms' “fundamentals” or future prospects. (If it were, then informational efficiency would conduce straightaway to allocative efficiency, more on which presently.) No, the "information" in question also can include misinformation, disinformation, incomplete information apt to be revised or more fully filled in later, and so forth. The idea animating the EMH, in other words, is simply that trading has become sufficiently rapid and easy, and the financial markets so liquid, that securities prices very quickly impound and reflect the beliefs of all market participants -- even beliefs that in the end prove ill-founded, incorrect, only partly correct, or what have you. (It probably bears noting here, while we are at it, that it is the so-called "semi-strong" form of the hypothesis that has thus far been pretty well empirically corroborated by Fama and his followers.)

Moreover, and possibly more crucially, the EMH has nothing what ever to say about facts bearing upon firms' future prospects of which no trader as yet has any knowledge or inkling at all -- facts that would fall under the Keynesian (and Knightian) headings of "uncertainty" as distinguished from "risk." Where we not only don't know which face of the die will land up, but also don't know what values are etched on the faces of the die to begin with, we cannot speak of probability distributions at all, hence cannot compute even "expected" values, let alone actual ones. Hence we speak less of “risk” than “uncertainty” in these settings. It is in this realm of complete informational void that Keynes thought conventions, rules of thumb, and ultimately "animal spirits" to play most freely. That fact, in interaction with the next observation I'll make – on "rationality" -- seems to me likely to have played a role in our recent bubbles and crashes. More on that in a moment.

Now the equivocation on "efficiency" to which I alluded above is just this: There seems to be a tendency for some participants in the current discussion to conflate informational efficiency (on the EMH’s permissive understanding of "information" just elaborated) either with allocative efficiency, whereby capital flows toward its most valued uses; or just plain "efficiency" understood as a rough sort of synonym for "the state of being really cool,” or “as good as it gets." The unexamined thought seems in other words to be that, because the capital markets quickly take in and impound all value-pertinent information that there happens to be, they also immediately and unambiguously direct capital toward where it adds most value. But the moment one reminds oneself that "information" is actually employed in a much looser sense than the "fundamental-value-pertinent" sense for purposes of the EMH, one spots a gap between informational efficiency and allocative efficiency.

Keynes, of course, famously spotted this gap, and virtually all of the General Theory's observations and recommendations are in effect situated within it. Keynes's interest in the gap between informational efficiency and allocative efficiency also underwrites the most droll, in my humble opinion, of all of his many droll quotable recommendations (thought this one for some reason seldom is quoted): I refer to the recommendation that we seek, in his words, "the euthanasia of the rentier." A frequent and accomplished trader in his own right (born to a humble academic family, he left what had been a secret estate of over $30 million upon his death), Keynes ultimately came to think the securities markets so prone to what we might call not just "informational efficiency," but also "misinformational efficiency," that he ultimately advocated that we consider some form of what he called "the socialization of investment." To employ the more contemporary Fischer Black lingo, he came to think "noise" trading more the rule than the exception much of the time, and recommended in consequence something in the way of a government-guided "industrial policy" -- something a bit like what MITI famously did in Japan in its "miracle" period, and what China's government does now.

Now to the conflated two senses of "rationality" that I have in mind: It is sometimes, though alas, apparently not often enough, observed that multiple acts of individual rationality can aggregate into collectively irrational outcomes. The most familiar instance of this phenomenon is of course the tiresomely familiar "prisoner's dilemma," but there are many more. All so-called "collective action problems" and probably most so-called "coordination problems," one reckons, are instances. Yet many commentators seem to speak of "market rationality" and "trader rationality" almost as if they were one and the same thing, or at least as if multiple acts of the latter always aggregate to the former. But it just isn't so. To take a recent salient case in point, there is plenty of anecdotal evidence to the effect that hedge fund managers and like folk strongly suspected in recent years that they were trading under asset price bubble conditions, and knew in consequence that at some point the whole thing was apt to come a cropper. They kept trading anyway, however (so some have told me), in view of (a) their knowledge that the closer you draw to the bubble’s inflection point before exiting, the more millions you win, (b) insistence on the part of their customers that they stay in the game to keep winning those millions for them, on pain of withdrawing and moving their money to competing funds, and (c) their oft-harbored and not altogether insane hope that they might manage to get out comparatively early once the inevitable but indeterminately dated collapse commenced.

What, then, to make of all of this? Well, these observations on efficiency, rationality, and the actual experience of those market participants who have reported from the trenches suggest, to me anyway, the following parable as possibly capturing what happened in the years leading up to the autumn of 2008 and its sequel. Consider a modified version of that game of drag race "chicken" you likely saw in the old James Dean film, Rebel Without a Cause. We’ll imagine that in this game James Dean and the punks who have provoked him are once again drag-racing toward a cliff side as in the film. In this rendition of the game, however:

First, the drivers cannot see more than a few meters ahead, and therefore mainly look only sideways and backwards. They basically see just one another and those who are at most a few meters ahead. Nor do the drivers know antecedently how far away the cliff side is. It might be blocks, it might be miles, they simply don’t know; there is as yet no information about that. The confident drivers, however, who by virtue of prior experience know themselves to have quick reflexes, expect to be able to bail the moment they feel a loss of elevation up front, or see someone incrementally ahead of them beginning to dip or plunge down. And as in the game in the film, no one loses by losing the car. (This image of course roughly models the uncertainty, as distinguished from risk, that the drivers face in respect of the date of an asset price bubble's collapse. It also captures what Keynes would call “beautiful baby” trading – trading with a view to what others are doing, rather than to any perceived “fundamentals” – more on which below. Finally, it also captures a bit of the “things might be different this time” idea that sometimes attends asset price bubbles when they are initially catalyzed by bona fide new value-adding inventions or technologies like desktop computing, an internet, or efficient new financial instruments.)

Second, the drivers win more money with each foot or meter they traverse en route to the edge. At least that is so until people begin to go over the edge. Once a certain – though presently indeterminate – number of them do begin driving over, drivers must reverse course and race back in the direction of the starting line as quickly as possible, with the last to arrive at some indeterminate point roughly between one-half and three-fourths of the way back losing more of their previous winnings than those ahead of them. We might picture it as the ground’s crumbling progressively from the cliff side on back, to around halfway to three-quarters of the way back to the starting point, once the critical number of drivers have gone over the cliff. (This roughly models continued winnings as the asset price bubble continues to inflate, and losses once the inflection point has been reached and the markets commence to return to the point actually warranted by the new value opportunity that got the spontaneous pyramid process that is the bubble started in the first place.)

Third, the drivers also can place side bets on their own performances and those of their competitors, both with each other and with other, non-drivers. Some of them can do so, moreover, without having to prove that they’re good for their possible debts on the bets. And some even can borrow in order to bet, some of these on the strength of their reputations, others on the strength of their winnings already won along the way in the race. Drivers might do such things with a view to hedging some of their risk of ultimately losing in the drag race, or with a view to “levering” their winnings, or some complex of these and related intentions. (This roughly models the swap and levering arrangements that also played a key role in the recent collapses.)

Fourth, in addition to the insurance that takes the form of the hedging bets, there are a number of variably wide nets down below the cliff, set at varying heights. There’s also one very wide net at the very bottom, though it’s not clear how far down that is. No racers will die, but there might be real butterflies for many on the way down. Racers also know from earlier drag race experience that many racers – those that land in higher nets – will get to keep some, perhaps even much -- though it's not clear how much -- of their winnings even after driving over the cliff. For the drag race chicken economy is very important in this James Dean world, and its continuance rests on the continued presence of a critical mass of people who remain willing to play. (This of course roughly models implicit guarantees by a lender of last resort ready to bail out, in various amounts and with various conditions, some of those who go over the cliff when ever many of them do.)

Finally fifth, everyone knows that if only cars, but not drivers, go over the cliff, the drag race chicken economy will retain more wealth than if otherwise. They accordingly expect that there might be some intervention by the race-administrator prior to anyone’s reaching the cliff side. For they have seen this happen before, in times when prior administrators spoke about “leaning against the wind,” “taking punch bowls away when the parties begin to get good,” and so forth. (This models another role, additional to that of lender of last resort, that has sometimes been played by systemic risk regulators – notably the Fed in the U.S. case, when chaired by the likes of William McChesney Martin and, later, Paul Volcker.)

Now, here’s the interesting thing – the punchline – of this fanciful story: There does not appear to be any generally accepted canon of rationality pursuant to which these drivers can be deemed "irrational." Indeed, it might even be tempting to say some of them would verge on irrationality (perhaps even violation of fiduciary duty in some cases) if they did not play for a while, assuming that this form of play constitutes their occupation, which of course it does for many professional market participants. Sure, some lily-livered types will find the whole thing a bit too wild and chaotic to indulge in, just like they find roller coasters loud and scary. These folk will stay home and read Jane Austen novels instead. But thrill-seekers and even just plain rational money-seekers will not be univocally irrational to enter this race, or even to stay in it until … when?

In short, this entire parable is just another instance of the not unfamiliar observation that multiple acts even of bona fide individual rationality (or at the very least non-irrationality), occurring against the backdrop even of an admittedly informationally efficient environment, can, if the "information" in question does not include anything one way or the other so far as the end date of an asset price bubble (the location of the cliff side) is concerned, readily aggregate to a pathological -- a collectively irrational and allocatively inefficient – outcome. That outcome in our latest iteration of this “individually rational, collectively irrational” game is the 2009-09 market crash and ensuing economic slump. In prisoner’s dilemma games it is imprisonment. In ultimatum and dictator games it is mutual harm. In arms races it is wasteful expenditure. And so forth.

The point here, I think, is important to make even if we are skeptical about whether all were rational in the latest round of the familiar bubble and bust story, or whether markets were informationally efficient over the course of the past several years. My point, in other words, is not that the markets were all the time informationally efficient and individuals were all the time individually rational, but that they didn’t have to lack these characteristics in order for what happened to happen. I think it important to make that point because I suspect that a misconception concerning rationality's and efficiency's compatibility with bubbles and bursts might at least partly account for "our" -- through our this time ineffectual collective agent, that systemic financial risk regulator known as the Fed -- not having looked out for and acted to forestall our most recent asset price bubbles and busts.

My conjecture, in other words, is that many people like Greenspan first mistakenly thought that asset price bubbles -- or at least their detectability in advance of their bursting, in Greenspan’s case -- are incompatible with individual rationality and market informational efficiency. Then, reluctant to part with the latter presumptions (which again are not altogether lacking in empirical corroboration, and are hard to part with for modeling purposes in any event), they concluded that there was no point in looking out for and acting to head off the bubbles. Had these people been clearer about bubbles' and bursts' compatibility even with individual rationality and informational efficiency as these are actually understood in the technical literature, they might have helped prevent the whole sorry thing's happening -- or at any rate lessened its severity. That seems a lesson worth bearing in mind as we turn now to updating our methods of financial regulation both nationally and internationally.

If anyone's interested in reading more on this, here's the link to an early draft from last year of an article on the whole matter that’s soon to come out. There you can read about “Beautiful Babies,” “Ponzi Processes with no Ponzis,” “Minsky Moments” and more here.