As the global economy appears to come out of recession, it is natural to worry. Is it a temporary uptick? Will we have another 2008-type meltdown in the financial markets? If not that, what? For the moment, the economics profession has no answers. None. And that makes the ideas of Nassim Taleb “interesting.”
The Taleb critique
Taleb has sparked a firefight of sorts in the killing fields of academia. In his book, The Black Swan, Taleb gives mainstream economics a very bad grade for its empiricals. Taleb claims that economics cannot predict because (a) either what it attempts to predict is unpredictable; or (b) it predicts using a wrong probability model. While I have no quarrel with Taleb, it does not follow that economics is no longer “useful.” Some parts of the discipline should survive.
It is no surprise that Taleb builds on the ideas of Hayek. Both believe that the Law of Large Numbers does not hold in economics (and Taleb clearly states that this Law applies to games of chance, biology, physics, etc.). Hayek argued that there cannot be enough observations in economic matters that would allow the invocation of the Law of Large Numbers. Taleb argues from an altogether different tack. He showed that from a tremendous amount of data, a small but highly significant number of “Black Swans” disconfirms the validity of assuming that the error in economic models is normally distributed.
My initial instinct was not to believe Taleb. If you start with a uniformly-distributed random variable (the one random variable where it seems that nothing is known about its probability distribution), but add up a sufficient number of such variables, you end up with a sum (of random variables) which is normally distributed. That means, of course, that the Law of Large Numbers must hold. If you get a large enough sample, your error drops sufficiently enough for meaningful scientific work, i.e. prediction and/or control. Of course, if your error rate remains high, particularly if the error is measured from careful experiment, then you remain within the territory that Richard Feynman calls “cargo cult” science. Taleb therefore properly scorns the use of “data mining” as a means of “proving” a particular economic model.
But it turns out that there are other random variables that behave very differently from a normal distribution. Taleb calls them fractally-distributed variables, and they have the property that they are not disconfirmed by the mountain of data in financial markets, or by data in other areas within the territory of economics. The lesson for those into econometrics is that they have to go back to square one: drop the normal distribution, and study other areas of probability and statistics, particularly those based on the work of Benoit Mandelbrot, that appear to have a much more promising concordance with empirical reality.
What of economics survives?
The debate is not about the deterministic element of economics. For example, many economic “truths,” such as the Law of Demand or the Law of Diminishing Returns do not apply to uncertainties. Importantly, the classical economics of Adam Smith and David Ricardo survive. The deterministic elements of Keynesian economics survive. Surprisingly, the Austrian critique of Keynesian economics also remains viable, particularly its emphasis on the limits of knowledge available to central planners or “policy makers.”
In the area of financial markets, the ideas of Hyman Minsky and Richard Shiller survive. These ideas build on irrational and (somewhat) unpredictable behavior of market participants.
Taleb makes the forceful point that future technological advances cannot be predicted. If they could, they would no longer be “advances” and there can be no discipline today called “growth economics.” But it is possible to make an educated guess as to what environments are conducive to innovation. Here, Taleb will find good company with institutional explanations of economic growth (see my review of Helpman’s The Mystery of Economic Growth).
So, should we continue to worry about future crises?
Here’s my best guess. Social contagion, typically involving irrational exuberance or despair, is likely to remain a perennial of human nature. That seems to be the lesson from the work of Kindleberger, et al., as well as by new research in behavioral economics. How such contagions develop is inherently unpredictable. Though they seem to have a self-correcting or “cyclical” mechanism, a nascent recovery can still turn out to be a “false” rally.
Taleb’s advice is akin to that from Shiller. The rational market participant, to the extent that he is right, stands to profit from the irrationality of the others. But that strategy is easier said than done because one can never be sure of being the rational one. A much simpler approach is Taleb’s barbell portfolio. He suggests placing the bulk of assets in almost risk-free instruments like T-bills, and a small fraction in potential positive Black Swans. This would avoid needless worry about future crises. But if (almost) everyone went this way, interest rates would likely fall or remain low, and we would see very few stock market bubbles. And that’s a big IF.