Friday, August 29, 2014

The test is how you do out of sample

Robert Schiller won (half ) a Nobel for his anti efficient markets critique - the committee fearing a backlash had they simply honored Gene Fama as they should have. Yet Scott Sumner points out ( for the umpteenth time) that the critics can't define any theory that makes a better prediction than EMH. Which of course is the definition of EMH. He points out that any idiot can run a bunch of numbers together (climate modelers call your service) and get a bunch of correlations. The test, however is how your model predicts out of sample. And there the anti EMHers are crap. Gee, just like Gene noticed when he formulated his theory ere long ago.

So even the chap who won the prize for being the best anti EMH/anti-Fama critic produced a model that turns out to be crap. Or in other words had you used Schiller's model or any other model to invest in stocks you would have made less money than if you'd taken a random walk down Chicago lane. Which is the definition of the long time reigning champeen of the world "It is the Greatest!" Efficient Market Hypothesis.

I think the problem EMH has is branding. If you simply argued that no model or investing system can consistently earn above market returns because then everyone would adopt that system and bid up prices until the excess returns disappeared then no one would disagree. Yet that's EMH in a nutshell: ultimately nobody knows more than anyone else about which way the market is heading because the market always knows far more than any of us can. Just like Fred Hayek pointed out long ago (and also won a Nobel for Chicago for).

Next! (economist who wants to beclown themselves or possibly win a Nobel when the committee beclowns itself)

Much more at the link.

No comments:

Post a Comment