Nassim Taleb’s Warnings About “Fat Tail” Dangers Were only 20 years Late

Posted on July 15th, 2009 by admin in Quant

Nassim Taleb’s brayings about the massive failure of Wall St. portfolio managers to account for the likelihood of Black Swans (unlikely, high cost unforeseeable events) have made him very famous. His point is valid. But he is 20 years late. Really.

We have been wondering why this technical issue in portfolio management has seemed so hard to solve. As far we can understand, one only has to change the definition of risk in the standard portfolio optimization algorithms and you can account for the Black Swan Events. Pick any measure that seems reasonable. (It’s been clear that using the variance of returns in Modern Portfolio Theory doesn’t capture the really big bad events.) Then do the numerical math and programming. It takes some very straight forward rocket science (OK, we indulge) to produce a more realistic approach.

So we were amazed to discover this problem was in fact solved for the first time in 1987. Even more amazing, one can buy off the shelf software (see This) to perform all the math computations. This whole approach even has a name: “Post-Modern Portfolio Theory (PMPT).

It turns out the Pension Research Institute at San Francisco State University developed the mathematical algorithms of PMPT that are in use today. These methods provide a risk measure that directly accommodates investors’ preferences for upside uncertainty over downside uncertainty in their returns. Many papers and programs have been developed since. But it is supremely obvious that this approach requires much more work and analysis than the original MPT required. Business schools didn’t teach this either. Thus, generations of advisors had no clue that the high risk events everyone is so sensitive to post-catastrophe could be at least numerically handled in portfolio designs.

There was a Devil’s Pact by investors, whose innumeracy avoided discussing issues of risk definition and who wanted to avoid math explanations, and their enablers, those investment advisors who themselves do not understand the basics of quantified portfolio design, let alone an approach as computationally oriented as PMPT.

But armed with this knowledge, we can progress to now use PMPT.

Leave a Reply

More News