A reader asked: “Do you have any insights on protecting one’s portfolio from the bad losses in the fat left tails of return distributions? Is there any data about stop-losses that actually work?”
The seminal work on non-normality of asset return distributions comes from Benoit Mandelbrot and Nassim Taleb. The empirical wildness of these distributions considerably devalues or invalidates the “normal” meaning of statistics such as mean, standard deviation and common significance metrics.
Nassim Taleb offers some vague guidance on avoiding wildly bad events. See:
“Surviving by Staying Out of the Fourth Quadrant”
“The Fourth Quadrant: No Realm for the Normal”
“The Black Swan: The Impact of the Highly Improbable (Chapter-by-Chapter Review)”
The guidance is not crisp because wildness is inherently not susceptible to much inference. It seems to be nature’s way of keeping us on our toes (or heels). See “Different Paths to the Same (Disconcerting) Destination?”.
For research on the effectiveness of stop-losses, see:
A reader commented:
“While I’m sure Nassim Taleb and Benoit Mandelbrot appreciate your conclusion that they were responsible for the seminal work in non-normal security returns, I think you are doing a disservice to finance professors who have been aware of the issue for at least four decades. Read, for example, ‘Q&A: Confidence in the Bell Curve’ by Eugene Fama where he writes that half his PhD thesis addressed the distribution of returns…back in 1964. Taleb doesn’t need more people inflating his ego any more.”
Perhaps “non-normality to the degree of intractability” is more appropriate.
Could it be that both practitioners and academia have a bias toward downplaying the degree to which asset return distributions may be intractable, because their work depends on tractability? See “Persistence of Diversity in Investor/Trader Beliefs” for related loose observations.
