Standard Risk Measurement Sucks (So Does Prediction)

Mike Shell reminds us all about poor risk measures…and the ugly stuff that happens when relying on faulty logic. An excerpt from Pensions & Investments magazine gets 50% right (the problem) and 50% wrong (the solution):

Value-at-Risk and other popular risk measurements are typically effective during calm markets but often times are quite ineffective during challenging periods such as market shocks. More predictive models need to be used now that market volatility is high.

Predictive models as a solution?


Jack White ain’t singing about VAR measures, but he does describe the end result if you rely on them or in the alternative “predictive models”:

“…stick a knife inside me, and twist it all around.”
“…grab my fingers gently slam them in a doorway put my face into the ground.”

If you want those feelings, go VAR and or prediction. Guarantee you will find all kinds of pain…in your account.

Note: Song is great.

One thought on “Standard Risk Measurement Sucks (So Does Prediction)

  1. Mandelbrot was a mathematician with no axes to grind in the markets…as pure an objective observer as could be found.  And he concluded after looking at overall market data that it has “infinite variability.” 

    That’s bad news for the analysts, economists, CNBC’s, Bloombergs, etc. etc. who have made a multi-billion dollar industry out of airy nothings.  If the public ever wakes up to the fact that they are being conned, these guys are all out of a job.

Comments are closed.