I read The Black Swan last month (it was OK), in which the author (Nassim Nicholas Taleb) repeatedly chastises the financial industry for using the Normal distribution for modeling and forecasting. I read in disbelief, thinking that there was no way that financial experts would use tools that make such simplifying assumptions. After all, a) financial data is not normally distributed, and the Normal distribution does not have “fat” tails to allow for rare events. I was wrong. Naked Capitalism explains why the Normal assumption is bad for financial data better than I can.
I’m feeling just a little nervous, because I taught introductory probability and statistics to a cohort of undergraduate engineering students last semester. We briefly discussed when it is appropriate to use certain distributions, so I hope my students absorbed the important messages.
Punk Rock Operations Research 
January 6th, 2009 at 2:16 pm
I don’t understand the details, but my impression is that although the financial folks did assume thin-tailed distributions when they should not have, it’s not as simple as it sounds. I don’t think they directly modeled market fluctuations as a normal distribution. At least I hope not. I agree that would be incredibly naive. My impression is that the normal distribution assumptions were buried a little deeper than that.
Nassim Taleb exaggerates and over-simplifies when he’s speaking to a general audience. (I’ve seen him speak to a room of statisticians and make his case using more carefully qualified language.) He may have intended some of the things he said in Black Swan to not be taken exactly at face value.