Fat Tails in the Stock Market

Had an interesting conversation with Dan diBartolomeo of Northfield yesterday, touching on, among other things, changing thinking about fat tails in the stock market. With that in mind, here is a nicely done survey of his (in presentation form) of recent work in the area. Fascinating stuff.

Link to original version of presentation.

Related posts:

  1. Longest Run in Recent Stock Market History
  2. Statistics, Cot-Deaths, and Stock-Market Correlation
  3. Materials from Google Day
  4. Can’t Keep the U.S. Stock Market Down
  5. Update: Enterprise 2.0 — Sheer Hype by Desperate Men

Comments

  1. Karthik says:

    Wow, that was some extremely interesting material. Thank you!

  2. Stephen says:

    This is interesting, in a geeky sort of way. But, I wonder about the validity of a basic assumption implied in models: that the price is primarily a proxy for market and company information. Markets and the transactions by which markets are measured, are much too complicated to assume that prices largely represent the market context behind each transaction. I think other factors are important because in any given transaction, at any moment for each party to a transaction, it is a good time to buy AND sell. Each party acts within a unique context that not only includes interpretation of market and company data, but also other individual factors such as past and future holding period and investment horizon, investment objective (one party wants short term growth, and another wants long term income), taxes (some are taxed and others aren’t), and availability of risk capital. Also, when looking at data over a long period, market factors change that would have an impact on prices and efficiency of the market (similar to the affect that it is said steroids has on baseball stats). Together, the effect of other factors may amplify market events, but are hard to isolate in statistical models of the price movements. As a result, the predictive value of models, and particularly the measurement of market and company risk in these widely accepted ways may be fundamentally flawed.