The Bell Curve May Have Been Rung

Lies, lies and more lies! Now the venerable Bell Curve (also known as the Normal Distribution Curve) may be found out as just another false god-at least in the field of investing. You see, a normal distribution defines probabilities among a set of random numbers. But investing returns and volatility aren't random. If this is so, then the validity of using normal distribution may very well nullify many important statistics used in the technical analysis of investments. For example, the measure of risk in a stock or option is measured by its variance from the mean of its prices. If a stock has a 15% variance from its mean, that means that price will vary-plus or minus- 15% from its average price. If you have a stock with an average price of $40 and a variance of 15% the price will normally move between $36-$46 about 68% of the time using one standard deviation. But there exists the probability that at least 2.1% of the time, the price could move out of that range (two standard deviations) and .1% that it would move even further away from the mean ( Three standard deviations). That's if the data is random. But are stock prices random? We hope not. If that's true, then IBM could range between 0 and infinity. If that's the case, it's a total crap shoot. But management, analysts and investors would beg to differ.

There are some economists and statisticians who believe that stocks move more like what is encountered in electrical power distribution. The real market experience over the recent years demonstrates that stocks move out of the three standard deviation range quite frequently. According to normal distribution, that should only happen in one out of billions of events. The online bubble and other highly volatile stock implosion episodes have pointed out that many stocks have strayed out of normal distribution expectations.

According to the power distribution curve, its not that uncommon for data to move out beyond a six standard deviation point. What this means is that risks can be much greater than those depicted by the normal distribution curve.

I guess you could liken it to how society and Wall Street seem to accept the bogus statistic we call inflation. Everybody pays attention to this most important number and accept it as fact. Incredible! Do you really think that housing prices, energy costs, insurance costs, medical costs should be ignored or given a "place holder number" way out of sync with reality? Its preposterous. Yet, the emperor still has new clothes. There are statistics and damn statistics. Maybe the seeming randomness of unexpected price movements is much greater and more frequent than we expect because we are using the wrong tool.