Monday, June 25, 2007


The REcapitulator has reviewed many prospectuses in his day.  Most of the statistical studies included therein are utterly depressing.  

The most glaring problem is the sample period of returns.  Mutual funds, as a rule, cannot help but fit their returns to the time period that produced the greatest period of compounded returns...fancy that.

And do not get the REcapitulator started on "risk" parameters such as beta or the sharpe ratio, both of which fall prey to difficulties with autocorrelation.

Since "Beta" (again:  WHY do we insist on using Greek variables in Finance?) is a measure of the covariance between an asset and the "market" in terms of the variance of the "market", autocorrelation, or the tendency for time-series data to be MORE dependent on adjacent values than values from the past, rears its ugly head.

The REcapitulator has never seen good mutual fund reports that include all the correct statistical studies...but that would not be good for marketing, now would it???

No comments: