Many investors and analysts use the Sharpe ratio (mean excess return per unit of risk) as a field-leveling measure of investment performance. Does this variable reliably indicate the best portfolio? In his brief January 2007 summary paper entitled “Beware the Sharpe Ratio”, Steve Christie applies the Generalized Method of Moments to test the portfolio discrimination power of the Sharpe ratio. Using two monthly data sets spanning 24 years for a set of multi-asset class portfolios created from index series and 18 years for a large group of mutual funds, he concludes that:
- The distribution of estimates for the Sharpe ratio has a variability many times larger than the variabilities of estimates for its components (mean return and variance of returns).
- Because of this relatively very large estimating error, Sharpe ratios for different investment portfolios (such as mutual funds) are very likely to be statistically indistinguishable at commonly used levels of significance, even for large data samples. There is simply too much noise.
The author notes two other important limitations of Sharpe ratios: (1) they are appropriate only when return distributions are normal; and, (2) they are not comparable when calculated for different investment periods.
In summary, the Sharpe ratio has such a high level of intrinsic variability that it is not a very reliable portfolio comparison tool.
