Improving on Morningstar's Ratings: Moving Beyond Past Performance
Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
In a December 8, 2009 Advisor Perspectives article, Robert Huebscher published evidence that, between 2004 and 2009, Morningstar’s “star” ratings were poor predictors of future fund returns:
“Our analysis found that Morningstar’s ratings lost virtually all of their predictive ability when measured over a full market cycle. … Advisors might as well flip a coin to decide whether to move to a fund with a rating that is one star higher.”
Huebscher’s results are consistent with other studies that show stars are poor predictors of future fund performance. As Huebscher observes, if advisors were looking for ways to make better investment decisions during this particularly volatile period, stars would have been of little help.
The reason stars fail is that they are based on a 20/30/50% weighted average of three-, five-, and 10-year historical fund performance. I have estimated the predictive power of these weighted historical returns using a survivor-bias-free sample of all active US equity mutual funds spanning the period from January 1980 through June 2008, which constitutes half a million fund-month observations. The fraction of the equity funds’ subsequent one-year return volatility explained by these weighted historical returns is .002 (i.e. r-squared). An essentially zero r-squared over the last 30 years is right in line with Huebscher’s finding that stars are not predictive of future fund performance.
Past returns provide little or no help in choosing the best fund going forward, and Morningstar’s stars are the best known example of this failure.In spite of this evidence, the vast majority of advisors still use past performance when making investment decisions. Furthermore, many believe that longer performance track records produce better predictions. After all, it is argued, superior skill will reveal itself over a longer time period and therefore highlight which funds are best. Unfortunately, the evidence shows just the opposite: predictive power declines precipitously when going from one-, to three-, to five-, to 10- year performance numbers. In fact 10-year returns provide the lowest predictive power.