'Star Ratings Are Not Predictive'

Except that they are.

John Rekenthaler 18.09.2013
Facebook Twitter LinkedIn

Say What?
I read a few weeks ago in "Picking Mutual Funds: What Works and What Doesn't" (a Huffington Post blog by Dan Solin) that the Morningstar Rating for funds contains no information about the future. The article stated:

 “Star Ratings Are Not Predictive
Recent research by Dimensional Fund Advisors casts serious doubt on the predictability of past star ratings on future returns. Dimensional looked at all 14,822 U.S.-based mutual funds and sorted them by their Morningstar rating at the end of 2007. It then plotted the performance of these funds over the subsequent five-year period based on each fund's percentile rank in its respective peer category. If a high star rating was predictive, you would expect high-star-rated funds to achieve high relative returns and lower-star funds to report lower relative performance.

Instead, the study found nearly 30% of the funds across all star ratings didn't even survive the subsequent five-year period. The funds that did survive displayed a broadly dispersed relative five-year performance. Based on this data, there does not appear to be reliable evidence that high star ratings are predictive of persistent outperformance."

Since I haven't yet found this DFA study, I don't know if it is the true source of the claim in the second paragraph, or if the author is (who does not work for the organization). But either way, the statement is a doozy. 

I initially interpreted "nearly 30% of the funds across all star ratings didn't even survive the subsequent five-year period" as suggesting that the star rating was somehow to blame. Look at that wicked rating, it killed 30% of the funds! But that cannot be the claim. By the same logic, expense ratios are equally at fault, since nearly 30% of the funds across all expense levels failed to survive the five-year period, too.  

Thus, I assume the meaning to be that the fund's star rating as of December 2007 was unrelated to its survival prospects--a finding that would be consistent with the author's initial claim that "star ratings are not predictive." Except ... that surely is false. There is no way on heaven, earth, or any other abode that mutual funds with the strongest long-term track records die at the same rate as those with the worst records. This cannot be so.

Indeed, they do not. Morningstar's Annette Larson duplicated the study, sorting all funds according to their December 2007 star ratings and then calculating their subsequent death rate. The results:

The star rating is very much predictive! In fact, it is more predictive of future fund deaths than is a fund's expense ratio. 

Here are the results from the same time period and the same percentile groupings (i.e. the distribution of the star rating, with the top 10% being assigned to the first group, the next 22.5% to the second group, the middle 35% to the third group, the next 22.5% to the fourth group, and the final 10% to the fifth group), only this time sorted by fund expenses:

A clear victory to the star rating.


My thought: People who write about the importance of expense ratios are useful and correct. That same article quotes Morningstar's Russ Kinnel as saying that expense ratios are the most reliable indicator of a fund's future returns. That is true; expenses, as we say, are not the best signal of a fund's survival, but they are the best signal of returns for the funds that do survive.

Everything else written by those who extol the virtues of expense ratios, treat with caution. Jack Bogle has always been careful with the facts, because his argument was unpopular at the time, and if he made a data error he would get shredded. His successors with the cost argument are under no such pressure, because they are viewed by the financial media (and most readers) as being on the side of good.

Fine, they are on the side of good. But they are often wrong, when they stray from the immediate subject of the benefits of low costs. 


John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own

Facebook Twitter LinkedIn

About Author

John Rekenthaler  is vice president of research for Morningstar.

© Copyright 2024 Morningstar Asia Ltd. All rights reserved.

Terms of Use        Privacy Policy          Disclosures