In this article, Kräussl and Sandelowsky take into account successive improvements in Morningstar’s fund ratings—originally established in 1985—in order to investigate the predictive performance ability of these ratings. Initially, funds were rated within broad asset classes. In July 2002, Morningstar changed the risk-adjusted return measure used to evaluate fund performance. Since then, they have also been rating funds within 64 narrower fund categories.

It appears that many investors base their choice of funds on the fund’s previous rating. For example, Del Guercio and Tkac (2001) identify a positive abnormal flow for five-star funds up to six months after the issuance of the initial five-star rating, while two-star funds experience a negative abnormal flow for three months after the fund receives its two-star rating, a finding confirming the influence of ratings on investor behaviour. Whether or not these ratings exhibit persistence is thus an interesting question.

Several studies have focused on the subject of Morningstar rating persistence, all of them having considered only US equity funds. Most of them were based on Morningstar’s initial methodology. Khorana and Nelling (1998) find that performance is persistent in the short-term. Blake and Morey (2000) find that Morningstar’s rating system is mediocre in terms of predicting future performance, while the Sharpe ratio does considerably better, with a higher ability to predict low-performing funds than high-performing funds. Morey (2003) finds that a 5-star Morningstar rating does not persist three years out-of-sample.

In the present article, the authors recommend considering old- as well as new-methodology ratings--with a period of study extending from March 1995 to September 2005--and including not only U.S. equity funds, but also international equity funds, taxable bond funds and municipal bond funds, the three latter categories not having been studied previously.

First, they analyse the predictive performance of Morningstar’s rating system considering all mutual funds as a single group and using 1995 ratings. According to their results, there is hardly any difference in performance between three-star and five-star rated funds and several four-star rated funds post significantly higher returns than do some five-star rated funds. However, Morningstar’s ratings system is excellent at predicting underperformance. These findings are consistent with those of Blake and Morey (2000).

Secondly—with data from 1997 onwards--they consider four broad asset categories for their analysis of the ratings. They find, for example, that five-star US equity funds significantly outperform one-star funds only 37.5% of the time; at the same time, these same funds significantly outperform three star-funds 18.75% of the time. It is clear then that—compared to a random walk--Morningstar’s ratings system offers no added value in terms of predicting mutual fund returns. It is not even able to predict underperformance when funds are classified in the various categories. For international stocks, the results are slightly better than those for U.S. equity funds, but the rating system still does not beat a random walk, as in 43.75% of the cases there is no significant difference between the performances of five- and three-star rated funds. For taxable bond funds, the predictive performance based upon Morningstar’s rating is extremely weak and will never exceed the returns yielded by a random walk. Lastly, the performance of all municipal bond funds appears alike, making no sense for the division into the five start groups.

Finally, Kräussl and Sandelowsky consider the recent period, starting in July 2002, during which funds are divided into 64 categories. For the categories related to U.S. stocks, they found that for 85.29% of funds there is no difference in the performance of one- and five-star rated funds, or that one-star funds significantly outperform five-star funds after one year. Concerning international stock funds, the predictive performance of the rating is good only for the Europe stock category. Morningstar ratings are also good at predicting fund performance in the taxable bond classifications of Intermediate Government, Intermediate-Term Bond, Multisector Bond and International Bond. For municipal bonds, predictive ability is found only in Muni Short and Muni Single State Intermediate categories.

Comparing new and old ratings on the same out-of sample-period, the authors of the study find that—in terms of predictive performance--the new ratings system is, at best, equal to the old one. In most cases, the results produced by the old ratings system are more significant than those produced by the new ratings system. They find that the old ratings system is superior in predicting short-term performance, while both systems are about equal in predicting longer-term performance.

They conclude that the new ratings system should be used as a source of investment decisions only for the few categories for which it has been identified as having predictive ability, and not for all other categories, for which it does not offer any value in terms of predicting future performance; in short, their conclusions support Morningstar’s assertions that its ratings are indicative of past performance alone.

References

Blake, C.R., and M.R. Morey, “Morningstar Ratings and Mutual Fund Performance”, Journal of Financial and Quantitative Analysis, vol. 35, n°3, 2000, p. 451-483.

Del Guercio, D. and P.A. Tkac, “Star Power: The Effect of Morningstar Ratings on Mutual Fund Flows”, Federal Reserve Bank of Atlanta, Working Paper 2001-15, 2001.

Khorana, A. and E. Nelling, “The Determinants and Predictive Ability of Mutual Funds Ratings”, Journal of Investing, vol. 7, n°3, 1998, p. 61-66.