Low-Cost 5-Star Funds Have the Edge

Not too long ago I wrote about the performance of the Morningstar Rating as a predictor of fund performance. I have long maintained that expense ratios are the best predictor of whether a fund will outperform or not, so I was eager to test the two together and to see what happens when you combine high star ratings and low expenses as fund-selection criteria.

As you may recall, the star rating study found that generally 5-star funds modestly outperformed their peers on an absolute basis and on a risk- and load-adjusted basis as measured from June 2002 to June 2005. Put another way, 5-star funds produced higher returns and received higher star ratings for the ensuing three years than did lower-rated funds.

So for this study, I broke funds within the same categories into quartiles based on their expenses as of June 2002. From there, I looked at returns, performance percentile ranks, and three-year star ratings over the ensuing three years. In addition, I combined the star ratings and expense quartile rankings to see what kinds of returns you would have earned if you had bought funds that were in their categories' cheapest quartile and rated 5 stars.

Expense Ratio Matches Star Rating's Predictive Value
It turns out that choosing funds from the cheapest quartile of their categories resulted in remarkably similar returns as buying 5-star funds. On the whole, the cheapest-quartile funds enjoyed higher returns and lower risk than the average fund. The results were quite similar to what you would have gotten if you had chosen 5-star funds.

Predictive Value of Expenses & Stars on Future Returns
Asset Class	3-Yr Return (%)
	Cheapest- Quartile Funds	5-StarFunds	Combined
U.S. Stock	9.98	10.10	10.61
International Stock	14.72	15.30	15.53
Taxable Bond	7.10	7.40	6.56
Municipal Bond	6.85	8.30	7.10
Data from June 2002 through June 2005.

For example, from June 2002 through June 2005 the cheapest-quartile domestic-stock funds produced an average return of 9.98%, an average performance rank of 41% (the lower the number, the better), and average three-year star rating of 3.3. By comparison, a strategy of buying 5-star domestic-stock funds would have produced an average return of 10.1%, a performance rank of 43%, and a star rating of 3.1 from June 2002 through June 2005.

Predictive Value of Expenses & Stars on Future Relative Returns
Asset Class	3-Yr Return Percentile Rank *
	Cheapest- Quartile Funds	5-StarFunds	Combined
U.S. Stock	41	43	41
International Stock	47	43	33
Taxable Bond	42	43	38
Municipal Bond	40	35	30
Data from June 2002 through June 2005. * On a scale of 1 to 100, with 1 being the best.

For taxable-bond funds, buying the cheapest-quartile funds would have produced an average return of 7.1%, a performance rank of 42%, and an average star rating of 3.5 over that three-year period. For 5-star taxable-bond funds, the average return over that stretch was 7.4%, the performance rank was 43%, and the average star rating was 3.4.

Predictive Value of Expenses & Stars on Future Star Ratings
Asset Class	3-Yr Morningstar Rating
	Cheapest- Quartile Funds	5-StarFunds	Combined
U.S. Stock	3.3	3.1	3.3
International Stock	3.1	3.4	3.4
Taxable Bond	3.5	3.4	3.7
Municipal Bond	3.6	3.8	4.1
Data from June 2002 through June 2005.

A Match Made in Heaven: Expenses and Stars
Things get more interesting when you combine low expenses and high star ratings. In nearly every case, you get equal or better performance than you'd get by basing your decision on one factor or the other.

For example, municipal-bond funds that pass the twin test of landing in the cheapest quartile and garnering 5 stars averaged a 7.1% return, a 30% performance rank, and a three-year star rating of 4.1 from June 2002 through June 2005. By comparison, focusing on funds in the cheapest quartile alone (and tuning out the stars) would have led you to a group of funds with an average 6.85% return, top-40% performance, and a three-year rating of 3.6 stars. Using only 5-star muni funds would have produced returns of 8.3%, top-35% performance, and 3.8 stars.

The results were equally compelling for international-stock funds. In that asset class, the combination of stars and expenses produced returns of 15.6%, top-33% performance, and an average three-year rating of 3.4 stars. Focusing on international funds in the cheapest quartile alone produced 14.7% returns, top-47% performance, and a three-year rating of 3.1 stars. You would have done better than focusing on expenses alone if you had bought only those funds with high star ratings, but not better than the combination. The average 5-star international fund returned 15.3%, produced top-43% performance, and averaged a star rating of 3.4.

In domestic stocks, the star rating and expenses were both helpful, though combining them didn't result in dramatic improvement. The average three-year return for domestic-stock funds in the cheapest quartile was 9.98%. For 5-star funds it was 10.1%, but that moved up to 10.61% when you combined top ratings with low costs. The percentile rank was 41% for both the cheapest funds and for the cheap funds with 5 stars--so each criteria produced above-average relative returns, but the combination didn't add anything. For the three-year star rating, the combination netted a rating of 3.30, which is modestly better than 3.27 for the cheapest quartile and 3.1 for 5-star funds.

For taxable bonds, we also saw improved results from the combination of high ratings and low costs. Five-star funds with bottom-quartile expenses netted a three-year star rating of 3.65 and relative performance in the top 38%.

The Class of 2003
I also ran the numbers from June 2003 to June 2005. I did so for two reasons. First, I simply wanted an additional time period to test. In addition, June 2002 was near a market inflection point marking the end of the bear market. After that, a lot of market trends reversed, thus making past performance less valuable than at most points in the market cycle. By June 2003, a rally was under way, and mid- and lower-quality credits had begun to rally in the bond market. For this two-year period, I'll focus on just relative performance because the absolute returns aren't from a comparable period and we need three years of data to calculate a star rating.

Over this shorter period, the combination of low expenses and 5 stars did even better than they did over the three-year period, most likely because of the turnaround in the markets. Buying domestic-stock funds that were both in their category's cheapest quartile and rated 5 stars netted top-39% returns. For international stock funds, it produced top-26% performance, for taxable bonds it was top-33% performance, and for the municipal-bond funds it was top-27% performance.

A version of this article appeared in the February 2006 issue ofMorningstar FundInvestor.