Morningstar Rating

The Morningstar Rating, commonly called the star rating, is a measure the research firm Morningstar assigns to funds and similar managed products, ranking each one from one to five stars against the other members of its Morningstar category on the basis of past risk-adjusted return. Morningstar's own description states that the rating "is calculated for managed products (including mutual funds, variable annuity and variable life subaccounts, exchange-traded funds, closed-end funds, and separate accounts) with at least a three-year history", that exchange-traded funds and open-end mutual funds are treated as a single population for comparison, and that it is "calculated based on a Morningstar Risk-Adjusted Return measure that accounts for variation in a managed product's monthly excess performance (not including the effects of sales charges, loads, and redemption fees), placing more emphasis on downward variations and rewarding consistent performance."

Three features of the method decide how the number should be read, and all three are stated in Morningstar's own published description of it.

The first is that the distribution is fixed in advance. In Morningstar's words, "The top 10% of products in each product category receive 5 stars, the next 22.5% receive 4 stars, the next 35% receive 3 stars, the next 22.5% receive 2 stars, and the bottom 10% receive 1 star." That makes the star count a rank within a group, not an assessment against any absolute standard. A category can have a decade in which every fund in it lost money and it will still contain the same proportion of five-star funds, because the shape of the curve does not depend on the results being good.

The second is the peer group. The comparison runs within a Morningstar category, which is Morningstar's own grouping of funds that hold broadly similar things. This is the single most useful correction to how the rating is usually read: five stars means a fund did well relative to funds doing the same job, and says nothing about whether that job was worth doing. A five-star fund in a category that trailed a broad market index over the same period both beat its peers and trailed the market.

The third is how the time periods are weighted. Morningstar computes separate three-, five- and ten-year ratings and blends them into the overall figure. The weights are 100 percent of the three-year rating for a fund with 36 to 59 months of returns, 60 percent five-year and 40 percent three-year for 60 to 119 months, and 50 percent ten-year, 30 percent five-year and 20 percent three-year for 120 months or more. Morningstar adds a caution that is easy to miss and worth repeating: "While the 10-year overall star rating formula seems to give the most weight to the 10-year period, the most recent three-year period actually has the greatest impact because it is included in all three rating periods." An investor who reads five stars as a verdict on a decade is reading a number in which the last three years are counted three times.

Two consequences follow that surprise people. A fund's rating can change without its own returns changing, because the rating is a position in a distribution and the other funds in the category moved, or because an old period rolled out of the window. And each share class of the same portfolio is rated separately, since the returns used are net of that class's costs. Fund filings routinely print the ratings class by class for exactly this reason: the same holdings, managed by the same person, can carry a different number of stars depending on which version of the fund an investor owns. Note what the returns are net of and what they are not: ongoing fund expenses are inside the figures, while sales charges, loads and redemption fees are excluded by Morningstar's stated method. A fund bought with a front-end sales charge is therefore rated on a return its buyer did not receive.

Because every input to the calculation is a return that has already happened, the rating is a description of a record rather than a forecast of one. That is not a criticism of the method, which does what it says. It is a warning about the use, since a table sorted by star count is a table sorted by what has already occurred, and buying the top of it is the mechanism of performance chasing.

The rating is produced in four steps.

1. Assign the fund to a Morningstar category of funds holding broadly similar investments.
2. Compute Morningstar Risk-Adjusted Return from monthly excess returns over the relevant period, weighting downside variation more heavily than upside variation so that consistency is rewarded.
3. Rank every fund in the category on that measure for each of the three-, five- and ten-year windows the fund is old enough to have.
4. Cut the ranking on the fixed curve and blend the period ratings into the overall star count.

A hypothetical example of what the fixed curve means in practice. Take a Morningstar category containing 400 rated funds. The top 10 percent, meaning 40 funds, receive five stars. The next 22.5 percent, or 90 funds, receive four. The next 35 percent, or 140 funds, receive three. Then 90 funds receive two and the bottom 40 receive one, and 40 plus 90 plus 140 plus 90 plus 40 is 400.

Those counts are the same next year, and the year after. If the whole category returns 12 percent a year, 40 funds still get one star. If the whole category loses 6 percent a year, 40 funds still get five. The rating is answering "how did this fund do against the others doing the same thing", and the arithmetic guarantees it an answer regardless of what the others did.

Pros

• Adjusts for risk rather than ranking on raw return, and penalizes downside variation more heavily, which matches how most investors actually experience a loss.
• Compares like with like by grouping funds into categories, so a bond fund is not scored against a stock fund.
• Uses a published, stable and checkable method, with the distribution and the period weights disclosed.
• Free, widely available, and shown on most fund research pages, which makes it a usable first filter.

Cons

• Every input is a past return, so the rating summarizes a record and does not forecast one.
• The fixed curve guarantees five-star funds exist in every category, including categories that lost money.
• Category-relative scoring says nothing about whether the category itself was a sensible place to be.
• The most recent three years dominate the overall rating even for a fund with a full ten-year history.
• Requires three years of history, so a new fund is unrated and an investor sorting by stars never sees it.
• Ratings move when peers move, so a change in the star count is not evidence that anything changed at the fund.
• The returns behind the rating exclude sales charges, loads and redemption fees, so the star count can flatter what a buyer paying those actually earned.