Search This Blog

Tuesday, February 24, 2009

About Those Morningstar Fund Ratings....They're Not Worth Much (Now They Tell US)





The WSJ reports that Morningstar has overhauled its system for analyzing mutual funds and I must admit it leaves me mystified.

Most academic research has found that the morningstar star system is of little use in predicting future mutual fund performance. The system went through a major overhaul in 2002. A 2006 academic studies after based on a very short number of data points ( 2 years of data in a strong stock market) found the new system to have a bit more predictive power.

But based on the sweeping nature of the current overhaul, the folks at Morningstar clearly thought it was time to reinvent the wheel again.

My confusion is primarily due to the fact that when the morningstar system reviews o active managers it is trying to measure manager skill in outperforming its benchmark in risk adjusted return (something close to what academics call "alpha"). One would assume that a managers alpha would come from his ability to make buy and sell decisions in adding or deleting holdings to the fund's portfolio. Yet it seems the morningstar analysis is meant to assess the holdings of the portfolio at a point in time. Why this will tell the potential investor anything about how the manager will perform over time is hard for me to comprehend.



My comments in parentheses and italics

• FEBRUARY 24, 2009
Morningstar to Take Fresh Look at Funds
A Move to 'Institutional-Style' Analysis
By SAM MAMUDI

Investment-research firm Morningstar Inc. has unveiled an organizational revamp that will change the way its analysts look at mutual funds.
Morningstar analysts will take a broader view of the funds they cover, for instance, considering asset allocations and offering more detail on the underlying stocks...


( I find this confusing: If they are evaluating actively managed funds based on their holdings and actively managed funds are only required to report the end of quarter holdings in arrears, then how useful could this analysis be ?)


....And fund analysts will reach out more to financial advisers to find out what is and isn't working. "Financial advisers are in the trenches and know what investors want," Mr. Phillips said. "We want to use our data to get better results for investors."


(Another confusing one: Up until now Morningstar research was supposed to give guidance to financial advisors and individual investors in choosing mutual funds for their portfolios. In other words morningstar research was designed to sift through funds and tell their clients which funds they expect "to work and not to work", Now the folk at morningstar tell us that they will go to the consurmer of the advice and ask them what is and isnt working. Wasn't Morningstar supposed to know that based on their own research ?

And here's another strange one, Advisors "know what investors want" and morningstar wants to use "our data to get better results for investors". Morningstar has long presented research based on fund flows (and other behavioral finance research has shown) what investors often "want" is to chase performance,which is not good for them. It would seem more logical that morningstar should be working to present objective research no research based on investors' biases.)


Here is more from Morningstar:


Mr. Phillips said part of the change will be to provide more institutional-style analysis. "We're going to take institutional-level insights and create streamlined offerings for retail investors," he said.
Analysts will consider asset allocation when evaluating a fund, for instance, looking at which funds might better suit particular portfolios and which funds might overlap with existing holdings.
"Many people, if they own several mutual funds, don't know the holdings of those funds," Mr. Phillips said. "We can help investors know if there are cumulative holdings in their portfolios."
Another example of the institutional approach is factoring in a stock's economic moat -- the ability of companies to keep rivals at bay.
"There's been a lot of talk about Fidelity Magellan Fund and figuring out its performance," Mr. Phillips said. The fund, which was once among the largest stock funds in the world with more than $100 billion in assets, was down nearly 50% last year and has seen its assets shrink to under $20 billion.
Using the economic-moat analysis can help investors understand Magellan's poor 2008, Mr. Phillips said. While the average stock fund has 46% invested in wide-moat companies, only 23% of Magellan's portfolio is in such companies.
"That's a huge underweight in companies with strong and enduring brands," he said. In 2008, wide-moat companies were down 28% on average, while narrow-moat companies lost 40%. No-moat companies -- in which Magellan was overweight -- were down 50%.
"That's a great snapshot of Magellan's performance," Mr. Phillips said.


(<
span style="font-style:italic;">that is all very interesting information on why Fidelity Magellan performed the way it did in 2008. I'm not sure what that tells me about investing in Magellan going forward. Am I supposed to assume that Magellan's managers will "learn from their mistakes" and change their portfolio. Or should I conduct independent analysis on whether "no moat" "narrow moat" or "wide moat" companies will perform best going forward and then invest (or avoid ) Magellan based on an expectation that Magellan will (or won't) continue its "no moat" strategy ?....my head is spinning)



...."My goal is for us to look at a fund portfolio through the eyes of a fund manager," added Mr. Phillips, speaking broadly of the changes. "We want to do all we can to tear a portfolio apart and understand its risk."
(While this sounds impressive, when applied to evaluating actively managed mutual funds it seems to be a futile exercise. To "tear apart" an actively managed portfolio at a point in time, based on old data is not a very productive way to "understand the risk" of the portfolio as it is currently constructed. In the case of actively managed funds you "never know what you own" no matter how sophisticated the analysis might seem.

Contrast that with analyzing a portfolio of index funds and/or etfs. Since the holdings of those portfolios is static (except when indices are rebalanced) it would be quite easy to monitor the holdings of the portfolio, and various metrics such as p/e price/book etc. And an exercise of "tearing a portfolio apart and understanding its risk" would actually be useful.


Ironically, Morningstar does have software that will allow very useful analysis of a portfolio of index funds or etfs. It is possible to use their software to produce a report that gives aggregate statistics for the portfolio like p/e and p/b for the portfolio as well as individual stock holdings, industry and country weightings. breakdown. Morningstar is an excellent compiler of data, it is their rating system that is of little utility. Yet despite their limited usefulness it is the ratings that get the most attention. ...And morningstar collects a fee everytime a mutual fund company uses their ratings in an advertisement....but of course that couldn't have anything to do with their persistence in revising a system that is not particularly useful in forecasting performance, as they themselves admit...sometimes.)



forbes has a little video piece on morningstar ratings

and morningstar's own explanation of how to use the ratings is here, although it seems that on their website articles they often dont follow their own guidelines and simply reccommend highly rated funds.

No comments: