How much performance is in the stars?
Frank Russell asks a few probing questions about managed funds star ratings and comes up with some telling conclusions.
Monday, October 15th 2001, 11:01PM
Star ratings for managed funds have been the rage for several years now. One of the questions is how good are the ratings and are they any use at predicting future performance.
To find the answers Frank Russell's Australasian manager director Alan Schoenheimer commissioned some research on the subject.
The first thing he points out is that star ratings are ubiquitious with consumer products and services, however there are different degrees of importance.
He notes that star rating systems are used for many services, however managed funds are probably the most high value of the lot.
For instance if the star rating of a movie attracts you to go and see the flick you only stand to lose about $10 if you're not satisfied.
If you choose a restaurant or hotel based on a star rating the potential loss is only a couple of hundred, or maybe a thousand dollars, and at least you still get a meal or roof over your head.
With a managed fund though the quantum is much greater because you're talking about someone's life savings.
"The cost of being wrong is in a different dimension," Schoenheimer says.
In Australia at least four organisations, Morningstar, Assirt, Investorweb and van Eyk do star ratings while Morningstar and FundSource provide them in New Zealand.
Frank Russell did its analysis using Morningstar data in Australia because that company provided the most comprehensive set of data and information.
"It's not because we have a vendatta," he says.
One of the first things Frank Russell found is that there were small sample sizes, and not a long history of data.
While Morningstar has a good classification system for putting funds into various categories, it results in small sample sizes.
For instance once you get inside Australian equities there are broad market funds, industrial shares, resource stock funds, property securities and small cap funds.
Each of these have different characteristics, and there aren't huge numbers of funds in each sector.
The next problem Schoenheimer identified was that the relationship between some of the items measured by the researcher and a fund's performance aren't clear.
For instance Morningstar splits its qualitative measures into five items including corporate strength (25%), administration and distribution strength (22.5%), investment management (22.5%), sector strength (20%) and product features (10%).
"What's not made clear is whether these variables have any predictive power.
"Is the paint (in a fund manager's) front office any more predictive?"
Schoenheimer says intuitively these factors have some bearing on a manager's performance, however he hasn't seen any research from the ratings agencies which shows the relationship.
His next problem is that the Morningstar system is a relative one.
That is it aims to have a set distribution of managers/funds in each of the star groups. For instance the top 15% of funds in a group will be five star, the next 20% will be four star and so on.
However, Russell's research over the period March 1998 to June 200 shows that the distribution wanders.
In Australian equities the bias is towards the top end with nearly 25% of funds being five star rated compared to the benchmark if 15%, while in fixed income the bias is the other way with just over 5% of funds having five stars.
Schoenheimer says there is also a risk of bias coming in as fund managers with poorly performing funds won't seek to get them rated.
Frank Russell's research shows that the ratings from the various Australian agencies are well correlated, so there seems little point in an adviser spending a lot of time deciding which one to use.
"You probably don't need to subscribe to more than one," he says.
Schoenheimer also suggests advisers should remember how ratings agencies make money, and consider whether it impacts on decisions.
He says some managers pay ratings agencies to research their funds (they will only pay for the good ones to be rated, therefore creating the potential for upward bias), some researchers charge managers big money to use their ratings in advertising.
While researchers freely give away ratings to the public, they also sell research, and finally they make money by consulting to the managers they also rate.
One of the big issues in the United States is that between 70 and 80% of funds go into four and five star rated funds, and money comes out of one and two star rated funds.
Schoenheimer says this doesn't appear to be the case in Australia at present, most probably because organisations with their own funds, also own their own distribution.
A bank for instance may not need to have a star rating to sell its funds, because they are sold internally through their advisory network using things like brand and service.
"It's the power of distribution rather than the power of star ratings (which sells funds)", Schoenheimer says.
The second big question Frank Russell asks is: Are star ratings merely a proxy for past performance?
To answer the question Frank Russell looked at performance before and after a star rating and worked out if there was a statistical difference.
"There is a hint that past performance is impacting on star ratings," Schoenheimer concludes.
Frank Russell also looked at the predictive power of star ratings. The conclusion was that "predictive power (of star ratings) was much less evident."
"There's not enough evidence here to say you should be relying on these four or five star funds," Schoenheimer says.
"Do not rely on these (ratings) as a future predictor of performance."
Footnote: Frank Russell also researches managed funds but its approach is different to the agencies referred to above. It charges its clients to research managers and that research is used to build multi-manager portfolios. Unlike other researchers it doesn't make its information publicly available. Schoenheimer says Frank Russell spends $20 mill a year on research.
To see what Morningstar boss Graham Rich has to say about Frank Russell's research click here.
| « Are hedge funds a seperate asset class? | King builds an empire » |
Special Offers
Commenting is closed
|
|
Printable version |
|
Email to a friend |
