International Week - A Chat with FTSE on its New Global Style Benchmarks By John Spence October 3, 2002

We've been dancing to salsa music and downing piroshki to keep with the spirit of International Week here at IndexFunds.com. Maybe we're just looking for a diversion from the U.S. stock market, which suffered the worst quarter in recent memory.

Indexers have been steadily calling for more international style funds, and the industry is taking notice. International index provider FTSE (that's pronounced "footsie") launched a new family of style indexes in September. We huddled with Jane Staunton, president of FTSE Americas, and Peter Wall, senior vice president of business development at FTSE Americas, to discuss the new benchmarks.

Q: The new style indices use 9 variables to determine style, 4 for value and 5 for growth. Index providers are using more sophisticated techniques to determine style. How is determining style an evolutionary process? Why did the old system become outdated?

A: Determining a company's style has evolved, but not because of major revisions to the principles of Graham & Dodd. Greater disclosure by companies - and in a sense, a change in accounting practice - and greater availability of computing power and databases have permitted analysts greater scope to look for combinations of financial ratios that together suggest value and growth opportunities. To some degree, greater conformity to International Accounting Standards has helped the process, too, making style analysis and investing across borders more feasible.

The older style index models used by some indexes use a single financial ratio - typically price-to-book - and rank all stocks by that variable, with a cut-off at the midpoint of the list. This is for example the approach used by MSCI from 1997 until today, and S&P/Barra from 1993 until today. More recent style indices, like the FTSE Global Style Index Series, use multiple financial ratios to determine styles. [Ed note - MSCI recently announced a new global style index methodology.]

Q: Under the new FTSE methodology, can parts of a company's market capitalization be in both the growth and value indexes? Is this preferable to simply pigeonholing a company in the growth or value index?

A: Yes, FTSE's Global Style methodology will allocate a stock's capital across value and growth indexes, to the extent our scoring system shows that the stocks are not "deep value" or "deep growth." During the research phase of our Global Style Index development, we consulted with many participants in style investing to get their views on this point, and the response was plainly that stocks can be both value and growth. At the same time, there was a strong preference that there shouldn't be overlaps of capitalization - that is, all of a company's capitalization shouldn't be in both value and growth. So FTSE devised a method to allocate portions of capitalization to value and growth indexes, with the portions determined by how strong the stock tended to lean toward one style or the other. A stock not demonstrating strong tendencies either way would have its capitalization split 50/50.

Q: Was reducing turnover in style indexes a major issue during your consultations with investment managers? What steps does FTSE take to cut down turnover in its indexes, since one of the strongest arguments for indexing is tax efficiency? Were there specific global regions or sectors where style turnover was a major problem?

A: Yes. The users of style indexes put turnover high on their lists of concerns during our consultations. FTSE's Global Style methodology limits turnover, or movement of capitalization between style categories, in three ways. First, the multi-factor approach, with the use of 9 distinct variables, adds stability to the value/growth outcomes at review dates. Second, FTSE's banding structure, which determines the allocation of a company's capitalization between value and growth, also acts as a buffer zone to keep companies from dramatically shifting between categories. Lastly, bi-annual reviews of the style variables avoids the "big bang" effect that is common in indices that just do annual reviews. Users also recognize that major changes in a company's business and financial characteristics will impact its style rating, and this should be recognized in style indexes as well. During our market consultations, concerns were not focused on specific sector or regional turnover, but rather on keeping overall index turnover down. We think our methodology has achieved a reasonable balance that keeps turnover to a minimum, but adequately reflects market conditions.

Q: It seems one of the major challenges for index providers is to balance index completeness (and accuracy) against investability. Can you explain how this relates to style indexes?

A: This is a challenge, and style adds its own wrinkles. In terms of style indexes, a user would want to be sure that the starting point is an index which already has balanced completeness of coverage and investability - like FTSE's World Index, which is the basis for our Global Style Index Series. The style "wrinkles" are achieving the best way to determine style. For example, do the metrics and methodology produce an accurate representation of style? We needed to find the right way to allocate capitalization between value and growth indexes, since it was important to users that all the FTSE World Index's capitalization be represented in the Style indexes, with no exclusions and no overlaps. There's also the search for the best way to manage the indexes and to maintain accurate style representation, while limiting turnover as much as possible.

Q: In his article on index methodology, Vanguard index fund manager Gus Sauter said index providers should let active managers do the work. I think another way of saying this is that it's active managers themselves who determine what style is, and indexes should reflect that process (rather than setting up arbitrary boundaries). Is that reasoning reflected in the new style index methodology?

A: To respond in the spirit of Mr. Sauter's article, yes and no. Yes, in the sense that FTSE spent a lot of time with style index users, who made us well aware that a value stock can also be a growth stock, and some stocks are neither value nor growth, and that two investors can see things differently and for very good reasons. No, in the sense that active managers have their own proprietary approaches to determining style and managing portfolios, and we can produce only one value and one growth benchmark series. However, index customization is our specialty and we anticipate a lot of requests to look at style differently.

Q: How often are the style indexes rebalanced? Have you done any research on the optimal frequency for rebalancing, for example quarterly vs. semiannually?

A: Most style indexes are rebalanced - that is, reviewed for stocks' style characteristics - once or twice a year. For a multi-country style index like the FTSE Global Style Index Series, twice a year is probably the most frequent basis on which it can be done, as new financial data aren't commonly available more than twice a year. As for optimal frequency, once a year - or longer - tends to produce a "big bang" effect, so we selected semiannual frequency.