Style makeovers

Investors looking at value and growth equity benchmarks often contend with conflicting signals. In the three months ended June 30, for example, the S&P 500/Barra value index gained 2.6 percent, but the Russell 1000 value index increased just 1.7 percent. The S&P 500/?Barra growth index, meanwhile, rose 0.1 percent; the Russell 1000 growth index climbed 2.4 percent.

"S&P/Barra says large-cap value stocks outperformed large-cap growth, while Russell says large-cap growth outperformed," observes Ron Surz, president of PPCA, an investment consulting firm based in San Clemente, California.

Wide divergences within the same investing style are commonplace for these prominent indexes, because the S&P/ Barra and Russell benchmarks use different methodologies and draw from different stock universes. Year to date through September 30, the rival value indexes were 2.2 percentage points apart; the growth indexes were separated by 0.2 percentage points. For the ten years ended September 30, the average annual spread between the two rival value indexes is 1.7 percentage points; the growth indexes, 1.8 points, says Surz.

"Historically, the various style indexes have captured the same long-term trends, but they have frequently ?diverged in the short term," notes Richard Whitney, ?director of quantitative equity at T. Rowe Price Associates in Baltimore. The conflicting benchmarks can make it difficult for managers to explain their performance.

Now Standard & Poor's has introduced a new batch of style indexes -- Standard & Poor's U.S. style indexes -- that seeks to remedy some of the short-comings of the soon-to-be-retired S&P/ Barra indexes.

The half dozen sets of indexes within the S&P U.S. style indexes track the growth and value segments of the S&P 500, the S&P midcap 400, the S&P small-cap 600, the S&P composite 1500, the S&P 900 and the S&P 1000.

As for the traditional S&P/Barra style indexes, they will cease real-time data dissemination on December 16, 2005, although end-of-day values will be available until June 30 of next year.

David Blitzer, chairman of S&P's index committee, says the new indexes were introduced to acknowledge users' need for a set of broad style benchmarks based on a more reliable delineation of growth and value stocks as well as for a set of purer benchmarks for style analysis. Some speculate, however, that S&P may also have been looking to shed the Barra name, as that firm is now owned by rival index provider Morgan Stanley Capital International.

Two years ago the Vanguard Group ?began using the MSCI indexes as the benchmarks for all of its index funds, with the obvious exception of its popular $106 billion-in-assets S&P 500 offering (Institutional Investor, May 2002). "This was a big indication to S&P that the old methodology wasn't working for practitioners," says a consultant.

The old indexes won't be greatly mourned. Developed in the late 1990s by Barra, the S&P/Barra benchmarks are based on a methodology that many practitioners have criticized as crude. In dividing the S&P 500 into growth and value stocks, the indexes use a single factor: the price-to-book ratio. Stocks with high price-to-book ratios are included in the growth index, while stocks with low price-to-book ratios go into the value index. A stock is deemed to be either 100 percent value or 100 percent growth.

"The use of a single metric is too simplistic," contends T. Rowe Price's Whitney. "The price-to-book ratio fails to capture forward-looking information."

Moreover, because S&P has always insisted on splitting the market capitalization of the S&P 500 equally between growth and value, more so-called core stocks -- those with a blend of growth and value traits -- tend to go into the value index, which has more room on a market-cap basis because value stocks usually have lower caps than their growth counterparts.

The methodology behind the new S&P indexes was developed by Salomon Smith Barney, which, as part of the Citicorp/Travelers Group merger, became part of Citigroup in 1998. S&P ?acquired the indexes from Citigroup in November 2003. This methodology represents a significant evolution. Rather than just one factor, it uses seven to distinguish value from growth. These comprise three growth measures (five-year earnings-per-share, sales-per-share and internal growth rates) and four value measures (book-, cash-flow- and sales-to-price ratios and dividend yield).

"The inclusion of multiple factors suggests the new S&P model will be more ?reliable," says Steven Cupchak, a vice president at IPEX, an investment consulting firm in Royal Oak, Michigan.

The value and growth indexes, in turn, come in two varieties. The "basic" style form offers market-cap-weighted indexes for use in performance measurement and benchmarking. As Russell has always done, S&P now apportions core stocks to the value or growth indexes based on the strength of their value or growth characteristics. A company might wind up, say, 30 percent in the value index and 70 percent in the growth index.

By contrast, the "pure" style series consists exclusively of stocks with pronounced growth or value traits weighted on the basis of their style scores. A stock with a growth score of 2 (on a scale of ­10 to +10) will have twice the weight in the growth index as a stock with a growth score of 1.

In a backtest conducted by S&P for the ten years ended June 30, 2005, the new S&P 500/Citigroup value index gained an average annual 14.0 percent, compared with 10.3 percent for the old S&P 500/Barra value index. The new growth index gained an annualized 12.8 percent, compared with 9.1 percent for the old growth index.

"With the new multifactor model, S&P is addressing one of the key criticisms of the old S&P/Barra methodology -- its sole reliance on price-to-book ratios," IPEX's Cupchak says.

While S&P has been tinkering, Russell has not fundamentally changed the methodology of its style indexes since they were launched in 1987. The firm relies on two factors: price-to-book ratio and projected earnings growth. Currently, 35 percent of Russell 1000 stocks are allocated in full to the growth index, and 34.6 percent to the value index; the remaining 30.4 percent are assigned to both in proportion to their mix of growth and value characteristics.

Some money managers see flaws with both the Russell and the new S&P indexes. PPCA's Surz complains that they don't treat core stocks as a separate asset class. "Core stocks usually perform somewhere between value and growth," he says, "but about a third of the time, they surprise by over- or underperforming." Hence they can distort conventional style indexes.

Surz argues that the underperformance of core stocks in the second quarter largely explains the putatively conflicting performance trends captured by the S&P 500/Barra and the Russell 1000 style indexes. Approximately 20 percent of the stocks that he currently defines as core were in the S&P 500/Barra value index; the other 80 percent were in the S&P 500/Barra growth index. In contrast, 70 percent of Surz's core stocks are included in the Russell 1000 value index, and 30 percent in the Russell growth index.

Surz was so unhappy with the regular style benchmarks, in fact, that back in 1986 he introduced his own indexes -- for growth, for value and for core stocks. In constructing his indexes, he ranked stocks on their "aggressiveness," as measured by three variables: P/E, price-to-book and dividend yield. Stocks in the top 40 percent went in the Surz growth index; those in the bottom 40 percent made up the Surz value index; and the middle 20 percent composed the Surz core index. In the second quarter of 2005, value gained 2.1 percent, growth rose 2.4 percent, and core lost 0.02 percent.

Though hardly a threat to the dominance of MSCI, Russell or S&P, Surz's ?indexes, published on the PPCA Web site, have their fans. "The Surz indexes provide a purer measure of style than ?the S&P/Barra and Russell indexes," says Barry Mendelson, a managing partner at ?Milwaukee-based Capital Markets Consultants. He says that they allow his firm to conduct more precise style analyses.

The style index sweepstakes has plenty of other entrants, of course. Two years ago Barra came up with a fresh approach to benchmarking style in U.S. markets on behalf of its new parent, MSCI. The brokerage firm now splits small-, mid- and large-cap stocks between value and growth according to eight factors, including price-to-book ratio, dividend yield and 12-month-forward P/E. Like the Russell and the new S&P style products, the MSCI U.S. large-cap growth index divvies up core stocks between the value and growth camps in proportion to their style characteristics, so that every core stock appears in both indexes.

"MSCI's methodology uses robust measures to delineate between growth and value stocks," says Paul Lohrey, a principal in Vanguard's quantitative equity group. "We also like that MSCI uses buffer zones to manage the migration of companies from one index to another."

The buffer zones are intended to prevent companies whose market capitalizations abruptly change, say as the result of mergers or spin-offs, from wreaking havoc on an index. For example, MSCI's large-cap index comprises all U.S. companies ranked one through 300 in market cap. (Small-cap is 750 to 2,500.) But if a stock falls out of the top 300, MSCI doesn't yank it from the large-cap index until it drops below the 450 mark.

In index design, there's a place for flexibility.