The S&P 500 incident was an extreme demonstration of the ability of indices to influence, rather than merely reflect, stock values. However, it was by no means an unparalleled occurrence. When Dimension Data Holdings joined the FTSE 100 Index, the price of the information technology services company’s shares soared from about £6.65 to £10 within a matter of minutes, only to recede to £6.60 the following day. The addition of Central European Media Enterprises to the MSCI Emerging Markets Index triggered a one-day trading volume of 520,000 shares, nearly 17 times the stock’s average for the preceding 108 trading days.
Daniel Broby, a member of the investment board of Denmark’s BankInvest Group, shows in A Guide to Equity Index Construction that inadvertent market impact is not the only worry index providers have. Their ostensible mission of simply reporting the performance of a universe of equities is far more complicated than it sounds. For example, the NASDAQ 100 Index is designed to prevent any sector from becoming dominant, yet in 2002, the software sector swelled to a 23.8 percent weighting. Index constructors have three weighting schemes to choose from—market capitalization, price, or equal weighting—but each entails formidable implementation problems. Even such a seemingly straightforward task as counting a company’s shares is obstructed by buybacks, directors’ share awards, and the existence of company treasury stock.
Lest anyone imagine these difficulties are solely of concern to theoreticians, Broby details important ramifications for portfolio managers. He points out, for instance, that Vodaphone Group’s weighting in the FTSE All-Share Index varied from 4 percent to 12 percent between 2001 and 2005. For a manager who liked the stock, what allocation would have represented an overweighting? Managers who wished to market-weight the technology company Baltimore had to contend with a rapidly moving target in 2000. The stock entered the FTSE in March, exited in June, then reentered in September. Yet another conundrum arises from overlap—the fact that, for instance, about one-third of the stock universe is included in both the value index and growth index published by Citigroup.
Startling assertions crop up throughout A Guide to Equity Index Construction . Broby says there is no correct answer to the question of how many stocks are optimal for inclusion in an index. He argues that style indices may be fundamentally flawed because of the shortcoming of book value as a measure of net assets. An index calculated on the widely used basis of a geometric mean, he notes, would collapse if a company within it went bankrupt. As for the less hypothetical phenomenon of correlation breakdown—a nightmare of index constructors—Broby contends that the best way to deal with it when it happens is to ignore it.
Considering the vast amount of valuable information conveyed in this book, it is a shame the editing was not tighter. The text contains such infelicities as the “Rhur” River, “per capital GDP,” and “the discounted cashflow of all future cashflows,” as well as simple spelling errors (e.g., “Sharia complaint” [in lieu of “compliant”], “boom” for “boon,” and “shear” for “sheer”). Warren Buffett appears in one paragraph and “Warren Buffet” in the next. A single paragraph refers correctly to “systematic” and incorrectly to “systemic” risk in connection with beta.
The bibliography is an amalgamation of errors in its own right. Certain articles cited in the text, such as “Aronson (1999)” and “Bookstaber (1997)” are nowhere to be found among the references. “Ramchand and Susmel (1998)” is rendered as “Ramchand, L. and S. Raul, 1998” in the bibliography. (The second author’s actual name is Raúl Susmel.) As a result of stylistic inconsistency, the bibliography mentions both “J. Treynor” and “J.L. Treynor,” “R. Stulz” as well as “R.M. Stulz,” and so forth. Louis Bachelier’s surname is misspelled.
Distracting though these blemishes are, A Guide to Equity Index Construction is highly worthwhile reading for both money managers and their clients, who must evaluate a staggering array of choices. The author reports that the number of indices in place already exceeds 250,000. Furthermore, the proliferation appears to be accelerating: In January 2007, MSCI Barra announced the launch of 20,000 new indices.
In this potentially chaotic situation, CFA Institute members can take pride in Broby’s strong endorsement of their organization’s Global Investment Performance Standards®. The 1999 introduction of GIPS standards, he says, at long last enabled clients to compare the performance of managers who previously cherry-picked the indices that made them look best. According to Broby, the global standards could not have been achieved by cooperation among the competitive index providers; therefore, the establishment of GIPS standards is “one of the most impressive pieces of self-regulation by the asset management community.”
—M.S.F.