It's Russell Index Reconstitution Time

Stock indexes aren't static. Indeed, as we've written before, they undergo regular changes, known as reconstitutions, during which securities are added and deleted. Market conditions change, index components drift up and down the market-capitalization ladder, new listings come on the market, and some stocks go away through bankruptcy or buyouts. So, if an index never added new equities and deleted others, it would eventually drift away from whatever universe of stocks it was meant to represent.

The subject of index reconstitution is of more than academic interest. After all, hedge funds can take advantage of an index's reconstitution to game the system in their favor--a practice known as front-running, which can erode the returns of both index funds and ETFs. And because not all indexes remake themselves in the same way, or follow universes of stocks with the same liquidity characteristics, reconstitution can have different effects on different benchmarks. The way an index rebalances itself can directly affect the returns of the investors in funds that track that index.

Standard & Poor's, compilers of the venerable S&P 500 Index, makes changes throughout the year. When a particular equity needs to be replaced in any of its indexes, S&P's selection committee approves a stock to replace it, only admitting companies with four straight quarters of positive earnings. Russell, on the other hand, compiles its various indexes once a year and uses a quantitative methodology based on float-adjusted market capitalization (the market cap of all shares available on the public market). Russell doesn't exclude unprofitable companies. Both index compilers, however, give advance notice of additions and deletions, which lets index-fund managers plan the trades they'll need to make to minimize the difference between their funds' returns and those of the index.

The Russell 2000 Reconstitution
That leads us to the particular problems associated with the Russell 2000's reconstitution. The Russell 2000, which consists of the 1,001st- through the 3,000th-biggest domestic equities by market capitalization, is the small-cap bogy most commonly tracked by indexers, with $43 billion invested in index funds that mimic it. That's a lot of money invested in a single index of small-cap stocks, which tend to be less liquid than their larger-cap counterparts. This year's official changeover will take place at the end of trading on June 25, 2007. And with dozens of index-fund managers clamoring to buy and sell several of the exact same illiquid stocks, index funds may have trouble getting favorable pricing. That's because buying securities en masse can drive up their share prices, and exiting them simultaneously and in large volume can mean selling at lower and lower prices.

This problem of so-called market impact is exacerbated by front-running. An academic study has shown that, because of the almost complete transparency in Russell's reconstitution methodology (relative to, say, S&P's closed-door, committee-driven process), it doesn't take a lot of effort to predict what stocks will enter and exit the Russell 2000 weeks before the changeover. That allows some traders to front-run the indexers. For instance, say a trader thinks a certain stock will be added to the Russell 2000. He can buy the stock in advance of the reconstitution and then reap a profit if the stock does get added, as all the Russell 2000 index funds drive up the shares' value as they buy. And as the front-runners can themselves bid up the value of the stock by buying it before it officially becomes an index component, the index funds are at times buying the new components at inflated prices.

Investors in Russell 2000-tracking funds can suffer as a result. Indeed, one study looked at "abnormal" price movements of stocks entering and exiting the Russell 2000 and concluded that they resulted from front-running. The study's authors estimated that from 1989 to 2002, front-running cost Russell 2000 funds 1.3% in annual returns. That may not sound like a lot, but it's a huge margin of error in the index-fund realm.

Nonindex mutual funds may also see their holdings affected by the upcoming Russell index reconstitutions. Take the case of the fine  TrendStar Small Cap  fund. Its managers, Tom Laming and James McBride, were puzzled late last June when four of the fund's holdings--CBRL Group  (CBRL), CACI International (CAI),  Affymetrix , and  Bisys Group --all dropped sharply during the last 30 minutes of trading. This took place on last year's Russell reconstitution day, and it happened to fall on the last trading day of the second quarter of 2006. TrendStar Small Cap is a fairly concentrated fund (with some 50 holdings), so those four stocks' collective drop contributed to its 5.3% loss that quarter.

As it turned out, these stocks sold off sharply because they were exiting the mid- and large-cap Russell 1000 Index and entering the small-cap Russell 2000. Much more money is invested in the Russell 1000 than the 2000, and these stocks' steep declines resulted from the big imbalance between Russell 1000 sellers and Russell 2000 buyers. Of course, these were merely technical forces weighing on the stocks and had nothing to do with the companies' fundamentals. All four stocks recovered, and the TrendStar fund turned in a respectable calendar year 2006. In any case, because active managers can buy and sell holdings when they choose, the effects of reconstitution on nonindex funds are usually fleeting.

Competing Small-Cap Indexes
To answer criticisms about the vulnerability of the Russell 2000's reconstitution to front-running, Russell recently said it would take steps to slow additions to and deletions from its indexes. If, on the rebalancing date, a stock is within a certain range above or below the Russell 2000's market-cap boundaries, it will stay in the index for another year. This will create buffer zones that limit the migration of stocks into and out of the index. That means fewer opportunities to front-run the index, fewer trades that could adversely affect prices for index managers, and ultimately lower transaction costs for investors.

Despite the methodology improvements, Russell 2000 is playing catch-up with other indexes that have already taken steps to combat front-running and market-impact costs. For example,  Schwab Small Cap Index  follows an in-house index that has a higher average weighted market cap--$1.5 billion versus $1.1 billion for the Russell 2000. Still, it's a fair approximation of the small-cap universe. Moreover, because this Schwab offering is the only fund that tracks its proprietary index, management needn't queue up with other indexers to buy and sell stocks on reconstitution day. And because the Schwab reconstitution takes place on an unspecified date (although the fund's purchase and sales data suggest it occurs sometime in the winter), hedge funds and day traders likely won't try to front-run it. Alas, the 0.57% expense ratio for the fund's investor shares is expensive and dims this fund's luster.

 Vanguard Small Cap Index (NAESX) and its identical twin,  Vanguard Small Cap ETF (VB), track the MSCI U.S. Small Cap 1750 Index. MSCI was an early adopter of buffer zones, which have limited turnover here but also have led to style impurity: Because stocks that graduate into mid-cap territory don't immediately leave the index, this fund can be left with a sizable mid-cap stake in the aftermath of a small-cap rally. Still, like the Schwab fund, this is the only fund that tracks its bogy, so front-runners won't likely be a problem here. And these funds are bargains: The traditional open-end shares charge 0.23%, while the ETF costs 0.1%. Given that these Vanguard offerings are cheap (unlike the Schwab fund) and track an index no one else tracks (unlike Russell-2000-trackers), they're arguably the best small-cap index offerings.