Selecting An Index
But what of the index itself? Does the choice matter? Of course it does.
The process may be complicated by the fact there are far more indexes than index providers, so there is more information to sift through before making a selection.
Continuing the pension example is useful here. Suppose a pension fund decides to benchmark its U.S. equity allocation to a large-cap index from one provider and a small-cap index from a different provider. There are more than a few funds that use the S&P 500 to benchmark their large-cap mandates and simultaneously employ the Russell 2000 to benchmark their small-cap mandates. For a long time, this was the accepted standard.
Even if both of the indexes were selected because of their popularity and broad use, the combination introduces gaps and overlaps in the fund’s aggregate exposure. The S&P 500 is a basket of 500 stocks that are selected by committee. Although they are generally large, the stocks are certainly not simply the largest 500. (If they were, S&P wouldn’t need a committee for selection.) Roughly speaking, the stocks are chosen subjectively based on how well they collectively represent the overall U.S. market, rather than strictly based on market capitalization. As a result, the S&P 500 contains the largest 350 or so, and then a smattering of smaller securities further down the spectrum of market capitalization.
In contrast, the Russell 2000 is a basket of stocks that includes the 2,000 stocks that fall below the top 1,000 stocks in the U.S. market, in terms of market capitalization. When paired with the S&P 500, there are hundreds of stocks that are simply not accounted for by either index—and a few that are, that appear in both indexes.
Although perhaps convenient, the combination is not an appropriate measure for a pension fund seeking to benchmark the total portfolio. Gaps and overlaps tend to create opportunities for managers to take risks and earn additional compensation for “beating the index” when, in fact, such risks are simply not being accounted for properly. A more appropriate benchmark could be created by using complementary indexes that provide seamless exposure to, and measure of, the U.S. equity opportunity set. In short, index selection creates unique risks if the indexes involved are not constructed using a consistent methodology, even if both indexes (and both index providers) are of the highest caliber.
Not surprisingly, an ETF provider that is more concerned with covering a specific segment of the market—as opposed to a particular asset class within the context of a broad-market investment scheme—is going to be focused on a slightly different set of considerations. When scrutinizing individual indexes that supposedly cover the same slice of the market, the seemingly significant similarities of purpose can obscure some of the more nuanced—but highly relevant—design differences. For example, is the index designed with an eye for complete coverage or for efficient representation? How many holdings does it have? What is the turnover? How does it weigh individual securities? The list goes on, with the user’s choice of index influenced by specific needs.
Users should look beyond the index itself to its governing principles, considering how the index defines its stated objective, how well the index addresses the user’s investment objective, how well it adheres to that objective, and whether it is transparent and rules-based. A firm grasp of these facets of the index is crucial to evaluating whether any perceived tracking error, performance shortfall and unnecessary costs are the result of poor index construction and maintenance or simply anomalies in that particular market’s performance.
These are not easy questions, and truly identifiable answers may not exist until an index is tested, as evidenced by what occurred in the markets during the dot-com bust and the financial crisis. During those periods, technology indexes, homebuilder indexes and dividend indexes faced serious and previously unanticipated challenges.