My latest story in the The Wall Street Journal’s “Investing in Funds” was about the intricacies of index fund tracking error. The stat, which can be hard to find and even harder to define, is a signal for how well an index fund tracks its benchmark. Some funds and indexes may have an inherent tracking error, due to the nature of the underlying index and the liquidity of the securities.
Using tracking error to compare funds is most useful in the rare situation of two funds tracking the same index and bearing the same cost. Outside of that, tracking error is useful only when viewed as a trend, if that.
The second level discussion around tracking error, however, can be helpful in understanding how index funds are managed. I touched upon some of this in “ETFs are People Too" and will reiterate here. To understand tracking error and index fund management (and why your fund may be better/worse than the next one), study these aspects of your funds management and processes:
Expense Ratio: Expense ratio is usually, but not always, the management fee charged by the fund’s adviser to run the fund. At times, the total expense ratio includes the management fee as well as fees charged by other funds or ETFs held by the ETF. (This pass through notion of “acquired fund fees” gets a little silly. After all, no mutual funds report through the net operational costs of the individual companies they hold.) Total expense ratios can also be driven down to a net expense ratio when the adviser waives or rebates back to the fund any cost above a cap.
Sampling: Also called optimization, sampling is an upfront admission by indexers that not all indexes are actually entirely investible. Bond indexes, in particular, can be hard to replicate based on the unavailability of certain bonds as well as the notion that some bonds are very hard and expensive to trade, regardless of their inclusion in a benchmark index. For stock funds, the ability to sample or optimize is largely dependent on the size of the fund and whether the manager, or market makers, can create baskets of the underlying securities effectively.
Rebalancing and Reconstitution: When an index is rebalanced, i.e. reweighted among its constituents, or reconstituted, i.e. screened for potential entrants or drop-outs, the index announces the changes in advance and then makes the changes several days (or weeks) later. Index fund and ETF managers, however, have to face the operational challenges of actually making those trades as close to real-time as the index shift. While ETF managers hope to make these changes mostly through market makers, at times they will actually have to buy and sell securities on their own.
Securities Lending: When interest rates were higher and demand (and appetite) to short stocks was greater, ETFs (and mutual funds) would enhance returns by lending out stocks and bonds. Due to the manner in which ETF shares are created or redeemed, whereby a market maker delivers a basket of securities or receives a basket of securities, securities lending by ETFs is generally more conservative than mutual funds. When securities are loaned out, they are over-collateralized, meaning that ETF issuer receives slightly more than the value of the security being lent. Net revenue from the loan, after lending agent fees, is income for the fund and can actually offset part of the expense ratio.
Dividends, Interest and Cash: Whether cash comes in through creation baskets or is paid by underlying holdings through dividends and interest, ETFs must reinvest cash in their holdings as well as set some aside to pay the fund’s expenses. For funds organized as Unit Investment Trusts, primarily SPY, QQQ or DIA, dividends are not reinvested in the fund and, therefore, create what is known as “cash drag” during upmarkets and a “cash buffer” during down markets. Regardless of structure, cash accrues daily to pay the manager and contributes to tracking error.
For more on my view of the market for ETFs and other asset management products, read ETFolution at Forbes.com.