Tracking difference (sometimes referred to as excess return) and tracking error are important metrics to consider, especially when evaluating traditional index ETFs.
Understanding what they measure can help you make smarter investment decisions.
Tracking difference for index ETFs
Tracking difference measures an ETF's performance against its benchmark index over a specific period of time. Calculating tracking difference is fairly simple: Subtract the index's total return from the ETF's total return.
Tracking difference can be positive or negative and reveals the extent to which an ETF outperforms or underperforms its benchmark index.
Factors affecting tracking difference
| Benchmark related factors | Negative | Negative/Positive | Positive |
| Index changes/turnover | x | ||
| Illiquidity of securities | x | ||
| ETF/fund related factors | Negative | Negative/Positive | Positive |
| Expense ratio | x | ||
| Other costs (e.g., commissions) | x | ||
| Sampling/optimization method | x | ||
| Management skill and experience | x | ||
| Securities lending | x | ||
| Purchase and redemption fees (paid to fund) | x | ||
| Fair-value pricing | x | ||
| Cash drag | x |
Tracking error
ETF providers define tracking error in different ways. The formal definition of tracking error is the annualized standard deviation of tracking difference. In other words, while tracking difference measures the amount by which an ETF's return differs from that of its benchmark over a specified period, tracking error measures the variability of tracking difference over time.
For example, if the tracking error is 50 basis points, about two-thirds of the time the ETF's excess returns are expected to be within 50 basis points of the average excess return. A lower tracking error would suggest lower variability of the excess return.
If your primary objective is seeking total return over a long-term time frame, then excess return is a more important measure than tracking error. However, over the short term, you may care more about performance consistency and want to minimize volatility, in which case you may wish to focus on tracking error.
In the hypothetical example shown here, investors seeking stronger long-term returns may find Fund A the better choice, despite its higher tracking error. However, investors who value returns that don't deviate too far from the benchmark may be attracted to Fund B, despite its lower average returns (i.e. greater negative tracking error).
When comparing funds in real life, you might not find such a clear-cut trade-off between tracking difference and tracking error. Other factors, such as asset allocation, index methodology and cost should also be evaluated before selecting an investment.
Key causes of tracking error and tracking difference
In an ideal world, ETFs would perfectly track their benchmark indexes and tracking difference and tracking error would not exist. However, from a practical standpoint, a number of factors work to make that ideal impossible to achieve.
Fees
By creating a drag on performance, fees are the most common contributor to negative tracking difference. Be sure to evaluate all of a fund's fees, including the trading costs, which are not included in the management expense ratio (MER). Swap fees associated with synthetic ETFs are also not included in the MER. They can vary over time and consequently can be an important cause of tracking error and tracking difference.
Management expertise
A good index fund manager will understand when to use a full replication approach and when a sampling or optimization approach may be more appropriate. A manager must also be adept at handling index constituent changes, index reconstitutions, fund cash flows and more.
Replication methodology
Replication methodology can be a major contributor to both tracking error and tracking difference. Fully replicated ETFs tend to have lower tracking errors than optimized and sampled ETFs.
Taxation
Depending on the domicile of the ETF, there may be withholding taxes both on dividends in the underlying portfolio and also on distributions of the ETF itself. Those taxes should be considered when analyzing and comparing the performance of ETFs. Withholding taxes on dividends in the underlying portfolio are reflected in the ETF's net asset value. As a result, they also have a direct impact on an ETF's performance and any tracking difference.
Fair-value pricing
Fair-value pricing is an adjustment to a closing price designed to correct pricing discrepancies caused by time zone differences among the global financial markets. Fair-value pricing is applied because the market value may be stale and not a real reflection of the securities' actual (fair) value.
Some ETF providers use fair-value pricing if the value of a security it holds has been affected by events occurring before the local market's close but after the close of the primary markets or exchanges on which the security is traded. The policy is most commonly used when valuing foreign securities.
Securities lending
Securities lending is a common practice among mutual funds, ETFs, pension funds and insurance companies, which lend securities from their portfolios to broker-dealers, hedge funds and investment banks in return for a fee. Collateral, generally in the form of cash and/or government securities, is delivered by the borrower in an amount greater than the loaned securities' value. In addition to the fee that a lender can generate from lending a security, the lender also earns interest by reinvesting the cash collateral.
Securities lending can benefit investors by offsetting some of the ETF's expenses, resulting in better performance. The degree to which securities lending benefits investors depends on how much of the lending revenue the fund company keeps. For Vanguard funds that lend securities, 100% of net lending revenue is returned to the funds.