Last Updated: 28 May 2021
According to Matthew Tagliani, head of European and Asian ETF product at Morgan Stanley in London, tracking error remains widely misunderstood by investors, while its measurement is a frustrating process.
At the same time, understanding tracking error is a fundamental consideration for investors in ETFs, particularly in Europe, where there are many similar funds based on the same index benchmark, says Tagliani in a recently released report (“Tracking Risk: Know Thy Enemy”).
The Morgan Stanley ETF specialist argues, however, that the traditional formula for measuring tracking error, the annualised standard deviation of return between a fund and its benchmark, measured at a series of evenly spaced observations, is misleading. In most cases, says Tagliani, such a tracking error measurement tends to overstate the actual deviation experienced by a fund, in addition to being highly unstable.
Many ETF and index fund observers tend to distinguish between “tracking difference”, the absolute deviation over time between a fund’s return and that of its benchmark, often explainable by management fees and other fixed costs, and “tracking error”, the variability of tracking differences, which can be due to difficulties in replicating a given benchmark or, in the case of an ETF, shortcomings in the arbitrage mechanism which is supposed to keep a fund’s net asset value (NAV) in line with that of its benchmark.
Tagliani, however, prefers to use the generic term “tracking error”, in turn subdividing the causes of tracking error into three distinct categories. “Systematic deviations”, the first type of tracking error, is produced by predictable, deterministic and recurring sources of error, and is present in every fund, he says.
An ETF’s management fee, swap-related or custody costs, or withholding taxes affecting either a fund or its benchmark are common types of systematic tracking error, says Tagliani. These systematic deviations are a major concern for long-term investors, as they suffer the cumulative effect of such tracking shortfalls, whereas for shorter-term investors such deviations may not matter at all, he adds.
The second type of tracking error, “mean-reverting noise”, represents small deviations caused primarily by trading-related factors, says Tagliani. Such deviations may be the result of many different factors: timing delays meaning that the index and the ETF tracking it are not calculated concurrently; prices being measured at bid or ask, which may differ from a fund’s net asset value; the effect of off-market price “prints” when market makers attempt to conduct arbitrage trades in a closing stock exchange auction; and timing differences between the FX rates used to calculate closing index prices and those used for local equity market closing times.
As an example of such mean-reverting deviations, Tagliani shows that the iNAV (“indicative NAV”) for a hypothetical ETF tracking a global equity index, computed at 14.00 London time, will have to rely on stale data for all the American and Asian stocks it includes. ETF traders, meanwhile, adjust their quotes to reflect the implied prices for underlying stocks in markets that are closed, using proxies such as futures (many of which trade around the clock) to measure the likely movement in equity markets that are closed.
Source: Morgan Stanley
While such mean-reverting “noise” can contribute significantly to the standard measure of tracking error, says Tagliani, it has little impact on the long-term performance of a fund and is therefore not a significant worry for investors.