If you sit down to choose between two S&P 500 ETFs available to a European retail investor, the comparison usually starts with the Total Expense Ratio. The iShares Core S&P 500 UCITS ETF (CSPX) has a published TER of 0.07%. The Vanguard S&P 500 UCITS ETF (VUSA) also has a TER of 0.07%. The SPDR S&P 500 UCITS ETF (SPYL/SPY5) has a TER of 0.03%. The Invesco S&P 500 UCITS ETF (SPXS) has a TER of 0.05%. On the basis of this number, SPYL looks like the obvious winner.
That is sometimes the right answer. It is not always the right answer. The TER is an input to a fund's performance, but it is not the only input, and it is not even the largest input for some funds. The number that captures what a buy-and-hold investor actually receives is the tracking difference, and over multi-year holding periods the tracking difference and the TER can diverge by amounts that exceed the headline fee gap many times over.
This piece explains both numbers, where to find them, why they can diverge, and how to use them together to make a more informed ETF choice than the TER alone allows.
What the TER actually is
The Total Expense Ratio, sometimes called the Ongoing Charges Figure under EU regulations, is the percentage of fund assets consumed annually by management fees, depositary fees, audit fees and other ongoing administrative costs. The figure is calculated under standardised UCITS rules and is required to be published in the fund's Key Information Document (KID), now harmonised across the EU under the PRIIPs Regulation. The methodology is set out in ESMA's PRIIPs Q&A guidance.
The TER is, by construction, a forward-looking estimate of the annual drag on the fund. A fund with a TER of 0.07% will, all else equal, return approximately 0.07 percentage points less per year than its benchmark index. Over a 30-year holding period, a 0.07% annual drag compounds to roughly 2.1% of cumulative wealth: not nothing, but small.
What the TER does not include is informative. The figure excludes:
- Trading costs inside the fund. When the fund rebalances its portfolio (for example, when a stock is added to or removed from the index), the broker spreads, market impact and any explicit commissions are paid out of fund assets but are not part of the TER.
- Withholding tax leakage on dividends. Many indices are calculated on a "gross" or "net" basis that assumes a particular withholding tax treatment, which the fund itself may or may not match depending on its domicile.
- Securities lending revenue (typically a positive contribution). Funds that lend securities to short sellers earn revenue, which is added back to fund returns. This is one of the few items that can make the actual return better than TER would predict.
- Sampling error for funds that do not hold every security in the index.
- Spread costs at primary market creation and redemption, which are absorbed by authorised participants and indirectly by other shareholders.
For two funds tracking the same index, with identical TERs, the items above can produce divergent actual outcomes. For two funds with different TERs, the items above can sometimes mean that the higher-TER fund actually performs better.
What tracking difference is, and where to find it
Tracking difference is the simplest possible measure of how well an ETF tracks its index: it is the actual total return of the ETF minus the total return of the index it is supposed to track, measured over a defined period. If the index returned 10.0% over a year and the ETF returned 9.85%, the tracking difference for that year is negative 0.15%.
Tracking difference is a backward-looking, observed number. It captures the net effect of the TER plus the trading costs, the withholding tax leakage, the securities lending revenue, and the sampling error, all combined into a single figure. For a buy-and-hold investor, tracking difference is the number that matters: it is, by definition, the gap between what the index did and what the investor received.
Tracking difference is published by some fund providers in their fact sheets, but the data is more easily aggregated and compared on third-party platforms. justETF and Morningstar both publish multi-year tracking difference data for European-listed UCITS ETFs. The pattern that emerges from the data is informative.
For broad developed-market equity index ETFs (S&P 500, MSCI World, FTSE 100, EURO STOXX 50), the absolute value of tracking difference is typically very close to the TER, sometimes slightly worse and occasionally slightly better. The differences between funds tracking the same index are usually a few basis points per year. For these large, liquid, well-arbitraged index funds, the TER is a good first approximation of the tracking difference, and a fund with a meaningfully lower TER will usually have a meaningfully better tracking difference.
For more specialised indices, the picture is more variable. Emerging market funds, smaller-cap indices, sector-specific indices, and ESG-screened indices often have tracking differences that diverge significantly from the TER, in either direction. Funds tracking the same emerging market index can differ by 30 to 80 basis points per year in observed tracking difference, swamping the typical 10 to 20 basis point TER gap between competing products.
Tracking error: a different (and less practical) statistic
Tracking error, despite the similar name, is a different statistic from tracking difference. Tracking error is the standard deviation of the daily or monthly tracking difference, annualised. It measures the consistency of the gap between fund and index, not the size of the gap.
A fund that tracks its index almost exactly each day, with a small consistent drag of 5 basis points per year, will have a low tracking error. A fund with the same average tracking difference but more day-to-day variability will have a higher tracking error. For two funds with similar tracking difference, the lower tracking error fund is more predictable, which is helpful for institutional investors running large portfolio overlays but is rarely material to a retail buy-and-hold investor.
The popular financial press sometimes conflates the two terms, using "tracking error" loosely to mean what is technically tracking difference. For a retail investor, the practically useful number is tracking difference: how much, on average, did this ETF underperform its index? Tracking error proper, the consistency measure, is informative but secondary.
The dividend tax dimension
One of the most important hidden drivers of tracking difference, particularly for funds tracking US-heavy indices like the S&P 500 and MSCI World, is dividend withholding tax. We discussed this in our earlier piece on ETF domicile. To recap briefly: an Irish-domiciled UCITS ETF holding US stocks pays 15% withholding tax on US dividends, while a Luxembourg or other European UCITS ETF pays 30%. The 15 percentage point difference applies to the dividend yield only (currently around 1.3% for the S&P 500), so the after-tax yield drag for an Irish fund is approximately 0.20% lower than for a non-Irish equivalent.
This is approximately three times the headline TER of the cheapest S&P 500 UCITS ETFs. A Luxembourg-domiciled S&P 500 ETF with a 0.05% TER will, in practice, often have a worse tracking difference than an Irish-domiciled S&P 500 ETF with a 0.07% TER, because the dividend tax effect dominates the TER difference. This is exactly the kind of pattern that the TER comparison alone cannot capture, and that the tracking difference data does.
How to actually compare two ETFs
For a retail investor choosing between two ETFs tracking the same index, a sensible workflow is:
- Confirm both funds track the same index. "S&P 500" is one specific index; "S&P 500 ESG" is a different index with different constituents and different historical returns. Like-for-like comparison only works when the benchmark is identical.
- Look up the tracking difference for both funds over the past 3 to 5 years. justETF and Morningstar both make this comparison straightforward. The fund with the smaller (less negative) tracking difference is the one whose investors received more of the index return.
- Confirm the tracking difference data is consistent year over year. A fund with a 0.10% tracking difference in three of the last five years and a 0.50% tracking difference in two of them is a different proposition from a fund with a 0.20% tracking difference in every year. The latter is more predictable.
- Use the TER as a tiebreaker, not the headline criterion. Where two funds have similar tracking differences, lower TER is better, because it is a forward-looking commitment to a lower charge. Where the TERs differ but the tracking differences contradict the TER ranking, trust the tracking difference data.
The TER is a useful, regulated, standardised number. It is also incomplete. The investor who has spent twenty minutes learning where to find the tracking difference data, and how to read it, has acquired a meaningful information advantage over the investor who is still picking ETFs purely on the basis of the headline expense ratio.