How to find out the true cost of an ETF: tracking difference

Being a passive investor requires being mindful of costs. Due to compounding, seemingly small differences in costs can lead to large differences in returns over long periods of time.

The Total Expense Ratio (TER) is only a fraction of the cost of owning an ETF. The Tracking Difference (TD) is a better indicator of the actual cost of holding an ETF. Though useful, tracking difference is a tricky metric: it isn’t available for all funds, it fluctuates, it is calculated differently depending on the fund.

“Tracking” what?

An index fund replicates the performance an index. This replication process is not perfect. 

An index is an ideal representation of the market, while a index fund is an actual implementation of it.

In short: a fund has costs while an index does not. A fund needs to pay transaction fees when it buys/sells assets. A fund needs to exchange currencies when buying assets. A fund may only maintain a representative sample of all the holdings in the index. A fund charges investors for its services through the TER. All this leads to the performance of the index fund diverging from the performance of the index it tracks.

This difference between the performance of the index and the performance of the fund is called Tracking Difference. For example: for a fund performance of 9.75% and an index performance of 10% we would have a tracking difference of -0.25%.

tracking_difference = fund_performance - index_performance 

The tracking difference is usually negative – the funds usually underperform their indexes by a small margin. It is possible for a fund to have a positive tracking difference (i.e. over perform its index).

A fund can over-perform its index because of profits from securities lending. Some index funds loan their securities and pass along the financial gains of this activity to the investors.

A fund can over-perform its index because of efficient management of tax liabilities. The index performance assumes that the maximum amount of withholding tax is paid for all shares. But a fund may be domiciled in a country with treaties that allow the fund to save money on taxes.

The Tracking Difference should not be confused with Tracking Error. The tracking error measures the variability of the tracking difference. It is calculated using the standard deviation of the tracking difference. In practice, the tracking error indicates the consistency of the tracking difference.

For example: an index fund which has had a tracking difference of -0.25% on every one of the past 10 years has a tracking error of 0% because the tracking difference is always the same (i.e. consistent).

Why tracking difference matters?

The tracking difference is the actual cost of the fund. You buy an index fund because you want to get the whole index performance. But (usually) you get less than the index performance because of costs.

fund_performance = index_performance - fund_costs
fund_costs = tracking_difference

We often focus on the TER because it is a visible cost clearly published by the fund provider. But the truth is that a fund may have additional (hidden) costs and you can only notice these costs through the tracking difference.

fund_costs = TER + hidden_fund_costs 

Tracking difference is important because there are index funds with a low TER but with a high tracking difference.

Imagine a hypothetical index fund with a TER of 0.1% but with a tracking difference (TD) of 0.5%. Just by looking at the TER, this hypothetical fund looks cheap, but after looking at the TD, you realize it is expensive. Inversely you may also find an ETF with a high TER but a low TD. Imagine a hypothetical index fund with a TER of 0.5% but with a tracking difference (TD) of 0.1%. Just by looking at the TER, this hypothetical fund looks expensive, but after looking at the TD, you realize it is cheap.

You get to keep what you don’t pay. You should optimize for a low tracking difference. If all else is equal, the fund with the lowest cost is the best one. Which means that funds with a low tracking difference are better.

My rule of thumb for TER is that any ETF should not be more expensive than 0.25%. Given TD is a cost we can apply the same rule for it. A fund should not have a TD lower than -0.25%.

Why is tracking difference a “tricky metric”?

Tracking difference is very unreliable on the short term. You need at least 3-5 years of data in order to be confident of a fund’s tracking difference.

Remember the tracking error metric discussed earlier? A fund with a high tracking error may have tracking differences all over the place. You look at it for a 1-year period and it looks great but in the coming year it looks terrible.

This means that for any recent funds you won’t have reliable data on tracking difference. Tracking difference is calculated for the past. There is no guarantee that we’ll observe those tracking differences in the future. You can reasonably assume that a fund with a low tracking error will have a similar tracking difference in the future. But that is all that is, an assumption!

Not all funds calculate tracking difference the same way. You may look at two fund factsheets from different providers and get data that isn’t comparable. Different funds may use currency exchange rates collected at different times. Different funds may use stock prices from international stocks collected at different times. Different funds may even compare themselves against different indexes even though they are supposed to track the same index. Tracking differences are calculated over specific periods. Distinct funds may calculate tracking differences over distinct periods (e.g. calendar year vs. June to June).

Some indexes are harder to track than other indexes. An index may be harder to track because the particular market has higher trading costs or the market’s stocks are hard to sell/buy (i.e. illiquid) or the market is hard to sample. For example, Emerging Markets index funds tend to have higher tracking differences than Developed Markets index funds. This means that the definition of “good” tracking difference may change depending on the particular index being tracked.

How to find the tracking difference for a fund?

The best place to compare tracking differences (TD) of different funds is on the website trackingdifferences.com.  The site is in German, so you may have to use Google Translate to navigate the site.

It is important to note that this site uses a different definition of tracking difference than what I have shown above. I used the following formula:

tracking_difference = fund_performance - index_performance 

But they use the following formula:

tracking_difference = index_performance - fund_performance

This means that for trackingdifferences.com a negative sign means the fund outperformed the index while a positive sign means it underperformed the index. It is the inverse of what we discussed above. The authors did this so it is easier to compare the TD with the TER.

The site provides a single table where you can see the TD of multiple funds over multiple years. You should compare TD across funds instead of seeing it in isolation for a single fund. Here is an example for funds tracking the MSCI World index:

On the table above, the iShares IWDA fund has an average TD of -0.01% over the past 7 years (column with name “TD). In 2018 and 2017, the fund had a TD of 0 (columns with names “18” and “17:)! Its tracking error is 0.02% (column with name “TDV”). You could argue that despite the TER of 0.2%, this fund is outperforming the index by 0.01%: these gains exist likely because of shares lending.

I find the TD column in this table a bit problematic because it does an average over the full period of data. When multiple funds are compared, you end up having averages done over different periods of data which may be an unfair comparison for the funds with the most data. Therefore, I often look at the TDs of the individual years to get a sense of how TD has changed recently.

Some claim that synthetic funds are guaranteed to have the lowest tracking difference because they don’t have to hold the equities of the index. This isn’t necessarily correct. A synthetic fund has costs too like any other fund. Some synthetic funds have low tracking differences and some don’t. The Xtrackers MSCI World Swap – 1C shown in the table has had a TD of 0.1% in 2015 (i.e. underperform the index) while the iShares IWDA fund (a physical sampling fund) has had a TD of -0.1% in that year (i.e. outperformed the index).

The data for the iShares IWDA fund (TER: 0.2%) shows that despite being more expensive than the new SPDR MSCI World fund (TER: 0.12%), it can be cheaper from a TD perspective – it is actually leading to gains over the index. Unfortunately we don’t have TD data for the SPDR MSCI World fund yet since it is a new fund. 

trackingdifferences.com will only add the data if the fund has been around for at least 1 year. For new funds you can only rely on their TER as an estimate of future costs.

The table shows a couple of funds that outperformed the index while others underperformed the index. For example the iShares MSCI World USD (List) had a TD of 0.2% in 2018. This means that the fund outperformed the index.

Are funds that outperform the index good? I don’t know. It depends on what you define as good. If good is a measure of highest performance then these funds are better. But if good is a measure of how well these funds track the index, then they aren’t. A fund with a TD of -0.5% is as “bad” as a fund with a TD of 0.5% when it comes to the ability of perfectly tracking an index. I prefer to choose funds with a TD as close to 0 as possible because index funds are supposed to track the index.

The data from the site on funds tracking the MSCI Emerging Markets index is also insightful. The chart below shows that many of these funds have large tracking differences. This shows why you always have to compare the TD against other comparable funds tracking the same index. 

trackingdifferences.com allows you to search for a specific fund by ISIN and it will show the details of the fund along with a table comparing all funds tracking the same index.

The TD data for the Vanguard FTSE All World is also impressive. It has a average TD of -0.03 for the past 6 years. The last 4 years have lead to a TD of 0! Its tracking error is 0!

Alternatively to trackingdifferences.com you can find details of an ETF’s tracking difference in the factsheet or official website. 

The factsheet shows the returns of the index (a.k.a. benchmark) and of the fund over different periods. You need to calculate the TD yourself by subtracting both.  Below is an example from iShares Core MSCI World ETF’s factsheet.

The information you find on the factsheet is also often available on the fund’s official website. Below is an example from iShares Core MSCI World ETF’s website (Performance section).

How to use tracking difference when selecting a fund?

Tracking difference is an additional data point you can use when selecting a fund. As I elaborated in my “Introduction to Passive Investing” e-mail course, my process for selecting funds involves two phases: Exclude and Prioritize.

In the Exclude phase I typically disregard funds that don’t meet my investment strategy (e.g. index, domicile, use of income).

For example, I exclude any funds with a TER higher than 0.25%. You may also disregard any fund with a TD lower than -0.25% or higher than 0.25%. You’ll need to use both TER and TD because new funds won’t have TD data. There are some funds with high TER (e.g. 0.50%) but a low TD (e.g. 0.1%) but I exclude them as well as I don’t see any point investing in funds with high TER when there are cheaper alternatives.

In the Prioritize phase, I already have funds that are eligible for my investment strategy but need to choose one. In this phase I rank the funds according to my preferences (e.g. fund size, transaction costs, trading volume). This phase ends up being a bit subjective. Often I bump into multiple good alternatives and need to choose one of them.

I typically rank funds with lower TERs in a higher position. You could start ranking funds with lower TDs in a higher position too. In this phase the usage of TD ends up being a sanity check that there isn’t a fund with a low TER but a high TD.

I also think the TD ends up being a useful information when comparing your existing funds with new funds. For example, the iShares Core MSCI World ETF has such a low TD that you may not even consider changing to the new SPDR MSCI World despite the new fund having a lower TER.

Do remember, that the TD is a tricky metric. And like anything in investing, there aren’t any absolutes. It is an additional data point you can use to make a more informed decision.