When it comes to choosing an index tracker, the great Tim Hale
Tracking error (or tracking difference) is the best measure of the true cost of your index tracker. It reveals by how much the returns of your fund have deviated from the index it’s supposed to match.
In other words, tracking error shows you the impact of all fund costs, including the charges that don’t show up in the Total Expense Ratio (TER).
To demonstrate how big a dent those hidden costs can make in your returns, I’ve compared the tracking error of three of the cheapest FTSE All-Share index funds.
Cheap FTSE All-Share trackers compared by TER
First, let’s compare our index funds by TER – the simplest match-up criteria for lookey-likey trackers.
| Index Fund | TER |
| Fidelity Moneybuilder UK Index Fund | 0.3% |
| HSBC FTSE All-Share Index Fund | 0.27% |
| Vanguard FTSE U.K. Equity Index Fund* | 0.15% |
* The Vanguard fund also levies a 0.5% initial fee for stamp duty.
The Vanguard fund is the cheapest of the three by TER (even with its initial fee on new contributions it will beat HSBC after 9 years). But the difference between the funds is nothing to lose sleep over.
If you stuck £300 a month away for the next 20 years then your Vanguard pot would be 1.2% bigger than a Fidelity pot.
You’d be about one and a half grand richer by the end. Woot!
Let’s see what happens when we bring tracking error into play.
Cheap FTSE All-Share trackers compared by tracking error
| Index / Index Fund | Return | Tracking error | TER |
| FTSE All-Share index | 0.7% | – | – |
| Fidelity Moneybuilder UK Index Fund | -0.12% | 0.82% | 0.3% |
| HSBC FTSE All-Share Index Fund | 0.2% | 0.5% | 0.27% |
| Vanguard FTSE U.K. Equity Index Fund | 0.63% | 0.07% | 0.15% |
Performance figures are for one year: 15/2/11 to 15/2/12.
The tracking error1 is purely the difference in performance between a tracker and its index.
The lag in performance is generally caused by TER, transaction costs2 and sampling error3.
Those extra costs make all the difference:
- The Fidelity fund didn’t cost 0.3% in the last year but 0.82% – getting on for three times as much.
- The HSBC index fund cost nearly twice as much as its TER indicates.
- At 0.07%, the Vanguard fund cost less than half as much as its TER. Magic? See below.
The Vanguard fund is still the cheapest (it outstrips HSBC after 3 years at the above rates), the Fidelity fund is still the most expensive but the gap between them has widened significantly.
And now the gulf in class is worth worrying about.
Put £300 away in each fund for 20 years and your Vanguard pot would end up 8.2% bigger than the Fidelity one – assuming the tracking error stays constant.
A false economy
Bear in mind I’m not saying cut and run from your current index tracker on the basis of the above. There are a number of other factors to take into account (as ever) and I’ve covered these in a follow-up post.
Also I must warn you that I plotted the self-same comparison via Hargreaves Lansdown and Google Finance and got completely contrary results. The performance data above comes from Hargreaves Lansdown but, when I checked on Google Finance, the result reversed – Fidelity won and Vanguard lost.
Hard though it is to diss the data-masters, I have strong reason to believe that the info on Google Finance is wide of the mark – specifically because it disagrees with Fidelity’s own published results (which tally precisely with the Hargreaves Lansdown data).
Hargreaves Lansdown takes its data from Financial Express, which is the same outfit that runs Trustnet.
Still, even though it’s harder to get a straight answer in the investment world than from a politician’s expense claim form, I’d urge all passive investors to investigate the tracking error of their funds, where possible.
We get caught out all the time as consumers by failing to account for the full cost of things. We buy big houses without regard for the big maintenance bills, dream cars without thinking about their insatiable demand for fuel, and cut-price printers that need their cartridges changed more often than a baby’s nappy.
The same is true for index trackers, so don’t stop at TER – use tracking error to nail the true cost of a fund.
Take it steady,
The Accumulator
Comments on this entry are closed.
could you post some links to the data sources on tracking error?
also, does anyone have an account with DataStream? Is it possible to get free access?
I’m keen to try and recreate some of the charts RIT (the person not the fund) put together last year
his blog seems to have gone quiet of late which is a shame as his data analysis was very uesful
Vanguard use some cautious security lending to help their returns.
Regards the error, the trackers tend to “jiggle” around the index, and that return was pretty slim, so it depends on the exact sampling times. Maybe!
@gadget
tracking error is probably averaged over the year from several measurements to avoid that sort of thing
@All — Next week’s (next Tuesday) follow-up post from T.A. has a bit more info on getting hands-on with tracking error for yourself. Watch this space! 🙂
Would be very interesting to know how the tracking differences above were calculated. I tried comparing tracking differences for the HSBC and Vanguard All-Share funds using the figures for fund and FTSE performance from the funds’ own factsheets, but they varied significantly depending on which month’s factsheet you used – in other words, depending on the data point. Your figures support my hunch, that Vanguard funds have lower tracking differences, but it is only a hunch. Did you go the extra mile, unlike me, and use multiple data points and average them out?
Anyone interested in sharing the pain I experienced trying to work this stuff out might be interested in these:
http://forums.moneysavingexpert.com/showpost.php?p=48160035&postcount=47
http://forums.moneysavingexpert.com/showpost.php?p=48797409&postcount=49
If you’re a real masochist, try reading my other contributions to the thread:
http://forums.moneysavingexpert.com/showthread.php?t=3567935
@TI,TA
that sounds even better – many thanks
It would be very interesting to see the same comparison over say 5 and 10 years, in case the results are different over longer term – yes, I hold one of these three trackers!
There are platform fees associated with Vanguard too, whereas it is possible for HSBC and Fidelity can be held for no fees.
Also, it appears that Vanguard ups its TER for a lifestyle type fund compared to holding the individual ingredients separately.
If I recall correctly, the Vanguard 100% equity fund TER is 0.33%, whereas holding the underlying funds separately would be only 0.24%. So looks like there’s an extra 0.09% to pay for automatic rebalance. That’s based on my own calculations which may have gone astray somewhere, but was the eventual conclusion I came to.
I’d be interested to know people’s thoughts. Keeping all fees to a bare minimum is obviously an important part of the strategy.
1. TA, as ever, thanks for this.
2. Here I think it’d be useful to remind us that ‘tracking error’ can sometimes even work in our favour – that is, fund total return > index total return.
3. For example, see the ETF XESC (which I own).
I don’t get why Vanguard performance is split off from the platform fees either. If you hold it in an ISA you get to invest up £10k a year so you have to take the hit on platform fees. If you have £100k in your back pocket and you aren’t looking for temporal diversity by dripfeeding you can go direct to them, but then you get to pay tax on the holding, which is one hefty fee of its own.
Vanguard looks like a classic bait and switch to me – yep, can’t argue with the exemplary performance. You just have to pay over the odds to get hold of it in practice 😉
Re tracking error, I recently came across an interview with the fidelity rep regarding the tracking error issue and he stated one of the reasons was that fidelity did not do stock lending, like some others. Also, I suspect there must have been some sampling error, rebalancing issues with fund flows affecting tracking possibly?
@ermine
i reckon its the platforms fault taking all those backhanders to make a lot of funds look like theyre free
in a perfect world i’d like to see a flat platform fee irrespective of holdings and no further charges (and miniscule dealing fees reflective of their true cost)
prob never happen though, even when we live in post RDR nirvana
@ben – On BestInvest the £50 (ISA) or £100 (SIPP) + VAT fee covers anything you care to hold. Of course, funds hit you for more fees, but that’s funds for you.
Dealing fees for equities are £7.50 a pop as long as you have £50k with them, and this can include all your accounts including family members.
@ Ben – as mentioned, I used HL’s charting tools. Unfortunately it doesn’t save comparisons so can’t link specifically. I go deeper into the methodology in my next post, though.
@ Saveonarola – Because I am a masochist, I’ve read all of that and I can only salute your dogged sleuthing. Great work. Thanks for the links, and if anyone wants to have their heads bended even more by tracking error then it’s worth your while reading Saveonarola’s journey into the heart of darkness.
@ Neil – I agree, sadly you can’t, at least not with Vanguard cos it didn’t exist.
@ SB – yes, you’re right you can’t avoid platform charges and dealing fees for Vanguard. Once you throw in those variables it depends on how much you invest and how long you hold for. The more you invest, the longer you hold, the more Vanguard outstrips its rivals.
Yes, you pay more for a LifeStrategy fund – I guess that’s the cost of convenience and Vanguard essentially managing your portfolio for you.
@ Alex – absolutely, every now and then there’s an upside. Though if the positive tracking error is too high it’s telling you the tracker is badly managed just as much as if it was deviating on the downside. It just hurts less 😉
@ Ermine – Ben speaks truth. You’re not paying over the odds, you’re just paying so that you can see it. One of the reasons VG has a lower TER is because they’re not paying trail commission to the platforms whereas HSBC are. That difference shows up in the tracking error.
@ Saveonarola – sorry, meant to say, I kept things simple and used a single data point for the above. You can see in the next post multiple data points that show the same general trend. For now I just wanted to show how investors can make a comparison to get a feel for how well their trackers can perform. And given the paucity of data available to us as retail investors I think a ‘feel’ is the best we can do. It’s a bit like the bid-offer spread – you need to monitor it over time to get an idea of the likely parameters you’re dealing with. What we need is an industry standard way of measuring tracking error that goes on all factsheets, and that investors can reliably use. Haven’t spotted that in any of the RDR reforms.
I may be wrong here, but couldn’t the tracking error work as a positive too? Could the tracker fund supposedly tracking the FTSE get the error in your favour?
‘teamdave’, you’re right – see my comment above (#9). Alex
@Alex @teamdave — As I understand it, a large versus a small tracking error is bad, even if it works in your favour in some particular year. That’s because something underlying is deviating your returns from the index, which is after all what you’re looking to follow. That something could turn negative/go away, in time, reversing your advantage. The ideal index tracker would therefore have an error of 0, reflecting perfect cost-less tracking.
But no, I wouldn’t send a £10 ‘outperformance through tracking error’ back to the fund manager in a brown paper envelope, either. 😉
1. Hi TI, thanks for your comment.
2. This raises some interesting issues, doesn’t it? Not least exactly how a tracker should, or should not, behave.
3. Consider the ETF XESC (see my disclaimer above).
4. This is a synthetic ETF, with TER p.a. = 0.00%. Not low-cost, but no-cost (well, at least by TER, anyway). We’d expect the tracking error to be zero, too (by definition).
5. Yet XESC consistently outperforms the relevant index – the tracking error is in our favour. Now, I understand how this outperformance could be perceived as undesirable in that it represents deviation from the underlying index – the performance we actually want.
6. Nevertheless, I’m happy here for two reasons. First, the outperformance is not really significant – we’re still getting the index total return, plus only a little extra. We’re not overweighting/underweighting index constituents, for example. Second, look at that TER: we know it must be zilch.
@Alex — Well, we’re into the realms of synthetics here, and T.A. is our man for that. But I’ll have a stab.
I am instinctively skeptical of the TER of 0%. I don’t think DB is offering this tracker out of the goodness of its heart! There will be fees for the underlying swaps etc. Presumably it has lodged the incoming money against a basket of something somewhere, and the fees are paid out of that, and for some reason DB is returning a bit of the leftover to owners of the tracker.
Perhaps its collateral isn’t performing as it expected, it is outperforming, and it is obligated under some agreement in the structure of the ETF to return something to the owners? (Or perhaps it happens via some other underlying mechanism of some sort). Please understand I am speculating — I am no expert on how synthetic ETFs are constructed, but in general this is just the sort of dubious opacity that sends shivers down my spine! 🙂
Anyway, if something isn’t performing as expected, then presumably it could revert the other way sooner or later. Probably a small matter if the tracking error is very small as you say, especially compared to other risks of holding a synthetic ETF. (The tracking error isn’t going to suddenly wipe the ETF out, like counter-party risk might, for instance).
I personally far prefer physical ETFs, though I’d not be averse to a synthetic for something I couldn’t access otherwise. (That would definitely NOT include tracking a basket of European large caps like XESC! I mean something like the Indonesian kumquat producer index).
But pays your money takes your choice etc. I’m not saying this ETF is dodgy at all, just that there are risks to it (as a synthetic) that personally I’m not partial to.
TI, why are you suspicious of something (TER p.a. = 0.00%) seeming too good to be true? Er… 🙂
One of the purported uses of Synthetic ETFs is to enable large banks to get illiquid assets off their balance sheets and neatly tucked away in the collateral baskets of their subsidiary’s ETFs. Meanwhile the ETF brings in a ready flow of cash that can be turned to other purposes. If that’s the case then it’s easier to see why a TER of 0% makes sense for a multi-tentacled, too-big-to-stop global bank.
This article is too superficial.
The HSBC trackers reduced their TER hugely last year and so any comparisons longer than 1 year are meaningless.
Why restrict the comparison to 3 funds? Have you never heard of L& G and Virgin? They are far bigger than the minnows you have chosen.
Are you aware that Vanguard require a min. of £100,000 to start with?
@Franco — Bit of an aggressive reply. These are three of the cheapest trackers in their class, so the comparison is perfectly warranted. It’s not possible to cover every tracker every time we do an article, and regular readers will know we’re very much “aware” of L&G and Virgin. This article illustrates a point, it doesn’t claim to survey the market. The last 18 months of coverage have been doing that just fine.
We know all about the £100,000 minimum. It can be circumvented by going through a broker, which presumably you are not aware of. See this article: http://monevator.com/2010/10/12/cheap-vanguard-index-funds/
Thanks for stopping by and commenting. Please be aware this is a friendly blog, and I’ve no intention of allowing it to turn into a stream of nastily-put commenting like on so many websites. My decision on what counts as ‘nasty’ is arbitrary and final. Comments in the tone you’ve used above from anyone are liable to be deleted, as I see fit.
Hi Franco, the idea was to show the effect tracking error can have and to enable readers to do their own comparisons with whatever funds they like. As such here’s part two:
http://monevator.com/2012/02/28/tracking-error-how-to-measure-it/
Also, are you aware that HSBC cut their index fund TERs in 2009, not last year as you state.
It would be very useful to have a table of the TER and tracking errors of as many index trackers as possible, with priority to the biggest ones, but including some with outrageous TER, if you dare. After all the financial industry is the one for which the phrase rip off Britain was coined and I think writers have a social duty to name and shame its greediest members.
I don’t think there’s any secrecy about who charges high TERs. Just go to these links and sort by TER:
http://www.investmentuk.org/investor-centre/fund-search
http://www.hl.co.uk/funds/index-tracker-funds/view-index-tracker-funds
I have a simple (possibly stupid) question, I would greatly appreciate advice for! If you already hold a less ideal index tracker outside an ISA, does it ever make any sense for you to trade for a lower cost alternative with lower tracking error? In my case, I have thought about trading a £40K holding in the Fidelity Moneybuilder UK Index Fund for the Vanguard equivalent given my 40 year horizon. But I might well incur capital gains tax, as well as possible losses if the market moves.
So, should I trade or is it generally more cost-effective/sensible to stop adding to the less ideal index tracker and start adding its better Vanguard equivalent as suggested in the Accumulator’s Word of Warning? With the latter “keep the so-so fund but add the better fund” approach I can foresee a rather complicated portfolio as more economical Vanguard funds hopefully become available.
(I also posted this question on the latest Slow and Steady portfolio post but it seems equally relevant to this post)
So if tracking error is the ultimate metric, incorporating all costs and profits of the fund, etc. I have a question. In the 2021 VWRL KIID, past performance is listed as:
Year Fund Index
2013 22.8 22.7
2014 4.3 4.2
2015 -2.2 -2.2
2016 8.0 8.0
2017 24.0 24.0
2018 -9.6 -9.6
2019 26.6 26.5
2020 16.0 16.0
From the above, tracking error is consistently either 0% or +0.1% (i.e. the fund returned more than the index) for every year since 2013 when the fund started. Does this mean that comparing stated annual fees and transaction fees is actually irrelevant? Because it doesn’t matter if VWRL costs a combined 0.22+0.03=0.25%, if those fees are somehow being offset elsewhere by something (stock lending?) that generates profit for the fund.
I’m just a little surprised that the tracking error implies an overall cost of 0.0% (or even a negative cost) when the management and transaction fees are 0.25%. The implication of this is that the fund is free, or even better. And that the rankings/arguments comparing annual fees, TER, etc. are irrelevant, and we should just be comparing tracking error.
Interesting discussion on this topic: https://www.reddit.com/r/eupersonalfinance/comments/ll4ean/the_secret_behind_vwces_022_ter_why_its_not_its/
It seems that regularly “outperforming the index” is attributable to the fact that the index figure is calculated net of dividend withholding tax, and incorporates worst-case withholding tax on dividends, whereas the fund may benefit from taxation treaties that mean it is actually paying a lower tax rate than the worst-case. The gain from securities lending is quoted as 0.03%, which is not enough to offset the management and transaction fees.
Quote: “They’re almost always compared to a net. index which assumes worst-case dividend taxes. (Which is how the funds OP compares can ‘beat’ the index.) It’s (practically) impossible for an index fund to beat the gross index after subtracting dividend taxes and management costs. Think of it like this… how is a passive index fund like IWDA able to get a tracking error of 0 compared to a gross index? They’ll have to compensate for dividend tax leakage (say 0.25% at a 2% dividend yield), management costs (0.2%) and internal transaction fees (usually very low, so let’s say 0.03%). The income from securities lending is about 0.03% too, so that would cancel out the latter. But where would the other 0.45% come from in a passive index fund if we would be comparing to a gross index?”
So a fund like VWRL could get maybe something like a 0.15% gain, versus the benchmark, just from paying less tax on US dividends.
Tracking difference and tracking error are required to be published in annual reports, e.g.:
iShares ETFs: https://www.ishares.com/uk/individual/en/literature/annual-report/ishares-v-plc-annual-report-november-2020-emea-en.pdf
Vanguard ETFs: https://www.vanguardinvestor.co.uk/rs/gre/gls/1.3.0/documents/959/gb
Vanguard funds: https://www.vanguardinvestor.co.uk/rs/gre/gls/1.3.0/documents/1524/gb
(One odd thing from the Vanguard fund report – the Vanguard FTSE Global All Cap Index Fund had a 0.634% tracking error, which was more than double the expected, and this was explained as “The difference in tracking error resulted from the 31 August 2020 Fund holiday; the underlying benchmark Index was open for trading as it does not observe the holiday”. Wow – Vanguard’s trackers don’t run on bank holidays…)