Update: There are some concerns about the reliability or our reading of the data source we’ve used to compile this feature, which it’s being suggested does not always take account of dividends in its returns. We have modified this copy to reflect this, and will investigate further.
Despite claims that the Total Expense Ratio (TER)1 is the best predictor of returns from passive index funds, it often only does half a job when pinpointing the real costs of a tracker.
Just as nature selects for big brains, sharp teeth, and the ability to waggle your dangly bits in an amusing fashion, smart passive investors look for low tracking error as a highly desirable trait when choosing the fittest index funds in the investment jungle.
But reliable tracking error data is devilishly hard to find. And summoning the will to trawl the internet can be harder still for investors of the couch potato kind.
Help is at hand
Never fear, the alarm has been sounded. Monevator’s sleeper agents have once again been activated and prodded into action to fill the yawning gaps on the ‘to do’ lists of Britain’s passive investors.
Our plan this time is to host a regular performance review of the main FTSE All Share trackers versus their index.
Hopefully this survey will:
- Highlight that TER or the Ongoing Charge Figure (OCF) isn’t automatically the best guide to your fund’s peer group performance.
- Expose shoddy tracker construction – we want to find out which funds are more costly than they look.
- Demonstrate how you can build up a better picture of the available index trackers with just a little extra effort.
Tracking error comparison table
The following table shows the returns from a variety of All Share index tracker funds, plus the index itself, marked in blue.
I’ve ranked the entries by order of their 3-year returns. Funds placed higher than the All Share index returned slightly more than it over three years. Funds placed below it lagged the returns from the index over that period.
| Fund name | 1 yr | 3 yr | 5 yr | 10 yr | OCF |
| DFA UK Core Equity | 17.7% | 10.9% | 5.5% | – | 0.36% |
| Edinburgh UK Tracker Trust | 17.2% | 10.8% | 5.8% | 9.5% | 0.30% |
| FTSE All Share index | 15.9% | 10.7% | 5.8% | 10.3% | – |
| Vanguard FTSE UK Equity Index | 15.2% | 10.4% | – | – | 0.15%2 |
| Royal London FTSE 350 Tracker | 14.8% | 9.8% | 5.3% | – | 0.12%3 |
| HSBC FTSE All Share Index Ret | 15.3% | 9.8% | 5.0% | – | 0.28% |
| F&C FTSE All Share Tracker 1 | 15.2% | 9.7% | 5.1% | – | 0.43% |
| M&G Index Tracker A | 15.4% | 9.6% | 5.0% | 9.7% | 0.46% |
| Virgin UK Index Tracking | 14.5% | 9.6% | 4.9% | 9.4% | 1.00% |
| Henderson UK Index A | 15.0% | 9.5% | 4.9% | 9.4% | 0.70% |
| Fidelity Moneybuilder UK Index | 14.9% | 9.5% | 4.8% | 9.5% | 0.30% |
| L&G UK Index Ret | 14.9% | 9.3% | 4.9% | 9.5% | 0.56% |
| db X-trackers FTSE All Share ETF (See note) | 11.7% | 9.0% | 4.7% | – | 0.40% |
Note: The db X-trackers data may not be representative — we’re investigating. Annualised performance figures
The tracking error calculation is simple. Just subtract the fund’s performance figure from the return of the index. (Some call this tracking difference).
For example, the tracking error figure for the M&G Index Tracker A fund over one year is:
Tracking error = (15.9-15.4) = 0.5%
The science bit
There are a few technicalities to note if you want a fully rounded picture of what I’m doing to get my table. (This section is eminently skippable if you just want to get straight to the point).
- The performance figures shown are annualised returns. They are brought to you by Trustnet’s excellent charting tool, which also shows discrete annual returns and cumulative returns.
- The OCF is the explicit cost of owning the fund now. The tracking error reveals the total cost including hidden charges such as turnover, taxes, and index replication deviancy.
- The OCF will have been higher for many funds in previous years.
- All funds are index trackers with the exception of the Dimensional Fund Advisors’ (DFA) UK Core Equity. It is formulated according to Fama and French’s special preparation H and is designed for passive investors.
- Most of the funds are accumulation flavour index funds with the exception of the db X-tracker ETF, and the Edinburgh investment trust tracker.
- All funds track the FTSE All Share index with the exception of the DFA fund and Royal London FTSE 350 tracker. There is a Royal London FTSE All Share fund, but Trustnet’s charting tool refused to recognise it. The 350 fund is very similar.
- Certain funds were excluded because they haven’t collected at least three years of data, namely: the BlackRock CIF UK Equity Tracker, new clean class funds, the SPDR FTSE UK All Share ETF, and SWIP Foundation Growth B.
- Chart options chosen: income reinvested and offer-to-bid. If you take advantage of Trustnet’s tool, beware the quirk: every time I added a fund, the numbers for the index went loopy. Just change your radio button selection to fix it e.g. % return to £ return and back again.
Analysis
The most important point this table makes is that time is a great leveller. The chart-topping DFA fund shoots the lights out over one year, but over five it’s settling back into the pack. Which is especially important to know as you have to pay expensive advisor fees to access DFA funds.
The same goes for the Edinburgh Tracker Trust that’s in a photo-finish with the index over five years but is deeply average over ten. If this fund was on my hotlist then my next step would be to find out whether its TER had been slashed, which may account for its recent rocketry, or whether there is some peculiarity in its index replication recipe.
On the face of it the stark underperformer amongst this lot is the db-X tracker ETF, though its OCF is reasonably competitive. (Update: This may be due to inaccurate or misread source data. We’re looking into this).
The Virgin tracker looks nowhere near as bad as its market-lagging OCF would suggest. Fidelity has a headline grabbing 0.3% charge but has somehow conspired to underperform Virgin’s 1% fund over the last three years.
Over ten years the best-performing fund we have data for is the M&G Index Tracker. Although, in the last three years it’s fallen behind the HSBC fund which aggressively cut its price in 2009 to combat Vanguard.
Intriguingly, Vanguard is flattered by a tracking error comparison because it charges a 0.5% initial fee to cover stamp duty. That means stamp duty doesn’t swell Vanguard’s tracking error whereas the other funds excrete the share-dealing tax directly via this route.
You can account for all these nuances by squaring up your favourite funds against each other using a fund cost comparison calculator. Use your tracking error figures as the annual charge to find out the real cost of the fund.
Perfect is the enemy of good
Over 10 years I wouldn’t worry about the differences between the funds in the table we have data for. The gap between top and bottom is a piffling 0.3%.
But I’d want to avoid the gulf that’s opened up between a very poorly performing fund and the top funds over five years. I’d worry that a fund that can miss its benchmark by over 4% isn’t doing its job properly, and would dig deeper into the data to try to find an explanation.
Avoiding a straggly bottom shouldn’t find its mirror in an obsession to own the top. Don’t waste time hunting down the very best fund and then lash yourself with the cat if it doesn’t ‘win’ one year.
Look for a fund that habitually hugs its index tight, has a low OCF, and check that its tracking error is reasonably low and consistent with its peer group over as many years as possible. That should do you nicely.
Take it steady,
The Accumulator
Thanks for this post, I think my comment a couple of weeks ago kicked off a lot of discussion on this topic and this post gives us an easy way to quantify what is going on.
I’ve now got to work out whether to bother ditching my Fidelity Moneybuilder for a more competitive tracker.
Thanks very much Accumulator. I think this project of keeping an eye on tracking difference / error will be a huge benefit to us blog followers.
I also agree that it is not wise to get involved in obsessive changing of funds for very marginal differences.
I will use the example of Fidelity UKMBI (annualised tracking difference over 3 years 1.2%) and Vanguard UK equity (annualised tracking difference over 3 years 0.3% + 0.5% explicit stamp duty charge). I have assumed 6% market return and a single initial investment of £1000. I have also assumed that the tracking differences are maintained which given the record of Vanguard (in the US) and Fidelity seems reasonable.
Vanguard Fidelity Difference
3 years 1,175 1,151 24
5 years 1,313 1,264 49
10 years 1,732 1,598 134
25 years 3,978 3,229 750
My take would be that for smallish sums over 10 years it isn’t worth the effort but (given it is the same amount of work) for much larger, pension-sized sums over 20 + years it is well worth changing.
Great analysis.
I did some digging into the db-X tracker ETF, as I looked at this as a possible investment a while back, and was surprised to see it had done that badly.
The figures I came across on Trustnet looked a lot more reasonable, although I suspect the period I’m looking at is a little different from your review:
http://www.trustnet.com/Factsheets/Factsheet.aspx?fundCode=B4FR8&univ=E
So this has the FTSE A/S TR as 34.5% and XASX as 31.8% over the past five years. That equates as 0.42% per year, which is the same as the charges.
The latest factsheet from db-X also suggest the tracking error is pretty small.
http://www.etf.db.com/UK/pdf/EN/factsheet/factsheetLU0292097747_2012_12_31.pdf
Any thoughts on how to reconcile these differences?
1. Thanks, as ever, for the post.
2. Have you asked Deutsche Bank for a response? It’d be useful to understand why the db X-trackers ETF seems to do so badly, if it does.
Nice idea.
I cannot duplicate your db X-tracker resuts, but think it may be the bid-offer spread.
@ Passive – I agree with your analysis. How did you account for Vanguard’s stamp duty charge as a proportion of the tracking error?
@ Stipe and Dave – using an income reinvested measure makes things slightly worse for the ETF using the Trustnet chart, but then its using the offer to bid measure instead of bid to bid which really opens up the gulf.
Excellent article well done.
My only observation is that the DFA and RL funds at £100k and £500k minimums are not really retails funds.
The best bit of the analysis is puting the spotlight on the ETF. These arcane creatures are extremely impervious to analysis.
@Accumulator I just adjusted the principal of £1000 by the 0.5% initial Vanguard charge and then compounded by 5.7% (6%interest rate – 0.3% tracking difference).
For Fidelity I compounded by £4.8% (6% interest rate – 1.2% td)
So VG £1000 –(yr 0)£995 —-(yr 1) £1052 ——–(yr2)£1,112 etc etc
Fidelity (yr 0) £1000 ——-(yr 1) £1048 ——–(yr 2) £1098 etc etc
As I said above [#4], we need a response from Deutsche Bank…
@ Passive – that makes sense, cheers.
@ Rob – you can get DFA funds through an advisor. Surely you can join in for less than 100K? As I understand it, you can get Royal London through TD Direct, but I’ve yet to get to the bottom of this. I was hoping someone else would know something about this one. If you’re right then I unreservedly strike them from the table.
There are some intermediaries who will allow you to invest in Dimensional funds at less than the £100,000 “minimum” investment.
I posted on another thread a couple of days ago regarding a similar analysis of tracking error: http://www.trustnet.com/News/381589/cheap-tracker-funds-thrash-expensive-rivals/1/
This shows a similar dispersion of results, although it looks at the last 10 years and therefore does not include Vanguard and the HSBC figures will reflect the much higher AMC for most of that ten-year period.
I understand how Dimensional could beat the index, because their whole raison d’être is that they try to track more intelligently, rather than mindless automatic tracking. It makes sense to me that this approach will sometimes beat the market, although when I looked at their performance a few years ago I concluded that in practice their results were no better than Vanguard, looked to have on average over a range of different sectors.
I am more puzzled about how the Edinburgh investment trust managed to beat the market. Perhaps they do not have 100% replication and they just got lucky. But I was wondering if the explanation lies in movements between discount and premium on the shares. If the figures are calculated on the basis of offer and bid price, then a swing from discount to premium in a particular period would mean that the investment trust shares performed better than the underlying assets. This could probably be regarded as a distortion, which does not reflect any underlying tracking error.
Doesn’t the Dimensional fund have a dilution levy, similar to Vanguard?
Am I right in understanding that the calculations in the table do not make any adjustment for the Vanguard stamp duty levy, so the real performance of an investment in the Vanguard fund would have been 0.5% lower than the figures shown in the table? (And if I am right about the dilution levy in Dimensional, the same would apply to this?)
According to this posting on TMF, Trustnet data or calculation is wrong for the ETF http://boards.fool.co.uk/their-numbers-are-probably-wrong-using-comparison-12736879.aspx
@ Ivan – re: DFA fund, no initial fee according to Morningstar.
Vanguard’s stamp duty levy means that stamp duty doesn’t show up in the tracking error unlike the other funds who don’t charge that transactional cost upfront. However, you don’t subtract 0.5% from all the figures to get a neat comparison. Imagine you make one lump sum investment in year 1 of £1000. That would lose 0.5% in stamp duty tax when you invest. But you wouldn’t lose anything to stamp duty in years 2 to n if you don’t trade.
@ NaeClue – the plot thickens.
The 2011 annual report for Dimensional says they do charge a dilution levy of 0.5%. See page 13 here:
http://www.dfaeurope.com/pdf/annual_reports/annual_report_icvc.pdf
I realise that the stamp duty levy is a one-off when you purchase. So, if you invest 1000, only 995 is invested in the fund. If Vanguard produced a gross gain of 10.4% over three years, are you saying that the 1000 had grown to 1104, or had it in fact grown to 995×1.104 = 1098.5?
Update re: the feedback above, we’re looking into the data source and our reading of it, as it seems there may be some discrepancies we need to investigate.
Obviously our only aim here is to inform not to confuse, so we’ll look into it and get back to you. I’ve modified the article above in light of this.
Some good news:
Looks like the European regulators are forcing transparency and making providers give tracking error predictions!
http://www.morningstar.co.uk/uk/news/105511/On-the-Right-Track-Measuring-Tracking-Efficiency-in-ETFs.aspx
The above includes a link the detailed study on tracking error and tracking difference in ETFs:
http://www.morningstar.co.uk/uk/news/105511/On-the-Right-Track-Measuring-Tracking-Efficiency-in-ETFs.aspx
I see that the db x-trackers ETF comes out very well!
@Greg
That Morningstar report is a great find.
Reading that and looking at tracking differences makes it much easier to compare ETFs with funds.
Comparing tracking difference between ETFs and funds (i.e. Fund/ETF return vs index return) is much better than just comparing TERs. And from the report ETFs seem to come out of this more precise comparison quite well for developed markets.
While tracking differences for emerging markets ETFS are high relative to TERs, I assume (but don’t know) the same is true of emerging markets funds.
In a TER comparison the cheapest funds usually (but not always see below) have lower TERs than the cheapest ETFs.
But in terms of the tracking difference I suspect it may be closer than the TERs indicate.
From a theoretical perspective when comparing (say) HSBC ETFs to HSBC funds, tracking differences bring in the negative effect of SDRT dilution on funds so there is theoretically a systematic error in just comparing TERs.
One particularly interesting case is the HSBC S&P 500 ETF which as well as a low TER of 0.09%, tracks the index closely (in fact there is outperformance that may in part be due to the index tax assumptions that the report mentions for the Euro Stoxx 50 ETFs?).
The TER of the HSBC S&P500 EFT tracker of 0.09% (same as for the Vanguard S&P500 ETF) is actually lower than the 0.2% TER for the HSBC S&P500 tracker fund. So you can put up a strong argument for using an ETF rather than a fund here (although you have to factor in other costs such as spreads, platform and broker fees). The Vanguard US equity tracker fund at 0.2% TER tracks the broader S&P Total Market Index index so might be preferred by some.
Using the trustnet chart today, and using the same chart settings as suggested (offer-to-bid and income reinvested), for 3yr annualised performance I get:
FTSE All Share 10.2%
HSBC FTSE All Share Index Ret Acc 10.0%
Fidelity Moneybuilder UK Index 9.8%
Virgin UK Index Tracking 9.4%
Vanguard FTSE U.K. All Share Index 8.7%!!!
If I set bid-to-bid in the chart options then I get a better showing from Vanguard (10.1% with all others remaining the same) but the article explicitly states that offer-to-bid was used.
The numbers of course could change in this time but 8.7% from Vanguard – really? I presume this is not correct and I’ve done something wrong, anyone care to point out what I’m missing here?
@Steve.
When I looked at this I though the 0.5% Vanguard stamp duty levy would be the explanation but isn’t, as the figures don’t work out.
I wonder if there is an error in the calculations as a 5% up-front charge does give figures closer to the 8.7% return.
Whatever the explanation it must surely be an error at Trustnet?
Also I don’t quite understand why some of the other funds have done so well. I think the Fidelity fund has a TER of 0.3% and a tracking error of 0.7% and last time I looked the Virgin fund had a TER of 1%.
Fund valuation timing can make quite a big difference (Fidelity is valued at midday, Vanguard at market close, Virgin I don’t know) but I don’t know really whether this is significant.
Interested if anyone else has comments
Figured out the difference between what I had and what the article had at least, I was using ‘Vanguard FTSE U.K. All Share Index’ whereas the article was using ‘Vanguard FTSE UK Equity Index Acc’ – as expected, my mistake.
Today’s figures therefore give:
FTSE All Share 10.5% (jump from yesterday, valuation timing?)
Vanguard FTSE UK Equity Index Acc 10.2%
HSBC FTSE All Share Index Ret Acc 10.0%
Fidelity Moneybuilder UK Index 9.8%
Virgin UK Index Tracking 9.4%
Vanguard FTSE U.K. All Share Index 9.0 %
This is also an interesting read from Fidelity and gives insight into why valuations timings can cause variations when looking into the figures:
https://www.fidelity.co.uk/institutional/news-insights/in-perspectives/tom-stevenson/details.page?whereParameter=tom_stevenson/trackers2180112
The HSBC tracker enhanced disclosure information (as per the IMA guidance) can be found at
http://www.assetmanagement.hsbc.com/uk/individuals/default.html
Makes interesting reading (for the true geeks) in terms of understanding what causes tracking error for the HSBC trackers (including the all share tracker)
Obviously stamp duty in all its forms is a major contributor to tracking error (as well as the OCF of course) although the removal of schedule 19 SDRT in April may reduce some of this in the future.
Similarly Black Rock enhanced IMA charge disclosures are here
http://www.blackrock.co.uk/individual/literature/fund-update/blackrock-authorise-unit-trusts.pdf
Similarly Legal and General enhanced IMA charge disclosures are here
http://www.legalandgeneral.com/investments/fund-information/charges-and-fees-explained/transaction-costs-and-taxes/
So of the big 4 in terms of the cheapest mainstream clean trackers (HSBC C, Vanguard, L&G I, Black Rock D) that just leaves Vanguard who are committed to enhanced disclosure but haven’t given a date when they will start see here.
http://www.investmentuk.org/current-topics-of-interest/charges/compliance/