Tracking error—the difference between the performance of a fund and that of its benchmark—is the best way to measure an index fund’s true cost. While many investors focus on MERs, a low fee means little if an ETF lags its index by an additional 30 or 40 basis points.
A couple of weeks ago, I reported on the tracking errors of iShares ETFs in 2011, which were mostly very low. Today let’s look at the 2011 performance of the most popular ETFs that formerly bore the Claymore name, all of which were recently rebranded as iShares.
Canadaian equity | Ticker | Fund | Index | Diff |
Canadian Fundamental | CRQ | -8.3% | -7.9% | -0.4% |
S&P/TSX Canadian Dividend | CDZ | 6.3% | 7.4% | -1.1% |
US and international equity | Ticker | Fund | Index | Diff |
US Fundamental (hedged) | CLU | -1.5% | -0.4% | -1.1% |
US Fundamental | CLU.C | 1.8% | 2.3% | -0.5% |
International Fundamental | CIE | -13.1% | -12.2% | -0.9% |
BRIC | CBQ | -22.5% | -22.0% | -0.5% |
Global Real Estate | CGR | -3.1% | -2.0% | -1.1% |
Global Monthly Advantaged Dividend | CYH | -5.5% | -5.7% | 0.2% |
Fixed income | Ticker | Fund | Index | Diff |
1–5 Yr Laddered Corp Bond | CBO | 4.7% | 5.0% | -0.3% |
1–5 Yr Laddered Gov’t Bond | CLF | 5.4% | 5.5% | -0.1% |
Advantaged Canadian Bond | CAB | 6.8% | 8.9% | -2.1% |
Advantaged High Yield Bond | CHB | 4.6% | 6.4% | -1.8% |
Let’s start with the positive results: the three core equity funds tracking the RAFI fundamental indexes had very low tracking errors. The Canadian Fundamental (CRQ) and the non-hedged version of the US Fundamental (CLU.C) both lagged by less than their MERs, which is always impressive. The International Fundamental (CIE) didn’t fare quite as well, but the higher costs of trading international equities makes this unsurprising.
Once again, currency hedging proved to be a major drag on performance: the hedged version of the US Fundamental (CLU) had a tracking error 60 basis points higher than that of CLU.C. Remember that this added friction will be there whether the Canadian dollar rises or falls during the year.
The performance of both CLU.C and CIE are very encouraging, given that both funds have experienced terrible tracking errors in the past. I currently recommend the US-listed PowerShares fundamental ETFs in my Über-Tuber portfolio because of this poor track record. But now that these funds hold all of the stocks in their indexes (as opposed to a representative sample, as in years past) they seem to be doing an excellent job of mirroring their benchmarks. Next time I update the model portfolios I will have to consider including CLU.C and CIE.
Not such an advantage
The Claymore/iShares Advantaged ETFs, which hold either bonds or foreign dividend paying stocks, are designed to be as tax-efficient as possible. I include three of them in my Yield-Hungry Couch Potato, which is designed for non-registered accounts.
The performance of these funds was mixed. The Global Monthly Advantaged Dividend (CYH) actually had a higher return than its benchmark, which is a pleasant surprise. But the Advantaged Canadian Bond (CAB) and the Advantaged U.S. High Yield Bond (CHB) lagged their indexes by 2.1% and 1.8%, respectively.
The Advantaged funds have higher expenses and should be expected to have bigger tracking errors than their plain-vanilla counterparts—even then the tax advantages might still put them ahead. But for investors who are not in the highest tax bracket, the “advantage” is likely to be slim to zero when tracking errors are this high. The comparable iShares DEX Universe Bond (XBB) returned 9.36% in 2011, while the iShares U.S. High Yield Bond (XHY) returned 6.83%. It’s quite possible the after-tax returns on these traditional ETFs may have been higher for investors who were not in the highest tax brackets.
I consider these Advantaged bond funds to be on probation: if they continue to lag like this, I will remove them from my model portfolio.
I find it surprising that CYH is the only example that outperformed the index it was tracking. I understand that the MER shifts the expectations to underperform the index by approximately that amount – but I assumed that because the MER tends to be small and these funds don’t typically hold the entire index that statistically these funds would leave out as many ‘losers’ as ‘winners’ and you would get more funds slightly outperforming the index they are tracking. 1 out of 12 just feels a little low, are there other systematic biases in how these funds are setup to track the indexes that are dragging the performance down?
@Brian: It’s not accurate to say that ETFs “don’t typically hold the entire index.” They usually do. For that reason, it’s very rare for any ETF too outperform its index, especially if the MER is 50 or 50 basis points. In some ways, it’s actually a red flag, because you have to wonder how and why it happened. :)
I notice that the report for CDZ shows a -1.1% difference.
If you have done previous reporting on this fund, is this about the same difference? I am tempted to find a similar fund, without such a seemingly large variance.
Hi Dan,
Any way you could do Vanguard in an upcoming post? I know their Canadian ETFs are relatively new though…
I found the point about CAB versus XBB especially interesting, since I’m currently holding CAB in my non-registered portfolio and XBB in my registered accounts. I ran the numbers using a hypothetical $1,000 and the Ontario tax rates for the second-highest bracket ( from http://www.taxtips.ca/taxrates/on.htm) and found the after-tax return to be within about 25¢ or 0.02%.
That’s certainly something I’ll keep my eye on, since there’s no point in over complicating things for the same return in the end.
@Pacific: CDZ had a tracking error of –2.2% in 2010, which is pretty terrible. No one notices these things when a fund returns 15.2% like CDZ did that year, but this catches up with a fund over the long term.
https://canadiancouchpotato.com/2011/04/26/tracking-errors-on-canadian-etfs/
@Adam: I’ve discussed Vanguard’s ETFs in several previous posts. Here’s an overview of my thoughts:
https://canadiancouchpotato.com/2011/08/25/meet-the-new-funds-same-as-the-old-funds/
I was a bit uncharitable in that post: that was before Vanguard announced their pricing, which turned out to be exceptionally good:
https://canadiancouchpotato.com/2011/11/10/vanguard-drops-the-gloves/
@Michael: Thanks for the real-world analysis, which is very telling. I agree with you: I would not use a complex ETF like this unless there was a significant tax advantage, and there is not likely to be any advantage as long as that tracking errors are that big.
@CCP:
VERY much appreciated!
Can you try to get iShares to justify that tracking error for its Advantaged ETFs?
Why is the index return for US Fundamental (hedged) different than the index return for US Fundamental? Is the index different when there is hedging? I’m trying to understand this better, thanks.
Potato,
The index for the hedged ETF is different than the index for the non-hedged ETF. The difference is due to the currency hedge.
It’s also worth noting that the index for the hedged ETF builds in some drag for rebalancing the currency hedge (which is not perfect), which will understate the underperformance of the ETF vs a perfectly currency hedged position in US equities.
@Andrew F: Sure, I can ask them for an explanation.
@Potato Salad: Yes, the two funds have different benchmarks. The hedged version will track the returns in US dollars (the local currency of the stocks), while the unhedged version tracks the returns in Canadian dollars.
This is why I love this blog! Thank you so much for this post! Always eye opening topics.
A couple of things I can wade in on. Further to what Dan said, most ETFs fully replicate, so tracking error should be pretty close to the ETF’s MER plus any portfolio trading expenses. Typically you get more tracking error where the index is replicated through a “sampled” strategy – which occurs most often with bond index ETFs. Owning 1000 plus bonds in even the largest etfs is very difficult. So what ETF companies do is choose representative bonds to own. For example, if an index includes 30 Govnt of Canada long maturity bonds, the ETF company only needs 1 or maybe 2 in order to get the beta associated with long Canada’s. same for Provi’s etc… Less sampling on the equity side because of company specific beta.
On the advantaged structure, the higher tracking error usually comes from a combination of the higher cost structure (the tax structuring costs between 0.45%-0.75% per year) and the tracking error from sampling the fixed income portfolios. Sooooo, the structure is most effective when wrapped around high coupon indexes like high yield or emerging market debt, because the tax efficiency better offsets the extra cost drag. The lower the natural level of income, the more the tax benefit gets eroded.
Thanks, Barry. I understand that there is added cost due to the swap counterparty taking a fee for these ‘Advantaged’ funds. What I would like is to hear iShares’ explanation. I have been uncomfortable with the lack of clear disclosure of this significant and material additional cost incurred due to the structure of the fund. It seems to me that the structure only works if all three of the following are satisfied:
-the investor is holding in a non-registered account
-the investor faces the highest marginal tax rate
-the current income of the underlying securities is over 6 or 7%/year
So, that rules out CAB for just about everybody, and CHB for most people. It may just break even versus a simpler vanilla high yield fund.
I like the concept in theory. In practice, the costs need to come down for the structure to be very compelling. The counterparty is the obvious area to squeeze to reduce fees, but the impetus to do this will only come when that cost is clearly and proactively disclosed to the investor.
Andrew F you have it exactly right. The cost is material and because it is considered a trading expense it ends up in trading expense ratio instead of MER. It is unfair to compare apple to apples unless the total cost incurred in delivering the tax benefit is properly disclosed. I leave it to you and others to determine the quantum of the benefit. As you say, unless the person is substantially taxable and the cash yield is above 6 percent, the structure loses a lot of its tax benefit. You should factor in time value of the extra money you (should) have in your pocket too. If you compound it, it can add up. The vast majority of investable cash in Canada is outside registered accounts, so there are plenty of people for who the structure should, I say should, provide a benefit, if the structure is employed properly.
As for iShares explaining it, they just inherited the ETFs from Claymore, so I am guessing they are just wrapping their heds around it themselves.
@Barry and Andrew: Many thanks for adding your insights here. I am definitely going to give some thought to designing a more tax-efficient portfolio for non-registered accounts. I ma not convinced that the Advantaged bond funds are really that advantaged. It does seem that CYH is doing a better job so far, but I will keep an eye on it, too.