When the BMO Discount Bond Index ETF (ZDB) was launched back in February 2014, it was unique: the first broad-market ETF in Canada made up primarily of bonds trading below their par value. By avoiding premium bonds, ZDB promised to deliver similar returns to traditional bond funds, but with greater tax efficiency, making it ideal for non-registered accounts. With a little more than two years of real-word performance, it’s time for a checkup. Has ZDB delivered on its promises?
Top of the heap
The first question we’ll examine is whether ZDB achieved pre-tax returns similar to other broad-market bond ETFs. The fund was designed to match the popular FTSE TMX Canada Universe Bond Index in credit quality, average term, duration and yield to maturity. But ZDB set out to achieve this profile using bonds with lower coupons to reduce the amount of taxable income.
As it turns out, ZDB outperformed all of its competitors in 2015. Here are the NAV returns for the calendar year:
|Bond ETF||Ticker||2015 Return|
|BMO Discount Bond||ZDB||3.60%|
|Vanguard Canadian Aggregate Bond||VAB||3.48%|
|iShares Core High Quality Canadian Bond||XQB||3.38%|
|BMO Aggregate Bond||ZAG||3.24%|
|iShares Canadian Universe Bond||XBB||3.15%|
Sources: BMO ETFs, BlackRock, Vanguard Canada, Morningstar
If you own ZDB you would have been pleased with that 3.60% return, but you should be curious about why the fund managed to finish ahead of its counterparts. And you’ll really be scratching your head when you learn that the FTSE TMX Canada Universe Discount Bond Index returned just 2.18% on the year. An index fund typically lags its benchmark by an amount roughly equal to its fee, so it’s unusual to see any outperformance at all, let alone 142 basis points. What’s going on?
The other side of the track
I contacted BMO for an answer, and they explained that falling rates in early 2015 made discount bonds extremely hard to find. As a result, the index (which was custom-designed for ZDB) became highly restrictive. Had the fund managers followed its rules slavishly, the fund would have been too concentrated in a small number of issuers. Moreover, its yield to maturity would have fallen well below that of the broad market. So the managers decided to include bonds outside the index in an effort to more closely match the characteristics of the much-followed FTSE TMX Canada Universe Bond Index while still keeping the bonds’ coupons as low as possible.
This kind of sampling can be expected to cause some tracking error, and in 2015 that turned out to be positive: even after fees, ZDB returned a little more than the Universe index, which delivered 3.52% on the year.
Were the managers right to ignore their own benchmark like this? Certainly it’s an unorthodox move for an index fund which, by definition, is supposed to follow a set of predetermined rules. And while it worked out last year, future tracking errors could easily be negative. But I think BMO made the right call here.
After all, investors who use ZDB expect it to deliver pre-tax returns in line with traditional broad-market bond ETFs, and last year that was about 3.5% before fees. Had the managers tracked their custom benchmark precisely and delivered about 2%, I’m not sure investors would have been pleased with that result. So BMO’s decision followed the spirit of the law, if not the letter.
I wouldn’t be surprised if BMO eventually changed the mandate of ZDB, dropping the custom index altogether and explicitly aiming to track the FTSE TMX Canada Universe Bond Index while keeping the fund’s average coupon as low as possible. That’s effectively what they’re doing now anyway.
How much did you keep?
OK, now let’s see how ZDB stacked up in terms of tax-efficiency. To answer this question, we turn to Justin Bender’s analysis, posted last week on his Canadian Portfolio Manager blog.
Justin has developed a methodology for estimating the after-tax returns on Canadian ETFs, which we introduced in a white paper back in 2014. In simple terms, we assume that all distributions from the ETF are taxed at the highest rate, and that that the remainder is reinvested in the fund. We then compare this after-tax return to the published pre-tax return and calculate the ETF’s tax cost ratio, which estimates the amount lost to taxes. The lower the number, the more tax-efficient the fund.
ZDB excelled here as well. If you held the ETF throughout 2015 and were taxed at the highest rate, you would still have kept more than two-thirds of the fund’s 3.60% pre-tax return. Meanwhile, the least tax-efficient bond funds in this group (ZAG and XBB) saw their returns reduced by more than half:
|Bond ETF||Ticker||After-Tax Return||Tax Cost Ratio|
|BMO Discount Bond||ZDB||2.52%||1.04%|
|Vanguard Canadian Aggregate Bond||VAB||1.87%||1.46%|
|iShares Core High Quality Canadian Bond||XQB||1.91%||1.51%|
|BMO Aggregate Bond||ZAG||1.61%||1.58%|
|iShares Canadian Universe Bond||XBB||1.39%||1.70%|
Source: Justin Bender, Canadian Portfolio Manager
But is it well diversified?
So we’ve established that the BMO Discount Bond Index ETF performed very well on both a pre-tax and after-tax basis in 2015. But did those higher returns come with greater risk?
One of the criticisms we’ve heard levelled at ZDB is that it is poorly diversified. After all, it holds just 77 bonds: by comparison, ZAG and VAB hold between 650 and 700, while XBB holds close to 1,000. One might conclude that these counterparts carry far less risk. But is that true?
First, remember that about 70% of ZDB is made up of government bonds, and the breakdown among federal and provincial issues is extremely similar to that of ZAG. Both funds therefore have virtually identical risk exposure on the government side, even though the latter holds hundreds of additional bonds. (Holding 100 Government of Canada bonds, after all, is no less risky than holding only 10 or 20.)
Among the remaining 30% that is invested in corporate bonds, ZDB’s holdings are indeed more concentrated, with significant holdings in only a small number of companies, including TELUS, Toyota, and several banks. To quantify this risk, Justin used a metric called the diversity index. This formula (based on a concept borrowed from ecology) produces a value for any portfolio based not only on the number of bonds, but also on the dollar amounts in each issuer. A portfolio with a diversity index value of 100, for example, would have only one-fifth the concentration risk of one with a value of 20.
When Justin compared the corporate holdings of the two BMO funds, he found that ZDB had a diversity index value of about 16.7, compared with 45.4 for ZAG. That’s a significant difference. But overall—when you consider the government and corporate holdings together—the diversity index values were 8.5 for ZDB and 8.9 for ZAG, a much more modest gap.
Overall, then, it seems safe to conclude that BMO Discount Bond Index ETF has performed as advertised. For investors who want broad-market bond exposure in a non-registered account, ZDB is an excellent way to get it.
I bought $157,500 of ZDB for my cash account in June 2015 and have been happy with the $250 per month it pays each month along with it being worth $159,000 today.
I only wish the $59K I put into BMO LADDERED PREF SHARE (ZPR) was half as good – its down 28% ?
Nice follow up Dan. I have been looking at some options for fixed income in taxable accounts and have read all your previous posts on the topic. Does the tool that Justin built allow for an estimation of after tax returns of the Horizons HBB? This is the swap based ETF that is designed for tax efficiency as you described here:
According to their website it returned 3.19% in 2015 pre tax. This is on the low end of the peer group you listed in the post which should be somewhat expected since it has higher fees associated with the swaps. I suspect the after tax returns may be ahead of the pack but I have no way to confirm on my own.
I would love to see a comparison of ZDB and HBB. I know you weren’t keen on HBB because it doesn’t actually hold bonds itself, but with returns so low right now the tax savings seem to be worth having. Especially when those tax savings are compounded over many years. I was on the fence for these two and other short term bond indexes. I decided to split my bonds between HBB and VSB….
@Bruce and Shaun: The after-tax return calculation is “pre-liquidation,” which means it only takes into account distributions and does not consider any capital gains taxes payable when the units are eventually sold. Since HBB has no distributions, its after-tax return is the same as its pre-tax return. In that sense there is not much to say about HBB except to acknowledge that it has tracked its index almost precisely after accounting for its fee, which is what you would expect from a swap-based ETF. As long as you are comfortable with the risks (described in the post that Shaun has linked to), HBB is also an excellent choice for non-registered accounts.
Bruce, please note that VSB is short-term bonds only, so the risk exposure is quite different compared with ZDB or HBB, which are broad-market bond funds. For tax-efficient short-term bond exposure, BXF is worth a look:
@Shaun and CCP
Thanks for the article Dan
Shaun: Justin did a previous analysis comparing HBB (with a YTM of 1.7% at the time, 1.9% now) with a 5 year GIC, which returns a higher interest rate, but has distributions that are taxed at the full marginal rate of income.
With his assumptions, it looks like the options came out about equal after a ten year holding period.
“Since HBB has no distributions, its after-tax return is the same as its pre-tax return.”
Technically you will eventually need to pay taxes on the capital gains though (as with ZDB to a lesser extent). That tax is both deferred and at half the rate of interest payments though.
“Followed the rules slavishly?” The manager needs to explain why they deviated from their investment policy. You won – great, nice return. BUT what if it was a loss? Funds have an investment policy for a reason. Not following it opens them up to all kinds of issues, and undermines my confidence buying BMO funds.
@Jeff: Certainly one needs to question why a manager would deviate form the index, but I think we have looked at that question. The idea was not to outperform the broad bond market: that was a lucky accident. The reason for the decisions was twofold. First, there was a recognition that the discount bond index was no longer tracking the broad market the way it was designed to because few discount bonds were available in the marketplace. Second, there were too few issues in the discount bond index provide adequate diversification. I think those are two good reasons to bend the rules. However, going forward, I think BMO will need to make a longer-term decision about how to handle this and then disclose their revised strategy so there are no surprises.
It would be good if you could include another table in your analysis that shows what the share price of the ETF’s have done over this period. It is common for bond ETF’s to pay out distributions but lose share price.
@rgz: Both the pre-tax and after-tax returns in this post are total returns, which include both the change in price and the distributions. That is the only meaningful way to measure an investment’s return.
How do CBO and CLF stack up to what you have in the tables?
@rgz: Because of their high coupons, CBO and CLF are among the least tax-efficient bond ETFs:
I used to hold ZDB but i switched to HBB as soon as it became available. Regarding ZDB vs HBB, M. Bender made the following comment in one of his recent post:
“The closest you could get is to calculate a post-liquidation return for all ETFs (i.e. assume that each one is sold at the end of the 1-year holding period). I would estimate that HBB would come out slightly ahead of ZDB post-liquidation (by about 0.22%).”
I know that PWL use ZDB with their clients but HBB is a true buy and hold and would beat ZDB by a large factor over a 5y-10y holding period.
By holding my bonds in my taxable account, i can use my RRSP/TFSA to hold foreign equities and REIT, thus achieving the best tax optimization. IMHO, HBB is probably one of the most useful financial instrument for investor with largish non-registered account !
Thank you for the interesting article.
Both ZDB and HBB hold around 70% government and 30% corporate bonds.
Do you know of any similarly tax-efficient bond ETF products which hold either all or a majority in corporate bonds?
I like to keep my bond holdings balanced between corporate and government bonds, and this would therefore be possible.
Similar to Lawrence W I held a large chunk of my large taxable account in BXF (First Asset’s 1-5 year Laddered Government Strip bond Index Fund), which in retrospect has also done really well, particularly on an after tax basis, but switched that entire holding to HBB as soon as it became available. I’m glad that BXF did as well as it did, and that it fulfilled its after tax promise, for the short duration that I held it, but really I share Laurence W’s opinion regarding the long term usefulness of HBB.
Thx for the reply
This is a very interesting concept that you have pointed out. Intentionally putting your fixed income in the form of HBB in a your taxable account to make room for REITs and foreign equities that are subject to full taxable rate on their dividends. It sure sound enticing.
In your past post that I linked above you discussed the risks associated with the Horizon swap structured funds. It seems that your impression is that the risks are small both in likelihood and potential impact. I agree. But I still get a sense that you have a lukewarm feeling towards these swap funds. It appears to me that the benefits of implementing a strategy like LawrenceW suggested would be very beneficial to investors who are out of room on their registered accounts.
Do you have any further thoughts or comments on the risks of the swap structured funds or LawrenW’s strategy?
@Shaun: Thanks for the comment. I think it’s accurate to say I am lukewarm on HBB. My concern is that I simply don’t understand what’s going on behind the scenes. If the counterparty holds the underlying bonds in the index in order to hedge its exposure (which is what you would expect with a swap), then it must be paying tax on the income. And if that’s the case, it’s not clear how they can profit from simply charging the small swap fee. And if they are not holding the bonds in the underlying index, then there is a risk investors do not know about. Unless you believe in free lunches, you need to be able to explain how you can convert bond interest into capital gains. If this was clearly disclosed I would me more comfortable with the product.
Second, as advisors we need to be aware that all of these risks are magnified. It’s one thing to have a small personal holding in a product like this: that’s your own decision and it affects only you. It’s another to invest tens of millions of clients’ money in it without a complete understanding of its inner workings.
Finally, let’s remember that with yields so low right now, the tax benefit of this structure is quite small. It would be a different story if bonds were yielding 5% to 6%. Then the additional risks might be worth it, but for now we are using GICs and low-coupon bond funds for taxable accounts.
Dan Hallet of high view wrote a very good article about swap based ETF structure, risk and counterparty compensation
@Lawrence: That article applies only to HXT. Any “dividend tax arbitrage” banks might be able to exploit does not apply to bond interest. That is specifically mentioned in the article:
you are totally right CCP, my mistake. However i still believe that HBB is safe and i am not concerned about any of Horizon tax saving sorcery ;)
I guess HBB might be interesting for those wanting to wait out a lowish average yield to maturity of only 1.9% over a long average duration of 7.9 years.
I’m still in the stages of learning the ropes about investing and finding the articles and comments to be very insightful-thank you. I currently have a portfolio thats a little aggressive with a mixture of equity funds and balanced funds with overall 68/32 equity bond ratio. I understand that it is important to have a portion of your portfolio as cash especially as I am about 5 +/- years away from retirement in anticipation of turbulent years in the stock market. My question is whether ZDB or BXF would be a better alternative than keeping the cash in GIC’s.
Thanks again for all the great information
@Jan: It’s important to understand the differences between these options so you can make the right decision. In general terms, ZDB is a broad-market bond fund (average maturity is 10 years), so it is fairly volatile and not appropriate as a short-term holding. BXF holds short-term bonds only (less than 5 years) and is more appropriate for shorter horizons, though it too will fluctuate and is not a replacement for cash. GICs do not fluctuate in value like bond funds, but they are not liquid: they cannot be sold before their maturity dates. So bottom line, cash is cash, and there are no substitutes. For longer time horizons, one of these other options may be appropriate for your fixed income.
@CPP, love your Blog.
I have another noob question about ZAG, when I look at the recommend bonds on CCP I can see ZAG shows-(BMO Aggregate Bond ZAG MER 0.10) but when I click on the fund and go to the BMO site and download the Factsheet (PDF)
This is what I see.
Maximum Annual -Management Fee 0.09%.
Management Expense Ratio 0.23%.
This is a bit confusing, can you elaborate? I though I understood MER’s but the huge difference in basis points (Still learning the terms) I am confused about. Thank you.
@Lintan: Good question, as this is pretty confusing. The published MER is a backward-looking number that indicates the total cost of running the fund over the previous year. When a fund lowers its management fee (as ZAG did very recently) that won’t show up in the MER until next year. But of course, anyone holding this fund today can expect the MER to be much closer to 0.09% (probably 0.10% after taxes are included).
I have gone back and read your posts on bond ETFs, and am still having trouble comparing CLF (1 – 5 year laddered gov’t bonds) which I own in my RRSP, and VAB (Vanguard Canadian Aggregate Bond) which I am considering switching to. They seem to hold similar products, except that CLF holds short term bonds (1 – 5 years) and VAB holds longer term bonds. And VAB has a 5 star rating, whereas CLF has only a 3 star rating. (I am not sure what this is based on.)
I have experienced, as you describe, a drop in capital value of CLF, while the interest has been good. I am now wondering if I should get rid of my holdings and swallow the loss, since it doesn’t look like the value will rise in the near or mid future. But will the same thing happen with VAB, or has it already lost what value it will likely lose with interest rates so low?
I find your posts extremely helpful, and hope that you will be able to shed some light for me on this conundrum.
@Ellen: Sorry for the delay in responding to your comment. CLF and VAB are actually quite different: the former is short-term government bonds only, while the latter holds both government and corporate bonds from all maturities. They have quite different risk-reward profiles: CLF is quite a bit more conservative.
You can safely ignore Morningstar ratings: they have no meaning for index funds. The most useful measure of an index fund is whether it performs as it is expected to, i.e. how well it tracks its benchmark index.
Pretty much all bond ETFs will see their prices fall gradually in the near future unless interest rates decline significantly. However, you have not suffered a loss in CLF, because the interest payments have more than offset the price declines:
Your decision should be, “Do I want the safety and low expected return of short government bonds, or am I willing to accept more risk with the expectation of higher returns from a broad-based bond fund?”
What about switching VAB for ZDB as a ‘perfect pairing’ for tax loss selling this year?
@FritesPaul: Sure, especially because it makes sense to stop using VAB in a taxable account anyway.
Why exactly is zdb more tax efficient than zag?
Also, if tax efficiency is the main push, what are the best recommendations for a balanced portfolio in a non registered account. And a registered account.
The ETFs I was looking and wanting to use are the ones recommended in another one of the money sense/couch potatoe articles.
XAW, XIC, ZAG (ZDB for non registered apparently)
Are xaw and xic both okay for registered and non registered?
@Steve: The equity ETFs are fine in non-registered accounts. The following should help explain why traditional bond ETFs are a poor choice in taxable accounts:
I pay capital gains on this fund every year even though I haven’t sold any units of the fund. For 2020 however, I’ve noticed that the capital gains for this fund are over $1200, even though the fund only paid out distributions of about $600 and the fund price per unit actually decreased. This is is contrast to 2019 taxable gains which were in the $300 range with the same amount of units. I’m trying to make sense of this. Can I do an ACB on a fund that I haven’t sold?
@Paul: ZDB saw huge capital gains (for a bond fund, anyway) in early 2020 when interest rates plummeted. That’s why it returned 9.3% last year. (In 2019, it retuned 6.97%, also surprisingly high.) In both years the fund managers had to rebalance the portfolio by selling premium bonds that were no longer part of its benchmark index, and these capital gains were passed along to unitholders at year-end. That’s why you saw such big capital gains on your T3 slip.
You do indeed need to adjust your cost base on this holding, or you will pay tax on these capital gains twice: once when reporting the amount on your T3 and again when you sell the holding. A distributed capital gain like this results in an increase to your ACB. In 2020, the capital gain was $0.63 per share (which is quite huge), so you should increase the ACB by that amount.
These articles explains the key ideas:
If you manage a non-registered portfolio, the website AdjustedCostBAse.ca has an excellent service that automatically calculates ACB to account for year-end distributions like this. You can use it to track your ACB over the years and then it will generate an accurate capital gain/loss report when you ultimately sell:
Thanks very much for the information and support. Although I have been a Couch potato investor for over seven years, I totally missed learning about this. I’ve been buying etfs and index funds but haven’t sold anything to date and haven’t been studying my brokerage statements. I’ve spent the week feeling overwhelmed trying to understand these documents/resourcesand beginning to enter all of this information, (I also strayed from the simplicity of the Couch Portfolio portfolios) into the Adjusted Cost base website. I’m still not sure I’m doing it correctly.
My annual Summary of Trust Income reports Capital gains but I’m not sure if these are capital gains dividends or reinvested capital gains distributions, according to the categories available on the Adjusted Cost base website. Also, I’m embarrassed to say that I’m not absolutely sure that I don’t have a DRIP with any of these funds -at least there’s nothing on my statements to indicate that so I’m gathering that I don’t have to input any reinvested distributions in calculating my ACB. I’ve been navigating the CDS Innovations website okay. Thanks!
@Canadian Couch Potato: With reference to your reply to Paul, and the recent (2019/2020) large capital gains that ZDB incurred. It seems to me that these capital gains are likely to have completely negated the tax efficiency of the fund over those two years. Is that a reasonable assessment or am I being too simplistic?
Going forward, I’m aware that many people are suggesting that there will be pressure on inflation over the coming years as global economies recover from the pandemic and that interest rates may rise as a consequence. Hence suggestions that I have read are often promoting short term bonds over longer term bonds. However, in a non-registered account is it still likely to be more beneficial buying ZDB (with it’s average term to maturity of 9.9 years) as opposed to a short term bond fund such as XSB (2.9 years) or vice versa?
@Paul: You need to include in your ACB calculations every single purchase of a security/fund. Therefore, if you have a DRIP that automatically purchases additional units of an ETF for example, then each and every one of those DRIP purchases needs to be included in your ACB calculation. The total number of units that you own shown in your ACB calculation should exactly match that shown by your brokerage.
Be aware, that as as far as I know there are only two items from the CDS Innovations sourced spreadsheets that affect your ACB and these are:
T3 Box 42 – Return of Capital
The latter are NOT captured on T3 slips and are often referred to as Phantom Distributions. It’s important to capture any of these as they increase your ACB ultimately reducing any capital gain you may have to pay upon selling that fund.
The adjustedcostbase.ca website has a very handy feature that can be used to import BOTH of the above items for each fund making the chances of missing them in your ACB calculations negligible.
@Andrew: It’s a good question. But remember that traditional bond funds also enjoyed large capital gains in 2019 and 2020: for example, ZAG returned 6.78% and 8.56% in those two years. The only difference is that ZDB distributed those gains to unitholders, while ZAG did not. If you held ZAG in a non-registered account, you would likely have a significant unrealized gain that will be realized at some point in the future (for example, if you sell bonds to rebalance).
We also need to remember that the reason ZDB delivers more of its return as capital gains is because it pays less interest, which of course is the whole point. A taxable investor should prefer a 1% capital gain to 1% in interest, because the former is taxed at half the rate.
As for the short-term v. long-term question, there are two issues here. First, the idea that you should switch to short-term bonds when you think rates will rise is a classic example of an idea that sounds great in theory but often backfires. It would have worked well in the last 12 months, but not during the previous period of rate increases:
Second, if you do decide to hold short-term bonds, an ETF such as XSB is an extremely tax-inefficient choice. Currently the yield to maturity on XSB is 0.74% and the average coupon is 2.2%. Whenever the coupon is much higher than the yield to maturity, the fund will be very tax-inefficient: this is the exact issue that ZDB was designed to address.
This blog is old, but it explains the concept in detail:
Thanks Andrew for the information about ACB and thanks to you both for furthering my knowledge about bond ETFs in taxable accounts.