Bonds are one of the least tax-friendly asset classes: most of their return comes from interest payments, which are taxed at the highest rate. They’re even less tax-efficient when their market price is higher than their par value: these premium bonds are taxed so unfavorably they can actually deliver a negative after-tax return. Unfortunately, because interest rates have trended down for three decades, virtually every bond index fund and ETF is filled with premium bonds. Enter the BMO Discount Bond ETF (ZDB), which begins trading tomorrow. This unique new ETF promises to eliminate the problem that has long plagued bond funds in non-registered accounts.
Let’s take a step back and review the important idea underpinning this new ETF. Consider a premium bond with a coupon of 5% and a yield to maturity of 3%. The bond will pay you 5% interest annually and then suffer a capital loss of 2% at maturity, for a total pre-tax return of 3%. Now consider a discount bond that pays a coupon of 2% and has the same yield to maturity of 3%: now, in addition to the interest payments, you’d net a 1% capital gain at maturity, and your total pre-tax return would again be 3%. In an RRSP or TFSA, therefore, these two bonds would be virtually identical.
But not so in a taxable account: the investor holding the premium bond would be fully taxed on the 5% interest payments and would suffer a capital loss—a double whammy. Meanwhile the holder of the discount bond would be fully taxed on just 2% in interest, and then taxed on only half the 1% capital gain. As a result, the discount bond holder would have a significantly higher after-tax return.
There are a couple of ways to hold fixed income in a non-registered account while avoiding premium bonds. One is to use GICs instead of bond funds: GICs always trade at par, so they have lower interest payments and never suffer capital losses. Another is to use strip bonds, which always trade at a discount to par value. Last year saw the launch of the First Asset DEX 1-5 Year Laddered Government Strip Bond Index ETF (BXF), inspired by Justin Bender’s search for a tax-efficient fixed-income ETF. Now BMO has entered the arena with the first ETF designed to give taxable investors exposure to the broad Canadian bond market, but with a portfolio that consists only of tax-friendly discount bonds.
Zigging along with ZAG
The new ETF will have characteristics very similar to the BMO Aggregate Bond Index ETF (ZAG), which could be a core bond holding in any balanced portfolio. The two funds will be very similar in average term to maturity, duration, credit quality, yield to maturity and management fee (0.20%). The key difference, however, will be that ZDB’s average coupon will be lower that its yield to maturity, resulting in much greater tax-efficiency:
According to BMO, the new fund will hold about 50 issues when it launches, and as the ETF gathers assets it will build to more than 70 bonds. By comparison, traditional broad-based bond index funds include hundreds of holdings, but remember, there just aren’t that many discount bonds available in the marketplace. A portfolio of 50 to 70 is more than enough to provide adequate diversification.
Doesn’t this depend on your holding period of the bonds? I am assuming that you buy and hold them so even paying a 1% markup means that you are likely better off buying the bond directly if you are going to be holding the bond fro five years or more. The other way of mitigating credit risk, but getting higher yields than Govt of Canadas is to buy Provincial bonds.
@Zaphod: I’m not an expert in bond pricing, but my understanding is the markup is proportional to the term of the bond: i.e. you pay a higher commission when you buy a 10-year bond than a 2-year bond. But in any case, I don’t mean to suggest it never makes sense to buy individual bonds, only that one needs to understand the trade-off between the two options.
@Zaphod, if you are worried about single issuer credit risk, Provincials are not a great substitute for Gov. of Canada Bonds. The Government of Canada has a much better balance sheet that many provinces, they have greater ability to tax and they also have the ability to print money. Municipal bonds are even worse as we’ve seen in the US with defaults not being uncommon. I personally am worried about the high debt levels of Ontario and Quebec for instance. I think it could be a problem at some point.
As long as you stay under the CDIC limits, CDIC insured GICs are as good as Gov. of Canada bonds because of their Government of Canada CDIC backing and their current returns are above Provincials (from what I can get.)
Alternately a diversified bond fund like ZAB or ZDB is very good too if you are willing to hold them long enough. The old adage of don’t put all your eggs in one basket is still a good one when it comes to risk control.
@Brian G – I agree that provincial balance sheets are not nearly as strong but I would argue that the probability of any province actually defaulting is extremely small and if a province were so weak that they were in danger of defaulting that they would be rescued by the Feds. In other words provinces, like the Canadian banks, are too big to fail.
There can be “technical” reasons why provincial bonds trade above governments, including bank capital charges, repo-ability of the bonds, etc. that aren’t directly related to credit quality.
Ontario 10 years are trading with yields of about 3.3% vs 2.55% for GoCs. I think the return is worth the risk.
@CCP: Regarding your earlier answer to @Jas that it was too early to judge the relative merits of BXF vs ZDB because of lack of sufficient information to date, I understand that in the strip coupon ladder ETF BXF the regular distributions to ETF holders, despite deriving ultimately from the capital gain on the discounted bonds is unfortunately treated by Revenue Canada as straight interest income. In ZDB, however, there appear to be two separate streams of return — the interest portion (adjusted “unnaturally” low to offset the gain on the discounted bonds, but taxed fully at the marginal rate) as well as the capital gain on the discounted bonds which I would hope would retain its character for the ETF holder to declare as a capital gain. If this is so (the many unknown other things being equal) this would seem to be an advantage. Am I understanding this correctly? Or is this still one of the unknowns that had not yet been clarified that you were referring to?
@Oldie: I would consider these unknown—at least unknown by me. If you have questions about these specific products, I’d encourage you to contact the ETF provider directly.
I’m noticing that ZEA only holds ~350 companies, even ZDM (which it will presumably catch up to eventually) only has about 450, with a stated target of 85% of all capitalization per country. The comparable Vanguard (VDU, >1000 companies including some small-cap) and iShares (XEF, >2500 companies of all caps) funds cover a far larger share and provide better diversification, in addition to the historic return advantage of small-cap equities.
You’ve previously stated that you like your funds as broad as possible. Given a tax advantage of 0.19%-0.27% in ETF costs as calculated in your white paper (depending on account type and ETF compared to), does this outweigh the disadvantages in diversification in your opinion?
The generalized answer might be a good topic for another blog post, where do I draw the line between savings and diversification? How much cheaper does a less-diversified ETF need to be to reasonably justify picking it over a more expensive but better-diversified one? This is of course an issue that also concerns Canadian and (especially) U.S. indexes. So far I’ve been dealing with this question by following your general advice and ignoring a few MER basis points up or down, but it would be nice to get an idea where these things actually start paying off.
@Jakob: BMO has assured me ZEA and ZDM will ultimately have the same portfolio, with the only difference being the hedge.
It’s hard to be too specific about the trade-off between added diversification and tax-efficientcy. My guess is you would find ZEA, VDU and XEF would have a correlation of over 0.95: once you have 400+ stocks, there’s no meaningful reduction in volatility from adding more. The added mid/small exposure in XEF might give it slightly higher expected returns, but that’s elusive and would only show up over the very long term: some years it will underperform because of that exposure. Tax-efficiency, by contrast, is essentially guaranteed.
For what it’s worth, with clients at PWL we’re mostly using VDU in taxable accounts for now, but VEA is worth keeping an eye on.
FWIW, here is the limited information I got back from my direct enquiry regarding the character of the expected returns on ZDB:
(My question):
Regarding ZDB, I understand returns are generated by 2 distinct mechanisms; firstly interest on the underlying bonds; and secondly capital gains when the underlying discounted bonds are sold. Please advise whether these two forms of revenue retain their character when returns are passed on to the ETF holders, that is, an interest portion and a capital gains portion which the ETF holder may then declare to Revenue Canada at a lower taxation rate.
(Their reply):
Thank you for contacting BMO ETFs. With reference to your inquiry, the distributions on the ZDB are expected to be paid primarily out of other income or gains, received from the underlying holdings less the expenses, but may also consist of non-taxable amounts including return of capital. The figures to be declared to Revenue Canada to will be reported on your tax slip at the end of the year.
Please let us know if we can be of further assistance…etc, etc
So, either everyone doesn’t know yet, or, if there is knowledge higher up about the relative ratio of the capital gains component, if any, based upon a known strategy, it will have to be pried free by a call from a personal contact.
I also got a reply from First Asset regarding the tax treatment of distributions of BXF. I was instructed to go the the website, look for distributions and download the indicated csv file. Here is a copy of the file contents:
(quote)”
=upper(“First Asset DEX 1-5 Year Laddered Government Strip Bond Index ETF”)
(1) The future distribution dates may be amended at any time.
(2) Reinvested distributions are not paid in cash but instead remain invested in the Fund. To recognize that these distributions have been allocated to investors for tax purposes the amounts of these distributions should be added to the adjusted cost base of the units held.
(3) The characterization of distributions for tax purposes (such as dividends/other income/capital gains etc.) will not be known for certain until after the Fund’s tax year end. Therefore investors will be informed of the characterization of the amounts distributed for tax purposes only for the entire year and not with each distribution. For tax purposes these amounts will be reported by brokers on official tax statements.
Record date,Payable date,Fund,Distributions per unit,Cash,Reinvested,Eligible dividends,Non eligible dividends,Other income,Capital Gains,Return of capital,Foreign Income,Foreign tax paid,
2013-12-30,2014-01-06,BXF.A,0.031250,0.031250,-,-,-,0.023530,-,0.007720,-,-,
2013-09-27,2013-10-03,BXF.A,0.038360,0.038360,-,-,-,0.028890,-,0.009470,-,-,
2013-12-30,2014-01-06,BXF,0.043750,0.043750,-,-,-,0.036550,-,0.007200,-,-,
2013-09-27,2013-10-03,BXF,0.053700,0.053700,-,-,-,0.044870,-,0.008830,-,-,
“(Unquote)
It appears that (if I am understanding this correctly) the 2013 distribution consists entirely of Capital Gain and Reinvestment of Capital, contrary to what I understood from earlier information derived from tax bulletins, which suggested that this was to be declared as interest, and therefore straight income.
I assume “the Fund’s tax year end” means December 31, 2013, in which case this is all the information we are going to get from First Asset. I suppose there always remains the possibility that the official tax statements from the broker to Fundholders (but not to the general public) might state the distributions to be identified as interest equivalents, or that Revenue Canada will unilaterally declare thusly. It is disconcerting not knowing which is what. Any definitive advice from anyone out there?
Correction for BXF Distributions file format interpretation: I read the comma separated value format incorrectly — on each line there is a blank spacer value between the second last and last given values, and if you count the headers, and then the corresponding values on the line you get the correct labels for the given values as, for instance, on the 2013-12-30 Record Date payment line:
…Other Income = 0.036550, (Capital Gains = Blank Value), Return of Capital = 0.007200,…(all amounts presumably in decimals of dollars per share).
So distributions, at least for 2013, are mostly “Other Income” which is fully taxable at marginal rate, with a small amount of “Return of Capital”. But that’s way better than what you’d get with a standard bond ETF — an inflated taxable interest amount offset by a capital loss (which you likely can’t exploit tax-wise).
I thought I’d note that BMO’s website is now showing that ZEA holds stocks directly.
I have been watching ZEA the last few weeks. As per Tyler’s note above, it held down to only about 1.3% iShares MSCI EAFE ETF, the rest was holding the international stocks directly. Last week I saw this was up to around 9%, and now just checking today, I see it has climbed all the way up to 49.28% iShares MSCI EAFE ETF. Is this variation only because it’s such a new ETF and they are trying to sort out its holdings still?
I have also noticed the Units Outstanding (000’s) has gone from 2,100 to 2,300 to now 4,100.
Is this cause for concern with this particular fund, or should it all eventually sort it outself out?
@MapleLeaf: These two bits of information are almost certainly related. If the fund went from 2.1 million to 4.1 million units, it means it virtually doubled in size because of a large inflow of new cash. That’s not a cause for concern, it’s exactly what you want to see, because it means the ETF is building scale.
It also explains why the ETF suddenly went from 9% to 49% invested in the iShares ETF: the managers were probably not able to efficiently invest all that new money directly in the stocks, so they have parked it in the iShares ETFs for now to maintain market exposure until they do.
Thanks CCP. Makes sense. And here I was thinking things were looking down and down for ZEA.
I am looking at adding a bond ETF to my rrsp and have close to 10000 to contribute to this (I need fixed income in my account). I would like a mix of government/corporate bonds and wondered between vanguard- VAB; XBB- ishares or ZAG (BMO) which one would be preferred? holding this for the long term. I have other fixed income/bond exposure through a balanced fund but specific to an ETF for fixed income, which may be the better choice? or is there a different core bond etf that would be preferred over one of these? I’m some what new to the fixed income side of things…..and my advisor is trying to get me to put more into my portfolio rather than having too much emphasis on growth oriented equities. Your thoughts would be appreciated. Thanks Dan.
@Sue: I can’t make a specific recommendation for you, but if you are looking for a broad-based bond fund with both government and corporate bonds, with a range of maturities averaging about 10 years, all three ETFs you mention are very similar. VAB and ZAG are somewhat cheaper that XBB. VAB also is a little heavier on the government bonds.
@ CCP: I am looking for advice in setting up a fixed income portfolio.
My aged and ailing mother is now in a care facility, and we are selling her home. I wish to invest those funds (let’s assume it’s $200K) in a non-registered low-risk portfolio. That money will be used gradually over the course of the next 8 to 12 years to pay for her care, and so preservation of capital is very important. A 3% to 5% annual return would be just fine.
I do not expect the tax implications between using GICs versus bond ETFs (not sure which yet) to be an issue, since my mother will have sufficient “attendant care expenses” to offset tax payable on the interest income. (I have read your blog, “Why GICs Beat Bond ETFs in Taxable Accounts”, that compares the tax efficiency of both products in taxable accounts.)
Question 1: Based on the above, I think it would be better to use bond ETFs rather than GICs; the ETFs are more liquid which is a key advantage I am seeking. Is my conclusion valid?
Question 2: I would set up a portfolio with an initial mix of 80% fixed income, and 20% equities, to try to get some “juice” from the equities in the early years; after 3 or 4 years, I would switch it all to fixed income. Is that a sensible strategy?
PS: I have purchased your Perfect Portfolio book, shut down my account with my advisor, am now a bona fide Couch Potato, and feeling good about it! Thanks for your sharing your knowledge, I have learned tons.
Best regards to you and this great community.
@Jean: I’m afraid I can’t offer any advice specific to your situation. Managing an ailing parent’s life savings is an important job, and I would encourage you to seek the help of a fee-only financial planner.
In general, if an investor has a marginal tax rate of 0% (because of low income and high expenses), then it’s true the “bond ETFs versus GICs” question is irrelevant. And, yes, the ETF would provide greater liquidity. However, an ETF can lose value over the short term, so there is always a possibility that it might be liquidated at a loss. Funds that are needed within a year or two are generally better kept in a high-interest savings account, which is likely to have a yield at least as high as a short-term bond ETF anyway, with zero volatility.
Excellent article.
From what i understand, for someone just starting with a new portfolio in a taxable account, it would make sense to use ZDB for the fixed income. I know that we cant time the market but is the ZDB advantage permanent ? Or will it dismiss in the future when interest rate rise and bond return at par.
@LawrenceW: If interest rates trend upwards for years then it will be increasingly easy to find discount bonds, and the tax inefficiency of traditional bonds ETFs will become less of an issue. But that won’t hard ZDB in any way. Discount bonds will always be more tax-efficient than premium bonds.
Hi Dan
Your article is great, and gives some good options for non-registered accounts. But I’m sure quite a few people are in my boat, where the non-registered account is used only after the registered ones are maxed out. To that end, I plan on regular monthly contributions in the non-registered account, so a MF would make sense. Do you have any suggestions on some good tax effective MFs? I bank with RBC so unfortunately the Mawer Tax Effective Balanced one (MAW105) is out. Some other balanced ones like the PH&N balanced fund (RBF1350) are almost 30% bonds, so not great for a taxable account.
I was thinking of a combination of a Canadian Index Equity fund like RBCs RBF556 (60%) along with a fixed income component (40%), like GICs possibly. Any suggestions?
@Sid: Your situation is actually extremely common. But the way to manage it is to treat the registered accounts and non-registered accounts as a single large portfolio and to use tax-efficient asset location. Holding balanced funds in a non-registered account is generally not a great idea because these funds hold a lot tax-inefficient bonds. A better solution is to hold the bonds in the RRSP and the equities in the non-registered account. If there is insufficient room to shelter all the bond holdings in the RRSP, then you could consider more tax-efficient fixed income options such as specialty ETFs or GICs.
I should mention that this idea is one of the more difficult aspects of portfolio management and very few DIY investors (and even advisors) manage it effectively.
https://canadiancouchpotato.com/2012/03/12/ask-the-spud-investing-with-multiple-accounts/
Dan,
Now that ZEA has a few months under its belt, are you ready to endorse it as an ideal ETF choice for your CCP portfolio? Given its very positive tax structure, would it not be considered particularly the best choice for an RRSP?
Thanks as always for the great blog
@Brian: I have no objection to ZEA at all. But I am interested in seeing what its tracking error and taxable distributions are after a full year. Our estimate of its total cost including MER and withholding taxes is just theoretical at this point.
Can anybody show me how to calculate or estimate the potential tax savings in an non-registered account if I hold ZDB rather than XBB or ZAG? Since ZDB is thinly traded, it has a larger bid-ask spread ($0.04/15.30) which represents a bigger loss each time I buy and sell ZDB. I’m guessing the loss is somewhere in the neigbourhood of half the spread, or half of $0.04/15.30 which = 0.13% on the purchase and similarly 0.13% on the sale say in 6 years from now for a total loss of 0.26% on the purchase and sale distributed over 5 years. This represents a loss of 0.04% per year due to a large bid-ask spread. What would the tax savings be due to the discounted bonds strategy? Would they offset this extra cost due to the larger bid-ask spread?
Hi Dan,
I”ve been researching ZEA and the tracking chart on the website looks pretty good so far, but I am not sure if I am reading it right. Based on your expertise, does the current tracking error look good compared to a Vanguard or iShares fund?
http://www.etfs.bmo.com/bmo-etfs/trackError?fundId=97630
Thanks!
@Brian: Ouch, from my eyeball look at the graph, the two lines seem to be separated by a constant interval (except within the past few days, which can be explained by a couple of days lag in the ZEA price) of about 3/4 to under 1 percent. I can’t confirm this with the data listing because my computer doesn’t support the method of display. But 0.75 to let’s say 0.9 per cent in the face of an MER of 0.20% doesn’t sound good. The 2 lines track the percentage return, so the tracking error is the ongoing difference between the two data lines, right?
Dan, what’s your take on this?
@Oldie, the constant difference was likely just caused by a tiny bit of tracking error right at start-up only. It would then just carry forward as a constant difference. I charted ZEA and VDU together on TMXMoney with re-investing dividends turn on and they are identical for the last 3 months. Interestingly, XEF had a tracker error event back in May and is a fair bit lower then these two.
All in all, ZEA looks good to me.
Oh heck. I just used the tracking error tool on iShares site for XEF. In the last 3 months, it is showing a huge tracking error. Anyone know what is going on?
Period : 23/04/2014 to 23/07/2014
Cumulative performance difference: 2.93%
Fund return: -0.08%
Index return: 2.85%
I would not read too much into website tools like this: errors in the data are common. The most meaningful way to measure tracking error is to wait for the annual Management Report of Fund Performance, which will give you accurate data over the last full year.
@CCP, you are correct. It looks like an error with the iShares tracking tool. Besides XEF, XUS is also showing wild tracking errors. My guess is they are comparing to the wrong benchmark (e.g. maybe the currency hedged one.) I have e-mailed Blackrock to point out the potential errors. I will post on here if I get a meaningful response.
I hold a substantial amount of XBB. When I get the statement of earning from Blackrock each year earnings are identified as dividends. Is that correct, or does Revenue Canada view the earnings as interest and tax them as such?
@Jim: XBB’s distributions are virtually all interest, with no dividends. So I’m not sure why BlackRock uses that terminology. The important paperwork is the T3 slip you receive, which should characterize the distributions properly.
Hi Dan – Would ZEA be your recommended ETF for international exposure in a non-registered account?
@PoiteClairePower: Both ZEA and XEF now hold their underlying stocks directly, so both would be a good choice.
I think the best tax friendly bonds ETF in Canada is HBB from horizon. I would be curious to see if Horizon will have a better after tax return than ZDB? I know that they have a swap structure with National Bank, and that ad a little bit more risk to it, but what do you think?
@Martin: I discuss HBB here:
https://canadiancouchpotato.com/2014/05/08/a-tax-friendly-bond-etf-on-the-horizon/
I have been periodically checking the BMO website for information on ZEA since this ETF was first introduced to us in these pages, and the amount of foreign equity held through an iShares ETF has not dropped to zero as I was anticipating; currently it is sitting just over 10%. So I sent a request for clarification from BMO and got this back:
“Thank you for contacting BMO ETF, reagarding to your inquiry about the holdings of iShares MSCI EAFE ETF, we use this as a liquidity vehicle to manage inflows or outflows into and out of the fund. It generally varies between 2% and 10% weight. The advantage is that we don’t have to trade international equities on each cash flow, which reduces trading costs and benefits tracking.
Best Regards,”
so I guess “somewhere between 2% and 10%” has been reached and is the best we can expect for ZEA. The largest holding in XEF currently is ISHARES MSCI EAFE ETF representing 3.67% of total capitalisation, so I suspect the situation is similar to ZEA.
Are there any Tax-Efficient ETF to hold US$/US ETF in a taxable account.
Thanks.