Your Complete Guide to Index Investing with Dan Bortolotti

Finding the Perfect Pair for Tax Loss Selling

2018-08-23T08:55:12+00:00October 24th, 2013|Categories: ETFs, Taxes|Tags: |82 Comments

If you’re using ETFs in a non-registered account, there’s plenty of opportunity to harvest capital losses and reduce your tax bill. Justin Bender and I tell you exactly how to do this in our new white paper, Tax Loss Selling: Using Canadian-listed ETFs to defer taxes on capital gains.

As I explained in my previous post, tax loss selling involves dumping an ETF that has declined in value to crystallize the loss, and then buying a similar (but not identical) ETF to maintain the exposure in your portfolio. The Canada Revenue Agency considers any two index funds tracking the same benchmark to be identical property. So you cannot, for example, claim a loss after selling the iShares S&P 500 (XUS) and replacing it with the Vanguard S&P 500 (VFV): if you do, it will be denied as a superficial loss.

Fortunately this year has seen a number of new ETF launches, including international equity ETFs from iShares and Canadian and US equity ETFs from Vanguard. These give Canadians much better options when tax loss selling, because it’s now relatively easy to find pairs of ETFs tracking similar indexes from different providers. That allows you to keep your market exposure essentially unchanged while staying on the right side of the tax laws.

When we put together the white paper, Justin analyzed 10 years’ worth of index data to find the pairs that most closely mirrored each other. For example, he compared the S&P/TSX Capped Composite Index to the FTSE Canada All Cap Index. Turns out the two indexes dance to the same drummer:

We can therefore be reasonably confident that the BMO S&P/TSX Capped Composite (ZCN) and the recently launched Vanguard FTSE Canada All Cap (VCN) make a suitable pair for tax loss selling. Swapping one for the other should allow you to harvest a loss without any meaningful difference in performance. Here’s a list of suitable Canadian ETF pairs to consider when tax loss selling:

Original ETFReplacement ETF
BMO S&P/TSX Capped Composite (ZCN)Vanguard FTSE Canada All Cap (VCN)
Vanguard US Total Market (VUN)iShares Core S&P U.S. Total Market (XUU)
iShares Core MSCI EAFE IMI (XEF)Vanguard FTSE Developed All Cap ex North America (VIU)
iShares MSCI Emerging Markets IMI (XEC)Vanguard FTSE Emerging Markets (VEE)



  1. Kenny October 30, 2013 at 11:34 am

    I’ve read both these posts and I’m curious what the risk is that this violates GAAR?

  2. Canadian Couch Potato October 30, 2013 at 11:54 am

    @Kenny: I always encourage people to consult a tax expert if they are concerned: I don’t profess to be an expert and I don’t give tax advice. But tax loss selling is an extremely common strategy, and the CRA has issued a bulletin specifically stating that two ETFs tracking different indexes are not considered identical property. The bulletin is from 2001, but the Canadian Financial DIY blogger contacted CRA to confirm in 2008, which he explains here:

  3. Tennis Lover November 6, 2013 at 5:32 pm

    I hold some XRB in a non- registered account and wonder if I could sell it and buy ZRR thus harvesting the loss on XRB. However both ETFs hold the same Govt. Of Canada real return bonds. The only difference between the two ETFs, in terms of holdings, is that XRB also holds some provincial real return bonds. Given that ZRR has all the same holdings as XRB would this be a superficial loss?

  4. Canadian Couch Potato November 6, 2013 at 5:50 pm

    @Tennis Lover: The two ETFs track different indexes and have different holdings (even if some of them overlap), so according to the CRA’s definition these would not be identical properties. But the larger issue here is that real-return bond ETFs are terribly tax-inefficient and should never be held in a non-registered account!

  5. Tennis Lover November 8, 2013 at 12:34 pm

    Thanks for the reply.

    Just want to clarify your comment about the terrible tax-inefficiency of real-return bond ETFs in a non-registred account. I agree that ETFs that distribute interest income should be held in a sheltered account if posssible. I thought that another factor is that the face value of real-return bonds increase each year with inflation and the holder could be stuck with paying capital gains taxes on these increases even though they don’t actually sell the bond (similar to strip bonds). I looked at the distribution history of XRB and found that from 2006 to 2012 the ETF distributed $5.58 with $4.89 (88%) of this being interest income and $0.40 (7%) being Capital Gains and $0.28 (5%) being Return of Capital. Based on these numbers the Capital Gain is not a significant portion of the distributions. Is there something I am missing?

    Assuming that a person is only able to set up a Complete Coach Potato Portfolio in a non-registered account, is there a recommended alternative to the 10% real-return bond weighting?

  6. ccpfan November 8, 2013 at 12:41 pm

    The most tax-efficient approach is to keep income below 10,000$. You won’t be rich, but you won’t have to pay taxes!

    Joke aside, I personally don’t subscribe to the “tax-efficiency first” approach to asset allocation/location; I’d rather take a “total after-tax return first” approach. This can (sometimes) lead to choices that differ than those most-often promoted on this blog.

    Just my humble opinion.

  7. Canadian Couch Potato November 8, 2013 at 12:46 pm

    @Tennis Lover: Yes, part of the issue is that some of the capital gains (the inflation adjustments) are taxed as they accrue, but the larger issue is that all RRBs are now selling at a premium, and premium bonds pay high coupons (fully taxable) that are offset with capital losses. This post explains this idea in detail:

    If you have no choice but to hold fixed income in a non-registered account, GICs tend to be the best choice. It may not be worth holding RRBs at all, since their after-tax return could be negative even in a year when they deliver positive pre-tax returns.

  8. Canadian Couch Potato November 8, 2013 at 12:55 pm

    @ccpfan: Opinions are always welcome, but I think I have bent over backwards to say tax-efficiency never comes first. The first step is asset allocation (which as about risk management, not tax management), and then asset location comes after that. Proper asset location can lead to a portfolio with the same level of risk but a higher after-tax return.

  9. Tennis Lover November 8, 2013 at 4:47 pm

    Thanks for the explanation and the link. As you mention the fact that all RRBs are trading at a premium means that an investor will be trading off fully taxable high coupons against capital losses. However will the RRBs necessarily have a capital loss when they mature in many many years? Won’t their value be adjusted upwards?

    I think that the advice is sound but wonder if there is a small element of market timing. Right now RRBs are unloved but that may change if inflation picks up.

  10. Canadian Couch Potato November 8, 2013 at 7:29 pm

    @Tennis Lover: RRBs may be a special case, because their maturity dates are so far away, unlike a fund like XBB where a good portion bonds of the bonds are nearing maturity every year. (Technically a bond never matures in an ETF: they are always sold with one year left to maturity, but there are still capital losses.) As for the market timing, the advice has nothing to do with the outlook for RRBs: I hold them myself and consider them part of a long-term strategy. The only question is where to hold them.

    To return to your original question, if you do want to hold RRBs in a taxable account, they are likely to be a good candidate for tax-loss harvesting now, being down almost 10% this year.

  11. ccpfan November 10, 2013 at 8:38 pm

    @CanadianCouchPotato: My apology; I did not intend to misrepresent what you write. I fully agree that you do put Asset Allocation first in your writings.

    What I intended to day was: On this blog, when you write about Asset Location, you do emphasize maximal “tax efficiency” as the goal. My opinion is that the primary goal driving the Asset Location decision should be different; it should be to maximize “total after-tax return”. In some cases, both approaches lead to similar Asset Location decisions, but sometimes, they can lead to different decisions.

    Both the return earned on the investment and the investment period have important impacts on the calculation. Unfortunately, the return is unknown in advance, so it is quite difficult to precisely locate assets so as to maximize after-ta return.

    Before you reject my argument (as you did with my tax-free RRSP grwoth comment, for which I then provided a proof yet go not reaction from you), here’s a numerical example.

    I’ll assume a hypothetical bond fund returning a 2% yield in interest every year and keeping constant NAV price. I’ll also assume a hypothetical stock fund that distributes no dividends and gains 5% NAV price every year. We have an initial $2000 portfolio that we want to invest 50/50 in stocks/bonds and we only have 1000$ contribution room in a TFSA. For simplicity, I won’t rebalance. Let’s invest for 10 years.

    The annual return for both funds is trivial in the TFSA.

    If we were to put the bond fund in taxable, its annual after-tax return would be 1% (1% gets eaten by taxes). At the end of 10 years, there would be no capital gains.

    If, instead, we were to put the stock fun in taxable, its annual after-tax return would be 5% (no tax). At the end of 10 years, 25% of the capital gains would be paid in taxes.


    * Bonds in TFSA, Stocks in taxable:
    Bonds = 1000$ x 1.02^10 = 1219$
    Stocks = 1000$ x 1.05^10 – taxes = 1629$ – 629$/4 = 1629$ – 157$ = 1472$
    After-tax Portfolio = 1219$ + 1472$ = 2691$

    * Bonds in taxable, Stocks in TFSA:
    Bonds = 1000$ x 1.01^10 = 1105$
    Stocks = 1000$ x 1.05^10 = 1629$
    After-tax Portfolio = 1205$ + 1629$ = 2834$


    After only 10 years, putting stocks into the TFSA instead of taxable, wins by by an additional 5%+ margin.

    The math is clear. The “tax efficient” approach does NOT maximize “after-tax return” in this example.

    Of course, one can build different examples with higher bond returns and get different results. I don’t claim that putting bonds in taxable is best. All I say is that making the best asset location decision depends on knowing the return of each asset, which is impossible at the time of investment.

    This is what I meant when I said that I don’t don’t subscribe to the “tax-efficiency first” to asset location promoted on this blog.

    Humble suggestion: Maybe putting a little of everything in every account would mitigate the risk of putting the wrong asset in the wrong location. Another reason to do the same would be Rick Ferri’s argument for keeping mirror asset allocation in every account: “[T]here is a hidden risk with having different allocations in taxable versus non-taxable, and we saw this risk turn into reality during 2008 and early 2009. A few clients terminated their higher risk taxable portfolio because that specific portfolio was losing more money than the more conservative non-taxable portfolio. In other words, they separated their portfolios in their mind and compared returns rather than looking at the big picture.”[]

  12. ccpfan November 10, 2013 at 10:32 pm

    Small mistake:

    After-tax Portfolio = 1105$ + 1629$ = 2734$

    This reduces the advantage to: 1.5% (instead of 5%), but “tax inefficient” still wins.

    (It would be so nice to be able to edit posts).

  13. Brad December 8, 2013 at 1:05 am

    Can you list ETF pairs for the Uber Tuber funds? Following the advice of your “Putting Your Assets in Their Place” I’m most interested in the Canadian ETFs (CRQ, XCS), International (EFV, SCZ) and Emerging (VWO).

    CRQ could likely use ZCN or VCN because CRQ tracks a different index. I wonder how their charts compare.

  14. Canadian Couch Potato December 8, 2013 at 12:31 pm

    – For CRQ, the closest is probably XCV, which has a similar value tilt.
    – There’s nothing comparable to XCS.
    – For international value there’s Wisdom Tree’s DTH and PowerShares PXF
    – For international small caps, there’s Vanguard’s VSS and Wisdom Tree’s DLS
    – For VWO there’s iShares EEM

  15. Elad December 8, 2013 at 11:47 pm

    So if you sell at loss and buy something that is pretty much identical to what you just sold, where is the benefit? I assume if one will continue to lose money, the other will lose too.
    I don’t see how you can offset the loss with profit from two funds that are almost identical.

  16. Ryan January 7, 2014 at 12:59 am

    This is a little off topic, but can ETF’s be donated to charity to avoid capital gains and get the full charity tax benefit from the donation?
    I donate to charities anyway, I’m thinking this may be a way to help me transition from VTI to VUN in my non-registered account without incurring capital gains.

  17. Canadian Couch Potato January 7, 2014 at 1:10 am

    @Ryan: Great question. You’ll definitely want to check with your accountant or tax advisor before acting, but yes, it may be possible to avoid capital gains taxes on securities (including ETFs) that are donated to charity:

  18. Erik July 1, 2014 at 7:39 pm

    For those of us holding EWC in our portfolio for US tax reasons (i.e. PFIC), can anyone recommend a pairing? I noticed that SPDR has a new MSCI Canada Quality Mix ETF (QCAN) that seems to fit the bill — same holdings as EWC but different weightings according due the factors in the index. Thoughts?

  19. Canadian Couch Potato July 2, 2014 at 7:59 am

    @Erik: The only way to test the pairs is to look at their correlations over several years. Since QCAN is so new (and I don’t know whether the index data is available further back) that’s impossible to do. But the sector mix of both ETFs is very similar, so it’s hard to imagine the difference would be significant, especially if you only plan to hold it long enough to satisfy the superficial loss rule.

  20. Erik July 2, 2014 at 9:18 pm

    @CCP: Thanks for the insight. Was pleased to see a second option for US-traded Canadian stock market exposure finally. And as you say, it’s only a month, so unlikely to do too much damage even if not an ideal fit.

  21. Erik August 6, 2014 at 3:07 pm

    Great article, but got me wondering whether there’s an equivalent CRA rule for superficial gains. i.e. if you sell an ETF at a gain but buy it back immediately, can you take the gain or is it superficial? Assuming no because CRA is usu happy to take tax now rather than defer, but seemed an appropriate place to ask.

    Reason for doing this would be an unusually low- or no-income year, where you can reset the ACB upward cheaply. If you have no other income, first $18k in gains is tax-free because of the personal exemption, and you’re at the lowest bracket after that, so it makes sense to give up the deferral if the tax is low enough.

    Also curious how you’d calculate the break even on this sort of thing…i.e. how low does the tax rate have to be to make up for loss of growth.

    – Erik

  22. Peter August 16, 2014 at 8:19 am

    @Erik: A good question, and I am wondering about that as well.

    Seems to me that if you have no other income in Canada you could sell a holding that had gained up to 10,000 and replace it with a similar holding and permennetly avoid the taxes each year. Your regional and sector balance would still be the same as would the value of your overall holdings. But after 5 or 6 years you would have been able to create and keep potentially up to 60K of tax free income. This would just become a part of your yearly rebalancing.

    Then when you start needing to draw down more than 10K a year for retirement income, your official capital gains claim would be much much lower saving you even more in taxes through your retirement years.

    Very very roughly-

    100K, 10K gain a year for 5 years- 150K and you start withdrawing with a claim of 50%+ capital gains.
    100K, 10K gain a year voided by tax harvesting for 5 years- 150K to withdraw from with 0% gain, and so no tax owed.

    Or do I have that wrong?

  23. Canadian Couch Potato August 16, 2014 at 10:10 am

    @Erik and Peter: There is no such thing as superficial gains rule: the government is more than happy to allow you to realize gains now rather than deferring them. If you have no other income you can actually realize a gain closer to $20,000, since only half of the realized gain needs to be reported as income and the basic personal amount is around $10,000. This should not be surprising. It’s the same as saying if you had no other income you can get a job that pays $10,000 a year and pay no tax.

    If you have carried forward losses when you reach retirement then, yes, you could indeed realize an equivalent amount of gains without paying tax.

  24. Peter August 16, 2014 at 10:35 am

    Thanks for the quick response,

    Just to be clear, you don’t even need to switch indexes when dealing with a gain? You can just sell the ETF, realize the gains and then buy exactly the same one again on the same day at the new price?

    I know this is would only be a benefit to a very few- but it looks like potentially a big benefit over the long term for those that can take advantage of it.

    I am not missing anything negative here am I?

  25. Canadian Couch Potato August 16, 2014 at 11:49 am

    @Peter: No need to switch indexes if you’re reporting a gain. It’s actually not that unusual of a strategy: indeed, one of the reasons to keep equities in taxable accounts is so you can defer gains until retirement, when they can be realized at lower tax rates.

  26. Tristan December 9, 2015 at 1:46 pm

    @CCP: I might soon be looking for a replacement for tax loss harvesting with XMD. Two contenders might be XCS and PZC (Powershares small and mid cap fundamental index). Which of those, or any others, would you favour? Thanks.

  27. Canadian Couch Potato December 9, 2015 at 2:33 pm

    @Tristan: There is really no direct counterpart to XMD. PZC would seem to be a good candidate, but since it is so new it’s impossible to know how closely the performance of these two ETFs will track each other. If an investor also held a large cap Canadian equity ETF (such as XIU) on solution might be to sell this as well as XMD and replace both with a broad-based ETF (such as XIC or VCN).

  28. Tristan December 10, 2015 at 11:01 am

    @CCP: The other Canadian equity I hold is HXT (in a 50/50 split). I’d like to be able to sell both and replace them with VCN (if I were setting up again I’d go with just VCN to reduce complexity), but the problem is, they’re in a taxable account and HXT has a capital gain of 15% from when the portfolio was set up in July 2013. So there’s a cost to doing that. With XMD down just over 4%, it shows what effect the big drop oil prices has done to that segment of the market.

  29. Tristan January 5, 2016 at 5:08 pm

    @CCP: I’m curious as to why you chose ZCN as opposed to XIC for tax loss selling with VCN? I’m thinking it would “six of one and half of the other”? Thanks.

  30. Canadian Couch Potato January 5, 2016 at 5:35 pm

    @Tristan: I’ve just listed one suggestion. Either one would be suitable.

  31. Tim September 16, 2017 at 3:05 pm

    @CCP: Great article. I use your model ETF portfolio and hold XAW as well as VCN in my non registered (TFSA and RRSP are full). Would you be able to suggest a good fund to tax loss harvest XAW? Or is that one so specific would I have to buy the underlying funds and find replacements for those individaully? Would prefer to keep it simple with the 1 fund if possible. Love the blog, thanks for everything!

  32. Canadian Couch Potato September 17, 2017 at 9:09 am

    @Tim: Thanks for the kind words. Vanguard’s VXC is an ideal replacement for XAW.

Leave A Comment