So far in our tax-loss selling blog post series, we’ve taken a pretty deep dive into how to engage in tax-loss selling in your non-registered accounts.
To review, tax-loss selling means selling an existing ETF at a capital loss, and then repurchasing a similar, but not identical holding to maintain similar asset class exposure in your portfolio.
Again, the new ETF must NOT be deemed identical to the one you just sold, or you’ll violate the CRA’s superficial loss rules. If you do, you’ll lose the tax savings you were hoping to gain.
To Err Is Human, To Pair Is Divine
To that end, it helps to have a list of ETF pairs that offer you similar, but different exposure to the same asset class. That’s what today’s post is all about. Here, we’ll suggest ETF pairs we believe would be good candidates. Fortunately, due to the seemingly endless supply of new ETFs being launched, tax-loss selling for Canadians has never been easier.
Actually, before we continue, we pause for an important public service announcement! The rest of this post is going to go deep into the weeds of tax-loss selling pairs. If your curiosity is itching, I am confident it will be deeply scratched by the end.
Are you ready? Let’s pair off! We’ll start with the most broadly diversified asset class ETFs for global (ex Canada), Canadian, U.S., international and emerging markets stocks. We then suggest a replacement ETF that will provide similar market exposure without being deemed identical property. Oh, and as always, you should speak with your accountant or other tax-planning specialist before engaging in any tax strategy.
Global (ex Canada) Equities Pairing: XAW Pairs with VXC (and a Surprise)
The iShares Core MSCI All Country World ex Canada Index ETF (XAW) is familiar fund among DIY investors, featured in the Canadian Couch Potato and Canadian Portfolio Manager model portfolios. XAW follows the MSCI All Country World ex Canada Investable Market Index. This index tracks the performance of large-, mid- and small-cap companies in the U.S., emerging markets, and developed markets outside of North America.
The Vanguard FTSE Global All Cap ex Canada Index ETF (VXC) is the best XAW replacement for tax-loss selling purposes, as it also tracks large-, mid- and small-cap companies across the world, excluding Canada. From 2009–2018, the ETFs’ underlying indexes had a low monthly tracking error of 0.065%, indicating similar expected performance going forward. Over the same period, performance for both indexes was nearly identical – so similar, in fact, you can barely make out the grey line of the MSCI ACWI ex Canada IMI in a graph charting both of them.
By the way, tracking error is often commonly used to describe the size of a fund’s outperformance or shortfall. Without getting too technical, I’m using a more technical application of the term here, based on the standard deviation of the differences in their monthly returns. As used in this post, the lower the tracking error, the more closely two indexes track one another.
Growth of $10,000: MSCI ACWI ex Canada IMI (net div.) (in CAD) vs. FTSE Global All Cap ex Canada China A Inclusion Index (net div.) (in CAD)
Sources: FTSE Russell Indices, MSCI, Morningstar Direct, Dimensional Returns 2.0 (2009 to 2018)
Now for the surprise. Until very recently, VXC had a higher management expense ratio than XAW (0.27% vs. 0.22%) and higher expected foreign withholding taxes in a taxable account (0.13% vs. 0.04%). This meant that most tax-loss selling investors would have preferred to switch back to XAW once the 30-day holding period had been satisfied. Plus, we would have encouraged investors to only realize a loss on XAW if it was at least $10,000 and 10% of the book value. By increasing the tax-loss selling threshold, you’d have been less likely to switch your VXC holdings back at a gain if markets recovered during the 30-day period, thus offsetting the benefit from your initial tax-loss selling trades.
But, then, just as we were preparing this post, Vanguard announced it was fixing VXC’s tax inefficiencies and lowering its fee. With this update, we can now change the loss threshold to $5,000 and 5% of book value, with no need to switch back to XAW. Way to go, Vanguard!
Canadian Equities: VCN Pairs with FLCD (which also pairs with XIC and ZCN)
If you’re currently holding the Vanguard FTSE Canada All Cap Index ETF (VCN), a suitable tax-loss selling alternative would be the Franklin FTSE Canada All Cap Index ETF (FLCD). Although both fund names suggest they track the same index, a closer inspection reveals they track slightly different ones. VCN follows the FTSE Canada All Cap Index, which tracks the performance of 200 large-, mid- and small-cap Canadian companies. FLCD tracks the FTSE Canada All Cap Domestic Index, which also tracks the same 200 large-, mid- and small-cap companies as VCN, but in slightly different proportions.
The term “Domestic” in the name indicates that this index does not adjust any company weights based on foreign ownership restrictions. In contrast, the FTSE Canada All Cap Index does adjust the weights of certain companies in the index, based on these restrictions. Governments impose limits to the foreign ownership of companies within certain industries when they feel it would not be in the public interest, such as in the aviation or telecommunications sectors.
For example, Bell Canada (BCE) is subject to the Telecommunications Act, which governs the ownership and control of Canadian telecommunications carriers. The Act restricts foreign investment in voting shares of BCE to a maximum of 33 1/3%. If we view FLCD’s holdings, Bell Canada has a weight of around 2.4% as of September 30, 2019. VCN has a weight of only 0.8%. That’s a third of FLCD’s weight, which is the maximum foreign ownership of the company allowed.
Another way to think about the difference between these two ETFs is that FLCD’s index is a better gauge of what Canadian investors would consider to be the “available Canadian stock market”, while VCN’s index is what foreign investors would consider as their “available Canadian stock market”. Knowing this, it’s a bit surprising that Vanguard Canada hasn’t shifted VCN to the domestic version of this index yet.
From 2009–2018, these two indices had a monthly tracking error of 0.106%. This is a relatively small tracking error, which indicates they can be interchanged with little expected impact on future market performance. By comparison, the monthly tracking error over the same period between the FTSE Canada All Cap Index and the S&P/TSX Capped Composite Index was slightly higher, at 0.177%.
If you’re considering a switch from VCN to FLCD to realize a capital loss, there’s little reason to switch FLCD back to VCN after 30 days, unless another tax-loss selling opportunity presents itself at the end of the required holding period. Both ETFs are expected to have similar returns going forward. Without knowing which ETF will slightly outperform over your future investment horizon, we would recommend saving yourself the extra cost and hassle, and hang onto the replacement ETF.
The graph below shows the growth of $10,000 in the FTSE Canada All Cap Index (grey line) and the FTSE Canada All Cap Domestic Index (blue line) between 2009–2018. This time, the very similar tracking produces a modest “shadow effect.”
Growth of $10,000: FTSE Canada All Cap Index vs. FTSE Canada All Cap Domestic Index
Sources: FTSE Russell Indices, Morningstar Direct, Dimensional Returns 2.0 (2009 to 2018)
FLCD is also the best available tax-loss selling replacement for either the iShares Core S&P/TSX Capped Composite Index ETF (XIC) or the BMO S&P/TSX Capped Composite Index ETF (ZCN). Both XIC and ZCN follow the S&P/TSX Capped Composite Index, which tracks 233 large-, mid- and small-cap companies in the Canadian stock market.
As the name suggests, index weights are capped at 10% (avoiding a future Nortel-like meltdown due to a single company dominating the index). If a single company were to one day become larger than 10% of the FTSE Canada All Cap Domestic Index, we might see significantly higher tracking error between the two indices. The S&P/TSX Capped Composite would cap the security at 10%, but the FTSE Canada All Cap Domestic Index would continue to allow the stock’s weight within its index to increase above 10%.
Another notable difference is FLCD excludes a number of Brookfield Limited Partnerships, while XIC and ZCN include them. This is because FTSE indexes exclude limited partnerships, while S&P Dow Jones indexes do not.
In terms of similarities, the S&P/TSX Capped Composite Index does not adjust company weights with regards to foreign ownership restrictions. This makes it more similar to the FTSE Canada All Cap Domestic Index than to the FTSE Canada All Cap Index. We can see in the graph below how this single difference has led to nearly identical performance from January 2009–2018. The grey S&P/TSX Capped Composite Index line is mostly indistinguishable from the blue FTSE Canada All Cap Domestic Index line.
Over the same period, these two indices had a monthly tracking error of 0.118%. As mentioned above, the monthly tracking error between the S&P/TSX Capped Composite Index and the FTSE Canada All Cap Index was slightly higher, at 0.177%.
What happens if Vanguard ever changes the underlying index for VCN to the Domestic version at some point in the future? All else equal, we’d then prefer pairing XIC or ZCN with VCN for tax-loss selling, so FLCD would no longer be needed.
Growth of $10,000: S&P/TSX Capped Composite Index vs. FTSE Canada All Cap Domestic Index
Sources: FTSE Russell Indices, S&P Dow Jones Indices, Morningstar Direct, Dimensional Returns 2.0 (2009 to 2018)
U.S. Equities: XUU Pairs with VUN
If you’ve decided on the iShares Core S&P U.S. Total Market Index ETF (XUU) for your U.S. equity exposure, we would recommend the Vanguard U.S. Total Market Index ETF (VUN) as the most suitable tax-loss selling replacement. XUU follows the S&P Total Market Index, which tracks the performance of around 3,800 large-, mid-, small- and micro-cap companies, making it one of the most diversified U.S. equity indexes. VUN follows an equally impressive index, the CRSP U.S. Total Market Index, which includes over 3,500 large-, mid-, small- and micro-cap companies.
After switching from XUU to VUN, cost-conscious investors may prefer to switch back, as XUU has an annual management expense ratio (MER) of 0.07%, while VUN has an MER of 0.16%. We don’t feel this is entirely necessary, but if you’re planning to switch back after the 30-day period has ended, consider only selling XUU if it has a loss of $10,000 and 10% of the book value. This will increase the odds you’ll still have an overall capital loss when you switch VUN back to XUU.
From 2009–2018, the underlying indices had a monthly tracking error of 0.065%. As with our last example, the blue and grey lines in the graph below mostly overlap.
Growth of $10,000: CRSP U.S. Total Market Index (in CAD) vs. S&P Total Market Index (in CAD)
Sources: CRSP, S&P Dow Jones Indices, Morningstar Direct, Dimensional Returns 2.0 (2009 to 2018)
International Equities: XEF Pairs with ZEA
Suitable tax-loss selling pairs for international equity ETFs are harder to come by than for Canadian or U.S. equity ETFs. Many investors may consider the iShares Core MSCI EAFE IMI Index ETF (XEF) and the Vanguard FTSE Developed All Cap ex North America Index ETF (VIU) to pair well with one another, as they both track over 3,000 large-, mid- and small-cap companies in developed countries outside of North America. However, index providers FTSE and MSCI differ in which countries they consider to be “developed”. The most obvious example is South Korea, which is classified as a developed market by FTSE, but an emerging market by MSCI. As VIU follows a FTSE index, it allocates more than 4% of its portfolio to over 400 Korean companies, while XEF (which follows an MSCI index) has no Korean companies in its fund. This difference has led to a higher monthly tracking error than our previous tax-loss selling pairs, at 0.226%.
Alternatively, investors could consider switching from XEF to the BMO MSCI EAFE Index ETF (ZEA). Both ETFs track an MSCI index, which avoids the South Korea country classification issue just discussed. But ZEA does not invest in small-cap stocks, which will result in return differences, relative to XEF.
From 2009–2018, the underlying indices had a monthly tracking error of 0.171%, which is lower than our last example (indicating a more optimal tax-loss selling pair). However, the index performance during this 10-year period was noticeably different. So, if you go this route, you may want to consider switching back to XEF after the 30-day period has passed, if you intend to maintain broad-market exposure. If you plan to switch ZEA back to XEF after thirty days, a $10,000 and 10% tax-loss selling threshold is recommended.
Growth of $10,000: MSCI EAFE IMI (net div.) (CAD) vs. MSCI EAFE Index (net div.) (CAD)
Sources: MSCI, Morningstar Direct, Dimensional Returns 2.0 (2009 to 2018)
Emerging Markets Equities: XEC Pairs with ZEM
The same country classification issue arises with emerging markets indexes as well. At first glance, you might think you could replace the iShares Core MSCI Emerging Markets IMI Index ETF (XEC) with the Vanguard FTSE Emerging Markets All Cap Index ETF (VEE).
Both underlying indexes follow thousands of large-, mid- and small-cap companies in developing markets across the globe. However, as XEC follows an MSCI index (which considers South Korea to be an emerging market), it allocates more than 12% of its fund to over 400 Korean companies. On the other hand, VEE (which follows a FTSE index) considers South Korea to be a developed country, and allocates these stocks to its developed markets ETF, VIU. This difference has led to a relatively high monthly tracking error from 2009–2018 of 0.544%.
Long story short? We believe a better tax-loss selling pairing for XEC would be the BMO MSCI Emerging Markets Index ETF (ZEM). ZEM follows the MSCI Emerging Markets Index, and includes large- and mid-sized companies from developing markets around the world. As both ETFs follow MSCI indexes, they both include South Korean companies in their funds. The monthly tracking error from 2009–2018 for their underlying indexes was 0.213%, which would indicate they make a better couple than VEE and XEC.
Growth of $10,000: MSCI Emerging Markets IMI (net div.) (in CAD) vs. MSCI Emerging Markets Index (net div.) (in CAD)
Sources: MSCI, Morningstar Direct, Dimensional Returns 2.0 (2009 to 2018)
Best of Both Worlds, with a PWL Twist
We’re the first to admit, our tax-loss selling pairs for international and emerging markets equity ETFs are not ideal. The replacement BMO ETFs exclude smaller companies, while the original iShares ETFs include large-, mid- and small-cap stocks. As such, we expect more tracking error among them than we’d like to see.
If MSCI upgraded South Korea to developed markets status, we wouldn’t have this problem. We could switch XEF to VIU, or XEC to VEE, knowing that the future performance of our replacement fund would likely closely track our original ETF. Until that happens, it would appear we’re stuck with these “next best” tax-loss selling pairs for international and emerging markets ETFs.
As we’re not fans of “good enough” at PWL, we’ve come up with a better (albeit, more complicated) solution. For tax-loss selling purposes, we have adjusted our rules slightly to treat XEF and XEC as if they were a single holding. If the net unrealized dollar losses on the combined holdings of XEF and XEC are greater than $5,000, and the total percentage loss of both holdings is greater than 5%, we consider the threshold met. We would consider selling all units of both XEF and XEC and repurchasing VIU and VEE with the proceeds – even if the loss on one of the two ETFs is not above our tax-loss selling threshold, and even if one of the ETFs has an unrealized capital gain.
This relaxed rule allows us to realize the overall capital loss for our clients, maintain similar broad equity market exposure, and keep the CRA from crying foul. The tradeoff: It involves more number crunching … and another example!
Switching XEF and XEC to VIU and VEE
Suppose you bought $80,088 of XEF and $28,315 of XEC in September 2018, for a total book value of $108,403. At year-end, the value of XEF and XEC fell to $72,924 and $27,076, respectively. You’d have a total market value of $100,000, and a total unrealized capital loss of $8,403.
XEF’s unrealized dollar loss is $7,164, and its percentage loss is 8.9%. It is now a potential tax-loss selling candidate; it could be switched to ZEA to realize the loss.
On the other hand, XEC’s dollar loss is only $1,239, and its percentage loss is only 4.4%, both of which are under our $5,000 and 5% tax-loss selling thresholds.
However, if we consider XEF and XEC to be a single position, their total dollar losses of $8,403 and percentage losses of 7.8% would meet our tax-loss selling requirements.
|Security||Book Value||Market Value (as of December 31, 2018)||Unrealized Capital Loss ($)||Unrealized Capital Loss (%)|
|iShares Core MSCI EAFE IMI Index ETF (XEF)||$80,088||$72,924||($7,164)||(8.9%)|
|iShares Core MSCI Emerging Markets IMI Index ETF (XEC)||$28,315||$27,076||($1,239)||(4.4%)|
Sources: BlackRock Canada Inc., MSCI Index Fact Sheets as of December 31, 2018
After selling XEF and XEC, VIU and VEE will immediately need to be purchased, but in slightly different proportions. To determine the appropriate market-cap weights, you’ll need to use the most recent month-end index fact sheets from MSCI and FTSE Russell Indices. Currently, the approximate splits are 73/27 for XEF/XEC, and 75/25 for VIU/VEE. For more information on this process, please refer to the blog post and video, Combining International and Emerging Markets Equity ETFs.
In this example, $75,909 of VIU and $24,091 of VEE would need to be purchased with the sale proceeds of $100,000 from XEF and XEC. These investment amounts provide market-cap-weighted exposure to international and emerging markets, as of December 31, 2018. The relative index weights fluctuate daily, but using the latest month-end index fact sheets should be close enough for our purposes.
MSCI Index Weights as of December 31, 2018
|Tracking ETF||Underlying Index||Market Cap (in USD)||Allocation (%)||Sale Amount ($)|
|iShares Core MSCI EAFE IMI Index ETF (XEF)||MSCI EAFE IMI||14,590,688||72.92%||$72,924|
|iShares Core MSCI Emerging Markets IMI Index ETF (XEC)||MSCI Emerging Markets IMI||5,417,489||27.08%||$27,076|
|Total||MSCI EAFE + EM IMI||20,008,177||100.00%||$100,000|
Sources: BlackRock Canada Inc., MSCI Index Fact Sheets as of December 31, 2018
FTSE Russell Index Weights as of December 31, 2018
|Tracking ETF||Underlying Index||Market Cap (in USD)||Allocation (%)||Purchase Amount ($)|
|Vanguard FTSE Developed All Cap ex North America Index ETF (VIU)||FTSE Developed All Cap ex North America Index||15,219,418||75.91%||$75,909|
|Vanguard FTSE Emerging Markets All Cap Index ETF (VEE)||FTSE Emerging Markets All Cap China A Inclusion Index||4,830,248||24.09%||$24,091|
|Total||FTSE Global All Cap ex North America China A Inclusion Index||20,049,666||100.00%||$100,000|
Sources: Vanguard Investments Canada Inc., FTSE Russell Index Fact Sheets as of December 31, 2018
What About Asset Allocation ETFs?
Vanguard’s Asset Allocation ETFs were launched in January 2018, and have become a popular choice for DIY investors. One of the downsides of these products (and there aren’t many) is that they’re not ideal for tax-loss selling.
Take the Vanguard Balanced ETF Portfolio (VBAL) for example. Let’s say you had purchased $1 million of the fund at the end of January 2018 (shortly after its launch). There would have only been 3 days in 2018 when the fund showed a loss of at least $5,000 and 5% of its book value. These days were December 20, December 21 and December 24. There were no days where the loss was at least $10,000 and 10% of the book value.
On those December days, you could have switched to a similar fund, like the iShares Core Balanced ETF Portfolio (XBAL), keeping in mind the “wet paint” was still drying on this fund too, at around the same time.
Unfortunately, had you switched XBAL back to VBAL after 30 days, you would have realized substantial capital gains due to the stock market recovery in January 2019. This would have offset the majority of the capital losses from your initial tax-loss selling trades. Also, the trading costs due to the ETF’s bid-ask spreads would generally be larger when trading an asset allocation ETF, relative to trading the individual underlying ETFs. You’re technically forced to sell and buy all of VBAL/ XBAL’s underlying ETF holdings behind the scenes, including the bond ETFs, which would not have been trading at large losses.
If you had instead purchased VBAL’s underlying ETFs in similar proportions, there would have been many more tax-loss selling opportunities to take advantage of throughout the year (and no need to switch back to the original ETFs). There were 33 days during 2018 where VCN would have been trading at a loss that breached our tax-loss selling thresholds. VUN also had a few days when it qualified for tax-loss selling. Combined, VIU and VEE spent an impressive 84 days in the proverbial danger zone.
2018 Tax-Loss Selling Opportunities (Number of Days)
Source: Vanguard Investments Canada Inc. (2018), FTSE Russell Indices as of December 31, 2018
On any of these days, you could have taken action to realize some capital losses. This would have been especially useful if you had already realized substantial capital gains when transitioning your $1 million portfolio to low-cost ETFs.
So, if you happen to be in a similar situation, and you don’t mind the added complexity, it could make sense to consider purchasing a handful of ETFs instead of an asset allocation ETF in your taxable account. It would increase your chances of stumbling across tax-loss selling opportunities to offset the capital gains realized in your portfolio transition.
That said, the hassle factor is worth calculating into your costs. For many, a simple asset allocation ETF may still be your preferred route.