This post is part of a series called Under the Hood, where l take a detailed look at specific Canadian ETFs or index funds.
The fund: Vanguard FTSE Canadian Capped REIT (VRE)
The index: The fund tracks the FTSE Canada All Cap Real Estate Capped 25% Index, which includes large, mid and small-cap companies in the Canadian real estate industry as defined by FTSE. The index is weighted by market cap with a limit of 25% on any single company. It currently has 19 holdings.
The cost: The management fee is 0.35%. Because the fund is less than a year old it has not published its full MER, but expect it to be about 0.40% after adding taxes and incidentals.
The details: Vanguard launched VRE last November and continued its tradition of being a cost leader: its management fee is about 20 basis points lower than its competitors.
VRE is not limited to REITs: some of its holdings are developers and real estate services companies that are not set up as income trusts. But even with this expanded definition, the real estate industry in Canada is extremely small. That makes concentration risk a problem in a cap-weighted fund, as a small number of companies will dominate. VRE’s top holding is the giant RioCan at more than 16%—more than the bottom eight holdings combined—followed by Brookfield Office Properties at almost 13%.
The ETF pays monthly distributions, and the most recent was about $0.04 per share, which works out to an annual yield of less than 2%. That’s unusually low for a REIT fund, so you should expect it to change once the fund has a longer history and its cash flows are smoothed out.
The alternatives: VRE is the newest of three real estate ETFs in Canada. The iShares S&P/TSX Capped REIT (XRE), which is also cap-weighted, is the leader with almost $1.4 billion in assets. The BMO Equal Weight REITs (ZRE) has about $375 million under management. Both ETFs have management fees of 0.55%.
A detailed look at the holdings of these three alternatives turns out to be very revealing:
| VRE | XRE | ZRE | |
| RioCan REIT | 16.3% | 18.9% | 5.0% |
| Brookfield Office Properties Inc. | 12.7 | – | 5.2 |
| H&R REIT | 9.4 | 10.9 | 5.0 |
| Dundee REIT | 7.2 | 8.7 | 5.2 |
| Calloway REIT | 6.2 | 6.8 | 5.2 |
| Boardwalk REIT | 6.2 | 6.1 | 5.1 |
| Canadian REIT | 5.9 | 7.3 | 5.3 |
| Cominar REIT | 5.6 | 6.1 | 5.1 |
| Primaris Retail REIT | 5.2 | 6.3 | 5.2 |
| Canadian Apartment Properties REIT | 5.1 | 6.0 | 5.3 |
| Allied Properties REIT | 4.0 | 5.3 | 5.4 |
| Chartwell Seniors Housing REIT | 3.7 | 4.4 | 5.2 |
| Artis REIT | 3.4 | 4.2 | 5.2 |
| First Capital Realty Inc. | 3.1 | – | – |
| Morguard REIT | 1.7 | – | 5.5 |
| FirstService Corp. | 1.6 | – | – |
| Extendicare Inc. (US) | 1.3 | – | – |
| InnVest REIT | 0.8 | – | 5.5 |
| Melcor Developments Ltd. | 0.4 | – | – |
| Granite REIT | – | 4.3 | – |
| Northern Property REIT | – | 2.4 | 5.2 |
| Crombie REIT | – | 1.8 | 5.1 |
| Pure Indstrial REIT | – | – | 5.7 |
Notice that if you lop off the top and bottom of the list, the holdings are not meaningfully different: the 10 REITs I’ve marked in blue have similar weightings and together make up approximately half of each fund. The truly important difference is in how each fund weights RioCan and Brookfield. The weight assigned to the smaller REITs also varies considerably: VRE includes several very small holdings that will have minimal influence on the fund’s overall performance.
What effect will these differences have? A lot, as it turns out. With a track record of just three months, there is no meaningful performance data for VRE, but FTSE has published index returns for the last five calendar years. Here they are, along with the performance of the indexes tracked by XRE and ZRE:
| 2008 | 2009 | 2010 | 2011 | 2012 | 5 Yr. | |
| FTSE Canada All Cap Real Estate | -42.8 | 46.5 | 31.6 | 8.0 | 16.4 | 6.8 |
| S&P/TSX Capped REIT | -38.3 | 55.3 | 22.6 | 21.7 | 17.0 | 10.8 |
| Dow Jones Cda Select Eq Weight REIT | -30.7 | 59.1 | 30.7 | 14.5 | 18.9 | 14.4 |
Notice the FTSE index returned just 8% in 2011, dramatically underperforming the other two. I don’t have access to the list of holdings from 2011, but it seems likely this discrepancy was driven by RioCan and Brookfield: the former gained over 23% in 2011, while the latter lost more than 8%. The S&P index has an even greater allocation to RioCan, and it does not include Brookfield at all, which worked out well that year. Meanwhile, the Dow Jones index holds equal amounts of each, so its performance fell between of the other two.
[Update: After this post appeared, Vanguard replied: “You’re correct in your reasoning for the relative underperformance in 2011. Until September 2011, non-REIT Real Estate holdings represented about 41% of the index, with one constituent (Brookfield Asset Management) capped at 25% itself. However, Brookfield (which was down 14% in 2011) was removed from the index at the September 2011 reconstitution as it was reclassified as a Financial Services company.” Please note Brookfield Asset Management and Brookfield Office Properties are distinct companies, and the latter remains in the FTSE index.]
Bottom line: One of the goals of passive investing is to capture the returns of entire asset class by minimizing the influence of any single security. Real estate has a place in a diversified portfolio because it has historically shown low correlation with equities, not because you want exposure to a couple of specific companies. Cap-weighted indexes are an efficient tool for broad equity markets, but not so for a tiny sector like Canadian real estate.
VRE is the cost leader in this category, but I feel it suffers from the same flaw as XRE: too much concentration.
@CCP
Could you comment on “Are REITs a Distinct Asset Class?” by Jared Kizer & Sean Grover (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2965146)? They suggest that “real estate has a place in a diversified portfolio because it has historically shown low correlation with equities” doesn’t actually translate to better diversification.
@CCP This is largely an academic question for me, i.e., more to reflect on the underlying considerations of what constitutes a distinct asset class and what makes a low correlation with another class likely (or not) to persist over time. But could you give one of your usual well-thought out opinions on this? Or is the answer as simple as “Inclusion of REIT is an attempt at Smart Investing, and Smart Investing is too clever for its own good.”
@oldie: For me, this is primarily one of those trade-offs between complexity and simplicity. It’s true that REITs have a relatively low correlation to equities, and so adding an allocation to a portfolio might offer some modest diversification benefit. But most would agree that an allocation to REITs should not be more than 5% to 10%, so now you have added another (small) moving part that needs to rebalanced. And in the case of Canadian REITs, you’re adding a lot of concentration in small number of companies. REIT ETFs are also significantly more expensive than plain-vanilla ETFs, and not very tax-efficient. Taken together, I think these cons significantly outweigh the pros.
@CCP: It occurs to me that of the underlying shares that make up the XIC basket, the shares that make up the VRE basket are already in there, and in roughly the same ratios to each other as in the VRE basket. So we already have bought VRE, except in a much smaller percent of our XIC basket than 5%. Am I correct on this?
So, rather than buy VRE directly as 5% of our total portfolio, and accepting all the higher costs and tax penalties etc., suppose instead we take the plain vanilla Single fund XGRO or VGRO route, to make a simplifying 80% equity allocation assumption, we are actually already getting this smaller percentage Canadian REIT allocation built into the Single fund XGRO or VGRO; and, in fact, if we wanted to keep track of whatever US domiciled REIT shares and Rest-of-World REIT shares that are also in there somewhere, we could drill down and find them and add those percentages to our above small percentage of Canadian REIT to get an estimate of what total world REIT equivalent we are holding as a percentage of our total Equity. Does this make any sense?
@CCP: In trying to find a meaningful lesson in reviewing the composition of VRE over 9 years of benign neglect I find that the Largest Component, RioCanREIT has gone from 16.3% to 10.3% and now sits at #3. Brookfield Office Properties has moved from #2 @12.7% to disappearing completely off the list, and H&R REIT has moved from #3 @9.4% to #7 @ 6.5%. The new Top 3 are now #1: Canadian Apartment Properties REIT at 14.7% (was #9 @ 5.1%), #2: First Services Corporation @ 14.3% (was #16 @1.6%), and #3: our old friend RioCanREIT as noted above.
Perhaps I’m exhibiting an unseemly interest in a niche product that is not generally recommended for sensible Couch Potatoes and might be distracting from the core message (Plan and Act Early, Diversify, Simplify) that you have tried to hone in on over these years, for which I apologize. But, does the seemingly unpredictable fortunes of the top 3 in VRE of 2013, and the surprising (to me) small origins of the current top 2 give us any lessons going forward? Does it indicate the volatility of such a niche market with few components of large proportions each? Or does it merely show us that the seeming top-heaviness of VRE is meaningless and will likely self correct harmlessly over time? And does this merciless pruning of the top few in VRE over time give us any cautionary lessons for even the “safe” equity sub-components of our XCNS/XBAL/XGRO portfolios (i.e. the top majority of the total markets of TSX, NYSE, Rest of world, etc) ?