It’s been a marvellous year for equities, but 2013 has not been kind to real estate investment trusts. Like other income-producing investments, REITs are sensitive to rising interest rates, and the sharp increase in the middle of this year hit them hard. But the three Canadian ETFs in this asset class have shown significant differences in performance. Notably, the Vanguard FTSE Canadian Capped REIT (VRE) has outperformed its iShares and BMO rivals by a wide margin. Here are the year-to-date returns of the funds as of December 18, according to Morningstar:

Vanguard FTSE Canadian Capped REIT (VRE)–2.47%
iShares S&P/TSX Capped REIT (XRE)–7.81%
BMO Equal Weight REITs (ZRE)–6.72%

REISs pieces

Whenever you see variance among funds in the same asset class, it’s important to determine why. In this case, the main reason is a significant difference in the underlying indexes. Shortly after VRE was launched, I wrote a detailed post describing its benchmark, the FTSE Canada All Cap Real Estate Capped 25% Index, which is quite different from those tracked by the iShares and BMO funds.

Like XRE, the Vanguard ETF is cap-weighted, so it is quite concentrated in the country’s largest REITs: RioCan and H&R. However, unlike either of its competitors, VRE also includes real estate investment and services companies (REISs), which make up 18.5% of the fund. The largest, Brookfield Office Properties, has increased more than 25% this year, while the fund’s small holding in FirstService Corporation is up over 60%. These gave a huge boost to VRE.

Higher return, lower yield

Despite VRE’s higher total return, the ETF’s yield is inconsistent and significantly lower than that of its competitors. Its monthly payouts in 2013 have ranged from about $0.01 to $0.07 per unit, with a 12-month trailing yield of 2.16% (according to Morningstar), while XRE and ZRE have both sported yields over 5%.

Vanguard’s David Hoffman generously provided an explanation. There are actually a few things going on here. The first is simply that some of the underlying holdings pay smaller distributions. “The REISs in the portfolio provide a lower yield than the remaining REITs,” Hoffman explains. Brookfield Office Properties, for example, has a 12-month yield of 2.82%, about half of what RioCan paid out. FirstService yielded 0.47%.

The rest of the story is a more complicated. Mutual fund trusts (including ETFs) are required to distribute their income to unitholders in the year it is received, but not necessarily in the same month. Many funds hold back some of the income they receive in an effort to smooth out the monthly distributions. They may even top up the distributions with some return of capital when necessary. XRE, for example, has paid out about $0.06 per unit every month for over three years. Meanwhile, ZRE has paid exactly $0.081 for the last 15 months.

Vanguard doesn’t do that, Hoffman says. “We generally pay the income as it is earned, so at each distribution period income received since the prior distribution period is paid. This in part explains the inconsistency in monthly income per share for VRE, as the underlying securities do not pay concurrently. Three securities pay quarterly dividends, while the remaining pay monthly. The quarterly dividends are typically reflected in the distributions of the months following the calendar quarter, thus increasing the distribution rate of the fund in January, April, July and October if all else is equal.”

They also do not add return of capital to the distributions. “Vanguard currently does not generate ROC to artificially maintain stable or index-equivalent yields. We have heard a variety of opinions from individual investors, financial advisors, and analysts regarding the preference for yield and ROC considerations, with no clear consensus.”

For what it’s worth, I tend to agree with Vanguard’s position. If you rely on the income from your REIT investments for spending money, then some return of capital would be welcome. But if you’re investing for long-term growth (and that includes anyone holding REITs in an RRSP), the additional yield is a nuisance that simply results in idle cash building up in your account.

Go with the flow

The other reason for VRE’s relatively low yield is probably temporary. When a fund sees large inflows shortly before a distribution date, the amount paid out to unitholders can be diluted. This can affect any investment fund, but the impact is likely to be highest when a new fund is small but growing rapidly. And as Hoffman says, “VRE has grown 470% in the 12 months ending November 2013, most of which is due to cash flow.”

To understand this idea, imagine an ETF has $10 million in assets, consisting of 500,000 units with a net asset value of $20 each. Now assume in a given month the fund receives $20,000 in dividends, which increases the NAV to $20.04. On the next distribution date, investors in the fund would expect a distribution of $0.04 per unit, after which the unit price should fall back to $20.

But let’s say that before the distribution date the ETF sees a huge inflow of new money, which leads to the creation of another 500,000 new units. Now the fund must distribute its $20,000 in dividends among twice as many shares, resulting in a payout of only $0.02 per share.

Before you cry foul, understand that your total return is not affected. “With a lower distribution, there is a correspondingly lower reduction in the NAV,” Hoffman explains, “so in this context investor returns are not harmed.” In the example above, the NAV would fall to just $20.02 after the two-cent distribution. The end result for investors is the same as it would have been without the inflow of new money.

My thanks to David Hoffman and Vanguard for explaining these ideas. Even if you don’t hold real estate in your portfolio, there are valuable lessons here for anyone making ETF comparisons.