Author Topic: Preferred Shares in Canada, are they worth it for someone who is FIRE?  (Read 1902 times)

Free Forever

  • 5 O'Clock Shadow
  • *
  • Posts: 23
I'll use the S&P/TSX PREFERRED SHARE INDEX (CAD) as my go to index of the entire Canadian preferred share market. It's useful because there is at least one ETF (iShares CPD) which tracks it.

Depending on your tax residence this analysis may apply to non-Canadian Preferred shares  in non-Canadian portfolios but that would be coincindental.

Preferred shares (as defined above) in a Canadian portfolio have the following 2 big benefits:

1. They offer diversification because they're not the same asset class as stocks and bonds (sort of a hybrid). In Canada most preferreds (about 75%) are of the rate-reset variety so the market value of an index of them would be expected to roughly go up and down with interest rates as opposed to bonds prices which generally move in the opposite direction of interest rates. In addition they have sginificantly safer dividends than common stock so there's an income factor there, although more variable and risky than bond income.

2. Distributions are taxed favourably due to the Dividend tax credit. During the withdrawal stage of FIRE I think most Canadian individuals will be withdrawing between 20-40K a year of which maybe half or more could be coming out of their RRSP and taxable accounts. Preferred shares are eligible for the dividend tax credit which for people with low taxable income means they often get to keep all of the dividend. Early retiree preferred share investors could very well end up with a negative marginal tax rate which they can apply to reduce taxes on other taxable income (like from your RRSP or savings account interest). I think this is one of the main reasons that preferred shares should get serious consideration from Canadians in the withdrawal stage of FIRE, especially if you've got some wiggle room for the variable nature of income coming out of a preferred shares index.

I also see some negative aspects as follows:

1. The main ETF that tracks the Canadian preferred market charges fee's of around .5% which is pretty high compared to an ETF charging around a tenth of that to track the Canadian common stock market.

2. Preferred share issues often have a call and/or expiry clause built into them allowing the issuer to get the preferred shares back based on certain terms and conditions (which are known in advance and therefore baked into prices according to the EMH). This means that characteristics of a portfolio of all preferred shares will change over time as shares are called back (or expire) and get re-issued under different terms.

3. There's no reason to expect the government to maintain the Canadian dividend tax credit. They could kill it or reduce it which would automatically cause a negative effect on the prices of preferred shares. The same argument could be made for cap gains, interest, rents etc. but it still is something to consider and a good reason for diversification in investment grade asset classes.

In a nutshell Canadian preferred shares (the total market) seem like a potentially good diversifier especially in the taxable portion of a Canadian portfolio for someone who has achieved FIRE.
« Last Edit: March 11, 2017, 02:24:44 PM by Free Forever »

scottish

  • Handlebar Stache
  • *****
  • Posts: 1435
  • Location: Ottawa
I'm also unsure.   Did you find the correlation coefficient between the preferred share index and the TSX composite somewhere?    This would help to understand the degree of diversification.

The dividend tax credit is appealing though.

Free Forever

  • 5 O'Clock Shadow
  • *
  • Posts: 23
I'm also unsure.   Did you find the correlation coefficient between the preferred share index and the TSX composite somewhere?    This would help to understand the degree of diversification.

The dividend tax credit is appealing though.

Good question. I don't think a correlation coefficient is going to be very useful because I don't think MVO analysis is very useful at forward looking analysis due to the extreme sensitivity it has to changes in input values. However I think it is a very useful theoretical concept and the long term past results that model outputs are useful in showing how diversification works and provides better risk adjusted returns. I think you can just overlay the total returns of the preferred shares index over a Canadian aggregate bond index and a Canadian stock index to see that they aren't perfectly co-related. Another problem lies in after tax returns, and it's difficult to see how the diversification effects change from an MVO perspective. MVO analysis almost always assumes pre-tax returns which I think is also a somewhat fatal flaw unless you adjust for that.

When it comes to actual allocations that's an implementation decision that's going to be based on everyones personal situation. I personally allocate about half my fixed income allocation (I consider it to be more an income asset then a growth asset like stocks) but others may want less or more more depending on their preferences. I know that's a very miasmic answer but it's the best I can come up with.
« Last Edit: March 11, 2017, 03:18:29 PM by Free Forever »

Al1961

  • Stubble
  • **
  • Posts: 225
  • Age: 58
  • Location: Alberta - B.C.
  • Dad of a husky Husky
For correlations as of 2005 to 2014, see table 3, page 6 of this:  https://www.pwlcapital.com/pwl/media/pwl-media/PDF-files/White-Papers/2015-02-20_PWL_Kerzerho-Bortolotti_The-Role-of-Preferred-Shares-In-Your-Portfolio_Hyperlinked.pdf?ext=.pdf

I admit that I like preferreds as part of my fixed income allocation. Especially given the poor correlation with bonds. I hold a preferred index etf, not individual preferred share issues.

Canadian Couch Potato doesn't really like preferred shares all that much. http://canadiancouchpotato.com/2015/03/17/does-your-portfolio-need-preferred-shares/ I hold more than the upper limit CCP recommends. But then, I also have a 60/40 fixed income to equity allocation because I do not need, or have much appetite for equity risk and volatility these days (pensions cover our base budget now, and, ten years from now, our entire budget).


Free Forever

  • 5 O'Clock Shadow
  • *
  • Posts: 23
@Al1961

I've seen those studies and the ZPR ETF (based on an index of Canadian rate reset preferred) they recommend is compelling to me and I think it should also be given consideration by investors.

I do have a few qualms with their analysis.

1. The PWL risk analysis is mostly pre-tax. A post tax analysis (maybe based on an 'average' Canadian) based on all assets might actually show the opposite results (bonds having worse risk adjusted returns) due to the tax credit on dividends and the cap gains treatment for stocks. It's even more complicated when you get different tax shelters (RRSP, TFSA, Taxable) involved.

2. They are able to show that the assets don't co-relate but it's hard to tell how much that co-relation will change going forward and how that will affect the diversification benefit. I personally invest a slightly higher proportion then what they recommend. I do fully agree with their recommendation to avoid investing in individual preferred shares, I use the CPD ETF because it's about as diverse as I can get in the Canadian market but ZPR also seems quite reasonable.

3. Some people care far more about the variability in after tax income then they do about the variability in the prices of their securities. This is particularly the case when they're in withdrawal mode and not prone to selling off much of the portfolio compared to taking income. So a drawdown scenario, after tax, showing results with bonds only and with various preferred share allocations would be useful as well.




scottish

  • Handlebar Stache
  • *****
  • Posts: 1435
  • Location: Ottawa
The low correlations in the PWL report look promising.   I'm pretty sure I've seen this before.   Every couple of years I go and look at preferred shares as a potential asset class.     Right now I have a pretty high income, so holding preferred shares in a cash account isn't very appealing because I pay lots of tax on dividends.    Once I stop working it'll be another story.

FF, I'm not sure why you think the correlation would change going forward?   In the efficient frontier, Bernstein was commenting on trying to make asset allocation decisions based on the statistics of the different asset classes and he found that his results were very sensitive to initial conditions.

But we're just looking to see if the asset classes behave in a correlated manner.    I don't think we're trying to decide how much of the portfolio should be in each asset class based on the statistics?

Free Forever

  • 5 O'Clock Shadow
  • *
  • Posts: 23

But we're just looking to see if the asset classes behave in a correlated manner.    I don't think we're trying to decide how much of the portfolio should be in each asset class based on the statistics?

You're right, sorry for the confusion.

RichMoose

  • Pencil Stache
  • ****
  • Posts: 966
  • Location: Alberta
  • RiskManagement
    • The Rich Moose | A Better Canadian Finance Blog
In retirement only I personally like some Canadian preferreds. In my view, they hedge against interest rate risk that is a big factor in bond returns.

Normally, we Mustachians would invest in a Universe Bond Index, so bonds can range from short to long term, governments to corporates. Basically everything except for high yields. While shorts are not really impacted by rates so much, mids and longs are.

Preferreds, especially rate resets, tend to move opposite of bonds when interest rates change (like you pointed out). By mixing preferreds into your normal bond allocation, you not only get great tax advantaged income that can offset fully taxed interest income, you also reduce fluctuations in the underlying values. This is important when you are no longer contributing to your investments.

That being said, I wouldn't go nuts on them. They don't carry near the same protection in downturns that bonds do. For example, in 08-09, diversified bonds remained pretty much flat. Preferreds (measured by CPD.TO) dropped nearly 35%, almost as much as stocks! But, maybe more importantly, their distributions stayed flat from 2008-2010, with the dividend portion dropping a bit in 2009 before returning in 2010.