Research: The Bond Portfolio Jigsaw Puzzle

Bond portfolios should be constructed from the top-down, adding pieces with useful characteristics relative to the rest of the portfolio as the dealers and issuers make them available.

Look for the research link!

8 Responses to “Research: The Bond Portfolio Jigsaw Puzzle”

  1. prefhound says:

    I always enjoy your articles, but am a bit surprised this one did not discuss the impact of taxes which are absolutely critical to bond portfolios.

    In an RRSP or RRIF, buying a premium bond for $120 is fine, but in a taxable account, the after tax yield is hurt because coupon income is taxed at income rates while the capital loss when the bond matures (or is sold) is only deductible to the extent of capital gains of the past 3 years and then only at a 50% inclusion rate.

    This makes premium bond after-tax returns much lower than par bonds. Indeed, some Canada bonds have negative after-tax returns. Taxable investors should prefer to buy discount bonds — but there are none these days. A strip bond attracts the same tax as a par bond but provides no interim cash flow and is very difficult to find for corporates. GICs are equivalent to par bonds, but sometimes the best rates require compounding and locking in.

    These are all arguments for working hard to find the right issues, and waiting to find them, as you say. However the “best” portfolio will be quite different for a taxable vs sheltered account. Furthermore, optimization trades will be quite different in the two accounts — a taxable account might sell a bond that rises to a premium after purchase in order to invest in something else with higher after tax return and give them “accelerated” return taxable as capital gain.

    Even ETFs have the same problem as bonds so taxable investors should be careful there as well. If the weighted average ETF bond holding is at a premium, taxable investors will earn a lower after-tax return because they should expect to sell the ETF at a lower price than purchased. Effectively some of “their” principal is being returned as taxable income while the premium erodes over time. This factor means we cannot use the ETF current yield to measure income potential, but should focus on the ETF yield to maturity (which is currently lower due to premium bonds).

    For investors without capital gains to offset losses (and unfortunately there are way too many of those), the situation is even worse — they have to wait years (or until death) to get the tax benefit of the capital loss.

    So there are really 3 portfolios: non-taxable, taxable with capital loss usability and taxable without capital loss usability. All have different after tax returns and hence different optimized portfolios. The latter two classes of investor may be well served by the simplicity of GIC investing — at least up to 5-10 years.

    Even advisors are not aware of all of these wrinkles, but the general rule of thinking about after-tax return still applies. There are no Excel functions for After-Tax Yield (although it is not hard to apply tax rates to the Yield () Function — just make sure to include the capital gain/loss tax effect).

  2. jiHymas says:

    Your point is well taken, but I only had 1750 words to play with! My main purpose in the article was to construct a sensible bond portfolio for a couple with the described situation that illustrated the idea that there is more to life than a 5-year GIC ladder and that buying a long-bond is not precisely equivalent to speculating on interest rates.

    I did refer to the fact that the difference in coupons between new issues and old will alter the risk/return characteristics.

    More insidious than taxation is the default risk of a premium bond: should Bad Things happen to the issuer, recovery will (except in very rare situations) be based entirely on par value, with no consideration for coupon.

  3. prefhound says:

    There were 491 words in my comment, which covered a lot of bases. Maybe we should write an article together showing how the 3 types of investor would choose among a given “universe” of (say) 10-15 investments.

    It would only be worthwhile if the portfolios were quite different, which I suspect, but have not proven.

  4. jiHymas says:

    Maybe we should write an article together

    You don’t need me! You’re perfectly capable of putting together 1,750 words on Premium Bonds in Portfolios.

  5. lyndon says:

    Prefhound “A day that will live in infamy, James Hymas lost for words!!!!.” That was an excellent post re building a bond portfolio, keep posting stuff like this, I challenge you to write the article yourself, I need the education/advice!!
    Lyndon

  6. pugwash says:

    Agree with lyndon – this is really an excellent post and well worth polishing up as an article.

    My take away is that during times of premium corporate bond pricing – $500k fixed income in a ‘taxable without capital loss usability’ account could be best structured as:

    30% 1-5 yr GIC ladder
    40% 6-15 yr Provincial/Corporate Strips A-AA
    30% Preferred Perpetuals PF1

    does this make sense or would a short term bond ETF have a role replacing 1- 3 year GICs?

    puggy

  7. jiHymas says:

    does this make sense

    The proposed allocation certainly has some attractive points, but is not particularly liquid. I claim that much of the point of holding short term (<5year) fixed income in the first place is to meet unexpected or uncertain cash needs - a feature that is lost with GICs.

    Liquidity is also restricted with strips and can be uncertain with Prefs. You would not have wanted to be a forced seller of Prefs in November 2008!

    I suggest that all investors are best off if they sit down and make a list of what they want their fixed income portfolio to do for them - then it may be implemented according to what's available.

  8. […] was referring to a comment by prefhound on my essay The Bond Portfolio Jigsaw […]

Leave a Reply

You must be logged in to post a comment.