PrefBlog

Malachite Aggressive Preferred Fund’s Net Asset Value per Unit as of the close February 26, 2021, was $9.4553.

Five issues held throughout the month returned between 0% and 1%: BMO.PR.Q, CU.PR.G, BNS.PR.H, FTN.PR.A and INE.PR.A, with a total weight in the portfolio (at month-end) of 14.1%. Star performers were SLF.PR.G, TRP.PR.B and IFC.PR.A, returning between 15% and 20%, with a total portfolio weight of 8.2%.

Insurance issues received a boost this month with the announcement of Manulife LRCNs (additional commentary February 11.

Quote quality was poor this month, with the difference in portfolio values when calculated with closing prices vs. calculation with bid prices increasing from 0.52% to 1.19%.

MAPF returns assume reinvestment of dividends, and are shown after expenses but before fees. Past performance is not a guarantee of future performance. You can lose money investing in Malachite Aggressive Preferred Fund or any other fund. For more information, see the fund’s main page. The fund is available directly from Hymas Investment Management.

The preferred share market continues to be underpriced relative to other capital markets, leaving a lot of room for outsized gains. The Seniority Spread (the interest-equivalent yield on reasonably liquid, investment-grade PerpetualDiscounts less the yield on long term corporate bonds) is extremely elevated (chart end-date 2021-2-12):

Note that the Seniority Spread was recorded at 330bp just before month-end, significantly narrower than the 355bp on February 3. As a good practical example of the spreads between markets, consider that CIU issued a long-term bond in early September, 2019 yielding 2.963%, about 411bp cheaper than the interest-equivalent figure of 7.07% for CIU.PR.A, which was then yielding about 5.44% as a dividend. CIU issued another bond in late September, 2020, yielding 2.609%, which was 399bp cheaper than the interest-equivalent figure of 6.60% for CIU.

… and the relationship between five-year Canada yields and yields on investment-grade FixedResets is still deep in ‘decoupled panic’ territory (chart end-date 2021-2-12):

In addition, I feel that the yield on five-year Canadas is unsustainably low (it should be the inflation rate plus an increment of … 1%? 1.5%? 2.0%?),and a return to sustainable levels is likely over the medium term.

It seems clear that many market players are, wittingly or not, using FixedResets to speculate on future moves in the Canada 5-Year yield. This is excellent news for those who take market action based on fundamentals and the long term characteristics of the market because nobody can consistently time the markets. The speculators will, over the long run and in aggregate, lose money, handing it over to more sober investors.

In this connection, it is useful to review some recent commentary regarding the preferred share market. Rob Carrick recently published a column with the title Rob Carrick: Does it make sense to use preferred shares as a bond substitute?, in which he gave a nod to the old shibboleth:

But does it truly make sense to substitute prefs for bonds? Not if you subscribe to the idea that the main point of bonds is not to generate income and returns, but rather to act as a portfolio life preserver when the stock market turns ugly.

… but then noted the crucial point:

Perpetuals are somewhat less touchy, but they’re not to be counted on in a stock market crash.

This distinction is meaningless to investors who are all about income and fine with portfolios that bounce around in price.

As I never get tired of repeating, Fixed Income portfolio management is a process that largely consists of balancing Security of Income vs. Security of Principal. These are opposing forces; the more you have of one, the less you’ll have of the other. And preferred shares are way, way over to the “security of income” pole. So in bad times … Shut up and Clip Your Coupons!

Financial Wisdom Forum’s long-running thread on preferred shares recently featured a series of posts which often deprecated preferred shares with some participants advocating market timing:

The thing about prefs is that timing is everything. You buy them when they’re being thrown out with the bathwater and you’ll be golden.
…
Yes, there’s money to be made, but only if the shares are in the order of a 30% discount.
…
If you care about total return, preferred shares probably aren’t for you, (unless of course you are interested in trading them or you have a crystal ball regarding interest rates).

If you want something that pays a decent, tax efficient dividend, in a non-registered account, and share price isn’t important, then straight perpetual preferred shares are probably fine (caveat for a few years with respect to NVCC – will no longer qualify as Tier 1 Capital post 2022). Suitable for the retired. Anything else – not so much. They all seem to favour the issuer.
…
Sure the market value is volatile but I don’t plan on selling so it doesn’t matter to me, barring default.

The part I’d like to focus on is:

Yes, there’s money to be made, but only if the shares are in the order of a 30% discount.

This is close – but not quite – to stating an essential point about Fixed Income investing: total returns are asymmetrical: you can lose a lot more than you can win. This is why credit quality is so important – a bankruptcy, or even a mere impairment, is going to take you a long way down the ‘bad side’ of the asymmetrical probability curve. But another point, often ignored, is that this implies that discounted issues are, all else being equal, better investments than premium issues: in good times, discounted issues can win a lot more than premium issues, which are tethered to their par value due to considerations of maturity or possible redemption. In bad times, they’ll lose a little more, since discounted issues have no ‘buffer’ to insulate them, while premium issues (and, to a lesser extent, issues trading modestly below par) have such a buffer that absorbs a portion of the shock – this is easier to understand in terms option values and volatility theory, as the negative value of the embedded option is reduced as the issue moves further away from par.

The implication for MAPF is that there is a bias towards holding issues trading well away from par value, which almost always means discounted issues. It’s only a bias, not a hard and fast rule, but this bias was deliberately built into the system in order to reduce the asymmetry of the projected returns curve. This means that MAPF will usually hold a portfolio more heavily weighted towards low coupon – low priced issues than the index and be correspondingly underweighted in high coupon – high priced instruments.

This in turn has the effect of introducing a bias in returns: MAPF will often underperform in poor markets (as it holds fewer of the better performing high-coupon issues than the index) and outperform in good markets (as it holds fewer of the worse performing high-coupon issues than the index). This is only a tendency, not an iron-clad rule, but the effect is there and it means that MAPF has a higher Beta than it might otherwise have, if you enjoy thinking in terms of investment models from the last century. And at all times, of course, the fund seeks to trade and exploit market inefficiencies, which mitigates but does not eliminate the effect.

On another FWF forum, there has been speculation that the fine performance of FixedResets since March, 2020, has been due to a conscious focus on spreads:

ZPR was up approximately 6% for 2020. Is this a reversal of the trend to follow interest rates downward? Perhaps investors will focus on the spread between preferred versus bond yields going forward rather than an irrational rush to the exits whenever prefs are resetting at lower rates.

[reply] It’s already happening. That’s why ZPR has finally had a positive year in a declining rate environment.

Well, this is something I would like to agree with. I’ve been pounding the ‘spreads’ drum for a long time…

Unfortunately, while one can only rarely point to a single mechanism explaining a change in relative prices and say, with any credibility whatsoever, that A caused B, I have to admit I’m more dubious than usual about this claim. I believe that the continued rally is due to continued interest rate anticipation, which is now (for some investors) considered to be on the way up rather than continuing the downward staggering of the past ten years; this in turn is derived from Central Bank ‘easy money’ policies and very loose government fiscal policies; which is considered to be on the verge of driving inflation upwards.

Who’s right? Well, we’ll know in ten years, at a time when half of the investing world will graciously explain at length how their uncanny ability to read global market forces allowed them to time the market and make enormous profits, while the other half will tell you the question is irrelevant because investing is about the future, not the past. My advice is, as always, to make asset allocation decisions based on the long-term characteristics of each asset class and how these characteristics interact with your long-term portfolio objectives.

It should be noted that I have been unable to provide an entirely satisfactory explanation for the relatively strong performance of Floor issues during the 2018-19 downdraft relative to their non-Floor counterparts. See the discussions on PrefBlog at LINK, LINK and LINK.

I believe the bear-market outperformance by the Floor issues is a behavioural phenomenon with very little basis in fundamentals. When interest rates in general move, FixedReset prices should not change much (to a first approximation, for issues priced near par), since in Fixed Income investing it is spreads that are important, not absolute yields. There should be some effect on Floor issues, which should move up slightly in price as yields go down since the ‘option’ to receive the floor rate will become more valuable. Adjustments due to this effect should be fairly small, however – and over the past year issues with a floor, that started the period being expensive, have simply gotten even more expensive, relative to their non-floored counterparts.

And the tricky thing about behavioural models of investing is that they can lose their explanatory power very quickly when an investment fashion shifts, whereas fundamentals will always be effective – sometimes it just takes a little time! Just to give an example from the preferred share market – until the end of 2014, FixedResets were priced relative to each other according to their initial dividend; when the reset of TRP.PR.A shocked a lot of investors, relative pricing became much more dependent upon the Issue Reset Spread, a much more logical and fundamental property. This paradigm shift was discussed extensively in PrefLetter.

FixedReset (Discount) performance on the month was +6.50% vs. PerpetualDiscounts of +0.84% in February; the two classes finally decoupled in mid-November, 2018, after months of moving in lockstep:

Floaters performed well, returning +10.03% for February and finally moving the trailing twelve months into positive territory at +17.65%. Look at the long-term performance:

Some Assiduous Readers will be interested to observe that the ‘Quantitative Easing’ decline was not initially as bad as the ‘Credit Crunch’ decline, which took the sector down to the point where the 15-year cumulative total return was negative. I wrote about that at the time. but it became worse in August, 2019! On August 30, 2019 the HIMI Floater Index (total return) value was calculated as 1906.6; the index first surpassed this value on 2003-8-13. Thus, cumulative total return (that is, including dividends) was negative over a period of slightly-over sixteen years. Worse, on March 31, 2020, the index level was 1454.8, a milestone first passed on 1997-7-30; a cumulative negative total return for 22 years and 8 months; at its low on March 18 the index level was 1253.7, first surpassed on 1996-1-4, a span of 24 years and over two months!

It seems clear that Floaters are used, wittingly or otherwise, as a vehicle for speculation on the policy rate and Canada Prime, while FixedResets are being used as a vehicle for speculation on the five-year Canada rate. In support of this idea, I present an Implied Volatility analysis of the TRP series of FixedResets as of February 26, which is comprised of six issues without a Minimum Rate Guarantee and two issues which do have this feature:

The two issues with floors, TRP.PR.J (+469, minimum 5.50%) and TRP.PR.K (+385, minimum 4.90%) are $0.48 cheap and $1.45 rich, respectively. These figures are much narrower than the 1.17 rich and 2.13 rich calculated with last month’s figures. The floors have become effective since five-year Canadas dipped below 0.81% and 1.05%, respectively – note that the floor on TRP.PR.J, which is no longer rich, is no longer assumed to be effective.

It will also be noted that the spread of a notional non-callable TRP FixedReset priced at par has narrowed from 376bp last month to 363bp this month, while GOC-5 has increased from 0.42% to 0.86%.

I also show results for the BAM series of FixedResets, which includes three issues with dividend floors: BAM.PF.H (+417, Minimum 5.00%); BAM.PF.I (+386, Minimum 4.80%); and BAM.PF.J (+310, Minimum 4.75%); these issues are all rich compared to their non-floor siblings, at 1.95, 1.96 and 1.94 respectively, very different from last month’s figures of cheap 1.29, cheap 1.79 and cheap 1.74. I note that it is normal for premium issues in a rising market to be cheap, since investors have an expectation for continued rising prices (and therefore an ‘automatic’ redemption), without considering the possibility of a market reversal, in which case the higher spreads will prove to be very useful.

It will also be noted that the spread of a notional non-callable BAM FixedReset priced at par has narrowed significantly over the month; 397bp last month to 354bp this month, while GOC-5 has increased from 0.42% to 0.86%.

Relative performance during the month was correlated (53%) with Issue Reset Spreads for the “Pfd-2 Group” and for the “Pfd-3 Group” issues (32%):

… with Pfd-2 Group correlation for the three-month period of 51% and 23% for Pfd-3:

This suggests to me that November commenced an actual theme, one based on an expectation of higher government interest rates in the future (perhaps due to renewed hopes that the introduction of coronavirus vaccines will return things to normal) and that this theme has continued through February, finally replacing the uncorrelated messes we’ve seen through much of the market’s recovery since the end of March, 2020.

As for the future, of course, it’s one thing to say that ‘spreads are unsustainable and so are government yields’ and it’s quite another to forecast just how and when a more economically sustainable environment will take effect. It could be years. The same caution applies for an end to the overpricing of issues with a minimum rate guarantee.

Yields on preferred shares of all stripes are high compared to those available from other investments of similar quality. As I told John Heinzl in an eMail interview in late November, 2018, the best advice I can offer investors remains Shut up and clip your coupons!

I think that a broad, sustainable rally in FixedResets will require higher five-year Canada yields (or a widespread expectation of them), since paradigm shifts generally require a trigger (a Wile E. Coyote moment, as they say!) … and although I’m sure this will happen eventually, it would be foolish to speculate on just when it will happen.

These calculations were performed assuming constant contemporary GOC-5 and 3-Month Bill rates, as follows:

I note that the sustainable yield calculated above is not directly comparable with any yield calculation currently reported by any other preferred share fund as far as I am aware. The Sustainable Yield depends on:
i) Calculating Yield-to-Worst for each instrument and using this yield for reporting purposes;
ii) Using the contemporary value of Five-Year Canadas to estimate dividends after reset for FixedResets. The assumption regarding the five-year Canada rate has become more important as the proportion of low-spread FixedResets in the portfolio has increased.
iii) Making the assumption that deeply discounted NVCC non-compliant issues from banks (and insurers, until November 2019), both Straight and FixedResets will be redeemed at par on their DeemedMaturity date as discussed above.

This entry was posted on Sunday, February 28th, 2021 at 1:58 pm and is filed under MAPF. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.