PrefBlog

The fund outperformed slightly in October

Click for Big

Click for Big

I continue to believe that the decline in the preferred share market remains overdone; the following table shows the increase in yields since May 22, 2013, of some fixed income sectors:

ZPR, is an ETF comprised of FixedResets and Floating Rate issues and a very high proportion of junk issues, returned +%, +% and +% over the past one-, three- and twelve-month periods, respectively (according to the fund’s data), versus returns for the TXPL index of +0.06%, +0.01% and +5.67% respectively. The fund has been able to attract assets of about $1,095-million since inception in November 2012; AUM increased by $21-million in October; given an index return of +0.06% an increase of less than $1-million was expected, indicating that money is still flowing into the fund. I feel that the flows into and out of this fund are very important in determining the performance of its constituents.

TXPR had returns over one- and three-months of +0.51% and +0.57%, respectively with CPD performance within expectations.

Returns for the HIMIPref™ investment grade sub-indices for October were as follows:

Malachite Aggressive Preferred Fund’s Net Asset Value per Unit as of the close October 31, 2014, was $10.5270.

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 either directly from Hymas Investment Management or through a brokerage account at Odlum Brown Limited.

A problem that has bedevilled the market over the past two years has been the OSFI decision not to grandfather Straight Perpetuals as Tier 1 bank capital, and their continued foot-dragging regarding a decision on insurer Straight Perpetuals has segmented the market to the point where trading has become much more difficult. The fund occasionally finds an attractive opportunity to trade between GWO issues, which have a good range of annual coupons (but in which trading is now hampered by the fact that the low-coupon issues are trading near par and are callable at par in the near term), but is “stuck” in the MFC and SLF issues, which have a much narrower range of coupon, while the IAG DeemedRetractibles are quite illiquid. Until the market became so grossly segmented, this was not so much of a problem – but now banks are not available to swap into (because they are so expensive) and non-regulated companies are likewise deprecated (because they are not DeemedRetractibles; they should not participate in the increase in value that will follow the OSFI decision I anticipate and, in addition, are analyzed as perpetuals). The fund’s portfolio is, in effect ‘locked in’ to the MFC & SLF issues due to projected gains from a future OSFI decision, to the detriment of trading gains particularly in May, 2013, when the three lowest-coupon SLF DeemedRetractibles (SLF.PR.C, SLF.PR.D and SLF.PR.E) were the worst performing DeemedRetractibles in the sub-index, and in June, 2013, when the insurance-issued DeemedRetractibles behaved like PerpetualDiscounts in a sharply negative market.

However, it will be noted, as discussed in the August report on Portfolio Composition that the month saw some swaps from the low-coupon SLF Straights to a low-spread SLF FixedReset … so there are some opportunities to trade, although they don’t happen often! There were similar swaps executed in June and July.

In October, insurance DeemedRetractibles outperformed bank DeemedRetractibles:

Click for Big

… and were about equal to Unregulated Straight Perpetuals.

Click for Big

Of the regressions shown in the above two charts, the Adjusted Correlation of the Bank DeemedRetractible performance is slightly negative, Straight Perpetuals come in at 4% and Insurance DeemedRetractibles are at 14%.

A lingering effect of the downdraft of 2013 has been the return of measurable Implied Volatility (all Implied Volatility calculations use bids from October 31):

Click for Big

Click for Big

However, while the fit for PWF is good and the Implied Volatility is high, there are many local minima for the spread:

Click for Big

Click for Big

It is disconcerting to see the difference between GWO and PWF; if anything, we would expect the implied volatility for GWO to be higher, given that the DeemedRetraction – not yet given significant credence by the market – implies a directionality in prices. On the other hand, the PWF issues are mostly trading above par, which in practice tends to add directionality although this makes no sense. The GWO data with the best fit derived for PWF is distinguishable from the best fit; the best fit has a lower Sum of Squared Errors (1.44 vs. 2.94):

Click for Big

In the September, 2013, edition of PrefLetter, I extended the theory of Implied Volatility to FixedResets – relating the option feature of the Issue Reset Spreads to a theoretical non-callable Market Spread.

Click for Big

Click for Big

These are very interesting results: The BPO issues are trading as if calls are a certainty, while FFH issues are trading as if calls are much less likely; this is probably due to the market’s over-reacting to the fact that all of the BPO issues are trading above par, while only one of the five FFH issues shares that happy status. The FFH series continues to be perplexing, this time with the four lower-coupon issues showing virtually no implied volatility – with the highest coupon issue (FFH.PR.K) being well off the mark … all I can think of is that the market has decided that FFH.PR.K, with an Issue Reset Spread of 351bp, is sure to be called in 2017, while the other four (highest spread is FFH.PR.C, +315) are not at all likely to be called. Note that FFH.PR.C will have its first Reset Date on 2014-12-31 and it would appear, given its bid of 24.62, that the market expects a reset rather than a call for redemption.

The Implied Volatility calculation for the TRP FixedResets is most interesting:

Click for Big

According to this calculation, TRP.PR.A is $1.26 cheap to theory, being bid at 21.74 compared to a theoretical price of 23.00. A portion of this difference is due to the approximations that have gone into the calculation, which assumes that all issues have three years to their call date and, critically, are all paying their long-term dividend rate right now. In an environment in which, given a GOC5 yield of 1.54%, virtually all dividend rates are expected to drop on reset, the time to reset and the degree of difference in the interim is important.

We can make some approximate adjustments to the theoretical prices:

So a more precise calculation could be performed by subtracting $0.07 from the actual bid of TRP.PR.A, since the calculation otherwise assumes the last dividend payment before reset will be $0.865/4 instead of $1.15/4; if we assume that the market is accounting for this, then subtraction of the excess from the market price will give a first-order approximation of what the market is actually paying for the expected future dividend stream (with the difference left undiscounted! That’s just another complication!).

However, making these adjustments doesn’t change the situation much, which is why I usually can’t be bothered:

Click for Big

And, with these adjustments, we still find that TRP.PR.A is cheap to theory, with an adjusted actual price of 21.67 compared to a theoretical price of $22.77 – so the adjusted calculation shows it being $1.10 cheap to theory, compared to the unadjusted calculation’s figure of $1.26 cheap to theory.

I suspect that the market is simply over-reacting to an expected change in dividends that is both significant and imminent. It will be most interesting to learn, as more data becomes accumulated, whether this is always the case, for junk FixedResets as well as investment-grade, for expected increases as well as decreases.

Those of you who have been paying attention will remember that in a “normal” market (which we have not seen in well over a year) the slope of this line is related to the implied volatility of yields in Black-Scholes theory, as discussed in the January, 2010, edition of PrefLetter. As has been previously noted, very high levels of Implied Volatility (in the 40% range, at which point the calculation may be considered virtually meaningless) imply a very strong expectation of directionality in future prices – i.e, an expectation that all issues will be redeemed at par.

It is significant that the preferred share market knows no moderation. I suggest that a good baseline estimate for Volatility over a three year period is 15% but the observed figure is generally higher in a rising market and lower in a declining one … with, of course, a period of adjustment in between, which I suspect we are currently experiencing.

Sometimes everything works … sometimes it’s 50-50 … sometimes nothing works. The fund seeks to earn incremental return by selling liquidity (that is, taking the other side of trades that other market participants are strongly motivated to execute), which can also be referred to as ‘trading noise’ – although for quite some time, noise trading has taken a distant second place to the sectoral play on insurance DeemedRetractibles; something that dismays me, particularly given that the market does not yet agree with me regarding the insurance issues! There were a lot of strongly motivated market participants during the Panic of 2007, generating a lot of noise! Unfortunately, the conditions of the Panic may never be repeated in my lifetime … but the fund will simply attempt to make trades when swaps seem profitable, without worrying about the level of monthly turnover.

There’s plenty of room for new money left in the fund. I have shown in PrefLetter that market pricing for FixedResets is very often irrational and I have lots of confidence – backed up by my bond portfolio management experience in the markets for Canadas and Treasuries, and equity trading on the NYSE & TSX – that there is enough demand for liquidity in any market to make the effort of providing it worthwhile (although the definition of “worthwhile” in terms of basis points of outperformance changes considerably from market to market!) I will continue to exert utmost efforts to outperform but it should be borne in mind that there will almost inevitably be periods of underperformance in the future.

The yields available on high quality preferred shares remain elevated, which is reflected in the current estimate of sustainable income.

Significant positions were held in DeemedRetractible, SplitShare and FixedReset issues on October 31; all of these currently have their yields calculated with the presumption that they will be called by the issuers at par prior to 2022-1-31 (banks) or 2025-1-31 (insurers and insurance holding companies) or on a different date (SplitShares) This presents another complication in the calculation of sustainable yield, which also assumes that redemption proceeds will be reinvested at the same rate.

I no longer show calculations that assume the conversion of the entire portfolio into PerpetualDiscounts, as the fund has only a small position in these issues.

I will also 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 (set at 1.49% for the October 31 calculation) to estimate dividends after reset for FixedResets.

Most funds report Current Yield. For instance, ZPR reports a “Dividend Yield” of 4.5% as of August 29, 2014, but this is the Current Yield, a meaningless number. The Current Yield of MAPF was 4.89% as of August 29, but I will neither report that with any degree of prominence nor take any great pleasure in the fact that it’s a little higher than the ZPR number. It’s meaningless; to discuss it in the context of portfolio reporting is misleading.

However, BMO has taken a significant step forward in that they are no longer reporting the “Portfolio Yield” directly on their website; the information is taken from the “Enhanced Fund Profile” which is available only as a PDF link. CPD doesn’t report this metric on the CPD fact sheet or on their website. I may have one less thing to mock the fundcos about!

It should be noted that the concept of this Sustainable Income calculation was developed when the fund’s holdings were overwhelmingly PerpetualDiscounts – see, for instance, the bottom of the market in November 2008. It is easy to understand that for a PerpetualDiscount, the technique of multiplying yield by price will indeed result in the coupon – a PerpetualDiscount paying $1 annually will show a Sustainable Income of $1, regardless of whether the price is $24 or $17.

Things are not quite so neat when maturity dates and maturity prices that are different from the current price are thrown into the mix. If we take a notional Straight Perpetual paying $5 annually, the price is $100 when the yield is 5% (all this ignores option effects). As the yield increases to 6%, the price declines to 83.33; and 83.33 x 6% is the same $5. Good enough.

But a ten year bond, priced at 100 when the yield is equal to its coupon of 5%, will decline in price to 92.56; and 92.56 x 6% is 5.55; thus, the calculated Sustainable Income has increased as the price has declined as shown in the graph:

Click for Big

The difference is because the bond’s yield calculation includes the amortization of the discount; therefore, so does the Sustainable Income estimate.

Different assumptions lead to different results from the calculation, but the overall positive trend is apparent. I’m very pleased with the long-term results! It will be noted that if there was no trading in the portfolio, one would expect the sustainable yield to be constant (before fees and expenses). The success of the fund’s trading is showing up in

• the very good performance against the index
• the long term increases in sustainable income per unit

As has been noted, the fund has maintained a credit quality equal to or better than the index; outperformance has generally been due to exploitation of trading anomalies.

Again, there are no predictions for the future! The fund will continue to trade between issues in an attempt to exploit market gaps in liquidity, in an effort to outperform the index and keep the sustainable income per unit – however calculated! – growing.

Turnover slowed in October to about 5%.

There is extreme segmentation in the marketplace, with OSFI’s NVCC rule changes in February 2011 having had the effect of splitting the formerly relatively homogeneous Straight Perpetual class of preferreds into three parts:

• Unaffected Straight Perpetuals
• DeemedRetractibles explicitly subject to the rules (banks)
• DeemedRetractibles considered by me, but not (yet!) by the market, to be likely to be explicitly subject to the rules in the future (insurers and insurance holding companies)

This segmentation, and the extreme valuation differences between the segments, has cut down markedly on the opportunities for trading. Another trend that hasn’t helped was the migration of PerpetualDiscounts into PerpetualPremiums (due to price increases) in early 2013 – many of the PerpetualPremiums had negative Yields-to-Worst and those that don’t aren’t particularly thrilling; speaking very generally, PerpetualPremiums are to be avoided, not traded! While market weakness since the peak of the PerpetualDiscount subindex in May, 2013, has mitigated the situation somewhat, the population of PerpetualDiscounts is still exceeded by that of PerpetualPremiums – most of which are trading at a negative Yield-to-Worst.

To make this more clear, it used to be that there were 70-odd Straight Perpetuals and I was more or less indifferent as to which ones I owned (subject, of course, to issuer concentration concerns and other risk management factors). Thus, if any one of these 70 were to go down in price by – say – $0.25, I would quite often have something in inventory that I’d be willing to swap for it. The segmentation means that I am no longer indifferent; in addition to checking the valuation of a potential buy to other Straights, I also have to check its peer group. This cuts down on the potential for trading.

There is no real hope that this situation will be corrected in the near-term. OSFI has indicated that the long-promised “Draft Definition of Capital” for insurers will not be issued “for public consultation in late 2012 or early 2013”, as they fear that it might encourage speculation in the marketplace. It is not clear why OSFI is so afraid of informed speculation, since the constant speculation in the marketplace is currently less informed than it would be with a little bit of regulatory clarity.

As a result of this delay, I have extended the Deemed Maturity date for insurers and insurance holding companies by three years (to 2025-1-31), in the expectation that when OSFI finally does provide clarity, they will allow the same degree of lead-in time for these companies as they did for banks. This had a major effect on the durations of preferred shares subject to the change but, fortunately, not much on their calculated yields as most of these issues were either trading near par when the change was made or were trading at sufficient premium that a par call was expected on economic grounds. However, with the declines in the market over the past nine months, the expected capital gain on redemption of the insurance-issued DeemedRetractibles has become an important component of the calculated yield.

Due to further footdragging by OSFI, I will be extending the DeemedMaturity date for insurance issues by another two years in the near future.

Sectoral distribution of the MAPF portfolio on October 31 was as follows:

The “total” reflects the un-leveraged total portfolio (i.e., cash is included in the portfolio calculations and is deemed to have a duration and yield of 0.00.). MAPF will often have relatively large cash balances, both credit and debit, to facilitate trading. Figures presented in the table have been rounded to the indicated precision.

There was a shift during the month from SplitShares (selling CGI.PR.D, mostly in the range of 25.25-30) into PerpetualDiscounts (CU.PR.F and CU.PR.G, in the range of 21.20-27). While CGI.PR.D has shown surprising strength since the sales, closing the month at a bid of 25.52, I find it difficult to believe they will increase any more, given that this bid implies a yield to maturity of 3.55%. The trade was a little worse than break-even at month-end, given bids for CU.PR.F and CU.PR.G of 22.40 and 22.48, respectively, but by no worse than normal transaction costs.

Credit distribution is:

Liquidity Distribution is:

MAPF is, of course, Malachite Aggressive Preferred Fund, a “unit trust” managed by Hymas Investment Management Inc. Further information and links to performance, audited financials and subscription information are available the fund’s web page. The fund may be purchased either directly from Hymas Investment Management or through a brokerage account at Odlum Brown Limited. A “unit trust” is like a regular mutual fund, but is sold by offering memorandum rather than prospectus. This is cheaper, but means subscription is restricted to “accredited investors” (as defined by the Ontario Securities Commission) or those who subscribe for $150,000+. Fund past performances are not a guarantee of future performance. You can lose money investing in MAPF or any other fund.

A similar portfolio composition analysis has been performed on the Claymore Preferred Share ETF (symbol CPD) (and other funds) as of August 31, 2012, and published in the October (mainly methodology), November (most funds), and December (ZPR) 2012, PrefLetter. While direct comparisons are difficult due to the introduction of the DeemedRetractible class of preferred share (see above) it is fair to say:

• MAPF credit quality is better
• MAPF liquidity is a bit lower
• MAPF Yield is higher
• Weightings
  • MAPF is much more exposed to DeemedRetractibles
  • MAPF is much less exposed to Operating Retractibles
  • MAPF is much more exposed to SplitShares
  • MAPF is less exposed to FixFloat / Floater / Ratchet
  • MAPF weighting in FixedResets is much lower