11 Responses to “July 10, 2023”

1. avocado says:

   July 11, 2023 at 12:23 am

   Interesting to see the volume on CVE.PR.A — 157,035 shares traded, dropping 9.8%; where as the other issues hardly moved in comparison.

2. skeptical says:

   July 11, 2023 at 6:42 am

   Why such big fall? I haven’t had this issue in a while, so not tracking. Anything with cve or just another regular day in the preferred market?

3. Brassens says:

   July 11, 2023 at 8:50 am

   CVE.PR.A yield was 4.92% . Compared to the other 5y reset from the same company, it is well below. C=6.34, E= 6.08, G=5.22%. With the drop from yesterday it now yields 5.38. To me it’s just a case of rebalancing. the A series is stuck with its dividend until 2026.

4. jiHymas says:

   July 11, 2023 at 10:13 am

   CVE.PR.A yield was 4.92% . Compared to the other 5y reset from the same company, it is well below. C=6.34, E= 6.08, G=5.22%. With the drop from yesterday it now yields 5.38.

   It appears to me that you are using Current Yield (current dividend divided by price) as a valuation measure – is this correct?

   Using Current Yield implies a belief that future resets will all be done according to the GOC-5 rate that prevailed at the time of thier last reset; i.e., that when CVE.PR.G next resets in 2025, the GOC-5 yield will be 0.415% (as it was in 2020; while HSE.PR.C will reset six months earlier, at the end of 2024, with a GOC-5 yield of 1.559% (as it was in 2019).

   This implies in turn a belief that the violent yield changes of the Pandemic and its aftermath will precisely repeat in a five year cycle forever, which seems to me to be very unlikely. Frankly, if I had such a belief in the future path of interest rates, I wouldn’t be fooling around with silly little preferred shares – I’d be mortgaging everything I owned to go long Five Year Canada futures in the expectation of enormous capital gains when rates decline with such ferocity from current levels. But I have no such belief.

   I suggest that when comparing issues you use the FixedReset yield calculator, which is described in the post What Is The Yield Of HSE.PR.A?. You can use any projection you like for future GOC-5 rates (see the discussion HERE) but I will assert that yields calculated using a consistent and fairly stable projection for GOC-5 will lead to a more accurate projected ranking of future returns than the ‘endless cycle’ projection of Current Yield.

5. Brassens says:

   July 11, 2023 at 12:26 pm

   Indeed I simply used the current yield. With a quick look ( which was maybe not a good idea), nothing made the A series attractive. It still had the lowest blended yield. The C series yielded a full 1.34%over the A with a closer reset date. For this to be an advantage you have to think that the rate in dec 2024 will be higher than in march 2026. I will look into the links you provided. thanks

6. jiHymas says:

   July 11, 2023 at 2:03 pm

   nothing made the A series attractive. It still had the lowest blended yield. The C series yielded a full 1.34%over the A with a closer reset date.

   I’m not sure how you blended the yield.

   According to my proprietary software, HIMIPref™, at the close yesterday and based upon a continuing GOC-5 yield of 3.95%, HSE.PR.A was bid at 11.95 to yield 10.38% while HSE.PR.C was bid at 18.55 to yield 9.23%.

   Numbers you obtain from the spreadsheet will differ a little from the above figures due to definitional and implementation differences.

7. jiHymas says:

   July 11, 2023 at 2:09 pm

   In my essay The Importance of Issue Reset Spreads, I point out that at a yield of 5%, the next five years of dividends represent less than 25% of the total value of a perpetual instrument.

8. Brassens says:

   July 13, 2023 at 6:44 am

   Thanks again for the clarification. I played a bit with your Excel sheet with the CVE.PR.A and C issues. I used a price of 13.10$ for the A as it was the price before the big drop in price and I get Current Yield= 4.92%, Annualized = 9.31%
   For the C series with a price of 18.55$ I get Currents Yield= 6.32%, Annualized = 8.99%.
   If I understand correctly, your point would be that the drop in price was uncalled for as the Annualized are very similar.
   If we play around with the call date, this is where we can see that for the A to be ”better” the call date needs to be very far. I tried with a 2030 call date for both and the C series was advantageous ( 8.46% vs 7.91%).
   I guess this is where the assumptions and time horizon for investment comes into play.
   Thanks again for taking the time, I will do my homework and keep reading your previous posts to better understand your point of view.

9. jiHymas says:

   July 13, 2023 at 9:37 am

   If we play around with the call date, this is where we can see that for the A to be ”better” the call date needs to be very far. I tried with a 2030 call date for both and the C series was advantageous ( 8.46% vs 7.91%).
   I guess this is where the assumptions and time horizon for investment comes into play.

   Given that the prices of these issues is so far below par, I suggest that the best estimate of end-date is ‘forever’; the spreadsheet imposes a limit of 30 years on the date used for computational purposes which is usually close enough, but you could modify the spreadsheet if you wish to a much longer term.

   One thing you could do is a ‘sanity check’ on your ending assumptions of price and date is to calculate the yield at that time for both instruments. If you accept the assumption of a constant GOC-5 yield, then after the first reset the issues will have a constant dividend rate forever afterwards and can therefore be treated as Straight Preferreds or, more specifically, PerpetualDiscoujnts.

   The yields on this future date should be more-or-less equal as they are from the same issuer; the higher-coupon, higher-price issue should have a slightly higher yield due to its increased call risk, but that will be a relatively small adjustment.

   I suggest that assuming that both issues will have a constant price forever given the assumption of constant GOC-5 is not logical; given the constant GOC-5 assumption, the relative prices should adjust to reflect this constancy. But this should become apparent once you calculate the future yields given a constant price after the first reset.

10. jiHymas says:

    July 15, 2023 at 9:15 am

    I guess this is where the assumptions and time horizon for investment comes into play.
    Thanks again for taking the time, I will do my homework and keep reading your previous posts to better understand your point of view.

    On reflection, I have become a little concerned about this statement. Regardless of what Brassens actually means, there is a chance that other readers might misinterpret the purpose of the yield calculator.

    This yield calculator is not a magic prediction tool for forecasting the total return that will be earned over any given period. The only thing it does:

    1) it is designed for application to FixedReset preferred shares that are of the form that has become conventional in the market.

    2) these instruments have features that force the investor to make assumptions about the future in order to calculate forecasted yield (forecasting yield always requires assumptions, some instruments more than others)

       2a) given current market conditions, virtually all issues may be expected to remain outstanding perpetually. This implies that if we perform our calculations over a finite time-period (which is what the spreadsheet does), we must make an assumption about the end-price. This end-price assumption will become less critical as the term to the end-date and the yield increase, as the end-price, when discounted, will become of less importance to the present value as these two estimated variables increase.

       2b) if it is assumed that the issue will remain outstanding through any reset date, then the reset yield must be forecast. Since in the conventional structure this rate is dependent upon the GOC-5 yield at the time of each reset. This set of assumptions can be as complex as desired; I prefer to take the simple route and assume a constant rate forever. Additionally, for ranking purposes, I prefer to assume that this constant rate will be equal to the current GOC-5 weight. It is critical, for ranking purposes, that the underlying GOC-5 assumptions be the same for each instrument examined; just what that assumption may be is a matter of taste and whatever the investor considers ‘reasonable’ (or he may calculate a variety of scenarios in order to get an idea of the sensitivity of the answers to the assumptions).

    3) The spreadsheet takes these assumptions and creates a table of the cash flows that ae predicted by assumptions. This table is in a form convenient for using the MS-Excel IRR() function

    4) The Internal Rate of Return implied by these cashflows is calculated and converted into standard format with the further assumption of quarterly compounding.

    5) That’s it. Plug in your assumptions, retrieve the overall effect of those assumptions in standard format. Full stop. It is recognized that the realized returns will almost certainly be different from the forecast returns, due to the differences between the assumptions and the realized experience; however, the provided these differences are not too great, the ranking of the relative yields at forecast-time should be a reasonable reflection of the ranking of realized returns at the ‘end-date’

11. stusclues says:

    July 15, 2023 at 12:58 pm

    “if it is assumed that the issue will remain outstanding through any reset date, then the reset yield must be forecast. Since in the conventional structure this rate is dependent upon the GOC-5 yield at the time of each reset. This set of assumptions can be as complex as desired; I prefer to take the simple route and assume a constant rate forever.”

    I have evolved to embrace this idea also, thanks to many years of reading/studying James’ work and following this blog. I was tempted to consider using Monte Carlo simulation using various GOC5 estimates, however this requires an estimation of the probabilities of various GOC5 levels, which I think is almost impossible to generate given global market complexity. Therefore, constant rate it is.