Sunday, 31 January 2016

H&R Reit, Strip Coupons vs Convertible Debentures vs Trust Units

I was searching TD's fixed income offerings on Friday and found the following strip coupon offering:

This is a March 2020 strip coupon, yielding 2.93% at TD's spread.  For almost the same duration, you can get a 300 BPS yield improvement by buying the June 2020 convertible debentures, as follows:

When I see a 300 BPS improvement, my gut tells me something must be up, and as there's no free lunch (EVER) in investing, the thought process behind buy the strip vs. buy the convert has to involve more than a simple cursory comparison of yield.

So here are my thoughts, in no particular order:

  • The strip coupon is the face value part of the Series F 4.45% senior debentures maturing March 2020 (i.e., face value matures at $100 in 2020)
  • With reference to H&R's most recent annual information form, H&R's senior unsecured debentures are "direct obligations of the REIT and rank equally and rateably with each other Senior Debenture and with all other unsecured and unsubordinated  indebtedness of the REIT"
  • Both this senior debenture issue and the June 2020 convertible debenture issue are unsecured and rank equally, so both unsecured obligations should trade similarly in terms of YTM as a whole (Par + coupons for the Series F vs. HR.DB.D YTM)
  • The difference in YTM must therefore be part attributable to the absence of PV'ing the residual series of coupons to be received under the Series F issue
  • If there is any remaining difference between YTM as a whole between the two issues, my thoughts are that the series D convertible debenture will show a higher yield to reflect the issuer's option to settle the redemption of principal in units of the REIT instead of cash at maturity
  • The next comparison is to risk free money.  I can earn a 2% yield risk free by buying 4 year GIC's, so by buying the convertible debentures of the same duration, I can earn almost 400 BPS over risk free GIC's

Here are my thoughts on possible issues which could impact H&R's overall business:

  • Exposure to Alberta - as at Sep 30, 2015, 30% of the portfolio is located in Alberta, and 12% of gross rents are derived from Encana
  • Encana occupied almost 11% of gross leasable area, but had a remaining lease term of around 22 years.  
  • Any significant change in Encana's business could have a material impact on H&R.  If Encana were to go bankrupt, H&R would be left with 2.1M sq ft of vacant leaseable office space in Alberta.
  • As the remaining 22 yr term of Encana's in place lease/s include  legacy negotiated rent escalation clauses which likely reflect pre-2014-2016 oil crash "market" rents for Alberta space, in the worst case, H&R would need to drop rents to below pre oil crash "market" in order to re-lease the space
  • I have no idea what may or may not happen with Encana, but certainly, any investor in either of H&R's debt or trust units should at least consider the possibility of this outcome (however remote)

So far, I'm considering the 2016 convertible debentures and not the 2020 Debentures or Trust Units. The reason being, these debentures mature on Dec 31, 2016 and yield 4.5% to maturity, so I can keep my duration short and earn 300 BPS over comparable risk free GIC's, but, I have to consider the possibility that if an adverse "Encana" event or an adverse credit market event occurred this year, H&R could redeem these debentures in units instead of cash.  I don't think it's highly likely, but there is always a possibility.  And if they did, hopefully the units would be depressed enough in terms of valuation that I wouldn't actually care.

I ran the following series of analysis using H&R's reported #'s over the last 8 years in order to assess the business overall:

First, an analysis of comparative valuation and composition of capital structure as per below:

What jumps out at me immediately is that the valuation comparison between mkt cap vs. net fair value of investment properties + investment in JV's (valued using the equity method) is slightly higher than the same relative valuation back in 08 and 09.  To be fair though, I used average market cap to assess the overall discount or premium, and at times during the 08/09 crisis, when H&R got down to below $10, the discount was in excess of 50%, so certainly, just because I've calculated -23% now doesn't mean much.  If access to credit or liquidity dried up again and H&R was forced to cut or suspend distributions, there's no reason why this discount can't widen to -50% (or more).

The next observation is the change in capital structure relative to 2008.  Back in 08, funding was predominantly mortgages.  Now it's 71% mortgages, 29% debentures.  My guess is that if convertible or senior debentures with a low enough strike price and lower coupon than conventional mortgages can be issued in order to minimize overall cost of funding, H&R will continue to do so.  While H&R cannot control market rents, it can control the coupons it chooses to pay (subject to market demand for unsecured obligations), so the overall spread between rents and funding costs is more a function of the latter than the former.  On that note, here's my calculation of this spread over time using property operating income as a proxy for NOI / FV of properties.  I figure that this measure is similar to EBIT in the case of non-REIT's, and surely, investors want to see ability to maintain a sufficiently wide spread between % NOI vs. cost of funding, as well as adequate interest coverage:

What jumps out at me here are the following observations:

  • Interest coverage has improved since 2008 on a property operating income / finance costs basis.  However, 2.59x is susceptible to change in the event of an adverse event 
  • I've calculated efficiency three ways, one as operating margin based on NOI / FV of properties (similar to cap rate), two as NOI / gross rents, and three as gross rents / FV of properties.  I suspect that all three measures are of difference importance
  • I have then compared cap rate to weighted avg cost of funding in order to determine a basis point spread over the last 8 yrs.  I suspect that this spread drives valuation over time, and over 8-10 years, I should be able to see a trend in spread.  My suspicion is that a 174 BPS spread today is a function of much lower weighted avg funding cost vs. 8 years ago
  • Occupancy at 96% through the 9 months ended Sep 30, 2015 is lower than at any time during the last 8 years, however, this may be a function Target vacancies in the last year in the retail (Primaris) portfolio.  
  • I have also calculated an at market margin spread  = gross rents / FV of properties, which has ranged from as low as 3% in 2011 to 6.3% in 2010.  This spread is a lot more variable vs. NOI / FV of properties, so I'm going to attach less importance to this spread vs. NOI / FV of properties
  • The avg lease term to avg mtg term spread has increased from 2.2 yrs in 08 to 4 yrs currently. This may be more a function of H&R increasing the proportion of debentures with 5 yr maturities in the capital structure relative to 2008 as noted above.
  • Finally, I've calculated NOI / gross rents as an alternative measure of analyzing efficiency. This operating margin shows improvement since 2008, however, 2008 may not be a fair comparable starting point.  It's entirely possible that due to the credit crisis in 2008, lessors (in general) had to offer below market rents in order to induce tenants to sign leases.  It's entirely possible that NOI at the time was abnormally depressed relative to what NOI would have been during economic expansion.  If this is the case, I may need to revisit the above analysis and adjust NOI as reported down to normalize for cyclicality in the business cycle.  It's possible that we haven't yet seen the full economic fallout of $30 oil in Alberta on business activity in Western (and Eastern) Canada as the initial price shock is still fairly recent.

Concluding Thoughts

Over 8-10 years, I should be able to analyze the trend in % NOI : Cost of Funding spread.  I think that this spread drives valuation and demand in the market for the REIT itself.  REITs are highly interest rate sensitive, so if the velocity of cost of funding increases (or is expected to increase) faster than noi/gross rents (a measure of how efficiently the properties are run before finance costs), then valuation should decrease.

We saw this in 2008, and it's interesting reading H&R's annual report in 2008, where they specifically speak about access to funding drying up as a result of the credit crisis, which caused massive compression in all REIT valuations.  When cost of funding spiked in 2008 (& H&R was & still is rated BBB), my noi/gross rents less cost of funding spread tightened, and valuation fell.  Even if they were never going to sell their interest in Scotia Plaza, the market repriced imputed cap rates in this scenario by extrapolating the impact of H&R having to refinance at peak cost of funding in 2008.  And really, this is the time to buy REIT trust units, when there is a massive disconnect between the market pricing renewal cost of funding vs duration of remaining mortgage term & probability of peak crisis cost of funding remaining elevated over time.

So, I don't think buying H&R trust units is the correct move now, because cap rates are low (& valuations are high), and there is no real fear in the market over limited access to credit.

The convertible debentures make more sense because the imbedded call options are near worthless the longer the units stay below the strike price, so they trade like straight bonds, and the shorter duration convertible debentures maturing Dec 31st this year will likely get redeemed & replaced by another issue with a lower coupon and a lower strike price so H&R can ratchet down their overall cost of funding for another 5 yrs, with minimal default or extension risk.

No comments:

Post a Comment