Constellation Software Bonds?
With all the excitement guided towards Constellation Software (CSU), I’m surprised that their bonds aren’t ever discussed. I believe that an investor can earn ~6.4% by holding CSU bonds to maturity, in addition to having significant protection against future inflation.
For all the CSU degenerates (myself included), ~6.4% doesn’t sound particularly appealing especially when the common equity has compounded at ~30% per annum. What I would argue though, is that equity investors are implicitly pricing in a trivial risk of default via the high multiple on CSU stock, and that this should be properly reflected in the bonds as well.
If you truly think that CSU is bulletproof, then the ~6.4% annual return on these bonds is essentially “risk-free”.
As I was making my way through a list of publicly traded bonds on the TSX, I noticed that CSU had some outstanding. This caught my eye because I typically knew CSU to be under-levered and financed with internal funds and/or bank debt.
These specific bonds trade under the symbol $CSU.DB and there are ~$360MM outstanding. Important features include:
Unsecured subordinate debentures (prospectus link)
March 31, 2040 maturity (~15.5 years left)
$100 par value (currently trading at $121.50, which is a premium to par)
Interest rate is reset on March 31st every year
How interest rate is determined: change in CPI index from the previous calendar year + 6.5%
Change in CPI index (aka: inflation rate) can be negative; however, the absolute interest rate on the bonds cannot go below 0%
CSU has an option to redeem them at par ($100), but they must provide notice today and subsequently wait 5 years before the actual redemption occurs
Investors have an option to put back the debentures to the company at par ($100), but they must provide notice today and subsequently wait 5 years before the actual put occurs
CSU cannot pay dividends or buyback shares unless they pay interest on the debentures first
CSU can elect to PIK (payment in kind) the debentures
PIK’ing the debentures still allows CSU to pay dividends or buy back shares
The napkin math on these is as follows: assume normalized CPI of 2% plus the 6.5% spread, for a yearly coupon of $8.50. Based on the price today, this amounts to a current yield of ~7.0% ($8.50/$121.50). Given that the debentures are trading at a premium today and will be redeemed at $100 in 2040, this means that the actual yield to maturity is ~6.4%. One could argue for a higher or lower assumed inflation rate, but 2% is what the government targets, and it is what has been historically achieved.
CSU currently pays an interest rate of 10.4% on these debentures, which will be reset relative to the new inflation rate on March 31, 2025. Investors will be able to collect a couple of these higher coupons before the reset occurs likely at a lower rate.
A yield to maturity of ~6.4% may seem low for equity-focused investors, so I have compared it to below to an Investment Grade Bond ETF and a Preferred Share ETF.
Canadian Investment grade bond ETF ($ZCB.TO):
Current yield to maturity of ~4.1%
Average term of ~8.4 years
Comprises 38% financials, 24% energy, 15% infrastructure, 10% communications, etc.
Because the ETF’s holdings are senior notes and not subordinated debentures, I would argue that the debentures should yield much higher than 4.1%. CSU is also a better business than the majority of the companies comprised in the ETF, which works to counteract a portion of this.
Canadian Preferred Share ETF ($CPD.TO):
Current yield of ~5.3%
Comprises 60% financials, 19% energy, 8% utilities, etc.
CSU’s debentures are much more similar to preferred shares. They are both subordinated, longer-term, and can be PIK’d. Once again, the 5.3% yield on this ETF is much lower than the CSU debentures and the issuing companies are much more levered.
One can also compare the debentures to the senior notes CSU issued in February 2024. The company issued both 5-year and 10-year notes, which had corresponding spreads over the Government of Canada (GoC) bond of ~1.50% and ~1.90%, respectively. As of today, I estimate that the CSU debentures have a spread of ~3.30% over the GoC, which seems reasonable given that they are subordinated to the senior notes.
One risk (not business-related) is that CSU has the option to redeem the securities at $100 five years from now. Given that the debentures are trading at a premium today, one would be guaranteed to lose $21.50 ($121.50 less $100). Fortunately, the expected coupons one would collect over that period would be ~$42.50 ($8.5 x 5 years), which means that you won’t lose money. You would just earn an inadequate return of ~3.9%.
The second risk that one is taking relates to CPI. If deflation occurs or if CPI is lower than 2%, the annual reset would mean that investors would earn much less than previously estimated. For example, if CPI were to be 0% over the life of the debenture, then I estimate that investors would earn ~4.7%.
I think it’s worth mentioning a few keys facts about CSU: In the last 12 months, it had ~$9.3bn in revenue and ~$1.6bn in EBITDA, relative to net debt of ~$2.5bn. This results in a leverage ratio of ~1.6x for a company where the majority of EBITDA drops to the bottom line. Leverage risk seems low based on this backward-looking metric.
This is not a “get-rich” investment after all, but I believe that earning ~6.4% for 15.5 years in a high-quality company is a decent risk/reward opportunity.
Indeed I have found the Constellation debentures interesting. Mark Leonard has said (in a letter? in one of those old Q&A's he did for a while? can't remember) that he considers these very cheap, flexible debt and would happily do more debentures like it, which I took as a good indicator that Constellation is unlikely to ever call these. And for a long time these traded such that getting called in five years would result in a very unsatisfactory return, vs now's merely low return. The capped downside now makes them more attractive (as does the reasonably high likelihood of being able to sell at a higher price at some point before maturity).
One thing I'd say is that the tax treatment of preferred shares (dividends vs interest) likely makes the spread of the yields on the preferred shares ETF and CSU.DB make more sense. I believe 1.3x is the number used by James Hymas (ie. a dividend of $1 is worth $1.3 of interest). That said, a 6.9% yield on the preferred ETF vs YTM of 6.4% for CSU.DB would still strike me as too narrow a spread.