Last Friday’s post about the death of poison pills and birth of investment-grade poison puts back in the mid-naughties left us plagued by thoughts about change of control covenants.
For those joining us now, change of control covenants are words thrown into bond prospectuses. They’re supposed to protect bondholders from the risk that the firms they lent money to are bought by other firms that aren’t as creditworthy. They allow bondholders to sell bonds back to issuers at around par value if that kind of thing happens.
In a post-rally bond world, they’re pretty worthless. The protection you get from being able to put a bond with a market price of 120 back at par (100) is worth … something? It’s not eminently clear what.
However, in a post-bond crash world (ie now) these covenants have at least the potential to provide a headwind to M&A deals. Why? Because, for an issuer, buying back a bunch of bonds trading far below par at par costs money. Sometimes, a lot of money — maybe enough to turn an otherwise appealing deal into something that looks plain stupid.
So we set out to try to quantify the headwind across US investment grade bonds in aggregate, and locate where this headwind might be most meaningful.
Immediate problem: bond prospectuses are notoriously idiosyncratic. So we’re going to start with some obviously false generalities and then work from there. Bear with us.
Suspend your disbelief
Our first assumption is that all US investment grade securities contain a change of control covenant. Why is this assumption wrong? While change of control covenants might very well be part of the standard boilerplate template that legal firms and banks start with when putting together a new issue, not every bond will have one.
The second assumption is that the covenant allows the bondholder to put the issue back at par any time there is a change of control. This is plainly nonsense, as bonds with change of control covenants typically define ‘Change of Control Events’ that can include all sorts of other requirements beyond a simple change of control — maybe a rating downgrade to below investment grade,maybe something else.
The third impossible assumption is that a change of control event is determined to have occurred anytime a change of control occurs. This is a nice idea, but we’re not super confident that such things are always so straightforward. Such questions might conceivably rest on the views of a specialist judge.
The fourth and final assumption is that the poison put is at 100. Often (where they exist) they will actually be at 101 plus accrued rather than par.
Why make assumptions that are plainly false? Journalistic licence. Because they allow us to make some charts without reading and digesting the prospectuses of the 10,160 bonds that are members of the ICE US Investment Grade index as of the start of 2024. And charts are good.
Some charts!
Our first chart shows the gap between overall market cap and par for all the bonds in the ICE US Investment Grade Index where the market cap (of the bonds) is below par, across a representative sample of years. Think about it this way – it shows the very worst case (from a shareholder’s perspective) of the amount bondholders would need to be paid beyond market value if a change of control occurred.
Sure, the number and face value of US investment grade bonds has increased a lot over the past ten years. But this estimate as to the best-case value (from a bondholder’s perspective) of these covenants has increased far far more. $675bn is a big number.
The second chart breaks this big number down on a firm-level basis. We count 171 firms whose debt’s face value was more than $1bn above its debt’s market value. These are all large firms with lots of debt, especially debt issued in recent years when yields were much lower. Using our made-up assumptions, the hypothetical covenants push up the (already frankly impossible) LBO price of Apple by around $10bn.
Of course, $10bn on Apple is a rounding error; a better way to look at the strength of deal headwinds could be to consider what percentage of market cap could be added to any purchase price. We did a quick estimate for the c1,300 firms with bonds in the ICE US Investment Grade index, again assuming (wrongly) that every security has a poison put at par.
Our third chart shows the twenty US-listed firms in the ICE US Investment Grade index that we reckon could have the largest potential deal headwinds (again, applying the assumptions listed above). To be clear (and preserve the author’s sanity), this approach looks only at bonds that are index constituents of the ICE US dollar investment grade index, and which also have a US listing. So if a US-listed firm has tonnes of Euro-denominated bonds trading far below par with change of control put options, these will be omitted from the calculation. And it certainly possible that while change of control covenants are super prevalent (the authors of this study reckon well over half of all bonds have them), they don’t actually appear in the bonds that contribute to these numbers. Is that caveated enough?:
Back to the real world
We’ve made some admittedly beautiful charts, but does any of this relate to the real world? To answer, we’ve got actually to look at some bond prospectuses. Oh joy.
We did some spot checks on bonds issued by the top six of our heuristic hit list. It turns out that the real world is messy.
We couldn’t immediately see a change of control covenant in this Hudson Pacific Properties bond prospectus, nor this one from Piedmont Operating Partnership LP. Sometimes prospectuses really go out of their way to be explicit about their absence, like this … ahem … reassuring (?) paragraph from Brighthouse Financial:
But we spotted change of control covenants in this Air Lease Corp prospectus, this one from Paramount Global, and this one from Warner Brothers/Discovery.
Stretching a meme joke beyond its worth
We took a deeper dive into the bonds of Paramount Global – number four on our list. Partly because it’s been subject to some MainFT scrutiny as a potential M&A candidate recently. But mostly because it distributed Raiders of the Lost Ark — the most famous film to be made about covenants, and source of great memes.
Paramount Global has got 18 bonds in the ICE US Investment Grade Index, with a collective face value of $13bn. The sum of the differences between par and market value for all those bonds that trade below par is just over $1.5bn, or around 15 per cent of market cap. We tracked down the prospectuses of sixteen of these eighteen bonds, and fourteen of them look to share this sort of change of control event language:
This is just a summary, but you get the gist. We couldn’t find equivalent language in the other two prospectuses that we located, bringing down our best-guess potential bondholder windfall/ shareholder cost in the event of a change of control from c$1.5bn to c.$1.2bn.
But the language is clear that before bondholders get their payouts, there needs to be not only a change of control, but also a downgrade to junk by each of the big three rating agencies.
We’re not going to run you through an analysis of the likelihood of this occurring, rating agency by rating agency. This would not only be far too much work, but also involve forward-looking statements and legal risk we’d prefer to avoid. We’ll just note that S&P currently rates Paramount Global one notch above junk with a stable outlook. Over to S&P:
The stable outlook reflects our expectation that Paramount will reduce its elevated leverage to about 4.0x in 2024 and 3.5x in 2025 as the macroeconomic environment improves and it materially decreased the losses in its DTC segment.
They go on to warn that they:
could lower … ratings on Paramount if we do not see a path for it to reduce its S&P Global Ratings-adjusted leverage below 3.5x on a sustained basis by 2025
Adding:
we could tighten our 3.5x leverage threshold for the rating, or even lower our rating, if the company’s business trends weaken significantly and it becomes clear its ability to monetize its content amid the changing media ecosystem has materially declined.
It sort of sounds like 3.5x leverage is a thing for them:
A 3.5x debt/EBITDA would bring EBITDA interest cover (the number of times over that interest can be paid out of earnings before interest tax, depreciation and amortisation) up from 3.6x to 4.9x in 2025 – a material credit improvement that would take the firm to within spitting distance of their pre-COVID metrics.
But the point is this: S&P believe Paramount is on a deleveraging path, and they’ll continue to rate them as investment grade as long as they can retain that belief. As such it seems reasonable that any takeover will need to be either credit-neutral or credit-positive to avoid adding the equivalent of $1.2bn of fresh debt.
Which brings us (finally) to our conclusions.
First up, bond prospectuses are all different, and they are no fun to read (I refuse to believe you took no enjoyment at all from reading them — Ed). There are folks out there who make a good living building and selling covenant databases, and we don’t envy them. We reckon that our attempt at building out a quick heuristic for the potential windfalls due to bondholders if change of control events are triggered for the entire universe of US investment grade firms will be right directionally, but it’s impossible to get a sense of the magnitude of the value of these covenants without doing the bond-by-bond work (or paying someone else to do it for you).
Second, this actually might be a thing to watch. It has been hard to avoid M&A dealmakers feeling sorry for themselves and expressing the hope that things pick up after a lousy 2023. Given that deals tend to involve debt finance, folks generally point at higher debt financing costs to explain the deal drought.
But change of control covenants – hardly in existence for investment grade firms prior to 2005 – are now reasonably abundant and could, maybe, be an additional headwind, at least for firms rated on the cusp of investment grade. Corporate bond veterans may recall dynamics that played out with so-called ‘credit rating cliffs’ back in 2000–2002. Well-intentioned bondholder-protection language triggered accelerated payments to creditors when agencies junked issuers, most famously and dramatically in the case of Enron.
To mix metaphors it feels reasonable to ask whether (hopefully less consequential!) credit rating cliffs exist today as moats around otherwise reasonable M&A deals that aren’t predominantly equity-financed.
