That Odd Sri Lankan Airline Guaranteed Bond
Mitu Gulati & Mark Weidemaier
After months of waffling, Sri Lanka’s head-in-sand government has finally acknowledged that it cannot pay its debts. The cavalry (IMF) has been called in and we guess that hordes of potential restructuring advisers are flying to Colombo to offer their services. Assuming they have done their homework, their proposals surely will consider both the government’s own debt and a Sri Lankan airline bond that the government has guaranteed.
Sri Lankan airlines used to be profitable. From 1998-2008, it was partially owned and run by Emirates. One of us recalls it being a special treat to fly on. But the government decided in 2008 to run the airline itself and, since then, it has performed terribly. There have been corruption scandals, accusations that Emirates was pushed out after the airline refused to bump paying passengers to make room for the royal family, and reports that local banks have been strong-armed into lending and will be in trouble if the airline collapses. Perhaps it’s no surprise that it needed a government guarantee to borrow money.
Sovereign guaranteed bonds often carry a higher coupon than a bond issued by the sovereign, perhaps because the sovereign is viewed as the safest credit. But this logic seems upside down. Unlike a pure sovereign bond, a guaranteed corporate bond is backed both by the sovereign’s credit and by a separate pool of assets (e.g., airplanes). Even if the company is literally worthless, there is still the full sovereign guarantee. Obviously there will be other factors that affect price, such as liquidity (the market for pure sovereign bonds may be much larger). But in crisis, when the bonds are sure to be restructured, there seems every reason to favor the guaranteed bond.
Another reason to favor a guaranteed bond is that these often have less effective restructuring mechanisms than are found in the sovereign’s own bonds. Oddly, then, a guaranteed bond that was viewed as riskier at issuance can end up being a safer bet. Greece’s 2012 restructuring imposed haircuts of over 50% on pure sovereign bonds but most holders of guaranteed bonds got paid in full. There is even some evidence suggesting that investors had figured this out towards the end game in Greece and favored guaranteed bonds.
Here are some of the provisions in the airline guaranteed bond that could cause Sri Lanka’s restructuring advisors a giant headache.
Most bonds issued after 2014 have modification provisions that allow a restructuring vote to be aggregated across multiple series of bonds. Aggregation allows for a more efficient restructuring (one vote for multiple bonds, as opposed to separate votes for each bond series) and is less vulnerable to holdouts (by making it very hard to obtain a blocking stake). As we’ve written recently, at least some versions of the aggregated modification clause—including, as best we can tell, the ones used by Sri Lanka—give an extraordinary amount of leverage to the issuer, assuming it is willing to use it. That leverage is absent in the Sri Lankan airline bond.
Although it was issued in 2019, five years after aggregation provisions became standard in sovereign bonds, the Sri Lankan airline bond does not contain this aggregation feature. Instead, it has what appears like an ancient restructuring provision of the type standard in older bonds governed by English law before 2004. These provisions were jettisoned from sovereign bonds starting around 2003-04, after a global consensus that international sovereign bonds needed a newer set of restructuring provisions that would protect against the risk of holdout creditors (for some histories, see here, here and here). The Sri Lankan airline bond doesn’t follow this practice. It has a clunky old pre-2004 restructuring provision that favors holdout creditors. For some reason, it is also issued under English law, while the sovereign’s own international bonds are governed by NY law.
Perhaps all of this is the product of careful drafting. Maybe investors in the airline bond were being smart and protecting themselves against the future possibility of an aggressive restructuring? Perhaps they cleverly wanted to ensure that their little bond (only $175 million) could not easily be swept into the bigger sovereign restructuring. Maybe, but we are skeptical that there was this degree of cleverness going on. The “clever drafter” story is hard to square with the fact that some other provisions of the airline bond seem quite restructuring friendly and, to be frank, goofy.
For example, sovereign bonds apply the highest voting threshold to modifications that will change payment and other important terms, often called “reserved matters.” In a modern bond, the reserved matter list is quite long, which helps prevent the issuer from doing sneaky things to holdouts with the consent of only a majority of bondholders. For instance, the issuer might persuade investors who favor a restructuring to accept new bonds, in the process voting to change the governing law of the non cooperating bonds (from foreign to local law). In a modern bond, the reserved matter list is broad enough to constrain such changes. But not in the Sri Lankan airline bond. (This is the “Exit Consent” strategy that is arguably disfavored under English law, but we won’t delve into that here).
Negative Pledge
Sometimes, when a country is in as deep of a crisis as Sri Lanka is, it needs to do emergency borrowing to buy time with creditors who are threatening to accelerate and sue (things always get much more complicated once creditors have gotten legal judgements and are chasing assets). Sri Lanka is in deep financial trouble, with a big payment coming due in the next couple of months. A country in this position might want emergency funding, potentially from a bilateral lender (India? China?). But that typically requires the ability to pledge assets to attract a loan on favorable terms.
In a standard sovereign bond, the issuer has limited room to pledge assets. The Negative Pledge clause in almost every international sovereign bond forbids the issuer from securing new lenders unless it also grants security to existing bondholders. In other words, under the standard clause, the issuer can’t give new lenders priority.
But Sri Lanka’s sovereign bonds seem to allow for the issuance of new, secured debt. There are at least two unusual provisions that arguably do this. First, in the part of the bond that lists exceptions to the Limitation on Liens (i.e., negative pledge) provision, there is a weird statement to the effect that Sri Lanka “believes” that its Central Bank is not bound by this restriction and may pledge International Monetary Assets.
Perhaps this is simply stating the obvious. The Limitation on Liens provision defines what the “Issuer” must do, and the “Issuer” means “the Government of the Democratic Socialist Republic of Sri Lanka.” The Central Bank is a legally separate entity. However, the Central Bank is close enough to the state that we can imagine a bondholder arguing that it also must honor the Limitation on Liens provision. If that argument is accepted, a statement that the issuer “believes” to the contrary might not be enough to carve out an exception. Still, the statement at least lets Sri Lanka argue that it can pledge international monetary assets through its Central Bank.
Moreover, in a separate provision (Waiver of Certain Covenants), Sri Lanka’s sovereign bonds let bondholders holding only a majority in principal amount of the bonds waive compliance with the Limitation on Liens provision. That’s a super low voting threshold. This gives the government an additional path to pledging assets, even if the Limitation on Liens provision would not otherwise allow it.
However, things get weirder still. Although Sri Lanka’s bonds arguably allow the Central Bank to pledge international monetary reserves as security for new lending, the bonds also say that it is an “event of default” if the Central Bank does not exercise “full ownership, power and control” of those assets. And at least some forms of a collateral pledge might trigger this provision. Why create this seemingly contradictory set of provisions? (Remember the goofy versus sneaky drafting hypotheses).
Anyway, back to the airline bond. Strangely, it has virtually none of these provisions. We can’t even find a Negative Pledge clause in the airline bond’s Offering Circular. Perhaps it is in the fuller trust document, although the Negative Pledge is an important source of legal protections, so it would be odd not to describe it in the Offering Circular. However, the airline bond does include an Event of Default, akin to the one in Sri Lanka’s sovereign bonds, triggered if the Central Bank fails does not have full ownership, power and control over international monetary assets. So, although the airline bond doesn’t seem to bar the issuance of secured debt (at least if we can trust the Offering Circular), it is at least possible that an asset pledge by the Central Bank might send the airline into default. Again, so weird.
All in all, the airline bondholders should have the upper hand in a restructuring. If nothing else, the older modification provision should make it easy for smart bondholders to secure the vote needed to reject a restructuring proposal. True, the issuer has other tools, but the relatively small size of the airline bond may allow investors to block even a restructuring method that requires only a majority vote. However, as best we can tell, the airline bond and the most comparable pure sovereign bond are trading at about the same price (thanks to our friend Ugo Panizza here). Maybe there are other important contract terms that we haven’t detected? Or maybe bondholders haven’t learned the lessons of Greece 2012?
Hi Mitu and Mark,
Thank you for your very insightful blog posts and podcasts. I’m a distressed debt analyst researching Sri Lanka’s bond complex and wanted to clarify what Andres de la Cruz meant by the lack of anti-discrimination provisions in the re-designation clauses in the government-guaranteed bonds. Does this:
1) Give the government more discretion to pool different series of bonds into different groups for the purposes of an aggregated vote. If so, how?
2) Does it provide the issuer with powers to re-designate groups ex-post if it is initially unable to obtain a requisite majority.
3) Does it refer to some roundabout way of undermining provisions which require comparable terms to be offered to dissenting groups in line with those accepted by others?
Would really appreciate it if either of you can shed some more light on this matter.
Regards,
Raza
Posted by: Raza | April 28, 2022 at 09:34 PM
hi Team
just wnat ot say today is the day...
Posted by: frank O Lehmann | June 27, 2022 at 03:17 AM