As we enter the New Year, we say goodbye to an old “friend” that has accompanied us (for better and for worse) since the inception of the UK ABS market in the late 1980s – Sterling LIBOR.
Beyond ABS, LIBOR has been widely used since the 1970s, and almost USD 400 trillion financial instruments, across every part of the global economy, are tied to LIBOR in some way [source: Bank of England]. This is quite a pivotal shift for the financial markets, but a necessary one, as the inter-bank unsecured lending which used to underpin LIBOR has virtually disappeared, leaving the rate mostly a theoretical exercise and making it unviable on a long-term basis.
However, this is not a blog about the end of LIBOR; it is a blog about how the UK’s transition to SONIA, the replacement benchmark, has once again highlighted the importance of ‘future proofing’ transaction documentation wherever possible.
Whilst they typically have shorter expected lives, securitisations are structured such that they can remain outstanding until after the very last loan in the underlying pool has paid if required; 25-30 years or more in the case of mortgages. As a result, securitisations can outlast their originators and sponsors and weather any unforeseeable market changes, regulatory or otherwise. Indeed, several transactions from the early 2000s remain outstanding almost 20 years later.
And so, following the first new ABS transactions referencing SONIA in early 2019, the stock of outstanding ABS deals also started moving to the new benchmark, with the pace picking up since Q1 of 2020. Its introduction required amendments to the bonds’ coupon formulas to reference SONIA instead of LIBOR; simple enough, one would think. The presumption of simplicity generally proved correct but, several deals lagged, and a minority have so far failed. These setbacks might be understandable for those pre-Financial Crisis deals where the originator or sponsor no longer exists, but it also includes a number from less ancient issuers!
Following deeper analysis, it became apparent that any problems related to the bondholder voting language incorporated into the deals’ documentation, especially for the older deals where no concept of LIBOR replacement had even been considered. Similarly, many deals issued whilst the new benchmark and the methodology were still under discussion also experienced difficulties.
In most of these transactions, the voting rules only provided for changes to occur when a significant majority of each and every class of notes within a transaction voted in favour. In those instances, what is less obvious is that many investors frequently don’t or cannot vote. This is often because notices and decisions get lost in procedural cracks between the custodian and the investors that actually own the bonds. For example, the investor may never actually receive the notice. Alternatively, a bond may be held in an investment vehicle or a dormant investment portfolio with no active investment team, or perhaps is being used as collateral in a repo with a counterparty that doesn’t take ownership of the voting rights.
The problem is exacerbated further in large transactions because a single investor often holds the smaller mezzanine and junior tranches, so a single individual may well hold up the entire vote on a multi-billion transaction – either inadvertently or deliberately.
A classic example is Honours No.2 Plc. This program financed a £1bn pool of the older-style UK student loans (where the UK Government essentially guaranteed ultimate payment of the loans) when it was first issued in 1999. However, its value decreased to around £420m when it was refinanced in 2006 and is not due to finally mature until 2029.
The Issuer and the Trustee have made a number of attempts to switch the benchmark, but the deal has six tranches, with most of the mezzanine and junior tranches now less than £15m in size. Unfortunately, all of those attempts have failed because they were unable to achieve the required quorum of responses in a number of those smaller tranches. This whole failed process has cost time and resources, all while the clock has been ticking.
Thankfully, the newly passed Critical Benchmarks Bill has come to the rescue, and so for at least the immediate future, the Issuer will be able to adopt the use of synthetic LIBOR. Otherwise, each of the tranches would have remained stuck for the rest of their lives at the final coupon rate set in 2021 – not ideal given the expectation for rising interest rates and a particularly bitter blow for those larger tranches that did achieve a quorum in favour of switching to SONIA, but were thwarted by those that didn’t!
Furthermore, it is a little embarrassing for a transaction that has an element of government sponsorship – albeit somewhat remote – to be unable to court enough investor attention to implement a government/regulatory policy change!
This situation couldn’t have occurred if the original documents had contemplated that a relatively benign (and in this case regulatorily imposed) change to the benchmark should not need a majority vote in favour for each tranche of notes, but simply no significant objections from noteholders at the overall transaction level.
To try to avoid these scenarios, particularly after the LIBOR cessation announcement but when the details of its replacement were still undecided, structurers adapted the voting rights language to include the concept of “negative consent”, whereby certain changes can occur unless a given percentage of noteholders vote against them. This follows the same process as before, albeit in reverse: investors must still vote, but only a negative vote impede change.
Clearly, the negative consent language is a practical solution in a context where garnering investor votes may be difficult and potentially a pillar for future proofing structures. However, whilst it has become practice, it is not really a market standard nor governed by regulation. Therefore whilst it may work in certain instances, it may not work in all.
While LIBOR transition may soon effectively be over, the myth of future proofing securitisation documents is probably on the same level as the squaring of the circle and may indeed represent an impossible task. One of the keys to achieving resilience, if not full impregnability, would be to make a structure as flexible as possible. To achieve that, the key is in the controls.
Whilst it’s unlikely that any documentation will ever be perfect, we hope that the lessons learnt from the LIBOR transition process will mean that the relevant industry bodies, issuers and their legal experts, will focus their energies and resources to design a set of voting rules that all issuers can apply in the future to allow changes to be made without excessive effort or expense while retaining protection of all noteholders’ interests.
Overall, the ABS market has done a pretty good job of implementing a change foisted upon it due to unforeseeable circumstances outside of its scope. Nevertheless, it does run the risk that without taking steps itself, the regulators may turn their attention to the subject of voting rights. As observed in many other areas of the market, the imposition of new regulation, even if it is well-meaning, can often turn out to be more complex and overbearing and ultimately hamper rather than help.