SONIA Transitioning to a New Era
11 April 2022
11 April 2022
The global financial crisis marked the beginning of the end for the London Interbank Offered Rate (“LIBOR”) – a benchmark interest rate which, as at the start of 2021, reportedly underpinned financial contracts totalling $265 trillion globally. Regulatory concern grew over the years that followed the global financial crisis, due largely to the lack of activity in the underlying markets on which the submissions made to generate the various LIBOR rates were based. Consequently, LIBOR was being sustained only by the use of “expert judgement” by the LIBOR panel banks, and not by reference to real market data. As a result, the panel banks became increasingly uncomfortable with providing submissions with so little actual borrowing activity on which to base those judgements. Indeed, the Bank of England reported that, in 2017, on average there was only £187 million of three-month deposits per day, notwithstanding that three-month LIBOR was the most widely used tenor for GBP LIBOR.
Given the concerns around the reliability and robustness of LIBOR, in July 2014 the Financial Stability Board (“FSB”, an international body that monitors the global financial system) recommended the development and adoption of nearly risk-free reference rates (“RFRs”) as potential alternatives to LIBOR. In response to those recommendations, working groups were set up in each LIBOR currency jurisdiction to consider potential alternative benchmark rates. These groups included the Working Group on Sterling Risk-Free Reference Rates (“Sterling Working Group”) in relation to GBP LIBOR, and, in relation to USD LIBOR, the Alternative Reference Rates Committee (“ARRC”).
In the UK, the Financial Conduct Authority (“FCA”), which had regulated LIBOR since April 2013, also had concerns about the representativeness and robustness of the rate in light of the lack of liquidity in the underlying market. In July 2017, it concluded that LIBOR was no longer sustainable, and set in motion the obligation to transition away from LIBOR by the end of 2021. Transitioning the market in such a relatively short time frame created an enormous task, with the working groups working closely with regulators, trade associations and market participants in order to drive the timetable for transition.
One key question in relation to the transition exercise was: which alternative rate would replace LIBOR in each relevant jurisdiction? In relation to sterling, the Sterling Working Group recommended the Sterling Overnight Index Average (“SONIA”) as its preferred RFR. SONIA was not a new benchmark; it had previously been used as a benchmark for overnight unsecured transactions denominated in sterling. It was chosen as an alternative rate to LIBOR by the Sterling Working Group on account of its robustness and, in stark contrast to LIBOR, the fact that it is based on an active, liquid underlying market.
SONIA differs from GBP LIBOR in a number of key respects, as discussed further below. Like other RFRs, SONIA is a backward-looking overnight rate. This was one of the more difficult early issues to be grappled with by the debt market, with the challenge being how to overcome the operational difficulties presented by using backward-looking rates for lending products that had been developed around the use of forward-looking term rates, like LIBOR.
As market participants worked towards the deadlines set by the various working groups for transition to RFRs, it became clear that these targets would only be achievable with the help of standardised industry documentation that addressed the operational challenges presented by backward-looking rates. In the UK, this process was driven by the Loan Market Association (“LMA”), who produced a number of “Exposure Drafts” of loan documentation referencing both SONIA compounded in arrears and SOFR compounded in arrears (SOFR being the secured overnight financing rate and the RFR for US dollars). Following market feedback over several months, the Exposure Drafts were amended and published in “recommended form” by the LMA.
Prior to the Exposure Drafts being produced, Ashurst worked on the first SONIA compounded in arrears loan in the market, acting for National Westminster Bank plc on a revolving credit facility for National Express. The methodology used in that facility formed the basis of the market norm, now supported by the LMA recommended forms for SONIA compounded in arrears and adopted through the major transition exercise undertaken in 2021 on English law-governed GBP LIBOR loans.
Ashurst also assisted in the early days of transition in the capital markets space, advising Santander on the transition of certain LIBOR-linked floating rate notes issued by its Holmes Master Issuer programme to SONIA.
At the time of writing in Q1 2022, the loan market has adopted SONIA compounded in arrears as its alternative to GBP LIBOR (save for certain loan products detailed below).
As noted above, LIBOR was a forward-looking term rate available for specified currencies and tenors (e.g. one, three or six months). A borrower of a LIBOR-based loan knew at the point of drawdown the rate of interest applicable to that loan and therefore the amount of interest that would be due at the end of the next interest period.
However, as SONIA is an overnight risk-free rate it does not factor into its pricing any term risk (i.e. risk attributed to the length of borrowing) or credit risk (i.e. risk of lending to a particular borrower). Accordingly, SONIA compounded in arrears was developed as a methodology to address the risk-free nature of SONIA.
For a loan based on SONIA compounded in arrears, the interest rate for each day of the interest period is based on the SONIA rate for that day and aggregated for the applicable period using a compounding formula. This means the interest payable on the loan increases in line with the tenor of the loan, thus incorporating an element of term risk.
SONIA for any particular business day is published at 9 a.m. (London time) on the following business day. Therefore, applying the SONIA rate for each day of an interest period would mean that the amount of interest owing for that interest period cannot be calculated until the business day following the last day of the applicable interest period.
In order to mitigate this issue, the period for which the daily SONIA rates are used in the formula is shifted back slightly by what is commonly referred to as the “Lookback Period”. The loan market has broadly settled on a Lookback Period of five “RFR Banking Days” (being a day on which SONIA is published) such that the amount of interest owed for an interest period can be calculated on or after the fourth business day prior to the end of the interest period. While this leaves a borrower with a relatively short notice period before it has to pay the accrued interest on the last day of the interest period, particularly when compared to a forward-looking rate, given that – to date at least – SONIA has been a relatively stable rate it has been possible for borrowers to estimate their likely interest cost for the relevant interest period using publicly available calculators.
While SONIA compounded in arrears introduces an element of term risk, it is not an exact replacement for LIBOR, and does not result in a similar rate when compared side by side.
A loan that was initially originated as a LIBOR-based loan will have been priced on this basis and this will be reflected in the margin. Accordingly, the pricing for a loan which is being transitioned from LIBOR to SONIA compounded in arrears will need to be adjusted to avoid any transfer of economic benefit between the lender and the borrower. This adjustment is commonly referred to as the “Credit Adjustment Spread”. A Credit Adjustment Spread is not required for loans originated on the basis of SONIA compounded in arrears but it may be included for commercial reasons. The methodology for calculating the Credit Adjustment Spread was subject to a lot of market scrutiny as lenders were very focused on ensuring customers were treated fairly when transitioning loans from LIBOR to SONIA compounded in arrears. The two primary methodologies for calculating the Credit Adjustment Spread which emerged were (i) the five-year historical median approach, and (ii) the forward approach.
The forward approach was calculated as the linear interpolation between differing tenors of GBP LIBOR vs SONIA swaps. A Credit Adjustment Spread based on the forward approach would be calculated on or shortly prior to the transition to SONIA compounded in arrears. However, following the cessation of GBP LIBOR on 31 December 2021, the lack of LIBOR data means the methodology underpinning the forward approach calculation can no longer be used.
In respect of the five-year historical median approach, the Credit Adjustment Spread is based on the median difference between GBP LIBOR for a particular interest period and SONIA compounded in arrears over the same period calculated over a five-year lookback period. Following the announcement by the FCA on 5 March 2021 in relation to future cessation and loss of representativeness of LIBOR benchmarks, Bloomberg published the rates using the five-year historical median approach for each tenor and currency of LIBOR as at 5 March 2021. These “Bloomberg” rates have seen widespread adoption by the market since their fixed nature provided parties with transparency and certainty.
At the early stages of the transition process, SONIA compounded in arrears-linked loans incorporated the Cumulative Compounded Rate (“CCR”) methodology, which only allows for the calculation of interest for the whole interest period using a single compounded rate. However, the CCR methodology did not easily cater for intra-period events such as prepayments or mid-interest period transfers by lenders.
As a consequence, the Daily Non-Cumulative Compounded Rate (“NCCR”) was developed. The NCCR is a daily compounded rate which is derived from the CCR, i.e. CCR as at the current day less CCR of the prior banking day. This allows for the calculation of daily interest for that day or days, which was required to easily facilitate prepayments and loan trading.
Historically, LIBOR-referencing loan agreements contemplated fallbacks to substitute a rate for LIBOR in the absence of its publication. The fallbacks were considered to be short-term solutions and included use of the most recently published historic LIBOR rate, interpolating a rate from the published rates for other tenors and, ultimately, each lender’s own cost of funds.
As the calculation of SONIA compounded in arrears relies on there being a SONIA rate on each RFR Banking Day, alternative fallbacks needed to be identified for a scenario where SONIA is not actually published on a particular date. The sterling market has adopted a primary fallback to the Bank of England’s Bank Rate, adjusted by reference to the spread between the SONIA daily rates and the Bank Rate(s) over the most recently preceding five RFR Banking Days on which both rates are available. Certain financing arrangements also include a further fallback to each creditor’s cost of funds, were the Bank Rate to be unavailable. Where this fallback has not been included, the relevant creditor is considered to be comfortable with the very low probability of both the SONIA daily rate and the Bank Rate no longer being available.
At the time SONIA was endorsed as the appropriate successor to GBP LIBOR, it was recognised that while a backward-looking compounded in arrears methodology was the appropriate successor to GBP LIBOR in many cases, there were some loan products that required a forward-looking term rate. Accordingly, it was recognised that a term rate based on SONIA would need to be developed for these limited circumstances.
Following consultation and discussions with the Sterling Working Group, three benchmark administrators made available “beta” Term SONIA reference rates. After six months of an initial “beta” rate, in January 2021 two providers began publishing Term SONIA without the “beta” status.
Ahead of the publication of these Term SONIA reference rates, the Sterling Working Group’s Term Rate Use Case Task Force had set out guidance as to the appropriate use of Term SONIA given the Sterling Working Group’s previous recommendation of a broad-based transition to SONIA compounded in arrears. In January 2020, the Sterling Working Group recommended a limited use of Term SONIA. The areas identified as being potentially appropriate for Term SONIA uses were:
In light of the limited uses identified above, the offering of Term SONIA-based products by lenders has been limited with the vast majority of commercial parties transitioning to SONIA compounded in arrears, as envisaged by the Sterling Working Group.
In contrast to Term SONIA, Term SOFR has seen fairly widespread adoption as an alternative to USD LIBOR. On 29 July 2021, ARRC formally recommended CME Group’s forward-looking SOFR term rates. ARRC did not impose similar restrictions on the use of Term SOFR as seen with Term SONIA.
In addition, financing transactions were entered into during 2021 using alternative term credit sensitive benchmarks, including the American Interbank Offered Rate (AMERIBOR), Bloomberg short-term bank yield index (BSBY) and HIS Markit USD Credit Inclusive Term Rate (CRITR), though these alternative credit sensitive rates have seen limited adoption in the US market.
For less sophisticated borrowers, the Bank of England’s Bank Rate (also known as base rate) has been adopted as an alternative rate to GBP LIBOR. The Bank Rate, published by the Bank of England, determines the interest rate the Bank pays to commercial banks that hold money with it. Typically in the lending market its use has been limited to retail customers and less sophisticated borrowers who prefer the use of this rate due to its simplicity and (historically) its lack of volatility.
Turning back to the methodology for calculating SONIA compounded in arrears described above, some initial discrepancies arose between the loan market and the derivatives market, tracking through from the documentation developed by the LMA and the International Swaps and Derivatives Association (“ISDA”) (in the case of the derivatives market). Two notable examples are:
Of course, the ISDA or lending documentation could be tailored to align the methodology, but this adds complexity to the relevant creditor or hedge counterparty in providing an off-market product.
Prior to GBP LIBOR’s cessation on 31 December 2021, regulators were also required to consider the likely scenario that significant volumes of loans and other financial instruments would continue to reference GBP LIBOR after the cessation of the relevant LIBOR rates. These so-called “Tough Legacy” contracts gave rise to concerns regarding the continued stability of the relevant markets and the potential for claims for breach or frustration of contract as a result of the cessation of the relevant benchmark.
In the UK, the FCA set up a Tough Legacy Task Force with a mandate to identify what might constitute “Tough Legacy” and propose options and recommendations for dealing with Tough Legacy contracts. By way of example of such a Tough Legacy arrangement, there are GBP LIBOR-linked mortgages with no fallback in the terms of the contract if GBP LIBOR ceases to be available. Such mortgages would be difficult to transition to a replacement reference rate due to the large number of customers and the challenge in using a complicated replacement rate while treating customers fairly. Similarly, there were loan books and bonds where, despite best efforts, the transition away from GBP LIBOR did not occur by 31 December 2021.
Under the UK Benchmarks Regulation (“BMR”), “supervised entities” were granted an exemption from the prohibition on using one-, three- and six-month GBP and JPY LIBOR in respect of any legacy contract (other than cleared derivatives) that had not transitioned from LIBOR by 31 December 2021. This exemption does not permit the use of LIBOR in any new contract. Using its powers under the BMR, the FCA was also able to compel the administrator of LIBOR, ICE Benchmark Administration, to publish these benchmarks on a non-representative, synthetic basis. These provisions are in place until the end of 2022, after which time it has been confirmed that the synthetic JPY LIBOR rates will be discontinued, and the use of synthetic GBP LIBOR rates may become restricted, noting the maximum duration for which synthetic GBP LIBOR rates could be continued is 10 years. These announcements were only made in Q4 2021, which encouraged the transition of as many GBP LIBOR-based contractual arrangements as possible.
In December 2021 the Critical Benchmarks (References and Administrators’ Liability) Act 2021 came into force, which applies the Tough Legacy regime outlined above to any English law contract that still references one-, three- or six-month GBP or JPY LIBOR. The FCA confirmed in February 2022 that, during the course of 2022, it will seek views on retiring one-month and six-month synthetic GBP LIBOR at the end of 2022, and on when to retire three-month synthetic GBP LIBOR. As a result, the same challenges noted above continue to apply and motivate creditors to transition away from this synthetic LIBOR rate.
This Act also incorporated safe harbour provisions protecting both users of synthetic LIBOR and ICE Benchmark Administration, restricting parties from asserting potential rights under “material adverse change” clauses and from otherwise claiming for breach or frustration of contract as a result of the cessation of the relevant LIBOR rates.
The synthetic rates are calculated using the methodology underlying the SONIA compounded in arrears rate, by taking the forward-looking term rate based on the relevant RFR (i.e. SONIA or, in the case of Japanese Yen, TONA) and then adjusting this rate by way of a fixed credit adjustment spread, using the Bloomberg five-year historical median credit adjustment spread for the applicable currency and interest period tenor. The term SONIA rate is calculated based on the fixed rates offered in the SONIA-referencing derivatives markets for the three forward-looking periods of one, three and six months.
The Bank of England has estimated that, as of February 2022 and across all asset types, less than 2% of the total GBP LIBOR legacy stock remains. The FCA also confirmed that, as of the same date, new SONIA lending now exceeds £100 billion across a diverse range of sectors and facility types. This suggests that the transition to SONIA is now largely complete. However, this is only part of the overall LIBOR transition process. While the UK’s Tough Legacy regime only applies to the above GBP and JPY LIBOR rates, it remains to be seen whether a similar approach will be taken to some or all of the USD LIBOR tenors when they cease to be published in June 2023. On the assumption that it will be, it is worth bearing in mind that any extension of the Tough Legacy regime under the BMR will only apply to English law-governed contracts. Solutions to the Tough Legacy problem in other jurisdictions will be determined by regulators in the relevant jurisdiction.
The UK approach can be contrasted with that of the State of New York in relation to New York law-governed Tough Legacy contracts. Rather than providing for a synthetic rate, any references to USD LIBOR in these contracts will be mandatorily replaced, by operation of law, with references to (adjusted) SOFR. However, these automatic replacement provisions will not apply where the documentation allows for the selection of (or interpolation between) unaffected LIBOR tenors. For the purposes of this legislation, Tough Legacy is much more narrowly defined than in the BMR, and will only include those New York law-governed contracts, securities or instruments that either contain no fallback provisions or fall back to a replacement rate based on USD LIBOR. Contracts which contain fallbacks to non-LIBOR-based rates will therefore be excluded.
The rates for the remaining USD LIBOR tenors will cease to be published after 30 June 2023 and for deals using these rates, they will need transitioning to a new benchmark rate prior to that date. We know that the synthetic rate for JPY LIBOR will cease to be available after 31 December 2022 and that the synthetic rate for GBP LIBOR will cease at some point in the next 10 years. We can expect these differing deadlines to impact the priority given by creditors to transitioning remaining legacy deals.
For now, EURIBOR continues to be the benchmark of choice for euro-lending, notwithstanding the availability of ESTR as the euro RFR, following the cessation of the publication of EONIA.
Benchmark reform remains a focus in other jurisdictions. For example, in Singapore the cessation of publication of the benchmark rates of SOR in June 2023 and SIBOR in December 2024 means the active transition of SGD-denominated financial instruments to SORA, the SGD RFR, will continue through 2022 and beyond. Similarly, in Canada, the Canadian Alternative Reference Rate working group recommended the cessation of the calculation and publication of CDOR after 30 June 2024. At the time of writing, Refinitiv (as administrator of CDOR) has issued this proposal for consultation.
The move away from LIBOR was one of the biggest transitions in the financial markets in decades. The experience changed the way in which the financial and legal markets operate now, and presented many lessons that can be used for any future regulatory overhaul.
Making an early start in alternative reference rates was critical. Yet being a “first mover” when clarity was limited and liquidity was lacking, risked creating unnecessary sunk costs. As noted above, NatWest entered into initial RFR lending transactions, assisted by Ashurst, on a bilateral basis. But to arrange a syndicated RFR loan required loan market education and consensus, and industry documents were essential to help lenders find common ground.
In the UK, the Sterling Working Group actively engaged with each of the other currency working groups and with the FSB’s Official Sector Steering Group on benchmark reform (which worked to coordinate action at an international level) to encourage and support the transition away from LIBOR.
In September 2020, the Sterling Working Group published recommendations for SONIA loan market conventions. Based on those conventions, the LMA developed documentation containing a framework for compounded RFRs. This included the Exposure Drafts mentioned above, which incorporated rate switch provisions enabling the switch to SONIA compounded in arrears at a future date, for market participants that were unable to enter into transactions based on SONIA initially but anticipated being able to do so at an agreed date in the future. As operational capacity to enter into loan transactions based on SONIA compounded in arrears increased, the LMA published additional draft documents that contemplated lending on SONIA compounded in arrears from the outset.
To ensure that the milestones set out in the Sterling Working Group’s roadmap for transition were met, a huge volume of transition activity was required in relation to legacy loan facilities referencing GBP LIBOR. Lenders had to identify all existing LIBOR exposures that would be impacted before determining which approach to take: renewal, which would require day-one RFR mechanics, or migration, which would require amendment (likely across multiple documents) in accordance with the contractual terms of each agreement. Consideration had to be given to the comprehension of the transition by less sophisticated customers, the form of the transition documentation employed, and the way in which later amendments would be made, if required. This created pressure on resources and on internal processes for all market participants.
As a result, override agreements were developed by some lenders for use with LIBOR transition and proved popular as a means of transitioning large numbers of less sophisticated customers. Under the override approach, the terms referencing the calculation of interest and all related provisions (such as default interest, break costs, market disruption and so on) were set out in the override agreement such that any provisions in the underlying facility agreement which were contrary to, or inconsistent with, the overriding terms were deemed to be amended and restated in accordance with those overriding terms. The override agreement could also be applied to multiple loan arrangements between the parties.
In order to further simplify the transition process, institutions also reconsidered requirements which would otherwise be typical for amendments, such as the provision of corporate approvals, legal opinions and the signatures of each guarantor. Most often, smaller banks classified their customers by size and proceeded with fewer requirements from lower risk customers.
As could be expected, not all customers were familiar with the fundamentals of reference rate reform, and they needed to understand how their lenders’ plans would impact them during and after the transition. Providing education and support for SME customers presented lenders with conduct risk concerns, in order to avoid mis-selling, negligent advice and poor management of conflicts of interest.
Both the Association for Financial Markets in Europe and The Fixed Income, Currencies and Commodities Markets Standards Board helpfully provided guidance on how to tackle conduct risk in the context of LIBOR transition, focused almost entirely on communications.
As part of the transition process, technology systems that referenced LIBOR needed to be made operation-ready for a post-LIBOR world. This included new calculation systems to ingest the daily SONIA rate, provide the compounded SONIA rate and support the new Bloomberg SONIA forward curve, to name a few. The Market Infrastructure sub-group of the Sterling Working Group established a list of infrastructure and operational readiness issues and engaged with technology and infrastructure firms via panels and roundtables to identify solutions to those issues.
A fundamental part of the LIBOR project was large-scale review of the legacy documents in order to determine how to execute the move away from LIBOR to an RFR. Legal artificial intelligence (“AI”) was used by law firms to analyse the content of large volumes of contracts that referenced LIBOR to find (for example) where those references appear, the process for amendments, consent rights and associated documents that may also require amendment. AI had been appearing in lawyers’ everyday workflows long before LIBOR transition started. For example, legal software was being used to streamline contract review and document analysis by quickly identifying common clauses and potential outliers, and to compare contract terms to market-standard provisions. The experience of LIBOR transition has proven beyond doubt that AI can nurture greater collaboration between clients’ operations and their law firms. Going forward, ongoing client needs will drive adoption of many further AI applications.
Authors: Tim Rennie, Darren Phelan, Katharine Tuohy, Sarah Curry