LIBOR Transition Clock Ticking
Work has continued over the last few months on the transition away from LIBOR, albeit with some recognition on the part of policy-makers that COVID-19 has led to some delay and loss of focus. There have been a number of developments in terms of data provision as well as in terms of documentation being proposed by banks.
This note is intended to update clients and set out what we see as key areas for consideration. We have previously published two versions of a more detailed paper and the more recent version, from October 2019, remains on the website by way of providing more historical context than is included here. It can be found at https://www.centrusfinancial.com/thought-leadership/white-papers/.
For the purpose of this paper we assume that it is understood that the move in Sterling is from LIBOR to a risk-free rate of SONIA, the Sterling Overnight Index Average, along with a level of understanding of the “mechanical” implications, for example the fact that SONIA is a daily backwards-looking publication. This paper is focussed on GBP but readers should be aware that similar changes are being made in relation to “IBOR” rates in other currencies.
Current state of play
The following table summarises our assessment of the current state of play:
The main developments in the course of 2020 have arguably been more of the nature of “thinking” than “doing”. However, the Bank of England has begun to publish a SONIA compounded index and its “Working Group on Sterling Risk-Free Reference Rates” (the “Working Group”) has continued to forge a consensus on how SONIA will be used in cash products. This primarily relates to bank loans, being the majority of the volume and given that floating-rate note (“FRN”) issuers were already making progress transitioning FRNs to SONIA. The extra mark in the above table for “legacy loans” represents the fact that a small number of major lenders are starting to make necessary amendments to documentation when legacy loans are being refinanced or restated, with prescriptive fallback language for example.
There is a risk that borrowers are “caught on the hoof” when making other changes, whilst at the same time in many cases we are seeing limited evidence yet of a pro-active strategy.
In relation to derivatives, the International Swaps and Derivatives Association (“ISDA”) has continued to work on methodologies such as the work on pre-cessation triggers and development of a “protocol” for parties to ISDA agreements to use to transition the very substantial volume of LIBOR-referencing ISDA transactions without needing to negotiate each one separately. This protocol is due to be launched shortly. ISDA have also started to publish a regular “RFR Adoption Indicator” which, as of July 2020, suggests that there is still quite a long way to go albeit Sterling shows the highest percentage of rates risk being traded using SONIA of all relevant currencies (at 32% with the global figure only 6.8%).
The key actions for market participants noted by the Working Group are:
- By end-Q3 2020 SONIA loans should be available, along with contractual conversion mechanisms in new or refinanced LIBOR products.
- By end-Q1 2021, banks should have ceased issuance of LIBOR loan products that expire after the end of 2021.
- Market participants need to establish a clear framework to manage the transition of legacy LIBOR products, to accelerate reduction of stock of GBP LIBOR referencing contracts by end-Q1 2021 and complete active conversion where viable by “end-Q2/3 2021”.
- The final date is that we collectively cannot rely on LIBOR being published after the end of 2021 and it remains reasonable in our view to assume that it will not be.
Based on our discussions with clients and lenders, our understanding is that some lenders are really now starting to engage while others have a long way to go. “End Q3 2020” is only a month or so away. In the mainstream market NatWest and Lloyds appear to be further advanced than others as of the date of this note and going by the evidence we see from clients. We are also seeing some evidence of lenders, perhaps with less sophisticated treasury functions or seeking to keep matters simple for some clients, considering a move to using the Bank of England Base Rate (rather than SONIA).
As lenders complete their own preparations, if borrowers do not play their part then in our view it is likely they will find themselves faced with a ticking clock and limited appetite from lenders to interrupt a standardised process. Having said that we have not seen significant evidence of banks seeking to use the transition from LIBOR to SONIA for their commercial advantage.
The risk of being a “taker” of terms is that the details may be more clunky or less clear for individual borrowers, and in some cases less attractive economically, than should be the case if “negotiated”.
- Term structure
The role of a “Term SONIA Reference Rate”, with a suite of forward rates published daily like happens for LIBOR (i.e. daily publication of “one / three / etc-month forward SONIA”) remains a sticky issue and the jury is still out on whether it will be that relevant. Conversely, the Bank of England “Base Rate” has been a comeback as a concept, albeit possibly just for smaller borrowers and/or a limited range of products for example trade finance.
Having said that, there are now three data providers who have started publishing forward-looking term rates, based on SONIA OIS contracts, with a fourth still work-in-progress. For borrowers who value the concept of a rate known at the start of each roll period more than they value really understanding where it comes from, it may be an option depending on where the views of their lenders settle. Lenders seem a bit sceptical, partly because it opens up somewhat similar abuse risks, or perceptions thereof, as the LIBOR model it replaces. Our advice is to exercise caution in this regard but we will know more in due course.
- Spread adjustment
Banks do not any more borrow on a term unsecured basis from each other in volume. But if they were to the credit risk is now recognised as being not zero. Daily SONIA is nearly risk-free so a margin on a SONIA loan should be higher than one on a LIBOR loan. For legacy loans an explicit margin adjustment (increase) is required.
This is, clearly, a meaningful commercial point. It is more material for some borrowers than others and the term of the remaining facility potentially exacerbates the materiality. For example, many borrowers have loans which are at a margin over LIBOR which could not be obtained in the current market (i.e., there is value for the borrower) but still with decades to run.
The following graph demonstrates the issue, comparing SONIA compounded on a one and three-month basis vs. the equivalent GBP LIBOR rates over the last decade:
Unsurprisingly, compounded SONIA is much the same whether compounded over one month or three (with small delays in the longer compounding period rate visible after major movements). LIBOR is significantly higher than either, with three-month LIBOR higher than one-month. Six-month LIBOR, not shown on the graph, is broadly as much higher again.
Based on the above, the 5-year median difference between compounded SONIA and one-month LIBOR is four basis points and for three-month LIBOR the difference is twelve basis points, using data as at 20th August 2020.
The commercial principle, if one is to borrow from the ISDA discussions in the derivatives market, should be to base the margin adjustment on the median difference between SONIA and the relevant LIBOR benchmark over the last five years, at the point of transition. According to the Working Group consultation has:
Working Group Consultation identified a strong consensus in favour of the historical 5 year median approach … in line with [ISDA], as the preferred methodology for credit adjustment spreads across both cessation and pre-cessation fallbacks for cash products maturing beyond end-2021.
We support this as a reasonable basis for agreeing the spread adjustment.
Having said that we should also recognise that some transactions have already taken place using the differences in the forward rates i.e. between “fix vs. LIBOR” swaps and “fix vs. SONIA” swaps. The numbers are similar – at the time of writing market data would imply the following using forward curves:
Looking at the data, for a loan which includes the right to roll on a one-month basis, the four basis points arising from the historical look-back is a reasonable outcome for the spread adjustment.
In principle, one would want to test the differential at the time of both negotiation and transition in order to ensure transparency and benchmarking of the spread adjustment. Having said that, in our view, for a loan which is not hedged, or where it is but one has managed to get comfortable with any relatively minor mismatch from an economic and financial accounting perspective, it is reasonable to take a practical approach to the precise timing. With a five-year look-back, it does not seem critical whether the adjustment is fixed now, or at the end of 2021, or at any time in-between. Some banks are seeking to agree the number now and we think that is ok so long as it is benchmarked on the above basis.
Therefore, for the majority of corporate facilities which are not hedged using standalone derivatives, it is possible to agree the commercial position now. All that is missing, for most banks and most borrowers, is operational and legal preparedness to implement.
- Lags + Shifts
For most borrowers and certainly larger ones, daily SONIA will be used on the basis of being compounded over interest periods, with a delay of five days between the “Interest Period” (ending with the payment date) and the “Observation Period” from which SONIA data will be drawn. This mitigates the operational challenge of instructing payments on the same day as the value becomes known.
However, there is a nuance around how weekends and holidays are dealt with and whether days with a “weighting” of more than one day have that based on the actual day the data came from or the equivalent day in the Interest Period. If it is agreed that the weighting is driven by the Observation Period then it becomes possible to produce a “compounded SONIA index”, like for RPI or CPI data, and people can use one day’s index number divided by the one from the start of the period to calculate the underlying compounded SONIA rate to which margins are added.
Happily, the Bank of England has said that it will publish exactly such an index. Indeed, it has started to and it can now be found on their website. By 20th August the index level was 101.3029551 from a base of 100 in April 2018. If this achieves status as a reference point for loans, then borrowers may wish to use that and as far as documentation goes we should all be “shifting” rather than “lagging” the calculation period.
We are now starting to see documentation and heads of terms being proposed by banks. From what we have seen to date we would flag the following points:
- Lenders do not appear in all cases fully to have signed up to the above methodology for setting the spread adjustment. As noted, this is a potential area of commercial discussion.
- We do not believe that very vague language allowing lenders to impose their standardised approach in due course is appropriate and it has been demonstrated by some lenders that clear drafting is now possible. Borrowers should ask their advisers, legal and financial, whether they are comfortable with proposed drafting.
- Lenders do not appear to have settled on shifts vs. lags but it makes sense, given the immateriality of the difference, either to focus on shifts or to permit whatever is easiest as agreed between the parties without needing to amend facility agreements again.
- Role of Bank of England: we think that it is reasonable to reference the Bank of England index (or rates derived from it) as a data source, and the Base Rate as a fallback if SONIA becomes unavailable. We have seen both feature in documentation now.
- It may be for smaller corporates that Base Rate takes the place of SONIA, but we see no substantive reason for such borrowers to favour this (other than perhaps where a lender is unable to offer SONIA and it is undesirable to refinance). We also do not really see a role for hanging on for Term SONIA to be properly established; shifted compounded SONIA (using the Bank of England index) will likely seem part of the furniture before long.
One further comment on fall-backs. It is important to document a fall-back to cope with SONIA not being published, for example the Bank of England rate. Most agreements contain provisions around “market disruption”, similar to the “increased costs” concept (re regulatory changes). Typically if lenders are not able to fund themselves at LIBOR they have been contractually entitled to pass on the cost to borrowers. There is an argument in our view that in a SONIA world, lenders should not be able to pass on costs associated just with their own position in the market. This is a commercial point which is not yet settled. The practical point is perhaps, as has been the case in the past, to be prepared to push back in future should a lender invoke market disruption and particularly if due just to their own financial circumstances.
What should happen now?
It does now make sense to be pro-active in seeking acceptable SONIA language on new and refinanced agreements. It also makes sense to raise the question of timing for transition of legacy documents. The time is fast approaching, even if any individual business may have its own “bandwidth constraints” for what is a somewhat technical and operational matter, for implementation across a portfolio.
For borrowers with derivatives the situation is complicated slightly by ISDA’s timetable and the nuances of hedge effectiveness and hedge accounting. It will be necessary for auditors to be comfortable with the details and it would make sense to seek their view before agreeing to changes; if nothing else that will smooth the audit path in future years. We do not really see material issues in substance and many corporates continue to wrestle with other implementation nuances relating to FRS 102 and hedge accounting which are more substantial than this. But on balance it is logical to start with loans which are not hedge accounted, or to coordinate the timing and commercial position across loans and hedging instruments (as some corporates have done).
Some lenders are and will remain “followers” of the market. If smaller lenders are simply not yet in a position to offer clarity borrowers have little choice but to wait. This does not however stop progress being made with counterparties who are more advanced in their preparations.
Lastly, we should flag that central government has recognised more of a role for central regulatory diktat. A May 2020 Working Group paper proposes that the UK Government consider legislation to address “tough legacy” exposures. We will see in due course whether that happens and how broad the net of “tough legacy” contracts is. The Chancellor Rishi Sunak also announced on 23rd June 2020 that government will legislate to increase FCA’s regulatory powers to manage an orderly transition. He contextualised by flagging that:
Rishi Sunak - The Chancellor Parties who rely on regulatory action … will not have control over the economic terms of that action. …. This reinforces the importance of parties who can transition away from LIBOR doing so on terms that they themselves agree with their counterparties.
So there appears to be recognition now that FCA may be required to adopt a role in clearing up any mess left from commercial inaction. But borrowers with non-standard positions (such as very long-dated loans) may find that their interests are not fully considered and we do not advise planning to rely on this last recourse.
In terms of practical next steps therefore we advise borrowers to:
- As above, be pro-active in seeking SONIA language on new and refinanced agreements (seeking guidance, as appropriate, from legal and financial advisers).
- Consider the materiality of any mismatch issues in relation to derivatives; perhaps focus on un-hedged loans in the first instance or coordinate hedging and hedged instruments.
- Formulate a view on timing in relation to legacy LIBOR agreements, covering the timing of (i) transition to SONIA and (ii) documentation of transition to SONIA. Documentation necessarily comes before action, but will you wait for cessation or pre-cessation “trigger” by the relevant authorities for the actual changeover, which probably means end-2021, or get on with the practicalities once the mechanics are settled?
- If lenders propose their own “standard” language, how demanding will you be on clarity? How important is complete clarity if making a change before all the dust has settled (how ‘valuable’ are your existing facilities and what is your view on setting precedents)?
- While the expectation is of a consensual process, it may be useful to review and document existing fall-back language in facility agreements, to understand the existing commercial position. This is not a major undertaking and we would advise doing this in any event.
Please contact us if you would like to discuss your preparedness for the LIBOR transition and how the team at Centrus can help.