FS21/11: Article 23D BMR decision for 6 sterling and yen LIBOR versions

Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Feedback Statement FS21/11 November 2021

2 FS21/11 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Moving around this document Use your browser’s bookmarks and tools to navigate. To search on a PC use Ctrl+F or Command+F on MACs. This relates to Consultation Paper 21/19 which is available on our website at www.fca.org.uk/publications Sign up for our news and publications alerts See all our latest press releases, consultations and speeches. Contents 1 Summary 3 2 Decision on whether to use the power 9 3 Decision on how to use the power 11 4 Other issues 18 Annex 1 FCA requirements to be imposed on IBA under Article 23D(2) 21 Annex 2 Abbreviations used in this paper 23

3 FS21/11 Chapter 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 1 Summary 1.1 In June 2021, we published a consultation on our proposed decision to use our Article 23D(2) power under the Benchmarks Regulation (BMR) for 1-month, 3-month and 6-month sterling and Japanese yen LIBOR settings (the 6 LIBOR settings). The current system of calculating these LIBOR settings based on submissions from panel banks will cease at end-2021. We proposed to require the administrator of LIBOR, ICE Benchmark Administration (IBA), to publish these 6 LIBOR settings under a changed, ‘synthetic’ methodology for a limited period beyond end-2021 for use in legacy contracts. This ‘synthetic’ methodology would be based on the relevant forward￾looking term risk-free rates (RFRs) and the ISDA spread adjustments for the relevant LIBOR settings. 1.2 Our proposal to use our Article 23D(2) power intended to ensure that the cessation of each of these 6 LIBOR settings is orderly, and to advance our objectives of ensuring consumer protection and the integrity of the UK financial system. 1.3 The proposed Article 23D(2) decision was contingent on our decision to compel the continued publication of the 6 LIBOR settings by IBA under Article 21(3) and our decision to designate the 6 LIBOR settings as Article 23A benchmarks. 1.4 A large majority of respondents to our Article 23D decision consultation supported our proposed approach to using the Article 23D(2) powers. 1.5 On 29 September, having considered consultation responses, we confirmed our decisions to compel and require IBA to publish these settings under a ‘synthetic’ methodology, based on forward-looking term RFRs and the ISDA spread adjustments, after end-2021 for the duration of 2022. Our requirements under Article 23D(2) will take effect immediately after the final Article 23D(2) Notice is given to IBA and after the Article 23A designations of the 6 LIBOR settings take effect on 1 January 2022. 1.6 This document summarises feedback we received to our Article 23D decision consultation and our responses. Who this affects 1.7 This Feedback Statement will be of interest to: • the administrator of LIBOR, IBA • providers of component inputs for the 6 LIBOR settings under the changed, ‘synthetic’ methodology • regulated and unregulated users of LIBOR

4 FS21/11 Chapter 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions The wider context of this feedback statement Our consultation on proposed Article 23D(2) decision 1.8 Article 23F(1)(d) of the BMR requires us to prepare and publish a Statement of Policy on how we exercise our powers under Article 23D. We must have regard to this Statement of Policy when exercising our powers, and we are obliged to explain how we have taken account of the relevant Statement of Policy. 1.9 We consulted on and subsequently published the Article 23D Statement of Policy in March 2021. 1.10 We further consulted on our proposed decision to use our Article 23D(2) power, in line with the Article 23D Statement of Policy, in June 2021. The consultation closed on 27 August 2021. How it links to our objectives 1.11 Article 23D(3) BMR provides that we may only exercise our Article 23D(2) powers if we consider: • it appropriate to do so having regard to the desirability of securing that the cessation of the benchmark takes place in an orderly fashion, and • it desirable to do so to advance either or both of our statutory objectives to: – secure an appropriate degree of protection for consumers – protect and enhance the integrity of the UK financial system 1.12 Our overview document sets out the background to the BMR and the amendments under the Financial Services Act 2021 to give us enhanced powers to help manage the orderly wind down of critical benchmarks and to support our objectives. End of LIBOR 1.13 We announced on 5 March 2021 that the sterling, euro, Swiss franc and Japanese yen LIBOR panels are ceasing at end-2021. The US dollar LIBOR panel will continue until end-June 2023, although publication of the 1-week and 2-month US dollar settings will cease at end-2021. 1.14 There is a pool of existing contracts referencing each of the 1-, 3-, and 6-month sterling and Japanese yen LIBOR settings that will not have been transitioned away from LIBOR by end-2021. To avoid disruption to these legacy contracts, we announced our decisions in September 2021 to exercise our powers under the BMR to ensure an orderly wind-down and to support our objectives. We said that we decided to: • use our Article 21(3) BMR power to compel the administrator of LIBOR, IBA, to continue the publication of the 1-, 3- and 6-month sterling and Japanese yen LIBOR settings after the panels cease at end-2021 for the duration of 2022 • designate the 1-, 3- and 6-month sterling and Japanese yen LIBOR settings under Article 23A BMR as they will be permanently ‘unrepresentative’ for the purpose of BMR at 00:01 on 1 January 2022

5 FS21/11 Chapter 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions • require IBA to change the way in which each of the 6 LIBOR settings is determined, including the input data, so that they will be published under a changed, ‘synthetic’ methodology from 1 January 2022 immediately upon the Article 23A designations take effect 1.15 Between 29 September and 20 October 2021, we consulted on our proposed decision to use our Article 23C BMR power to permit legacy use of the 6 LIBOR settings by supervised entities other than in cleared derivatives (whether directly or indirectly cleared). Following the consultation, we confirmed our final decision. 1.16 We also confirmed, following consultation, our decision to prohibit most new use of US dollar LIBOR by BMR supervised entities after end 2021. Our decisions to require publication of some sterling and Japanese yen LIBOR settings on a 'synthetic' basis are not determinative of any future decisions in respect of US dollar LIBOR from end-June 2023. 1.17 We encourage market participants to continue to transition away from LIBOR and not to rely on any ‘synthetic’ LIBOR solution. Further Treasury Legislation 1.18 On 8 September 2021, the Government introduced the Critical Benchmarks (References and Administrators’ Liability’) Bill to Parliament. This legislation is designed to address any remaining risks to contractual certainty, or of disputes, in respect of legacy contracts referencing ‘synthetic’ LIBOR rates after end-2021. Outcome we are seeking 1.19 We have decided to use our Article 23D(2) powers in a way that is: • appropriate to secure the cessation of an Article 23A benchmark in an orderly fashion, and • desirable to advance both our consumer protection objective and integrity objective Summary of feedback we received 1.20 We received 36 responses from banks, asset managers, mortgage lenders, trade bodies, 1 national working group and 1 infrastructure provider. 23 respondents are based in the UK while 13 respondents are from overseas. 1.21 All 36 respondents supported our proposal to use the Article 23D(2) power to intervene for the 6 LIBOR settings. A large majority of respondents agreed with all aspects of our proposals. A few respondents expressed different views on some aspects, and we discuss these in Chapters 2 and 3.

6 FS21/11 Chapter 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Whether to use the Article 23D(2) power 1.22 We consulted on whether we should use our Article 23D(2) powers for the 6 LIBOR settings at end-2021 to secure an orderly wind-down of each of the settings and to advance our consumer protection and integrity objectives. 1.23 While all 36 respondents supported our proposal to use the Article 23D(2) power to intervene for the 6 LIBOR settings, 4 respondents suggested that we should also intervene for the 12-month sterling LIBOR setting. 1 respondent suggested that we should intervene for all sterling and yen LIBOR settings. Based on the information and data available to us and the evidence provided by respondents, we do not consider that there are grounds to justify intervention for sterling and yen LIBOR settings other than the 6 settings we proposed in our consultation. We discuss our considerations in Chapter 2. How to use the Article 23D(2) power 1.24 We consulted on our proposal to require IBA to change the way the 6 LIBOR settings are determined using a modified methodology based on the 2 following components: a. the relevant forward-looking term rates published by IBA in the UK for SONIA, and by QUICK Benchmarks Inc. (QBS) in Japan for TONA, plus b. the fixed spread adjustment that also applies as part of the ISDA IBOR fallbacks for the relevant LIBOR currency and tenor setting and that is published for the purposes of the ISDA IBOR Fallbacks Supplement and Protocol 1.25 We sought views on whether our proposed methodology is appropriate, taking into account the following factors as set out in our Article 23D Statement of Policy: • fair approximation of the value LIBOR would have had • least disturbance or disadvantage to affected parties • market support on a fair way of calculating a replacement value for LIBOR • availability to the benchmark administrator of robust and transparent inputs • impact on the administrator • appropriate length of time over which we require or may require a changed methodology • likely effect outside the United Kingdom of exercising the power 1.26 Overall 31 out of 36 respondents agreed with our proposed methodology for the 6 LIBOR settings. Among the supportive responses, 3 respondents suggested that our proposed methodology for each of the 3 Japanese yen LIBOR settings should be adjusted to account for day count differences between yen LIBOR and forward-looking term TONA, ie TORF. 1.27 With regards to our proposal to use the relevant forward-looking term RFRs, 1 respondent did not comment while 4 respondents expressed contrary views. We discuss these views and our considerations in Chapter 3. 1.28 With regards to our proposal to use the ISDA spread adjustment, 3 respondents did not comment while 2 respondents expressed contrary views. We discuss these views and our considerations in Chapter 3.

7 FS21/11 Chapter 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions What we are changing from our proposal 1.29 Following the consultation responses, we decided to adjust our proposed methodology for each of the 3 Japanese yen LIBOR settings to account for day count differences between yen LIBOR and TORF. Therefore, in using TORF as a component for the calculation of each of the 3 Japanese yen settings, IBA will be required to multiply the value of TORF for the relevant setting published for an applicable London business day by 360/365. Further detail on this may be found in Annex 1 of this Feedback Statement. We do not consider there are grounds to make other changes to our proposal in the consultation. 1.30 We said in our consultation that there were certain operational matters associated with the process of publishing each of the 6 LIBOR settings under the proposed methodology. For example, matters relating to data sufficiency and availability. We did not consult on these matters as we did not consider them to be part of the substantive proposal for the methodology. Some respondents commented on these operational matters, ie it is important to ensure that each of the 6 LIBOR settings can continue to be published on each applicable London business day. We have clarified these matters in ‘Annex I technical specification’ in our draft Article 23D Notice published in September. We also include a copy of the ‘technical specification’ in Annex 1 of this Feedback Statement. Equality and diversity considerations 1.31 We have considered the equality and diversity issues that may arise from our decision to use the Article 23D(2) power. 1.32 Our decision may result in a differential impact on consumers with different protected characteristics under the Equality Act 2010. But overall we consider any such a differential impact to be justified as we do not consider that there is any less impactful way to achieve the objectives that we seek to achieve by the proposed use of the Article 23D(2) power.  Next steps 1.33 We have published a draft version of our Notice to confirm our decision to use the Article 23D(2) power with regards to the 6 LIBOR settings (or ‘versions’ as defined in Article 23G(2) of the BMR). 1.34 A final form of the Notice as required by Article 23D(2) will be given to IBA on 1 January 2022. The Notice will take immediate effect after it is given to IBA and after the Article 23A(4) Notice takes effect at 00:01 on 1 January 2022. 1.35 We will publish the final Notice on 1 January 2022 on our website.

8 FS21/11 Chapter 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions What you need to do next 1.36 While the underlying methodology for each of the 6 LIBOR settings will change after end-2021, IBA will be required to continue publishing these settings at the scheduled publication time and for all applicable London business days as is currently the case for panel-bank LIBOR. We expect that the 6 LIBOR settings should continue to flow through to existing contracts that might have to rely on these settings after end-2021, as previously noted in our Feedback Statement on Article 23D Statement of Policy and our Article 23D decision consultation. We do not expect the changes in the underlying methodology to affect how LIBOR should continue to be displayed on the relevant screen pages. 1.37 We have now confirmed our decision to permit all legacy use of the 6 LIBOR settings by supervised entities other than in cleared derivatives (whether directly or indirectly cleared). 1.38 It is important that users of each of the 6 LIBOR settings take all necessary steps to ensure that they understand how their contract terms interact with our decision. Other than the 6 LIBOR settings, all euro, Swiss franc, 1-week and 2-month US dollar and the remaining sterling and Japanese yen LIBOR settings will cease at end￾2021. Users should take all necessary steps to prepare for the cessation and avoid disruptions to their contracts. 1.39 We discourage the use of permanently unrepresentative benchmarks where appropriate robust alternatives are available. Users should seek to move away from using ‘synthetic’ LIBOR, wherever practicable and fair to all parties. This informs our current supervisory approach to the use of benchmarks by firms.

9 FS21/11 Chapter 2 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 2 Decision on whether to use the power 2.1 We considered that using the Article 23D(2) power would be appropriate to secure an orderly wind-down of each of the 6 sterling and Japanese yen LIBOR settings, taking into account the likely existence and scale of outstanding legacy contracts referencing these settings at end-2021. 2.2 We also considered that using the Article 23D(2) power would be desirable to advance both our consumer protection and integrity objectives, considering the likely disruption to consumers and the wider market should the 6 LIBOR settings cease at end-2021. We asked: Q1: Do you have any views, information or data which suggest that we should or should not use our Article 23D(2) powers for the 6 LIBOR settings to secure an orderly cessation and to advance our objectives of consumer protection and integrity? 2.3 All 36 respondents supported our proposal to use the Article 23D(2) power to intervene for the 6 LIBOR settings to ensure an orderly cessation in line with our consumer protection and integrity objectives. A number of respondents provided further comments as set out below. 2.4 Four respondents suggested that we should also intervene for the 12-month sterling LIBOR setting. 3 of these respondents said this would be helpful for the wind-down of a very small number of bond exposures but they provided no data. 1 of the respondents provided data on their exposure to a fixed-to-float bond that will start to reference the 12-month sterling LIBOR in 2040. 2.5 One respondent said we should intervene for all sterling and Japanese yen LIBOR settings but provided no data. 2.6 Four respondents suggested that the 3 Japanese yen LIBOR settings under a changed methodology should continue for more than a year to facilitate transition where it might require more time, for example, through consent solicitation. 2.7 One respondent asked for the 3 sterling LIBOR settings under a changed methodology to continue for more than 10 years due to their exposure to long-dated contracts that reference these settings. Our response Intervention for other LIBOR settings Based on the information and data that are available to us, including views and evidence presented by the respondents, on balance, we do not consider that the conditions have been met for us to intervene for sterling and Japanese yen LIBOR settings other than the 6 settings as proposed in our consultation.

10 FS21/11 Chapter 2 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions We set out in our Article 23D decision consultation and the draft Article 23D Notice the information, data and views we took into account to decide to require the continued publication by IBA of each of the 6 LIBOR settings under a changed, ‘synthetic’ methodology. Length of our intervention Our intention is to intervene for no longer than it is desirable to do so to protect consumers and maintain market integrity. Further, insofar as we exercise our Article 23D(2) powers in conjunction with our compulsion power under Article 21(3), the maximum period for us to compel continued publication of a critical benchmark, such as LIBOR under a changed methodology, is 10 years from when we start exercising the Article 21(3) power at end-2021. This means that any existing contracts that only start to reference one of the 6 LIBOR settings after end-2031 are not relevant for our consideration of whether to use or renew our Article 21(3) power. Any contracts that mature after end-2031 that are not converted away from LIBOR will also outlive the maximum period for our Article 21(3) power. We said in our Article 23D decision consultation and the draft Article 23D Notice that should market participants want to extend their use of ‘synthetic’ LIBOR beyond the end of its required publication, the two components of ‘synthetic’ LIBOR will remain visible and available. Any use of the relevant forward-looking term RFRs should be in line with limited use cases identified by national working groups in the UK and Japan. We are required to review whether it would be necessary to renew our Article 21(3) decision to achieve an orderly cessation of LIBOR before the end of each compulsion period. For the 3 Japanese yen LIBOR settings, our intention is not to renew our compulsion so they will cease at end-2022. We have not received evidence which suggests that a longer period would be necessary for an orderly cessation of the 3 Japanese yen LIBOR settings. We expect market participants to continue transition of relevant contracts between now and end-2022.

11 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 3 Decision on how to use the power 3.1 In our Article 23D decision consultation, we proposed to impose the following requirements on IBA so that each of the 6 LIBOR settings would be determined under a changed, ‘synthetic’ methodology after end-2021 for a limited period. This would be required immediately upon any designation of these benchmarks as Article 23A benchmarks taking effect. • 1-month sterling LIBOR is to be calculated as the sum of the 1-month Term SONIA Reference Rate provided by IBA and the fixed spread adjustment that applies as part of the ISDA IBOR fallback for 1-month sterling LIBOR and that is published for the purposes of the ISDA IBOR Fallbacks Supplement and Protocol (hereafter referred to as ‘the ISDA spread adjustment’ for the relevant LIBOR setting). • 3-month sterling LIBOR is to be calculated as the sum of the 3-month Term SONIA Reference Rate provided by IBA and the ISDA spread adjustment for 3-month sterling LIBOR. • 6-month sterling LIBOR is to be calculated as the sum of the 6-month Term SONIA Reference Rate provided by IBA and the ISDA spread adjustment for 6-month sterling LIBOR. • 1-month Japanese yen LIBOR is to be calculated as the sum of the 1-month Tokyo Term Risk Free Rate provided by QUICK Benchmarks Inc. and the ISDA spread adjustment for 1-month Japanese yen LIBOR. • 3-month Japanese yen LIBOR is to be calculated as the sum of the 3-month Tokyo Term Risk Free Rate provided by QUICK Benchmarks Inc. and the ISDA spread adjustment for 3-month Japanese yen LIBOR. • 6-month Japanese yen LIBOR is to be calculated as the sum of the 6-month Tokyo Term Risk Free Rate provided by QUICK Benchmarks Inc. and the ISDA spread adjustment for 6-month Japanese yen LIBOR. 3.2 We did not invite views on our choices of the relevant forward-looking term RFRs provided by IBA and by QUICK Benchmarks Inc. respectively. We asked: Q2: Do you agree that a forward-looking term RFR is an appropriate component for producing a ‘synthetic’ LIBOR to measure on an ongoing basis the expectation of interest rates over a fixed term that is reflected in LIBOR itself? 3.3 Thirty-one of 36 respondents supported our proposal to use forward-looking term RFRs. 1 respondent did not comment. 4 respondents expressed contrary views. 3.4 Among the supportive responses, several respondents provided further comments as set out below. 3.5 Two respondents asked for clear communications that outside ‘synthetic’ sterling LIBOR, term SONIA should be limited to specific use cases in line with relevant industry standards and guidance. 3.6 Two respondents commented that a ‘synthetic’ LIBOR based on a forward-looking term RFR would not be appropriate for derivatives.

12 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 3.7 One respondent noted that liquidity in both sterling and Japanese yen markets is sufficient to support use of the relevant forward-looking term RFRs as components for ‘synthetic’ LIBOR. 3.8 One respondent noted that our periodic review of the Article 23D(2) powers should assess the robustness and suitability of the component inputs for ‘synthetic’ LIBOR. 3.9 One respondent commented that they would support using RFRs ‘in-arrears’ as components for ‘synthetic’ LIBOR if that allows wider scope of permitted use of ‘synthetic’ LIBOR in legacy contracts. 3.10 Three respondents suggested that in calculating a ‘synthetic’ Japanese yen LIBOR, the published TORF rate on a given day should be adjusted as TORF * 360/365 to take into account day count differences between TONA/TORF (for which day count is 365) and yen LIBOR (for which day count is 360). 3.11 The contrary views from 4 respondents are set out as below. 3.12 Two respondents objected to using forward-looking term RFR as a component for ‘synthetic’ LIBOR and suggested using RFR ‘in-arrears’ to fully align with ISDA fallbacks. 1 of the 2 respondents was concerned about creating a dual market for professional bonds and loans (ie that in those asset classes, there would be a different reference rate for new issuance on the one hand and legacy contracts on the other), as well as between cash and derivatives products. They highlighted the additional operational and hedging costs due to the basis between forward-looking term SONIA and SONIA ‘in-arrears’. They also pointed out that using forward-looking term SONIA may disincentivise ongoing bond transition through consent solicitation and disrupt future roll-off of legacy contracts to SONIA ‘in-arrears’ at the end of our requirement of a ‘synthetic’ sterling LIBOR. 3.13 One respondent said that ‘synthetic’ LIBOR should not be a ‘one size fit all’ solution. In particular, retail customers may not understand the concept of a term SONIA Reference Rate (TSRR) and large corporates may prefer RFR ‘in arrears’. They expressed concerns that the proposed methodology could be a de facto endorsement of TSRRs for transition, which could potentially lead to disputes where firms have been transitioning customers away to alternative rates. 3.14 One respondent said that a forward-looking term RFR is not appropriate for ‘synthetic’ Japanese yen LIBOR due to the lack of a deep and liquid underlying TONA swaps market. They said that according to cleared TONA swaps transaction data reported to the Depository Trust & Clearing Corporation Swap Data Repository (DTCC SDR) during July-August 2021, there was no reported transaction for 1- and 3-month settings and there were 3 transactions for the 6-month setting. Based on information available to this respondent, they said that for many business days there were no executable quotes for short-term TONA swaps in the inter-dealer market. They expressed concerns that TORF would therefore likely often fall back on the last published rate as the contingency plan provided by the administrator of TORF. As alternatives, they suggested using TONA ‘in-advance’ as a component for ‘synthetic’ Japanese yen LIBOR or a fixed average of a few last published yen LIBOR numbers at end-2021 instead of a ‘synthetic’ yen LIBOR. They also suggested that should TORF be used as a component for ‘synthetic’ Japanese yen LIBOR, the value of TORF should be adjusted given day count differences between yen LIBOR and TORF/TONA.

13 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Our response Use of forward-looking term RFRs outside ‘synthetic’ LIBOR for transition We communicated our views in our statement in June that outside our choice of the relevant forward-looking term RFR for a ‘synthetic’ sterling LIBOR, use of the available TSRRs should be in line with certain niche use cases as identified by the RFRWG guidance. We also said in our consultation that there are certain limited use cases in the cash market for forward-looking term RFRs as recommended by national working groups in both UK and Japan to facilitate transition away from LIBOR. Market participants should have due regard to the established industry guidance when considering any use of forward-looking term RFRs. For LIBOR transition in the sterling market, our joint Dear CEO letter with the Prudential Regulation Authority pointed out that SONIA ‘in-arrears’ is the most robust alternative rate, and that using SONIA ‘in-arrears’ in derivative and bond markets would align with well-established market practice, and with use cases identified in loan market by the RFRWG. In considering whether it is appropriate to use a forward-looking term SONIA, firms should take into account RFRWG’s use cases and the work undertaken by the FICC Markets Standards Board. Our Dear CEO letter also emphasised that it is important that firms select an alternative rate that meets customers’ needs, and that customers understand the properties and implications of the rates to which they are moving. Forward-looking term RFRs compared with RFRs ‘in-arrears’ for ‘synthetic’ LIBOR We have communicated on various occasions that ‘synthetic’ LIBOR is a time-limited bridging solution only to help with contracts that cannot practicably be amended or renegotiated to transition away to alternative rates by end-2021. The choice of forward-looking term RFRs as components for the 6 sterling and Japanese yen LIBOR settings is intended to secure an orderly wind-down of these contracts referencing each of the 6 settings. As we discussed in paragraph 3.54 of our consultation, RFRs ‘in-arrears’ are operationally unsuitable for these legacy contracts which identify the interest rate payable at the beginning of the relevant interest period or well in advance of the end of the relevant interest period. This is because using an RFR ‘in-arrears’ means the interest rate would only be known at the end of the relevant interest payment period. As many of the legacy contracts cannot realistically be amended to work on the basis of a rate only identifiable 1-month, 3-months or 6-months after the date envisaged when entering the contract, using RFRs ‘in-arrears’ could potentially cause disruption and/ or disturbance to affected parties. Where contracts can feasibly be converted by end-2021 (eg through consent solicitation for bonds), we expect parties to take steps to transition these contracts (including transition to RFRs ‘in-arrears’ as the most robust alternative) and not to wait and rely on a ‘synthetic’ LIBOR. The short-term publication of ‘synthetic’ methodology will

14 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions provide further time to complete consent solicitation and customer engagement processes if necessary. As we set out in paragraph 3.55 of our consultation, using a forward￾looking term RFR means that outstanding legacy contracts that continue to reference LIBOR under the changed methodology would have the same expected value of interest payments as those that are being amended to use RFRs ‘in-arrears’ over the same calendar period. Any basis between a forward-looking term RFR and RFR ‘in-arrears’ can be hedged. TONA swap market Based on data and information available to us on cleared TONA swap transactions, there has been a significant uptake in the short-term TONA swap market since the launch of TONA First initiative in August 2021. As of October 2021, around 75% of new Japanese yen swap transactions and executable quotes, by notional value, referenced TONA compared with fewer than 10% of new swap transactions and executable quotes, by notional value, that referenced TONA before the initiative was launched. We expect liquidity in TONA swaps market to continue to increase, contributing to the robustness of TORF. We do not consider TONA ‘in-advance’ as appropriate for ‘synthetic’ Japanese yen LIBOR as we discussed in paragraph 3.35 of our consultation. This is because RFR ‘in-advance’ is a backward-looking measure of the relevant RFR, eg it reflects the interest over the previous interest period, so it would be less effective in measuring the first economic component of LIBOR which is the forward-looking expectation of the relevant RFR. We also do not consider it appropriate to continue publishing Japanese yen LIBOR as a fixed average of a few last published yen LIBOR rates after end-2021. This would be unlikely to achieve as reasonable and fair an approximation of the value panel-bank LIBOR might have had as is achievable using a roust floating rate measure. Using a fixed average means that legacy yen LIBOR contracts would reference a fixed, instead of a variable rate under this approach. This may lead to increased value transfer (particularly if there are not symmetrical expectations of future interest rate moves) and cause disruption to parties to these legacy contracts that specifically sought a floating rate exposure. Adjustment to TORF in calculating ‘synthetic’ yen LIBOR We agree with the respondents who raised this point. We have decided to adjust the methodology for each of the 3 Japanese yen LIBOR settings to account for day count differences between yen LIBOR and TORF. So in using TORF as a component for the calculation of each of the 3 Japanese yen settings, IBA will be required to multiply the value of TORF for the relevant setting published for an applicable London business day by 360/365. Please see detail in our draft Article 23D Notice and in Annex 1 of this Feedback Statement.

15 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Periodic review of our use of the Article 23D(2) powers We are required to review whether the form in which we use our Article 23D(2) powers has advanced either or both of our consumer protection and integrity objectives, at least every 2 years under Article 23E. As part of the review, we will consider the appropriateness of the changes we impose on IBA to determine each of the 6 LIBOR settings in line with our objectives. Scope of permitted use of ‘synthetic’ LIBOR This is a matter for our power under Article 23C BMR. We consulted on our proposed decision for permitting use of ‘synthetic’ sterling and Japanese yen LIBOR in legacy contracts and have now published our final decision following the consultation. Q3: Do you agree that the fixed spread adjustment that applies as part of the ISDA IBOR fallbacks for the relevant LIBOR currency and tenor setting and that is published for the purposes of the ISDA IBOR Fallbacks Supplement and Protocol is the fairest and most robust way to calculate the replacement value for LIBOR? 3.15 Thirty-one of 36 respondents supported our proposal to use the ISDA spread adjustment. 3 respondents did not comment. 2 respondents expressed contrary views. 3.16 Among the supportive responses, some respondents provided further comments as set out below. 3.17 Twelve respondents highlighted that our proposed decision is consistent with the wide market acceptance and international authorities’ endorsement of the ISDA spread adjustment. 3.18 Three respondents commented that there is no alternative robust approach to calculating the spread differently from the ISDA approach. 3.19 One respondent said that using the ISDA spread adjustment contributes to the transparency and understandability of ‘synthetic’ LIBOR. 3.20 One respondent noted that since the ISDA spread adjustment was fixed in March 2021, the market has already been converging around the use of the ISDA spread to transition away from LIBOR. 3.21 One respondent noted that the proposed methodology is likely to result in an increase in LIBOR rates for various tenors compared with LIBOR rates at the time of consultation. 3.22 A retail mortgage provider noted that using the ISDA spread smooths out any spikes in the future that may otherwise unduly prejudice retail customers. 3.23 Contrary views were expressed by 2 respondents as set out below.

16 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 3.24 One mortgage provider said that using the ISDA spread might be unfair to retail and/ or less sophisticated customers who are parties to legacy contracts that might rely on a ‘synthetic’ LIBOR because the ISDA spread adjustment for 3-month sterling LIBOR is higher than the credit premium embedded in LIBOR at the time of the consultation response. 3.25 One respondent considered that adding the ISDA spread adjustment to a forward￾looking term RFR would create value change for issuers who have hedged the issuance with swaps. Our response We agree with the large majority of respondents. We consider that the ISDA spread adjustment, which is calculated from taking a median average of the relationship between LIBOR and RFRs that has existed in the past, is the most appropriate, and also a fair way of approximating the relationship that may reasonably be expected to have existed in the future had panel bank LIBOR continued. As respondents noted, it has been widely adopted by mutual agreement of interest payer and receiver across very large parts of financial markets, and endorsed by international authorities. We do not think there is a more appropriate number that would have been unambiguously fairer for both payers and receivers of interest for the remaining life of contracts that rely on ‘synthetic’ LIBOR. There has been a recent period of an unusually low spread between panel-bank LIBOR and RFRs, widely considered to be due in large part to central banks’ monetary policy in response to Covid-19. During the economic stress in April 2020, by contrast, the spread was 0.49 percentage points or 49 basis points above the ISDA spread. As we discussed in paragraphs 3.40-3.43 of our consultation, the ISDA spread adjustment which is a fixed 5-year historic median reflects a range of economic and market conditions. We do not know what the spread would have been in the future, but we think the 5-year historical median is a reasonable and fair approximation of what it might have been. The fixed spread is not vulnerable to manipulation. It also has the benefit of protecting customers paying LIBOR interest from the effects of spikes going forward. The lending bank rather than the mortgage borrower will in future have to manage the risks of the bank’s future funding costs rising more than the interest they receive on their mortgage lending because of changing market liquidity or perceptions of credit risk. We consider banks better placed than retail consumers to manage this risk. For consumers with LIBOR-linked mortgages that may have to rely on ‘synthetic’ LIBOR after end-2021, under the fixed ISDA spread adjustment they would benefit from no longer being exposed to the risk of the spread increasing at times of economic stress (for example during March/April 2020). More broadly, our decision to require the 6 LIBOR settings to continue under the changed, ‘synthetic’ methodology is to avoid disruption to consumers who are parties to legacy contracts, including some mortgages, that cannot be moved away from LIBOR by year end. We have consistently encouraged firms to transition away from LIBOR where they can as this is the best option to remove reliance on LIBOR after

17 FS21/11 Chapter 3 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions end-2021. There are a range of solutions for transition that lenders may choose from, including compounded SONIA ‘in-arrears’ or the Bank rate, with or without a spread. Providing any spread added to SONIA over the relevant term is no higher than the ISDA spread for the same term, we think all of these arrangements are fair to the consumer and welcome a degree of competition between lenders in what they are offering. Where there is no active transition, ‘synthetic’ LIBOR is likely to be relied upon by legacy contracts. We have communicated to firms, including in the Dear CEO letter, that when replacing LIBOR with an alternative rate, they should understand customers’ needs and ensure that they are treating consumers fairly. More information on our expectations of firms when replacing LIBOR can be found in our LIBOR conduct Q&A. Synthetic LIBOR is only intended as a temporary bridging solution for legacy contracts that cannot realistically be transitioned within the timeframe available. We continue to encourage parties to transition their contracts to alternative rates wherever practicable between now and end-2021, as well as after end-2021, as that is the only way parties can have certainty and control over contractual terms of their contracts.

18 FS21/11 Chapter 4 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 4 Other issues 4.1 We asked: Q4: Do you have any representations or points about how our proposed 23D(2) decision would impact you? 4.2 Twelve respondents commented that each of the 6 LIBOR settings, albeit under a changed, ‘synthetic’ methodology, should continue to be published on the same screen pages by redistributors and at the same publication time as panel-bank LIBOR. 4.3 Two respondents said that in considering using TORF as a component, it was important to ensure that the continuing 3 Japanese yen LIBOR settings can continue to be published on all London business days. 4.4 Seven respondents asked for early clarity on the outcome of our annual review of the Article 21(3) decision, ie whether we would extend our compulsion of IBA to continue publishing the 6 LIBOR settings after end-2022 for another time-limited period. 4.5 Four respondents asked for clarification on the operation of cessation fallbacks in contracts that continue to reference one of the 6 LIBOR settings after end-2021. They expressed concerns that the relevant cessation fallbacks may not function as intended if the 6 LIBOR settings continue under a changed, ‘synthetic’ methodology. 4.6 One EU-based respondent asked for international coordination so that statutory replacement rates are in line with ‘synthetic’ LIBOR. 4.7 One respondent noted that it would be helpful for us to restate the unrepresentative nature of ‘synthetic’ LIBOR for the purpose of ISDA IBOR Fallback Protocol. 4.8 Many respondents asked for clarity on who would be allowed to use a ‘synthetic’ LIBOR. They also called for international consistency on any permitted use of a ‘synthetic’ LIBOR to avoid market fragmentation. 4.9 Some respondents commented on the need for legal protection to ensure contractual continuity for legacy contracts continuing to reference the 6 LIBOR settings, albeit under a changed, ‘synthetic’ methodology after end-2021. 4.10 One respondent called for further legislative solution for remaining legacy contracts, at the end of our maximum compulsion period under Article 21(3), which is up to 10 years.

19 FS21/11 Chapter 4 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Our response We expect that the 6 LIBOR settings under a changed methodology should continue to flow through to existing contracts that might have to rely on these settings after end-2021, as previously noted in our Feedback Statement on Article 23D Statement of Policy and our Article 23D decision consultation. While the underlying methodology for the 6 LIBOR settings will change after end-2021, IBA will be required to continue publishing these settings at the scheduled publication time and for all applicable London business days as is currently the case for panel-bank LIBOR. We do not expect the changes in the underlying methodology to affect how LIBOR should continue to be displayed on the relevant screen pages. The review of our Article 21(3) decision is separate from our Article 23D(2) decision. We will consider how best to keep the market informed in a clear and timely manner about the results of our Article 21(3) review before the end of our compulsion at end-2022. For the 3 Japanese yen LIBOR settings, we do not intend to renew the compulsion and their publication will cease at end-2022. We note the concerns around the inoperability of cessation fallbacks in contracts. For the 6 LIBOR settings, they will continue for a further time-limited period after end-2021 to provide a bridging solution for contracts to transition away in an orderly fashion. We encourage market participants to reach agreement where feasible to transition away from LIBOR if they want to avoid relying on a ‘synthetic’ LIBOR rate. This is the only way for parties to have certainty and control over their contractual terms when LIBOR ceases or is no longer representative. Respective EU and UK legislative solutions give scope to align interventions so that they point to the same economic outcome and minimise market fragmentation where this is desirable. We continue to engage closely with our EU counterparts. However, the choice of a statutory replacement rate under the EU BMR is properly a matter for the EU authorities. In the event that EU authorities decide not to impose a statutory replacement rate, ‘synthetic’ LIBOR should flow through to legacy contracts, referencing one of the 6 LIBOR settings, governed under the EU law (and other non-UK law) subject to any applicable local legislation and any relevant contractual terms. Our announcement in September made it clear that the 6 LIBOR settings under the changed, ‘synthetic’ methodology will no longer be representative for the purpose of the BMR. The permitted use of ‘synthetic’ LIBOR in legacy contracts is outside the scope of Article 23D(2). We consulted separately on our proposed decision under Article 23C BMR in relation to the permitted use of the 6 LIBOR settings after end-2021. Following the consultation, we have published our decision. We engaged with our international counterparts both bilaterally and through the Financial Stability Board’s Official Sector Steering Group and IOSCO on these matters. We will continue to engage through the relevant international fora.

20 FS21/11 Chapter 4 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions The Government introduced supplementary legislation that seeks to reduce yet further any risks to contractual certainty and disputes in respect of legacy contracts referencing ‘synthetic’ LIBOR. While the maximum period of our Article 21(3) power is 10 years, we are required to review whether it would be necessary for an orderly cessation of LIBOR to renew our decision to compel IBA beyond the end of each compulsion period. Any further legislation to provide solutions for contracts that mature after the end of our compulsion power is a matter for the Government.

21 FS21/11 Annex 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Annex 1 FCA requirements to be imposed on IBA under Article 23D(2)

1. The Financial Conduct Authority (“the Authority”) requires IBA to change the way the following 6 LIBOR Versions are determined, which will take immediate effect after their Article 23A designations become effective on 1 January 2022:
2. 1-month sterling LIBOR is to be calculated as the sum of the ICE 1-month Term SONIA Reference Rate and the ISDA Spread Adjustment for 1-month sterling LIBOR.
3. 3-month sterling LIBOR is to be calculated as the sum of the ICE 3-month Term SONIA Reference Rate and the ISDA Spread Adjustment for 3-month sterling LIBOR.
4. 6-month sterling LIBOR is to be calculated as the sum of the ICE 6-month Term SONIA Reference Rate and the ISDA Spread Adjustment for 6-month sterling LIBOR.
5. 1-month yen LIBOR is to be calculated as the sum of the QBS 1-month TORF *(360/365) and the ISDA Spread Adjustment for 1-month yen LIBOR.
6. 3-month yen LIBOR as the sum of the QBS 3-month TORF*(360/365) and the ISDA Spread Adjustment for 3-month yen LIBOR.
7. 6-month yen LIBOR as the sum of the QBS 6-month TORF*(360/365) and the ISDA Spread Adjustment for 6-month yen LIBOR.
8. Each of the 6 LIBOR Versions under the methodology in paragraph 1 should continue to be published at or around 11:55 am London time on each applicable London business day, except for London public holidays.
9. IBA should use the latest available value of the required input data, under the methodology in paragraph 1, as published by its provider prior to the scheduled publication time for the relevant LIBOR version on an applicable London business day, subject to the below.
10. IBA should use the required input data, under the methodology in paragraph 1 for the relevant LIBOR Version, in the same form as they are published by the respective provider.
11. Each of the 6 LIBOR Versions under the methodology in paragraph 1 should continue to be rounded and published to 5 decimal places. Where any input data is published to fewer than 5 decimal places, IBA should continue to publish the relevant LIBOR Version to 5 decimal places, by adding zero(s) at the end of the decimal value of the relevant LIBOR Version.

22 FS21/11 Annex 1 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions 6. In the event where the required input data, under the methodology in paragraph 1 for the relevant LIBOR Version, is not published by its provider for an applicable London business day in accordance with its normal publication schedule, IBA should use the last published value for that input data to produce the relevant LIBOR Version under the methodology in paragraph 1 unless the Authority decides and notifies IBA otherwise. 7. In the event where the required input data under the methodology in paragraph 1 for the relevant LIBOR Version are not available prior to the scheduled publication time for the relevant LIBOR Version on an applicable London business day, IBA should act in accordance with the requirement in either subparagraph 1) or subparagraph 2) as below.

1. Where IBA has been informed by the relevant provider of the input data that the required input data will become available within 1 hour after the scheduled publication time for the relevant LIBOR Version on that applicable London business day, IBA should delay the publication of the relevant LIBOR Version and publish as soon as practicable once the required input data becomes available to IBA for that applicable London business day. If the required input data do not become available to IBA within 1 hour after the scheduled publication time for the relevant LIBOR Version on that applicable London business day, IBA should use the last published value for that input data to produce the relevant LIBOR Version under the methodology in paragraph 1 for that applicable London business day and publish as soon as practicable after the 1 hour delay, unless the Authority decides and notifies IBA otherwise.
2. In other circumstances except the one described under subparagraph 1), IBA should use the last published value for that input data to produce the relevant LIBOR Version under the methodology in paragraph 1 for that applicable London business day unless the Authority decides and notifies IBA otherwise.
3. In the event where after the publication of the relevant LIBOR Version on an applicable London business day, IBA is notified that the relevant forward-looking term RFR is subsequently republished by its provider on that day, IBA should re-calculate and re-publish the relevant LIBOR version using the relevant republished forward-looking term RFR for that applicable London business day.

23 FS21/11 Annex 2 Financial Conduct Authority Feedback Statement: Article 23D BMR Decision for 6 sterling and yen LIBOR versions Annex 2 Abbreviations used in this paper Abbreviation Description BMR Benchmarks Regulation IBA ICE Benchmark Administration IBOR Interbank Offered Rate IOSCO International Organization of Securities Commissions ISDA International Swaps and Derivatives Association QBS QUICK Benchmarks Inc. RFR Risk-Free Rate RFRWG Working Group on Sterling Risk-Free Reference Rates SONIA Sterling Overnight Index Average TONA Tokyo Overnight Average Rate TORF Tokyo Term Risk Free Rate TSRR Term SONIA Reference Rate All our publications are available to download from www.fca.org.uk. If you would like to receive this paper in an alternative format, please call 020 7066 7948 or email: publications_graphics@fca.org.uk or write to: Editorial and Digital team, Financial Conduct Authority, 12 Endeavour Square, London, E20 1JN Sign up for our news and publications alerts

© Financial Conduct Authority 2021 12 Endeavour Square London E20 1JN Telephone: +44 (0)20 7066 1000 Website: www.fca.org.uk All rights reserved Pub ref: 007233