Amendments to the FPC’s framework for the O-SII buffer

Overview

1. The Financial Policy Committee (‘the FPC’ or ‘the Committee’) must have a framework for the O-SII buffer and review this framework at least every second year.^{footnote [1]}

2. Following its December 2020 review, the FPC was not required to review the framework again until December 2022. The Committee decided, however, to undertake a review in 2021 Q3 in line with its commitment in the 2020 Q3 Record and December 2020 Financial Stability Report to consider information that became available during the Covid-19 (Covid) shock about how the framework operates in stress.

3. In response to the 2021 Q3 review, the FPC decided to consult on a proposal to amend its framework as follows:

(i) To change the metric used to determine other systemically important institutions (O-SII) buffer rates from total assets to the UK leverage exposure measure.

(ii) To recalibrate the thresholds used to determine O-SII buffer rates to prevent an overall tightening or loosening of the framework relative to its pre-Covid level.

4. The proposal aims to ensure that the framework still addresses the key systemic risk intended by the FPC: the risk that a distressed ring-fenced bank or large building society disrupts the supply of credit to the real economy. It achieves this as follows:

(i) First, it excludes from the framework central bank reserves,^{footnote [2]} which grew significantly during the pandemic but do not reflect a bank’s potential to disrupt the credit supply. This exclusion mitigates the risk that future changes in central bank balance sheets inadvertently affect banks’ lending decisions by interacting with the O-SII buffer. It also allows banks to draw on central bank liquidity as necessary without becoming constrained by the associated effect on buffer requirements.

(ii) Second, the proposal brings into the framework committed but undrawn credit facilities. Experience during the pandemic suggests that these can form an important part of the credit supply in stress.

5. If the proposal is adopted, the changes would come into effect in time for the Prudential Regulation Authority (PRA) to assess rates under a revised framework in December 2023, based on end-2022 financial results. Rates set in 2023 would then apply from January 2025.

6. This consultation is relevant to PRA-regulated ring-fenced banks and large building societies that are subject to the O-SII buffer.

Responses and next steps

7. The consultation will close on 15 February 2022. All responses should be emailed to osiibufferframework@bankofengland.co.uk.

The FPC and its regulatory powers

8. The FPC was established under the Bank of England Act 1998, through amendments made in the Financial Services Act 2012.^{footnote [3]} It is responsible for protecting and enhancing the resilience of the UK financial system, including identifying, monitoring, and taking action to remove, or reduce, systemic risks. But the FPC is not required to achieve resilience at any cost. Its actions must not, in the provisions of the legislation, have a ‘significant adverse effect on the capacity of the financial sector to contribute to the growth of the UK economy in the medium or long term’.^{footnote [4]} Subject to achieving its main objective, the FPC is required to support the Government’s economic policy, including its objectives for growth and employment.

Background

Systemic banks and capital

9. The economy depends on critical financial services provided by financial institutions, including large banks and building societies. It is important that systemically important institutions have levels of capital that are sufficient to absorb losses in stress and continue to maintain critical financial services to the real economy, particularly the provision of credit.

10. The international Basel framework for global systemically important banks (G-SIBs) has been implemented through European and UK legislation. This framework sets higher capital buffers for G-SIBs. The Basel and European frameworks, and the UK ring-fencing regime, also recognise that banks can be systemically important in a domestic context and may likewise warrant higher capital to absorb stress. This additional capital may take the form of an explicit loss-absorbing buffer for domestic systemically important banks (D-SIBs).

11. The O-SII buffer is the UK’s capital buffer for UK D-SIBs and applies to ring-fenced banks and large building societies. It raises the capacity of these banks to withstand stress, thereby increasing their resilience. This reflects the additional damage that these banks could cause to the UK economy if they were close to failure.

History and legal framework

12. In 2011, the Independent Commission for Banking recommended higher regulatory capital for ring-fenced banks to reflect the additional risk they pose to the UK’s real economy.^{footnote [5]} While the European Capital Requirements Directive IV was in force, HM Treasury enacted this recommendation by implementing the Systemic Risk Buffer (SRB). Following the transition to the Capital Requirements Directive V, the PRA implemented the O-SII buffer in order to replace the role previously performed by the SRB, but the economic substance is unchanged.

13. Under the Capital Requirements (Capital Buffers and Macro-prudential Measures) Regulations 2014, the FPC must maintain a framework for setting O-SII buffer rates that reflects the extent to which the failure or distress of a ring-fenced bank or large building society might pose a risk to the UK financial system.^{footnote [6]}

14. When the FPC instituted this framework in 2016 it judged that the size of a bank’s O-SII buffer should specifically reflect the greater costs to the economy if that bank fell into distress relative to a smaller, non-systemic bank. The FPC considered that the main channel by which distressed banks could cause damage to the financial system and real economy was through contraction of their household and corporate lending.

15. As regards other channels of transmitting systemic risk, there was little evidence to suggest that banks in distress (outside of failure) disrupt deposit‐taking and payments activities, and the structural measures in the ring-fencing regime, as well as progress on resolution, should ensure continuity of critical functions following failure.

16. The power to set an O-SII buffer for a given bank in a given year falls to the PRA. The PRA may deviate from the FPC’s framework for a given bank, if it makes a sound supervisory judgement to do so.^{footnote [7]} Under current PRA policy, the O-SII buffer rates set by the PRA in a given year are based on financial results from December of the previous year and apply from January the year after next. (For example, rates set in December 2018 were based on December 2017 balance sheets and applied from January 2020).

The FPC’s framework for the O-SII buffer

17. The FPC’s framework must contain the following elements:

(i) a set of criteria for assessing the extent to which the failure or distress of a relevant O-SII might pose a risk to the financial system;

(ii) a methodology for measuring the criteria and giving a relevant O-SII a single score in relation to the criteria; and

(iii) in relation to each score that an O-SII may receive, a buffer rate that corresponds to the score.

18. As set out above, the FPC judged in 2016 that, for the purposes of the O-SII buffer, the key criterion for assessing the risk a bank might pose to the financial system was its potential to disrupt the credit supply in distress. The Committee decided to use banks’ ‘total assets’ as a proxy to measure this criterion. It considered this an appropriate proxy metric as banks’ total assets were well-correlated with their share of the lending market; and lending was expected to comprise a large share of relevant banks’ total assets going forward. The Committee also considered that total assets were a simple, easily understood metric and would not create incentives for banks to place assets that incur a higher capital charge outside the ring-fence.

19. Under the FPC’s current framework, a bank’s total assets correspond to O-SII buffer rates as follows:

Table A: Current thresholds for determining O-SII buffer rates

20. Under the relevant legislation, a firm’s O-SII buffer rate must apply across all risk-weighted exposures not just on marginal exposures above the threshold. Hence, the size of a firm’s O-SII buffer therefore increases as a step-function with total assets. Marginal rates are not possible under the legislation.

The FPC’s latest review of its O-SII buffer framework

21. In 2021 Q3, the FPC reviewed its O-SII buffer framework in line with its commitment in the Q3 Record and 2020 Financial Stability Report to consider information that became available during the Covid shock about how the framework operates in stress.

22. Following this review, the FPC judges that total assets may no longer be a desirable metric to determine O-SII buffer rates.

23. Total assets across ring-fenced banks grew significantly in 2020 (Chart 1). This was not driven by lending (which grew only marginally) but by very high growth in central bank reserves. Such growth does not reflect an increase in a bank’s potential to disrupt the credit supply, the key externality that the FPC intends the O-SII buffer to address.

24. Balance sheet movements during the Covid shock therefore demonstrated that total assets and lending do not necessarily move together over time in a stress. This suggests that total assets may no longer be the most appropriate proxy for a bank’s potential to disrupt the credit supply.

Chart 1: Aggregate assets and central bank reserves for ring-fenced banks ^{(a) (b)}

Footnotes

• (a) Ring-fenced banks of major UK banks and Nationwide.
• (b) Reserves shown here are based on the definition used in the UK leverage ratio.

25. The growth in central bank reserves also had the effect of pushing banks towards O-SII buffer rate thresholds, irrespective of their potential to disrupt the credit supply. This risked creating an incentive for banks to manage the size of their balance sheets by constraining lending. It also created a risk that banks drawing on central bank liquidity could become constrained by O-SII buffer requirements.

26. In light of these factors, the FPC now considers an alternative proxy that excludes central bank reserves would be more appropriate.

27. Experience during the pandemic also suggests that committed but undrawn credit facilities can form a key part of the credit supply. In its May 2020 Financial Stability Report, the FPC observed that in 2020 Q1, the major UK banks had expanded their net lending by around £20 billion, as business drew down committed credit lines. This compared to a reduction in net lending of £3 billion over 2019.^{footnote [8]} Committed but undrawn credit facilities are not captured by total assets, which again suggests that it may no longer be the optimal metric to determine O-SII buffer rates.

28. The FPC has therefore decided to consult on a proposal to change the metric used to determine O-SII buffer rates from total assets to the UK leverage exposure measure, as this would both exclude reserves and bring committed but undrawn credit facilities into the framework. The FPC announced its intention to consult on this proposal in its 2021 Q3 Financial Policy Summary and Record.^{footnote [9]}

FPC proposal

(i) Changing the metric from total assets to the UK leverage exposure measure

29. The first part of the FPC’s proposal is to change the metric used to determine O-SII buffer rates from total assets to the UK leverage exposure measure. Details of how the UK leverage exposure measure is calculated can be found in the UK leverage ratio framework.^{footnote [10]}

30. The FPC’s framework^{footnote [11]} for the O-SII buffer intends that the buffer should reflect the risk that a bank disrupts the supply of credit to the real economy if it falls into distress. The FPC now considers the UK leverage exposure measure to be a better proxy for this risk than total assets, for the following reasons:

31. First, it excludes from the framework central bank reserves, which grew significantly during the pandemic, but do not reflect a bank’s potential to disrupt the credit supply. This exclusion mitigates the risk that future changes in central bank balance sheets inadvertently affect banks’ lending decisions by interacting with the O-SII buffer. It also allows banks to draw on central bank liquidity as necessary without becoming constrained by the associated effect on buffer requirements.

32. Second, the proposed metric brings into the framework committed but undrawn credit facilities. Experience during the pandemic suggests that these can form an important part of the credit supply in stress.

33. The UK leverage exposure also meets the FPC’s original criteria for an appropriate proxy metric as described in paragraph 18 above. In particular, it is well-correlated with banks’ share of the lending market and lending is expected to comprise a large share of the metric going forward.

(ii) Recalibrating the thresholds that determine O-SII buffer rates

34. The second part of the FPC’s proposal is to adjust the thresholds that determine O-SII buffer rates, alongside the change in metric, to prevent an overall tightening or loosening of the framework relative to its pre-Covid level. The FPC therefore proposes calibrating the thresholds based on financial results from 2019, before the large expansion in central bank reserves during the pandemic.

35. Current thresholds are expressed in terms of total assets. Under the FPC proposal, the revised thresholds would be expressed in terms of the UK leverage exposure measure.

36. As at December 2019, the UK leverage exposure measure for banks subject to an O-SII buffer was on average c.£15 billion lower than those firms’ total assets.

37. The FPC therefore proposes that following the change in metric, each of the O-SII buffer thresholds should reduce by £15 billion. This should ensure that on average, banks maintain 2019 levels of headroom below a threshold.

38. The proposed thresholds are set out in Table B below:

Table B: Current thresholds and proposed thresholds for determining O-SII buffer rates

39. The FPC considered alternative policy options, consistent with the requirements set by the Buffer Regulations described in paragraphs 12 to 15.

40. For example, the FPC considered simply excluding reserves from total assets for the purpose of determining O-SII buffer rates. The Committee ultimately preferred the UK leverage exposure measure as it had the additional benefit of bringing committed but undrawn credit lines into the framework.

41. The FPC also considered maintaining the existing framework until it was next legally required to review in December 2022. The Committee preferred to consult this year on a proposal to amend the framework to give banks earlier clarity for decisions relating to capital planning, lending and their use of central bank liquidity.

42. In its original policy statement on the framework, the FPC noted that thresholds ‘could be adjusted in the future (for example in line with nominal GDP or inflation) as part of the FPC’s mandated two‐yearly reviews of the framework.’^{footnote [12]} The FPC has not ruled out indexing in future should it judge this appropriate.

Implementation

43. In order to give firms time to adjust to the proposal, the FPC has decided that the changes, if adopted, would not take effect until after the PRA’s December 2022 review of O-SII buffer rates. The changes would instead come into effect in time for the PRA to review rates under a revised framework in December 2023 based on end-2022 financial results. Rates set in 2023 would apply from January 2025.

44. The FPC also reminds firms that in October 2021, the PRA announced its intention to continue to freeze O-SII buffer rates at pre-Covid levels for a further year.^{footnote [13]} The FPC welcomed this announcement. The PRA is required to review the O-SII buffer rates once a year but, barring an unforeseen change in circumstances, has confirmed that it does not currently expect the 2022 review to result in any changes.^{footnote [14]} The ongoing rate freeze, together with the FPC’s proposals for the changes to the framework, should give banks clarity and allow them time to adapt to the proposed changes should they be implemented.

Links with other FPC policies

45. The FPC will review the UK leverage ratio framework on a regular basis in line with its statutory obligations (Section 9T of the Bank of England Act 1998). This would include reviewing the calculation of the UK leverage exposure measure, as appropriate.

46. The FPC must review its framework for the O-SII buffer at least every second year, including its methodology for measuring its criteria for assessing the risk that banks subject to the O-SII buffer could pose to the financial system. The Committee will take into account any future changes to the calculation of the UK leverage exposure measure (which may be made via future reviews of the leverage framework).

Cost-benefit analysis

47. The proposals in this consultation aim to ensure that the O-SII buffer continues to achieve the aims set out in its current framework without creating unintended costs. The proposal also retains the simple, transparent approach of the current framework. The Committee therefore expects the cost-benefit profile of the O-SII buffer to remain broadly in line with the analysis set out in the current framework. This includes the FPC’s assessment of the impact of the O-SII buffer on matters relevant to the FPC’s secondary objective, including competition and economic growth. The FPC does not expect any costs associated with implementing the proposal to be material.

How the FPC’s proposal supports the Committee’s original intention for the O-SII buffer

48. The proposal aims to ensure that the metric used to determine O-SII buffer rates more accurately reflects the risk that a distressed ring-fenced bank or large building society disrupts the supply of credit to the real economy. Moving from total assets to the UK leverage exposure measure achieves this by excluding central bank reserves, which do not reflect a bank’s potential to disrupt the credit supply, while bringing in committed but undrawn credit facilities which can form a key part of the credit supply in a stress.

49. The UK leverage exposure also meets the FPC’s original criteria for an appropriate metric as described in paragraph 18 above.

50. Under the proposal, the FPC would recalibrate the thresholds that determine O-SII buffer rates to avoid a loosening or tightening of the overall framework relative to pre-pandemic levels. The Committee therefore does not expect the proposed new thresholds to introduce any new costs overall on banks beyond those set out in the FPC’s 2016 framework.

How the FPC’s proposal addresses potential unintended costs

51. By excluding central bank reserves, the proposal mitigates the risk that future changes in central bank balance sheets inadvertently affect banks’ lending decisions by interacting with the O-SII buffer. The exclusion of reserves also removes a potential unintended constraint on banks’ decisions around use of central bank liquidity facilities. For example, it bolsters the ability of banks subject to the O-SII buffer to cushion shocks to the financial system, by allowing them to draw on central bank liquidity as necessary without becoming constrained by O-SII buffer requirements. This removes a potential impediment to the transmission of monetary policy.

Potential additional effects of implementing the framework

52. While the FPC intends to broadly maintain the original cost-benefit profile of its framework, its implementation could potentially create certain costs, but the Committee considers it unlikely that these will be material.

53. For example, the FPC recognises that some banks have a slightly higher UK leverage exposure measure relative to total assets, than other banks. For banks where the UK leverage exposure is relatively high compared to total assets, the recalibrated thresholds might reduce their headroom from the next rate threshold.

54. The FPC also notes that banks’ O-SII buffer rates determine their additional leverage ratio buffer (ALRB) so any change in O-SII buffer rate could increase the resources needed to meet the UK leverage capital buffer requirements. However, changes to the ALRB resulting from the FPC’s proposals are unlikely as no bank’s buffer rate is expected to change.

55. The FPC has performed an assessment of the policy proposals in this Consultation Paper and does not consider that the proposals give rise to equality and diversity implications.

56. The FPC is required by law to review its O-SII buffer framework at least every second year so will assess its impact on an ongoing basis.

Annex

The FPC would implement the proposals in this Consultation Paper by adding a new box to the framework.

• Box 1: Amendments to the FPC’s framework for the O-SII buffer following the FPC’s review in 2021

  1. The Financial Policy Committee (‘the FPC’ or ‘the Committee’) must have a framework for the O-SII buffer and review this framework at least every second year.^{footnote [15]}

  2. Following its December 2020 review, the FPC was not required to review the framework again until December 2022. The Committee decided, however, to undertake a review in 2021 Q3 in line with its commitment in the 2020 Q3 Record and 2020 Financial Stability Report to consider information that became available during the Covid-19 (Covid) shock about how the framework operates in stress.

  3. In response to the 2021 Q3 review, the Committee amended its framework as follows:

  (i) Changed the metric used to determine O-SII buffer rates from total assets to the UK leverage exposure measure.

  (ii) Recalibrated the thresholds that determine O-SII buffer rates to prevent an overall tightening or loosening of the framework relative to its pre-Covid level.

  4. Details of how the UK leverage exposure measure is calculated can be found in the UK Leverage Ratio Framework.^{footnote [16]}

  5. The amendment ensured that the framework still addressed the key systemic risk intended by the FPC: the risk that a distressed ring-fenced bank or large building society disrupts the supply of credit to the real economy. It achieved this as follows:

  (i) First, it excluded from the framework central bank reserves, which grew significantly during the pandemic but do not reflect a bank’s potential to disrupt the credit supply. This exclusion mitigates the risk that future changes in central bank balance sheets inadvertently affect banks’ lending decisions by interacting with the O-SII buffer. It also allows banks to draw on central bank liquidity as necessary without becoming constrained by the associated effect on buffer requirements.

  (ii) Second, it brought into the framework committed but undrawn credit facilities. Experience during the pandemic suggested that these can form an important part of the credit supply in stress.

  6. The FPC decided that the changes would come into effect in time for the PRA to assess rates under a revised framework in December 2023, based on end-2022 financial results. Rates set in 2023 would then apply from January 2025.

  7. The FPC consulted on these amendments in an FPC Consultation Paper^{footnote [17]} which was published on 15 November 2021. The Consultation Paper gives a detailed explanation of the change, including a cost-benefit analysis. The FPC’s Response sets out feedback on responses to this consultation and confirms the FPC’s final policy decision.

  Motivation for the amendments

  8. Total assets across ring-fenced banks grew significantly in 2020 (Chart A). This was not driven by lending (which stayed broadly constant) but by very high growth in central bank reserves. Such growth does not reflect an increase in a bank’s potential to disrupt the credit supply, the key externality that the FPC intends the O-SII buffer to address.

  Chart A: Aggregate assets and central bank reserves for ring-fenced banks ^{(a) (b)}

  

  Footnotes

  • (a) Ring-fenced banks of major UK banks and Nationwide.
  • (b) Reserves shown here are based on the definition used in the UK leverage ratio.

  9. The evolution in balance sheets demonstrated that total assets and lending do not necessarily move together over time in a stress, suggesting that total assets were no longer the most appropriate proxy for a bank’s potential to disrupt the credit supply.

  10. The growth in central bank reserves also had the effect of pushing banks towards O-SII buffer rate thresholds, irrespective of their potential to disrupt the credit supply. This risked creating an incentive for banks to manage the size of their balance sheets by constraining lending. It also risked influencing banks’ decisions about their use of central bank liquidity.

  11. In light of these factors, the FPC considered an alternative proxy that excluded central bank reserves would be more appropriate.

  12. Experience during the pandemic also suggested that committed but undrawn credit facilities can form a key part of the credit supply. In its May 2020 Financial Stability Report, the FPC observed that in 2020 Q1, the major UK banks had expanded their net lending by around £20 billion, as business drew down committed credit lines. This compared to a reduction in net lending of £3 billion over 2019.^{footnote [18]} Committed but undrawn credit facilities are not captured by total assets, which again suggested that total assets was no longer the optimal metric to determine O-SII buffer rates.

  13. The FPC therefore changed the metric used to determine O-SII buffer rates from total assets to the UK leverage exposure measure, as this both excluded reserves and brought committed but undrawn credit facilities into the framework.

  Detail of amendments to the O-SII buffer framework

  (i) Amended metric to determine O-SII buffer rates

  14. The FPC has changed the metric used to determine O-SII buffer rates from total assets to the UK leverage exposure measure.

  (ii) Recalibrated O-Sll buffer rate thresholds

  15. The FPC has adjusted the thresholds used to determine O-Sll buffer alongside the change in metric, in order to prevent an overall tightening or loosening of the framework relative to its pre-Covid level. The FPC has calibrated this adjustment based on financial results from 2019, before the large expansion in central bank reserves during the pandemic.

  16. Previously, thresholds were expressed in terms of total assets. Following the FPC’s amendments, the revised thresholds are expressed in terms of the UK leverage exposure measure.

  17. As at December 2019, the UK leverage exposure measure for banks attracting an O-SII buffer was on average c.£15 billion lower than those banks’ total assets. The FPC has therefore reduced the previous O-SII buffer thresholds by a constant £15 billion. This should prevent an overall tightening or loosening of the framework relative to pre-Covid levels. The revised thresholds are set out in Table 1 below:

  Table 1: Previous thresholds and recalibrated thresholds for determining O-SII buffer rates

  Buffer rate	Previous total asset thresholds (£ billions)	Recalibrated UK leverage exposure measure thresholds (£ billions)
  0%	<175	<160
  1%	175 to <320	160 to <305
  1.5%	320 to <465	305 to <450
  2%	465 to <610	450 to <595
  2.5%	610 to <755	595 to <740
  3%	≥755	≥740

  The Financial Policy Committee

  Andrew Bailey, Governor
  Jon Cunliffe, Deputy Governor responsible for financial stability
  Ben Broadbent, Deputy Governor responsible for monetary policy
  Dave Ramsden, Deputy Governor responsible for markets and banking
  Sam Woods, Deputy Governor responsible for prudential regulation
  Nikhil Rathi, Chief Executive of the Financial Conduct Authority
  Sarah Breeden, Executive Director for Financial Stability, Strategy and Risk
  Colette Bowe
  Jon Hall
  Anil Kashyap
  Elisabeth Stheeman
  Carolyn Wilkins
  Charles Roxburgh attends as the Treasury member in a non-voting capacity