FS1/23 – The prudential liquidity framework: Supporting liquid asset usability

Executive summary

This Feedback Statement (FS) provides a summary of the responses to the Bank of England (‘the Bank’) and the Prudential Regulation Authority’s (PRA) Discussion Paper (DP) 1/22 –‘The prudential liquidity framework: Supporting liquid asset usability’. The global financial crisis of 2007/2008 exposed a number of cases where banks^{footnote [1]} did not hold an adequate quantity of sufficiently liquid assets. In response, the Liquidity Coverage Ratio (LCR) was introduced to promote the short-term resilience of the liquidity risk profile of banks. The LCR requires banks to hold a large enough stock of high quality liquid assets (HQLA) to meet their payment obligations in the case of a severe short-term stress. The Bank also stands ready to use its balance sheet to provide liquidity insurance as appropriate. Taking the two improvements together, banks can draw on significantly more liquidity, in a more reliable and timely manner, than was the case going into the global financial crisis. These changes have played a central role in making the banking system safer.

DP1/22 sets out that the UK’s prudential framework is calibrated to ensure that banks have sufficient liquidity to continue their activities through severe stresses. It is important that banks feel able to draw on their liquidity, as appropriate, to reduce the risk of destabilising actions that could cause unnecessary adverse impacts on the wider economy and financial system. Therefore, while in normal times banks maintain LCRs of 100% or more, firms may draw down their HQLA even if it may mean that LCRs decline below 100% in stress.

However, the Bank and the PRA have been concerned for a number of years that banks may be overly reluctant to draw on their HQLA in periods of unusual liquidity pressures. Results from the Bank’s 2019 Liquidity Biennial Exploratory Scenario (LBES) stress testing exercise as well as international supervisory intelligence gathered during the Covid-19 stress^{footnote [2]} reinforced these concerns. Therefore, the DP sought views from banks, wider market participants, and other interested parties to understand issues surrounding HQLA usability.

This FS pulls out broad themes from the responses to the DP with the intention of providing an overall summary in an anonymised way. It summarises the responses to DP1/22 as far as they concern the matters raised in that DP. It does not include policy proposals, nor does it signal how the PRA is considering to support banks in prudently using their HQLA when facing liquidity pressures in future. In publishing this FS, the Bank and the PRA aim to contribute to improving understanding of why and to what extent banks are reluctant to draw on their stock of HQLA when facing liquidity pressures and how HQLA usability could be improved.

Most respondents agreed with the evidence presented in the DP that banks are reluctant to draw on their stock of HQLA in periods of unusual liquidity pressures. Most respondents commented that banks are concerned about regulatory reactions to initial falls in their LCRs, with such reactions including more intensive supervisory oversight and heightened regulatory reporting. Many respondents also mentioned concerns about regulatory views on the amount of time that is appropriate to rebuild HQLA buffers following a drawdown. In addition, most respondents noted that banks allowing LCRs to fall would be perceived by the market as a signal that a bank is experiencing a liquidity stress.

There was a range of views on how the Bank and the PRA could improve HQLA usability. The majority of respondents suggested that future regulatory communications in a liquidity stress should clarify the extent to which LCRs can fall and the time banks have to rebuild their stocks of HQLA subsequent to such falls. Many respondents suggested possible adjustments to how the LCR is calculated in stress, for example, adjusting the LCR’s calibration and design to reduce expected liquidity outflows in the LCR stress, and expanding the range of assets eligible as HQLA. Some respondents recommended that regulators codify in the PRA Rulebook explicitly defined reductions in liquidity requirements in a stress. Some respondents proposed these policy recommendations as part of a ‘liquidity stress playbook’, in which pre-defined adjustments to the liquidity regulatory framework would be triggered in a market disruption.

Many respondents suggested simplifications to liquidity-related disclosures in a liquidity stress, as well as recalibrations of the LCR to account for pro-cyclicality in the metric. Many respondents also emphasised that harmony between regulatory guidance and supervisory actions is required to positively impact HQLA usability. Lastly, the majority of respondents advocated greater international coordination to avoid conflicting regulatory guidance in different jurisdictions.

1. Introduction

1.1 This Feedback Statement (FS) provides a summary of the responses to the Bank of England (‘the Bank’) and Prudential Regulation Authority’s (PRA) Discussion Paper (DP) 1/22 – ‘The prudential liquidity framework: Supporting liquid asset usability’. The DP, published in March 2022, set out the Bank and the PRA’s views on the issues surrounding liquid asset usability.

1.2 The aim of this FS is to contribute to improving understanding of why and to what extent banks are reluctant to draw on their stock of high quality liquid assets (HQLA) when facing liquidity pressures, and how HQLA usability could be improved. This FS does not include policy proposals, nor does it signal how the PRA is considering to support banks in prudently using their HQLA when facing liquidity pressures in the future. This FS summarises comments on potential factors which might impact HQLA usability and potential approaches which could enhance banks’ use of HQLA in a liquidity stress. Responses are summarised for each question asked in DP1/22.

1.3 The responses described in this FS are presented in an anonymised way. This FS pulls out broad themes from the responses to DP1/22 with the intention of providing an overall summary of the responses. Not all respondents answered all questions, and the grouping of responses into themes reflects the PRA’s judgements about how responses can be grouped into different categories. While this FS does not refer to all of the comments in the responses nor reflect the level of detail in the responses, the PRA will be making use of all the responses as it continues to consider issues around HQLA usability.

Background

1.4 DP1/22 considered the usability of banks’ stocks of HQLA and sought views from banks, wider market participants, and other interested parties to improve understanding of the issues around HQLA usability. The Bank and the PRA have been concerned for a number of years that banks may be reluctant to draw on their HQLA in periods of unusual liquidity pressures. It is important that banks feel able to draw on their HQLA as appropriate to reduce the risk of contractionary or destabilising actions.

1.5 The Bank and the PRA welcomed comments to DP1/22, including answers to the questions laid out in it. The DP stated that the Bank and the PRA may publish a summary of the comments received.

Responses

1.6 The Bank and the PRA received 22 written responses to DP1/22 from PRA-regulated banks and trade bodies, in addition to a number of comments received in meetings with PRA-regulated banks, trade bodies, credit rating agencies, and investors.

2. Evidence on banks’ willingness and ability to draw on their stock of HQLA in a stress

2.1 Chapter 3 in DP1/22 outlined the Bank and the PRA’s evidence on banks’ willingness and ability to draw on their stock of HQLA in a stress. The results from the Bank’s 2019 Liquidity Biennial Exploratory Scenario (LBES) as well as subsequent domestic and international supervisory intelligence gathered during the Covid-19 stress suggested that banks may be reluctant to draw on their stock of HQLA where this would result in falls in the Liquidity Coverage Ratio (LCR), particularly to below 100%, the regulatory standard in normal times. Instead, banks may prefer taking actions that support the LCR but that may negatively impact on financial markets and the real economy.

Q1: How do your perceptions of banks’ willingness to draw on their stocks of HQLA compare with the evidence presented in this section?

2.2 Most respondents responded that, in general, banks are unwilling to draw down their stocks of HQLA, especially if this caused significant falls in the LCR, particularly to below 100%, and were in broad agreement with the evidence presented in the DP.

2.3 In the context of Covid-19, the majority of respondents commented that banks were either hesitant to allow their LCR to fall and/or they took actions to bolster their liquidity buffers in response to the uncertainty surrounding the nature and length of the stress. However, many respondents flagged that Covid-19 was not a pure liquidity stress and that banks did not have to actively defend their liquidity position, with a few respondents noting that banks’ liquidity profiles benefited from the monetary and fiscal measures, as well as central bank liquidity facilities, that were in place. As a result, evidence from the Covid-19 stress of banks’ reluctance to draw on their stocks of HQLA is somewhat limited.

Q2: To what extent would market participants be comfortable with banks drawing on their stock of HQLA to meet unexpected liquidity demands, and with accompanying falls in LCR?

2.4 Most respondents noted that a falling LCR and/or a reduction in HQLA would be perceived by the market as a sign that a bank is experiencing a liquidity stress. Some respondents commented that there is a heightened focus from market participants on regulatory metrics during a stress.

2.5 Respondents differed in their views on the scale of the LCR change that would trigger a negative market reaction. Many of the respondents commented that markets would react negatively to an LCR below 100%. Some respondents noted that market reactions could impact on bank funding costs and on a bank’s ability to raise liquidity. A few respondents also noted that market reactions could exacerbate liquidity stress on the bank. Some respondents mentioned that a substantial drop in LCR, even where it remained above 100%, would have the same effect. A few respondents expanded on this point and mentioned that it could be interpreted as a signal of liquidity stress and/or poor liquidity management. Some respondents commented that market participants would be wary of an LCR close to, but above, 100%.

2.6 Many of the respondents stated that market participants consider a bank’s LCR in relation to its peer group. If one bank allows its LCR to fall, it would look like an outlier and consequently be penalised by the market. More feedback on this is provided under question 13.

2.7 Many respondents highlighted that a bank’s liquidity position is considered by credit rating agencies (CRAs) and therefore expressed concerns that falls in the LCR could be reflected in a lower credit rating.

2.8 CRA respondents noted that they consider liquidity and funding as part of their bank rating methodologies. In normal times, liquidity and funding tend to be less of a ratings differentiator than solvency, but attention would probably turn to them in times of liquidity stress. The LCR metric is only one aspect of these assessments, as CRAs recognise its limitations and look beyond regulatory metrics to form a holistic view of banks’ liquidity and funding profiles. CRAs noted that HQLA is intended to be usable in stress, but understood banks’ apparent reluctance to allow LCRs to fall given concerns around potential regulatory and market reactions. Two CRAs had released statements during the Covid-19 stress addressing associated concerns, including by clarifying that a bank’s use of its liquid asset buffer would not by itself necessarily trigger ratings actions.^{footnote [3]}

2.9 One respondent mentioned that consistency would be needed in communications across all CRAs.

3. How could HQLA usability be improved?

3.1 Chapter 4 in DP1/22 noted that banks may be constrained in their willingness and ability to draw on their HQLA due to concerns around regulatory reactions and market reactions to falls in the LCR. It welcomed views on the nature and extent of these concerns, to understand better if there are ways in which the Bank and the PRA could consider supporting banks in prudently using their HQLA when facing liquidity pressures in the future.

Concerns around regulatory reaction

3.2 DP1/22 noted that banks may be concerned about uncertainty around regulatory consequences of usage of HQLA, and/or about known regulatory consequences of usage – such as more intensive supervisory monitoring.

3.3 During the Covid-19 stress, regulatory authorities in a number of jurisdictions made public statements and took public actions intended to support HQLA usability. While these actions appeared to be supportive of banks drawing on their HQLA to some extent, the majority of banks were reported as leaving the internal targets and limits they have in relation to the LCR unchanged.

Q3: How do banks’ internal LCR targets affect HQLA usability? To what extent would it be feasible and appropriate for banks to adjust or tailor internal targets for different scenarios?

3.4 The majority of respondents to this question commented that internal targets are a buffer to ensure that a bank stays above the regulatory standard, and confirmed that banks’ internal LCR targets inhibit HQLA usability. Many of the respondents noted that a bank will take defensive actions when it is at risk of breaching internal targets. Some respondents emphasised that extensive internal governance processes are triggered once internal LCR targets are breached and so are to be avoided. A few respondents mentioned that for smaller banks, internal risk appetites affect HQLA usability more than market-related reasons, with one respondent commenting that this is a result of not being rated externally.

3.5 The majority of respondents to this question provided views on what informs internal targets. For example, many respondents commented that internal LCR targets are informed by regulatory standards, and they, therefore, felt that greater regulatory guidance on when the LCR can dip below 100% could help a Board tailor internal risk appetites for different scenarios. Some respondents commented that internal risk appetites are informed by internal stress tests, which assume a more severe scenario than underlies the LCR. Some respondents commented that risk appetites are informed by recovery plans, which generally include LCR triggers of above 100% for recovery options. Recovery options refers to potential actions that firms could take to rebuild liquidity buffers.

3.6 Of the respondents that provided an opinion on how feasible it would be to adjust internal targets, the majority believed that it would not be easy. Some respondents emphasised their Board’s hesitancy to adjust internal risk appetites, especially in the absence of regulatory guidance concerning the extent to which it would be appropriate to lower the LCR in a stress. A few respondents commented that factors contributing to such hesitancy include heightened uncertainty during times of stress and individuals’ responsibilities under the Senior Managers Regime. One respondent noted that it might be difficult to adjust targets for short-term scenarios given that internal risk appetites are informed by the need to comply with the Overall Liquidity Adequacy Rule (OLAR).^{footnote [4]}

Q4: To what extent did authorities’ communications around HQLA usability during the Covid-19 stress support banks in using their HQLA?

3.7 Respondents provided a wide range of opinions on the extent to which authorities’ communications around HQLA usability during the Covid-19 stress were helpful. Some respondents commented that authorities’ public communications were welcome and helpful to steer internal conversations on using HQLA. Some respondents commented that authorities’ communications were only partially effective because they did not overcome banks’ hesitancy to draw on their stock of HQLA in a stress. Some respondents were of the view that regulatory communications did not result in banks using their HQLA, with one respondent noting that, in hindsight, regulatory communications could have been issued earlier. One respondent felt that several banks had increased liquidity conservation in response to regulatory communications.

3.8 In the context of the regulatory framework, some respondents to this question attributed banks’ reluctance to draw on their stock of HQLA, despite regulatory communications, to the regulatory framework, which in their view suggests that banks must always comply with 100% LCR.^{footnote [5]} A few respondents were of the view that regulatory communications did not reduce the market stigma associated with a fall in the LCR. A few respondents noted that the uncertainty surrounding the Covid-19 stress incentivised banks to maintain their stock of HQLA.

3.9 Some respondents commented that the guidance from regulatory authorities on capital buffers was more effective than that on liquidity because it was more detailed and clearer. A few respondents emphasised the need for consistency between regulatory communications and supervisory messaging.

Q5: What forms of communication and guidance were, and would be, most effective?

3.10 The majority of respondents to this question suggested that, to make authorities’ communications more effective in supporting HQLA usability, they should clarify the extent to which banks’ LCRs can fall without regulatory consequences. For example, the majority of respondents suggested that the PRA could communicate a temporary reduction in the LCR standard and referred to approaches taken by other central banks to lower it during the Covid-19 stress, eg the Reserve Bank of India. However, a few respondents mentioned that they were unsure if banks would make use of a temporary reduction in the LCR standard, given that there could still be an issue of stigma for individual banks with lower LCRs compared to peers (see question 13).

3.11 Respondents also provided a variety of other suggestions. For example, many of the respondents suggested that regulators issue guidance on the time over which banks are expected to restore their liquidity buffers following a drawdown in stress. Some respondents expressed the view that regulatory communications announcing that a liquidity stress has begun would be helpful, especially if there was a liquidity stress playbook delineated in the regulatory framework (more information is provided under question 15).

3.12 In terms of how regulatory communications should be delivered, in addition to emphasising clarity, some respondents asked for widespread communication from supervisors to ensure that regulatory guidance is reflected in supervisory actions. Many respondents argued that communications would be more effective if coordinated internationally. Some respondents expressed the view that regulatory communications must be widely communicated to reach all market participants while also being targeted at banks, for example by taking the form of a ‘Dear CEO’ letter.

Q6: How much are banks concerned about regulatory reactions to initial falls in LCR, and how much about potential regulatory views on the timeline for rebuilding HQLA stocks?

3.13 The majority of respondents to this question commented that banks are concerned about regulatory reactions to initial falls in LCR. Regulatory reactions mentioned by respondents included: more intensive regulatory and supervisory oversight; the need to develop a restoration plan; and heightened regulatory reporting. A few respondents commented that banks believe that there would be more intensive supervisory engagement following a fall in the LCR, even if it remained above 100%.

3.14 Many of the respondents emphasised that there is a high degree of uncertainty surrounding the consequences of breaching the regulatory standard, which contributes to banks’ reluctance to use buffers, even when HQLA usage is being promoted by authorities. Respondents put forward a variety of reasons behind this uncertainty. For example, a few respondents emphasised that the capital regulatory framework – in contrast to the liquidity framework – more clearly defines what part of the buffer is usable.

3.15 In addition to concerns surrounding regulatory reactions to initial falls in LCR, many of the respondents commented that regulatory views on the timeline for rebuilding HQLA buffers would impact banks’ reluctance to use buffers, and that currently there is a lot of uncertainty surrounding this timeline.

Concerns around market reaction

3.16 DP1/22 noted that banks’ willingness to use HQLA and allow LCRs to fall may also depend on how they think the market would react to such falls. In a liquidity stress, there are potential drawbacks as well as potential benefits to prudential regulatory disclosures. Transparency reduces the risk that market participants’ fears about one bank spread to other banks. However, market concerns about an individual bank’s LCR may prompt a self-fulfilling liquidity crisis. There may also be a collective action problem – a ‘stigma’ attached to a bank being an outlier compared to peers.

Q7: What may be driving a potential stigma around banks allowing LCRs to fall? How much do you think a potential stigma may be around falls in LCR in and of themselves, and how much around concerns around how quickly LCRs can be restored?

3.17 Most respondents to this question attributed the stigma to the disclosure of the LCR. Specifically, respondents noted that the LCR is a highly visible benchmark metric that indicates to the market the liquidity profile of the bank and allows for comparison across peer banks. Some respondents attributed the stigma to the regulatory framework, in which the market may interpret a 100% LCR as a regulatory minimum in all circumstances. Some respondents also noted that market participants lacked the necessary context to interpret falls in the LCR, and whether they indicated that a bank is facing liquidity stress. For example, a few respondents emphasised that the market has less information than regulators.

3.18 As discussed under question 2, many respondents were of the view that there is a market stigma associated with falls in the LCR, especially to below 100%. Very few respondents provided feedback to the second part of the question but of those respondents that did, the majority believed that market participants are predominantly concerned about falls in the LCR, with the time required to rebuild liquidity buffers a secondary concern. A few respondents stated that both factors are important to market participants.

3.19 A few respondents commented that the extent of market concerns around how quickly LCRs can be restored is conditional on factors such as regulatory guidance, and the ability of the bank(s) in question to rebuild liquidity buffers. Respondents noted that the latter might be more difficult for smaller banks.

Q8: To what extent is it challenging for market participants to interpret the signals they receive from LCR-related disclosures when banks are facing liquidity pressures?

3.20 Of the respondents to this question, most answered that it is challenging for market participants to interpret LCR-related disclosures when banks are facing liquidity pressures. For example, a few respondents observed that the calibration of the LCR means that the pre-stress and post-stress LCR are not equivalent metrics. This is partly because the LCR applies constant outflow rates at all times, even if some outflows have crystallised in a stress. As a result, the LCR in a stress could be a conservative measure of a bank’s liquidity position, and market participants could incorrectly interpret this figure.

3.21 Some respondents commented that the LCR is an inherently volatile metric (discussed further in question 10). A large drop in the LCR may not necessarily reflect realised outflows and liquidity pressures. One example given by respondents included a situation where the residual maturity of liabilities enters the 30-day window used in LCR calculations – causing a drop in the LCR – and then moves out when the liabilities are rolled over – causing an increase in the LCR. Another example was the ‘ratio effect’, where there can be a similar change to the numerator (HQLA) and denominator (net outflows) but a sharp drop in the overall LCR ratio.^{footnote [6]} A few respondents noted that the LCR components are difficult for analysts to understand.

Q9: What impact do regulatory liquidity disclosures have on HQLA usability? How might regulatory liquidity disclosures be improved?

3.22 The majority of respondents to this question commented that regulatory liquidity disclosures have a negative effect on HQLA usability. As discussed under question 7, the LCR is a highly visible regulatory metrics that the market uses to assess a bank’s liquidity soundness. Consequently, banks want to show a strong liquidity position through their regulatory disclosures and will be hesitant to reduce their HQLA if it would cause a substantial drop in the LCR.

3.23 Many respondents noted that the 12 month-end average used in the disclosure of the LCR in the UK smooths out small dips in a bank’s LCR, which supports HQLA usability. However, a few respondents mentioned that the benefits associated with the 12 month-end average LCR are reduced by banks’ legal obligation to disclose inside information without delay under the Market Abuse Regulation (MAR), with one respondent noting that senior managers may feel duty-bound to disclose the spot LCR if their LCR falls significantly or below 100%. A few respondents also mentioned that market participants can deduce when a firm has significantly drawn on its HQLA from financial statements. For more information on this see question 11.

3.24 Most respondents who provided suggestions on how regulatory liquidity disclosures could be improved recommended that liquidity-related disclosures be simplified. For example, many respondents suggested reducing the amount of detail in public disclosures to improve HQLA usability. A few respondents suggested that LCR disclosures be suspended in a stress. A few respondents suggested a change in disclosure requirements so certain subsidiaries are not required to disclose their liquidity position. This is because, in their view, a fall in the LCR of a subsidiary could result in a negative market reaction that, in turn, could have knock-on impacts on other subsidiaries and the parent bank. A few respondents suggested that disclosure requirements be modified so that there is a longer lag in banks’ disclosures. This, in their view, could give banks more time to assess the situation and draw on their liquidity buffers without concerns around potential short-term negative market reactions.

3.25 A few respondents promoted the opposite approach and suggested additional liquidity disclosures to reduce the emphasis on the LCR.

Q10: How do factors that drive LCR volatility contribute to concerns around HQLA usability? Which factors are most important?

3.26 The majority of respondents to this question observed that the LCR is a volatile and/or pro-cyclical metric. Many of the respondents also noted that LCR volatility negatively affects HQLA usability in times of disruption.

3.27 Of the respondents that provided more details on the drivers of LCR volatility, most highlighted large movements in banks’ net outflows, which is the denominator of the LCR and is a measure of the expected outflows of a bank’s liabilities minus its expected inflows in the LCR stress. Specifically, the majority emphasised the role of the Historical Look-Back Approach (HLBA) within the LCR calibration. The HLBA captures the risk of future variation margin outflows associated with derivatives, and is calculated using margin flows over the previous two years. If a bank experiences larger margin inflows or outflows than in the previous two years, LCR net outflows increase, which reduces the LCR, such that the LCR behaves cyclically. Many respondents attributed LCR volatility to the calibration of outflow rates, which are used in the calculation of overall net outflows, in the LCR. For example, respondents mentioned that the LCR continues to apply the same outflow rates in normal times and times of stress, including when banks have experienced actual outflows.

3.28 In line with these concerns, most respondents to this question suggested adjustments to the LCR calibration within a liquidity stress to reduce the pro-cyclicality of the metric. For example, most of the respondents proposed that regulators freeze the HLBA calculation at pre-stress values during a liquidity stress. The majority of respondents suggested a recalibration of outflow rates once a bank has experienced unusual liquidity demands to reflect that the outflows have crystallised and the risk of further outflows has fallen. However, one respondent expressed concern about the recalibration of outflow rates post-stress. Specifically, that it might make the metric too complex and market participants might find it difficult to interpret this metric.

3.29 A few respondents also mentioned that a bank’s LCR volatility is conditional on a variety of factors. For example, running a higher LCR or having a smaller balance sheet could be accompanied by higher volatility.

Q11: Why do some banks disclose spot LCR? How does the practice of spot LCR disclosures affect HQLA usability in times of liquidity pressure?

3.30 Respondents provided a variety of reasons for why some banks disclose spot LCR. Some respondents noted that they choose to disclose spot LCR to signal their strength to the market. This is because the spot LCR provides transparency on their liquidity position and provides up-to-date information. Some respondents commented that they choose to disclose spot LCRs to reduce uncertainty, as they believe increasing routine disclosures offers greater transparency for the market. Some respondents commented that they disclose spot LCR because other reporting disclosures, such as those in financial statements, are done on a spot basis and together they give the market a holistic picture of their financial position. A few respondents noted that banks have a legal obligation to disclose inside information without delay under MAR (or equivalent overseas regulation), with one respondent noting that senior managers may feel duty-bound to disclose the spot LCR if their LCR falls significantly or below 100%. A few respondents also believed that routinely disclosing the spot LCR would lessen the surprise for market participants if their spot LCR had to be disclosed under MAR.

3.31 The majority of respondents to this question believed spot LCR disclosures negatively affect HQLA usability in times of liquidity pressure due to the market stigma of disclosing an LCR below 100%. A few respondents also thought that if a bank did not disclose their spot LCR, the market would be able to estimate it from other disclosures (eg Pillar 3 disclosures), perhaps inaccurately, which could have negative consequences.

3.32 A few respondents reiterated that because the LCR is a volatile metric (see question 10), spot disclosures do not convey helpful information. Additionally, a few respondents mentioned that, in order to report a high spot LCR, firms might ‘window dress’ the metric.

Q12: What would the potential costs and benefits be of changing prudential regulatory LCR disclosures to be more in line with the Basel (typically 90-day) averaging approach?

3.33 DP1/22 sets out that, while prudential regulatory disclosure rules require banks to disclose relevant figures as averages to reduce volatility, averaging requirements can differ. UK and EU prudential regulatory disclosure rules require that figures are averaged over the previous 12 month-ends; while the Basel disclosure standard entails daily averaging over the previous quarter (typically 90 days). 

3.34 The majority of respondents to this question believed that the costs of implementing the Basel averaging approach for LCR disclosures would outweigh the benefits. This is primarily due to the time and operational costs involved in calculating the LCR on a daily basis to an auditable standard. The only benefit, mentioned by a few respondents, was that a 90-day average could improve transparency.

3.35 In terms of HQLA usability, several respondents highlighted that the Basel approach could further discourage banks from operating below 100% LCR. This is because it would reduce the averaging period relative to the existing 12 month-end average approach, and would therefore be more likely to reveal when a bank has fallen below 100% LCR.

Q13: How and to what extent may a collective action problem be a factor in limiting HQLA usability?

3.36 All respondents to this question were of the view that the collective action problem is a factor in limiting HQLA usability. For example, many respondents noted the ‘first-mover’ disadvantage ie where the first bank to report a fall in LCR would be perceived negatively relative to its peers. A few respondents commented that this is likely to result in a negative market reaction, which could trigger a liquidity stress on the bank.

3.37 A few respondents commented that overcoming the collective action problem would require a coordinated fall in the LCR across banks and that regulators would have to play a role in organising this action.

Q14: What other factors may impact on HQLA usability?

3.38 Factors that may impact on HQLA usability mentioned in response to the previous questions included:

• market stigma (including the collective action problem);
• internal risk appetite, which is informed by regulatory metrics and recovery plans;
• the inherent uncertainty of the nature and duration of a liquidity stress;
• firms’ concerns about regulatory reactions;
• the design and calibration of the LCR (including volatility of the metric); and
• liquidity-related regulatory disclosures, as well as other disclosures, eg spot LCR and financial statements.

3.39 The majority of respondents commented that banks consider the wider regulatory framework when making liquidity decisions, including for example the overall liquidity adequacy rule (OLAR) and internal liquidity stress testing. Respondents also commented on the importance of broader prudential risk management in affecting decisions regarding HQLA usage. For example: firms would have to take into account the impact on capital positions of drawing down HQLA; and actions motivated by market risk, credit risk or capital management concerns could bolster the LCR. A few respondents commented that banks are cautious about their liquidity positions given previous financial crises, which made firms highly risk averse concerning liquidity. One respondent mentioned that a group’s willingness to draw on its stock of HQLA will also depend on the entity facing liquidity pressures and whether that is the same entity that holds surplus HQLA. One respondent commented that a bank is less likely to draw on its HQLA in an idiosyncratic stress, compared to a market-wide stress.

Q15: What other approaches could enhance banks’ use of HQLA in times of unexpected liquidity needs?

3.40 Factors that could enhance banks’ use of HQLA in times of unexpected liquidity needs mentioned in response to the previous questions included:

• regulatory communications during a market-wide stress, communicating a reduced LCR requirement, and providing guidance on appropriate timelines for rebuilding the LCR;
• adjustments to LCR-related disclosures; and
• recalibrating the LCR to account for pro-cyclicality in the calculation in times of market disruption, for example, by freezing the HLBA at pre-stress values and adjusting outflow assumptions to recognise the risks that have already crystallised.

3.41 The majority of respondents to this question made various policy recommendations to enhance banks’ use of HQLA. For example, some respondents suggested codifying in the PRA Rulebook a reduced LCR standard in a stress. Some respondents suggested reductions in Pillar 2 liquidity add-ons in a stress, moving more of a bank’s Pillar 1 requirements to Pillar 2 (which are not disclosed), and providing clarity on how actions taken by banks in a stress (eg debt buy-backs) would feed into setting Pillar 2 add-ons in the future. Respondents to this question also endorsed recalibrations of the LCR components during stress. For example, many respondents recommended adjustments to the calculation of net outflows, including loosening the ‘inflow cap’ – which limits total expected cash inflows to 75% of total expected cash outflows – or reducing outflow rates by applying a percentage scalar. By decreasing the denominator of the LCR calculation in a stress, in their view, banks could have a lower value of HQLA but still maintain the same LCR. The majority of respondents recommended changes to the HQLA-eligibility criteria to include a wider range of assets, for example, pre-positioned collateral at central banks. A few respondents suggested adjusting HQLA haircuts to reflect realised changes in market prices. A few respondents suggested an easing of downgrade triggers^{footnote [7]} or suspending the price decline test.^{footnote [8]} A few respondents suggested the introduction of a ‘counter-cyclical buffer’ in the LCR, akin to the capital framework.

3.42 Many respondents made these policy recommendations as part of a ‘liquidity stress playbook’ to provide banks with greater assurance on the regulatory consequences of dipping below 100% LCR. They recommended that this pre-defined playbook could be an adjusted liquidity regulatory framework triggered in a liquidity stress. Some respondents emphasised the importance of clear guidance and the potential benefits of a playbook across the whole liquidity framework and broader regulatory framework. One respondent expressed the view that changing the mind-set around HQLA usability would be challenging, and would require sustained regulatory engagement with the broader set of market participants and bank stakeholders.

3.43 Building on their suggestions for how to enhance banks’ use of HQLA in times of stress, respondents provided details on how, in their view, regulators should implement these recommendations. For example, some respondents recommended setting out explicit criteria in the PRA Rulebook to determine what constitutes a liquidity stress that would trigger the adjusted playbook. Many respondents emphasised that in a liquidity stress, they would require confirmation that regulatory communications, guidance, and temporary changes to the regulatory framework would be reflected in supervisory actions. Most respondents advocated for greater international coordination to avoid conflicting regulatory guidance in different jurisdictions.

3.44 Outside of a liquidity stress, some respondents suggested that UK regulators could change the rules and permit the LCR to fall below the regulatory standard on average or for a few days, before activating heightened regulatory disclosures and monitoring.

4. Questions

4.1 The questions listed below are all the questions that appeared in DP1/22.

Q1: How do your perceptions of banks’ willingness to draw on their stocks of HQLA compare with the evidence presented in this section?

Q2: To what extent would market participants be comfortable with banks drawing on their stock of HQLA to meet unexpected liquidity demands, and with accompanying falls in LCR?

Q3: How do banks’ internal LCR targets affect HQLA usability? To what extent would it be feasible and appropriate for banks to adjust or tailor internal targets for different scenarios?

Q4: To what extent did authorities’ communications around HQLA usability during the Covid-19 stress support banks in using their HQLA?

Q5: What forms of communication and guidance were, and would be, most effective?

Q6: How much are banks concerned about regulatory reactions to initial falls in LCR, and how much about potential regulatory views on the timeline for rebuilding HQLA stocks?

Q7: What may be driving a potential stigma around banks allowing LCRs to fall? How much do you think a potential stigma may be around falls in LCR in and of themselves, and how much around concerns around how quickly LCRs can be restored?

Q8: To what extent is it challenging for market participants to interpret the signals they receive from LCR-related disclosures when banks are facing liquidity pressures?

Q9: What impact do regulatory liquidity disclosures have on HQLA usability? How might regulatory liquidity disclosures be improved?

Q10: How do factors that drive LCR volatility contribute to concerns around HQLA usability? Which factors are most important?

Q11: Why do some banks disclose spot LCR? How does the practice of spot LCR disclosures affect HQLA usability in times of liquidity pressure?

Q12: What would the potential costs and benefits be of changing prudential regulatory LCR disclosures to be more in line with the Basel (typically 90-day) averaging approach?

Q13: How and to what extent may a collective action problem be a factor in limiting HQLA usability?

Q14: What other factors may impact on HQLA usability?

Q15: What other approaches could enhance banks’ use of HQLA in times of unexpected liquidity needs?