Note to support the Bank of England’s Climate and Capital conference

Background

The Bank of England has a role to ensure that the financial system is resilient to climate-related financial risks (climate risks). Capital frameworks are a key part of the supervisory and regulatory toolkit to ensure that banks, building societies and insurers maintain adequate financial resources against the risks to which they are exposed. The Bank does not think capital frameworks should be used to address the causes of climate change. But, as set out in the PRA’s Climate Change Adaptation Report 2021 (CCAR)^{footnote [1]}, and as with any other risk – it does think the capital framework could be a useful tool within the broader regulatory frameworks to ensure that PRA-regulated firms are resilient to climate risks.

The current regulatory capital framework already captures climate risks to some extent through existing methodologies, for example through credit and market risk in the banking framework. But this risk capture is potentially incomplete due to difficulties in estimating climate risks (capability gaps). And there may also be challenges in capturing the risks in the existing capital regimes themselves (regime gaps). The CCAR highlighted issues fundamental to understanding the size and materiality of these gaps, including understanding:

• where these risks are best reflected in the capital regime, i.e. the appropriateness of microprudential firm-specific versus macroprudential system-wide capital requirements;
• the appropriate time horizon over which the risks should be reflected in capital requirements; and
• how to understand and deal with uncertainty over that time period, including the use of scenarios to more accurately measure and calibrate capital requirements.

The Bank published a call for research on the issues identified in the CCAR earlier this year, recognising that it is an area that would benefit from learning from a diverse set of perspectives. The Bank received a large number of submissions to this call for research, touching on many interesting points and would like to thank all those that took the time to submit their research. Submissions show that research on the core questions posed exists, but also highlighted that it remains limited.

The Bank of England’s Climate and Capital conference will explore each of these issues further to help inform the debate on the role of capital in the management of climate risks. It is not a given that changes to the capital frameworks will be required as the risks might already be adequately addressed in the capital regime, or could be addressed through the embedding of supervisory expectations on climate risks and improvements to accounting and disclosure. This conference should therefore be seen in the context of broader work on climate.

This document synthesises key points from some relevant research and broader public commentary, for each of the questions that will be discussed at the conference. These short notes are designed to provide background that can help to structure, motivate and inform the discussions at the conference, including by anchoring individual research presentations in the broader debate. As a result, the notes in particular reference those papers that are due to be presented at the conference.

The notes, which are set out in order of the conference agenda, represent a synthesis of key points from notes of participants to the conference in order to aid the debate. They do not represent Bank views or current thinking on each of the questions and are not intended to be comprehensive.

1: The appropriate place to reflect climate risk into the regulatory capital frameworks

Two sessions on 19 October - Understanding the conceptual route from climate change to capital for banks and insurers and Understand how and whether to reflect climate-related financial risk into the banking microprudential and macroprudential frameworks – consider where climate risks could be addressed in the regulatory capital frameworks. For example, for banks, should climate risks be reflected into firm-specific microprudential capital requirements or would it to be appropriate to set capital to address macroprudential risks?

Understanding the conceptual route from climate change to capital for banks and insurers

This session will conceptually explore where it might be most appropriate to reflect climate risks into regulatory capital frameworks. One piece of research that will be presented to the conference sets out a stylised framework to assess how climate risks could impact the banking regulatory capital framework.^{footnote [2]} The session will explore the themes and issues raised in this research, including how to take a coherent approach to assessing the potential links between components of the regulatory capital framework and climate change and introduce some of the conceptual challenges to overcome when considering this issue (such as uncertainty, the relevant time horizon and the interaction between minimum requirements and risk-weighted assets for banks). To provide a foundation for the subsequent conference discussions, this session will also draw out similarities and differences between banking and insurance regulatory regimes.

• Understanding the conceptual route from climate change to capital for banks and insurers

Understand how and whether to reflect climate risks into the banking microprudential and macroprudential frameworks

This session will explore research submitted on where climate risks could be incorporated into the banking capital framework. It will also take a more detailed look at how the calibration of capital for climate risks could be explored within different parts of the capital framework. The session will be structured in three parts:

• First, specific research will be discussed on how microprudential capital requirements could be adjusted to reflect climate risks if that were deemed appropriate,
• Second, it will cover how the macroprudential framework could be adjusted to reflect climate risks if that were deemed appropriate, and
• Finally, the session will cover research on where it could be appropriate to reflect climate risk across the banking capital framework.

• Understand how and whether to reflect climate risks into the banking microprudential and macroprudential frameworks

Where might be appropriate to reflect the risks across the capital framework

A foundational question is where climate risks would be optimally reflected in the capital framework.^{footnote [3]} This is a live discussion, within which, the following perspectives have been put forward:^{footnote [4]}

• Firm-specific microprudential capital elements: The microprudential capital framework aims to ensure that firms are resilient and adequately capitalised for the risks to which they are exposed. Climate risks manifest through traditional risk channels and as a result they should naturally be captured by parts of the microprudential regime. The question is therefore not whether risks should be reflected in these parts of the framework, but whether they are already adequately captured – a point explored in research submitted. And if not, whether any adjustments might better capture these risks.
• System-wide macroprudential elements: The macroprudential capital framework aims to ensure that the financial system is resilient to system-wide risks, which could include those from climate change. As well as the prudential rationale for macroprudential buffers, some research submitted observed that the role of macroprudential policymakers in the climate transition is not yet clear but (depending on remits) could provide a basis for system-wide capital requirements.
• Both micro and macroprudential elements: Climate risks are multi-faceted, diverse and will crystallise both within firms and across the system. As a result there is an argument that they will manifest in ways that make both microprudential and macroprudential tools appropriate.
• Neither microprudential or macroprudential: Some research submitted to the Bank suggest that current levels of uncertainty mean that banks and insurers should focus on building their risk management capabilities - for example in relation to valuation, accounting, and other means of capturing and measuring risks. This would better create a foundation for understanding risks that would naturally feed into microprudential tools over time.

The conference will primarily focus on the conceptual arguments for where risks might optimally sit in prudential frameworks. There are a number of practical considerations though to keep in mind as these questions are worked through. A key one is institutional remits, and the trade-offs that they provoke. For example, an institution with a microprudential remit might make a different choice to one that also has a remit that covers financial stability.

Research on climate risks and the banking capital framework

These cross-framework considerations were not addressed in much of the research received. Instead, most research focussed on specific areas of the capital framework. This section summarises research submitted, broader external views and publications on where it could be appropriate to reflect climate risks into different components of the capital framework.

Microprudential: risk-weighted assets (RWAs)

Research submitted noted that if climate change affects the riskiness of assets, it may be appropriate to reflect this in how RWAs are calculated. The challenges here are extensive though as data and models remain limited. Research that will be presented at the conference explored whether there is a risk differential between green and carbon-intensive assets in the UK residential mortgage portfolio.^{footnote [5]} The research papers assess the relative riskiness of residential mortgages (measured by credit risk/default risk) in relation to the energy efficiency (EPC) of the underlying properties. The analysis suggests that customers with an energy-efficient property are less likely to default on their mortgage.

More broadly, the Network for Greening the Financial System (NGFS) recently published a progress report on “capturing risk differentials from climate-related risks”.^{footnote [6]} It concluded that further analytical work is required to better quantify potential climate-related credit risks differentials, explore forward-looking methodologies to better identify and assess climate risks, including a better understanding of climate-related credit factors on credit ratings, and determine whether and where a potential adjustment of the existing capital framework is justified and feasible in relation to climate risks.

The session will look to explore the implications of this issue for policymaking further, including the questions of which, if any, asset classes might be prioritised for future research and any challenges to overcome.

Microprudential: minimum capital requirements

Minimum capital requirements reflect the required capital to cover unexpected losses to ensure firms remains viable with a certain level of confidence. Under Basel III, the capital adequacy ratio measures a bank’s capital in relation to its RWAs and the minimum is set at 8%. Research submitted highlights that if climate change results in a structural and permanent increase in unexpected losses, one implication may be to increase this minimum requirement.^{footnote [7]}

Changing minimum capital requirements would be a fundamental change to the framework, which would require substantial evidence of the structural shift in losses. Any changes to these requirements would be additional to the extent that climate risks are already reflected in the existing framework. This in itself is hard to determine and is an area of ongoing research and debate, which will be explored further at this conference. If climate risks would not be captured by the existing capital framework, this poses the headline question of this session, where might be appropriate to reflect the risks into the capital framework?

Microprudential and/or macroprudential: pillar 2

If it is established that aspects of climate risks are not captured within the existing microprudential framework, then an alternative to increasing the minimum capital requirement would be to make changes to Pillar 2. For example, supervisors could establish a methodology to set Pillar 2A capital for climate risks that were not captured, or not captured adequately on a microprudential basis in Pillar 1. Pillar 2 could also be used to reflect macroprudential concerns, for example, through firm buffers alongside the broader buffer framework discussed below.

Macroprudential: buffers

The banking capital regime (but not the insurance regime) contains regulatory capital buffers to ensure banks can absorb losses in times of stress without breaching minimum requirements. These buffers include the capital conservation buffer, countercyclical buffer, sectoral capital requirements and additional buffers for institutions that are deemed to be systemically important. In considering whether it would be appropriate to reflect climate risks into the macroprudential buffer framework, some research flags the importance of matching any policy action to the underlying buffer’s purpose.^{footnote [8]}

Work will be presented to the conference on papers that recognised the systemic nature of climate risks – the risks are foreseeable, irreversible and have the potential to further amplifly the shock and spillovers within the financial system.^{footnote [9]} These papers note that this may motivate a macroprudential policy response that goes beyond individual firms and ensures a consistent system-wide approach. Where this is undertaken, a consistent macroprudential approach to other systemic risks is deemed valuable in the research to help alleviate possible spillovers between the banking and non-banking sectors, reduce cross-sector arbitrage and limit risk shifting to parts of the financial system outside the regulatory perimeter. However, research highlighted that existing macroprudential tools may be less suited to recognise the systemic impact of climate risks.

These papers argue for refinement to these tools and/or development of new tools before they can be effectively used. Proposed solutions include novel hybrid instruments that combine borrower-based exposures (akin to large exposure limits) and a systemic (sectoral) risk buffer. However, the research also acknowledged that several challenges remain to any operation of tools (including modelling capabilities, dealing with uncertainty and avoiding unintended consequences) so the idea of a ‘sandbox’ has been suggested as a way forward.

2: The appropriate time horizon to reflect climate risks into the regulatory capital frameworks

Climate risks could crystallise over the short, medium and longer term. But regulatory capital requirements tend to be set over a short or medium time horizon. This raises the question as to whether the time horizons for setting of regulatory capital requirements are appropriate for climate risks – or if there could be disconnect in how the risks are captured by the prudential regimes. The panel discussion on the appropriate time horizons to reflect climate risks into the capital framework is going to get to the core of this issue, by drawing on views of regulatory, industry and academic experts.

This issue can be broken into underlying themes and questions that will each be discussed. First, what are the time horizons over which capital is set within the existing regulatory framework? Second, are there parallels between the management of climate risks and other long-term risks that are captured by the existing regulatory framework and that can inform our thinking? Finally, are the existing time horizons over which capital is set are appropriate for climate risks? And related to this question, what would be the benefits and challenges of changing them?

Background on regulatory capital time horizons

The prudential frameworks for banks and insurers are designed to ensure that firms have sufficient resources to mitigate the risks to which they are exposed to a high degree of certainty. Within the prudential frameworks, it is important to understand the interactions between different quantitative elements, including the approach to valuations, capital requirements and capital resources.

Different time horizons are considered for capitalising against the risks of unexpected losses. These regulatory time horizons take into account both the horizons over which firms might need to respond to stresses, as well as a number of practical considerations.

Both the banking and insurance regulatory frameworks largely calibrate capital requirements over a one-year time horizon. This one-year time horizon for capital-setting can be linked to assumptions around the time it might take for firms to respond to unexpected losses, including through recapitalisation, as well as data availability for modelling purposes.

But regulatory time horizons also consider risks that might crystallise over longer time horizons to ensure that firms maintain adequate financial resources against those risks. For example, stress testing frameworks often consider a time horizon of 3-5 years. This horizon is intended to be consistent with the duration of losses that firms might face from the macroeconomic shock that the test is designed to explore.

The capital frameworks also capture forward-looking information about risks beyond these time horizons which inform capital adequacy. For example, the regulatory capital framework might capture future climate risks through:

• Reference to credit ratings. Banks and insurers use internal and external credit ratings to assess the riskiness of assets, which are designed to incorporate future risks. So they may already reflect potential future changes in risks from climate change.
• Market consistent valuation approaches. For example, under Solvency II, the value of an insurer’s liabilities reflects the value of expected future cashflows over the duration of the insurance contract (which in some cases can be over 30-50 years) and asset values reflect market expectations of the value of the asset. The insurance balance sheet is then stressed to calculate the capital requirement.
• The accounting regime. Under IFRS 9, banks recognise expected credit losses on loans from default events that are possible within the next 12 months, or over the lifetime of loans where credit risk has increased significantly since initial recognition or loans are credit impaired. Accounting numbers feed into regulatory capital metrics.

While the capture of forward-looking information for climate risks is currently imperfect within the existing framework, for example as a result of data or modelling capabilities, this could improve over time with better information and disclosures. The extent to which forward-looking information will be captured by the existing regime is both complex to unravel and important to understand before a judgement can be made on whether changes to the regulatory time horizons may be warranted.

Regulators also expect firms to carry out both short-term and long-term assessments of their exposures to risks as part of scenario analysis (for example, within the internal capital adequacy assessment process and own risk and solvency assessment), in order to understand the impact of those risks on their solvency. These assessments take into consideration that over the longer term, firms may adjust their business planning and strategy to manage future risks they might be exposed to.

Climate risks and regulatory time horizons

Climate risks are likely to crystallise over a range of time horizons but how much is known about those horizons? Some climate risks are already starting to crystallise but there is uncertainty over when others will. While some risks may materialise within the regulatory time horizons, some may materialise over a much longer time horizon.

The time horizon used for assessing the risks from climate change therefore affects both regulators’ and firms’ assessment of the size and materiality of the climate risks. To illustrate this, under the late action scenario used in the Bank’s climate biennial exploratory scenario (CBES),^{footnote [10]} around 40% of the additional £110 billion of losses for participating banks are concentrated in the first five years of transition – whereas loss rates over a one-year period would be much less. This scenario was not a stressed scenario, but it shows that the time horizon chosen affects estimated losses.

Banks and insurers already capitalise against unexpected losses from other long-term financial risks under the existing regulatory framework. For example, insurers capitalise against long-term risks such as longevity and liability risks, and banks consider long-term shocks to interest rate risk in the banking book. And, as discussed above, conceptually climate risks may be increasingly captured within the existing framework through forward-looking elements. Therefore, this poses the question about whether there is something idiosyncratic to climate risks that suggests greater and earlier recognition of the risks would be required. This will be explored further during the panel.

Equally, this session will also explore what can be learnt from how firms manage other long-term risks within the time horizons in the existing framework. The related question is exposed of whether there is anything idiosyncratic to the uncertainty of climate risks and how this uncertainty would be addressed if time horizons were extended. This will be discussed further under the session Understanding uncertainty in climate-related financial risks and its implication for capital.

Should regulatory time horizons for capital be changed for climate risks?

The hurdle for changing the regulatory horizon is high and would require an evidence-based rationale on why the idiosyncrasies of climate risks make them different to other risks. If the risks from climate change were to be reflected into capital requirements over a longer time horizon, the policy intervention would require justification on the basis of its costs and benefits. This is an area of live debate and will be explored further as part of the panel session.

Submissions to the call for research papers recognised the question on what is the appropriate time horizon. While more limited submissions on this question were received, submissions were generally not supportive of changing the time horizons for calculating capital requirements. For example, research noted that the impact of this would likely be an increase in capital requirements today, which would front-load costs potentially long before the risks may materialise.^{footnote [11]} Further, some external commentary on this issue also highlights that there could be unintended consequences of reflecting financial risks associated with climate drivers over a longer period than for other drivers of those risks. As a result, it would also be important to consider whether the action was aligned with facilitating a smooth transition to net zero or whether there would be implication for provision of insurance services for certain risks. This session will look to explore some of the conceptual challenges with extending the relevant time horizons further.

While there was no widespread support for changing the time horizon for calculating capital requirements, there was however backing in the submissions for firms to consider the impact of climate risks over a longer period for business planning and risk management, for example as part of climate scenario analysis. But this in turn raises questions about how far firms should look into the future as part of their risk management? And second, how should regulators ensure that firms have credible plans to manage and mitigate the risks that might crystallise outside their usual business planning horizon?

Annex: further detail on regulatory time horizons

There are a range of time horizons in the regulatory capital frameworks. Examples of the different time horizons in the regulatory framework are set out in the Figure 1 and Figure 2 for both the banking and insurance regulatory capital frameworks respectively.

Figure 1: stylised diagram of examples of time horizons considered within the banking regulatory capital framework

In addition, under IFRS 9 banks recognise expected credit losses on loans from default events that are possible within the next 12 months, or over the lifetime of loans where credit risk has increased significantly since initial recognition or loans are credit-impaired. Timely recognition of climate-related risks in accounting valuations will support timely recognition in regulatory capital metrics.

Figure 2: stylised diagram of examples of time horizons considered within the insurance regulatory capital framework

Market-consistent approach to the valuation of an insurance company’s balance sheet under Solvency II:

Solvency capital requirement (SCR)

The SCR is calibrated to ensure all quantifiable risks are taken into account and is calibrated as a stress over a one-year horizon. It can be thought of as three periods:

3: Understanding uncertainty in climate-related financial risks and its implication for capital

Introduction

For this conference, we define uncertainty as the lack of certainty around the timing and severity of an event and what that event occurring means for losses in regulated firms and their counterparties.

Firms and financial regulators already manage uncertainty from risks within the existing prudential framework. To a degree, as touched on in previous notes from this conference, this already captures climate. Financial regulators and firms use forward-looking analytical techniques to understand and estimate losses from different risks, including through the use of scenario analysis. Firms also set risk appetites to calibrate the levels of risk and uncertainty to which they are prepared to be exposed, and then manage that uncertainty through their business planning, strategy and risk management. And regulatory capital requirements ensure that firms are able to absorb unexpected losses to a certain degree of confidence. The quantification of losses typically rely on historic data that are either used as input to models or to validate the output of models.

Analysis of how risk drivers can affect banks, such as by the Basel Committee on Banking Supervision, suggests that while climate risks can translate into traditional financial risks, more understanding is needed on the impact of climate risk drivers on firms’ exposures.^{footnote [12]} And so an important question is whether there is something inherently different about the uncertainty associated with climate risks. And if there is, what does this mean for how firms and regulators understand and manage that uncertainty.

The session Understanding uncertainty in climate-related financial risks and its implication for capital will explore these different questions and the implications for how firms and financial regulators then look to understand and deal with uncertainty around climate risks in the regulatory capital frameworks.

Does the uncertainty of climate risk differ from other financial risks?

Climate events are expected to be largely unprecedented in their frequency, severity and the speed in which they occur – giving rise to material uncertainty. Another source of uncertainty relates to the variety of transition paths that could unfold. For example, while a number of governments have committed to specific net zero targets, the transition paths are largely not set out in detail and may be subject to change. There is also uncertainty linked to whether aspects such as technological innovation (including for carbon capture) will materialise to support the transition.

The prudential framework is already intended to help manage uncertainty. This raises the question as to whether it adequately captures uncertainty in climate risks. If not, what are the idiosyncratic traits of climate risks that make them different and what do they tell us about how they should be managed?

A number of the submissions to the call for research on this topic highlighted characteristics that distinguish uncertainty in climate risks from other financial risks. These can be summarised into three main themes below:

• Historic data is not a good predictor of how climate risk affects losses of firms in the future. This makes it more difficult to understand and anticipate whether climate risks change, for example: the mean or variance of a loss distribution; the probability of default or the loss given default; and if it should be considered as additive to or fundamentally changing the financial cycle. This lack of relevant data makes quantification of risk and validation of predictive models difficult.
• There are non-linearities around physical and transition risks and the actions taken to reduce them, meaning their relationships are indirect and the impact of one on the other is difficult to predict.
• Some climate risks compound after tipping points are crossed, making them irreversible to a large extent. And while it is certain that risks will arise, there is no certainty over how and when.

To help understand the extent that climate risks are captured within the existing framework, this session will explore whether there are further differences to the above and how significant those differences might be.

How firms and financial regulators assess and deal with uncertainty around climate risks

Some of the research submitted to the call for papers suggest there are significant idiosyncrasies to the uncertainty exhibited by climate risks. So what are the implications for how firms and financial regulators should approach them? And what does this mean for relevance of existing tools and techniques when assessing resilience against risks in firms and the broader system? This session will look to explore this further, including through presentations on research submitted.^{footnote [13]}

At a high level, supervisory processes are designed so that risks are identified and mitigated. Importantly, this does not always necessitate the use of capital. Instead, there is often a heavy reliance on firms’ pre-emptively identifying and managing risks. To ensure firms are building their capabilities to do this, many regulators have issued supervisory expectations for firms to build these capabilities - the PRA’s version of these expectations were set out in its supervisory statement SS3/19.^{footnote [14]} These expectations will not be discussed in detail at this conference, but it will address the extent that they can provide comfort over how the risks are understood and managed, which could negate the need for additional regulatory capital.

The potential role of scenario analysis

The session will ask what forward-looking analytical techniques could be used by regulators to gain a better understanding of climate risks. The majority of academic submissions received focussed on scenario analysis and/or stress testing as the best available tool. These exercises seek to understand the magnitude and drivers of losses that might occur if an identified economic downturn or risk event crystallises.

Climate scenario analysis has been widely used by firms and financial regulators to explore future paths of risk channels, and build a better understanding of how they might arise. For example, the Bank recently published the results from its first exploratory climate scenario that covered banks and insurers (the 2021 CBES). For now though, there are limitations to these exercises. While the CBES delivered valuable insights about the challenges ahead, uncertainty around the impact of climate change remains extremely high and the exercise did not cover all possible risks and regulated institutions.

Some commentators have suggested that scenario analysis can not only be used to better understand uncertainty, but also be used to calibrate capital. This is a nascent area and most regulators have flagged that their exercises will not be used to directly set regulatory capital owing to: the frameworks used (e.g. fixed balance sheets); the variation in calculations; risk coverage; and the intensity of scenarios. The latter point could be important, as the existing capital framework suggests any scenario would need to provide a stress that was severe but plausible. This calibration would help to ensure that capital is set at an amount that allows for it to absorb unexpected losses for an event that was not anticipated through regular risk management practices. For climate risk, this might require a new way of thinking about appropriate scenarios given that risk appetites are difficult to formulate under uncertain transition paths. If scenarios were deemed to be the appropriate analytical tool to deal with this significant uncertainty, this session will ask what would need to be in place before it could be used to set capital requirements.

Some commentators argue that levels of uncertainty are too pronounced and could give rise to forecasting errors and a structural understatement of risks. They suggest that a more fixed, cross-system macroprudential tool is appropriate. On the other side, some researchers proposed that, given prudential measures in their view relied on relatively precise estimation of risks, a more targeted or sequenced proportional implementation of policies was a potential way forward. A minority view was that climate risk is so uncertain and complex to measure that it is impossible to manage and should therefore be minimised at all cost. In the absence of any consensus, this session will explore the merits of these tools in addressing elevated uncertainty.