Speech
Thank you for inviting me to speak today at the ABI’s Climate Summit, to talk about the results of our Climate Biennial Exploratory Scenario exercise – or CBES in case you want another regulatory acronymfootnote [1]. An exercise that wouldn’t have been possible without the considerable time and effort of the life and general insurers (firms) that participated in the exercise over the last 18 months – helping to complete what is regarded as one of the most comprehensive climate scenario exercises of any regulator to date.
The objective of the CBES exercise was to examine climate-related financial risks that might unfold over a timescale of thirty years or more. The exercise considered two possible routes to net-zero carbon dioxide emissions globally by 2050: an Early Action scenario and a Late Action scenario. In both of these scenarios, climate risks have been managed, reaching net-zero by 2050. A third scenario explored the physical risks that would begin to materialise if governments around the world fail to enact policy responses to global warming, hence No Additional Action is taken to address climate change. These scenarios are not ‘worst case’ scenarios. For example, they did not include potential additional losses caused by mass migration or conflict.
The idea was to examine the industry’s resilience to those risks, to see how the industry might respond to that evolution, and to consider the possible impact of this response to the wider economy. This should provide insight for industry, regulators and government - by shining a spotlight on very long-term and otherwise opaque risks - to help decision-making and drive better risk management today.
So in this speech today I want to focus on drawing out the lessons learnt both for insurers and for the Bank of England (Bank) and Prudential Regulation Authority (PRA) across each of the three CBES core objectives – climate risk management, sizing the impact of climate risks and understanding longer-term implications on business models.
Climate risk management
The CBES was both deep - requiring a demanding level of modelling detail - and broad - including a wide-ranging questionnaire on risk management and business model responses. It was always about more than just the numbers. And, though the CBES has several novel features, we should acknowledge that this is not all new – certainly not to insurers.
Many of you will have participated in the Climate Insurance Stress Tests in 2019footnote [2], or contributed to the PRA’s industry survey starting in late 2020, which assessed insurers against the expectations set out in our supervisory framework on managing the financial risks from climate change (Supervisory Statement3/19)footnote [3]. The CBES results have shown that the largest insurers are making good progress in some aspects of their climate risk management. For example on risk governance there was substantial evidence of senior accountability, on risk appetite, climate risk is now common for many firms, and most firms have included, or are planning to include, climate scenarios within their own risk and solvency assessment (ORSA).
That said, there is still much more to do to understand the exposures to climate risks. The two most common issues that insurers face in managing climate risks are as follows:
The lack of comprehensive and high quality data. There were two areas that insurers found challenging to obtain sufficient data: information on corporate emissions across value chains (also known as Scope 3 emissions) and geographical corporate asset location data. Addressing data gaps for climate analysis is a priority if insurers are to deliver effective climate risk management, and to innovate and develop products to support the transition to a more climate-sustainable pathway.
Insurers might need to invest in better processes to improve data capture across their operations, customers and investments. At the same time, regulators have their part to play in setting standards, such as those put forward by the International Sustainability Standards Board (ISSB).
Use of third party models. Nearly all firms were quite heavily reliant on third party models to complete their CBES submissions. While that is not a problem in and of itself, some firms struggled to adapt these models to the specific CBES scenario, or to effectively understand and challenge the outputs of these models. This applies both to the asset models of life insurers, and to the catastrophe models of general insurers.
When modelling general insurance liabilities, the better approaches incorporated a wide range of physical perils, beyond those of the most readily available catastrophe models. They identified limitations of the third party models used, and made adjustments to address these limitations. We have included a box in the publication, which sets out some more examples of good practice we saw in the CBESfootnote [4].
Sizing of risks
We wanted to use the CBES to get a picture of the size of climate financial risks in the insurance industry. The CBES was not meant to be a capital setting exercise, but an exploratory exercise designed to improve capabilities of both the Bank and the CBES participants. Expertise in modelling climate-related risks is still relatively immature, and unsurprisingly, insurers’ results spanned a wide range. This reflects the uncertainty, the sensitivity to assumptions as well as differing methodologies.
With these words of caution in mind, I would like to turn to looking at what the numbers tell us. The aggregate results show that, for life and general insurers, the No Additional Action scenario would be likely to have a more significant impact than either of the two transition scenarios.
Life Insurers’ Exposure
I would like now to turn to sizing the exposures of life insurers, where we looked at invested assets across their entire balance sheet, including unit-linked and ring-fenced funds. We estimated the size and scale of the insurance industry in each scenario, compared with a hypothetical scenario where economic growth is unaffected by climate risk.
Chart 1 shows that the loss of life insurers’ asset values, and how these are largely mirrored by falls in liability values. This is partly because matching assets and liabilities respond to interest rate effects in the same way, and partly because policyholders bear the investment risk in unit-linked funds. What we found is that only a minority of the investment losses act to reduce the capital resources of our insurance firms.
There are two important insights here. The first is that that policyholders may be more exposed to climate risks than the insurer, and this could have a wider macroeconomic effect. The second is that a life insurer’s mix of business explains most differences in their climate vulnerability.
General Insurers’ Exposure
For general insurers, we focused on property classes and the losses stemming from the direct impact of physical climate risk. Chart 2 indicates the size of the impact in Year 30 of the No Additional Action scenario, compared to current weather conditions. You can see here:
- US average annual loss is dominated by Tropical Cyclone, and increases by around 70%; and
- UK average annual loss is dominated by Inland and Coastal Flood, and increases by around 50%.
As I mentioned, results were sensitive to assumptions, and not all models could incorporate all assumptions – for example the rainfall intensities that we prescribed. To better assess the impact of those assumptions, we undertook a sensitivity on the UK flood results by overlaying a consistent approach. The results in Chart 3 show that, with this closer alignment, insurers could have experienced materially more severe losses – up to doubling of their estimates.
Summary on Sizing the Risk
So to summarise the insights from sizing the risk for life and general insurers:
- Models are sensitive to input parameters, and there is a potential for results, to be more severe.
- Across all three CBES scenarios, the projections show that if insurers do not respond effectively, climate risks are likely to cause a persistent and material drag on their profitability. In the No Additional Action scenario we saw an annual drag on insurance profits of around £1.2bn. Losses of this size would make individual insurers, and the financial system overall, more vulnerable to other future shocks.
- But overall costs to insurers from the transition to net-zero should be absorbable, partly because some losses are passed to policyholders through lower returns in savings and retirement products. Those costs will likely be much lower if early, well-ordered policy action is taken.
Climate Litigation
The CBES exercise also incorporated an investigation into litigation risk. Climate-related litigation is increasing globally; claims are being initiated against businesses - and their directors - with the aim of changing corporate behaviour, or seeking compensation for consumer or investment losses. The aim of this exercise was to illustrate the potential scale of climate litigation risks, and help firms develop processes to monitor and manage such risks. In the exercise, we set out seven hypothetical legal cases to explore possible exposures of the London Marketfootnote [5].
In Chart 4 we can see that Directors’ and Officers’ policies were the most likely to pay out. These policies were considered particularly vulnerable to three of the hypothetical legal cases:
- the greenwashing claim,
- the claim based on a breach of fiduciary duties, and
- the claim based on the indirect financing of carbon emissions.
This was the first time many insurers had attempted to draw together the necessary information, and some firms faced difficulty identifying and aggregating policy exposures according to specific contract wording and industry sector classifications. We encourage all London Market insurers to develop their data extraction and modelling techniques to enable regular scenario testing to examine whether coverage intent is aligned with contract wording.
Longer term implications for business models
We also used the CBES to explore the resilience and adaptability of firms’ business models in response to climate risk. Firms generally expressed confidence that their fundamental business models were robust.
Most life insurers showcased their existing plans around the transition to net zero, and showed relatively little responsiveness to the specific features of the three different scenarios. These plans involve some reallocation away from carbon-intensive sectors, but not to the same extent as banks. The focus was more on engagement with investees to support the transition of the wider economy. The Bank will continue to investigate the potential for knock-on effects to the wider economy as a consequence of any withdrawal of finance to specific sectors.
General Insurers had two main strategies for responding to increased physical risk:
- They expressed confidence in the ability to reprice in the face of increased risk on the basis that most business is written annually.
The physical risks evolve smoothly in our CBES scenario, but a future path with a more sudden change in physical conditions, or a rapid shift in the legislative environment, could bring an increased risk of mis-pricing and under-reserving.
- They planned to purchase more reinsurance in the No Additional Action scenario to mitigate increased tail risk.
Few considered the implications on longer-term profitability from increased reinsurance premiums and commissions, and the increased risk to reinsurers under this scenario. This highlights one dependency the industry should explore further.
Beyond their own business models, insurers responses indicated that future insurance cover could become prohibitively expensive for households or business, or in extremis not available at all – particularly if insurers feel they can no longer accurately assess the risk. As an example, we investigated the potential for these effects on home insurance in the UK. Analysing the results shows that losses are concentrated in certain postcodes. This is not surprising where flood risk is the dominant risk, since this peril is highly localised. Insurers’ responses indicate that around 7% of UK households could be forced to go without insurance in the No Additional Action scenario. And those households most exposed to flood risk are likely to face additional challenges when re-mortgaging their properties – either because the properties have fallen in value, and/or because they have become uninsurable.
This example comes from residential coverage, but one could equally imagine that a reduction of insurance coverage for directors and professional advisors coverage could reduce innovation and risk-taking in the corporate sector.
Insurers noted in their responses that these risks could, in part, be mitigated by investment in flood defences, increasing flood resilience measures for properties, and encouraging flood-resilient repairs. They also noted their support for a continuation of a publicly supported UK flood reinsurance pool in such a scenario, and an extension to include properties built after 2009. This highlights the importance of financing the adaptation across future climate scenarios – and how the nature of that adaptation will have an impact on the financial resilience of households and the wider economy to climate risks.
Conclusion
The CBES is not the end of our work, it is the beginning. This exercise has shown that the losses to both life and general insurers appear manageable under the three CBES scenarios. Their resilience means that they should be able to play their vital role in financing the transition to net zero, and driving improvements in resilience to physical risks. This will help reduce the (much bigger) risks to businesses and households, and will also bring opportunities in financing or insuring the new industries and infrastructure needed in an economy responding to climate change.
Our focus is therefore on the safety and soundness of our firms in response to the risks arising from climate change. The CBES has acted as a powerful lens, but two key gaps remain:
- Firstly, ‘regime gaps’, where the design, methodology or scope of the capital framework does not adequately cover risks from climate. We are planning a conference in Q4 2022 on climate change and capital to facilitate discussion on this complex issue.footnote [6]
- Secondly, ‘capability gaps’ – i.e. do firms and regulators have the data and modelling abilities to ensure climate risk is captured in practice. There is a great deal of uncertainty, and a collective interest in managing climate-related financial risks in a way that supports a genuine and smooth transition.
The PRA will continue to work closely with industry, and with our international partners, to help accelerate the development of these capabilities and our collective understanding of the risks and opportunities from climate change and the transition to net-zero. We will run a webinar later this year to help non-participating firms gain insights from the CBES.
At the same time, insurers will need to continue to develop more advanced capabilities to identify, measure and manage climate risks, including through incorporating climate risks in scenario analysis and investing in the development and scrutiny of models. We will incorporate these lessons from the CBES into our ongoing supervision of firms against the aforementioned expectations set out in SS3/19, and we look forward to engaging with all of you in that endeavour.
I am grateful to Jethro Green, Giorgis Hadzilacos, Daniel Curtis, Nikolaos Petalas, Anooj Dodhia and Miranda Hewkin Smith for their assistance in helping me prepare this speech.
Thank you.