Regulation – the next innovation super-charger for climate change risk assessment?

Twitter icon
Facebook icon
LinkedIn icon
Matthew Grant

Heavy-handed regulation can choke the growth of new businesses, particularly in financial markets, but regulation can also be the catalyst that accelerates the demand for new analytics.

Forget the 2021 Olympics, the race to watch this summer is the speed with which company founders and leaders react to a raft of new regulations that raise the stakes for insurers and banks when measuring and mitigating climate change impacts.

The most recent, and the most significant of the regulator’s demands was announced in the UK on 8th June when the PRA (Prudential Regulatory Authority), part of the Bank of England, released its 2021 "Climate Biennial Exploratory Scenarios" (CBES).

These are "stress tests" intended to explore how major insurers and banks will be impacted by, and the measures they will take, to reduce climate change risk. The stress tests build on, and significantly expand, climate risk disclosures that were first required by the PRA in 2019. Major financial institutions must now provide extensive information in two areas:

  • how they will undertake rigorous measurement of climate change risk
  • what actions they will take to mitigate their potential exposure to, and influence on, climate change.

The president of the influential Network of Central Banks and Supervisors for Greening the Financial System (NGFS) explains the problem in a recent report:

“The world is at a critical juncture where climate pathways could move in materially different directions: from a successful transition to net-zero emissions by 2050; to a hot house world with global warming of 3˚C or more by 2100. In the face of such uncertainty, climate scenario analysis is a vital tool that helps us to prepare for a range of future pathways”.

The NGFS goes on to explain that climate scenario analyses will:

“focus minds on a variety of different outcomes, challenging users to consider what risks and opportunities might arise – and crucially, what action might be required today in light of these potential challenges tomorrow”.

New solutions, old problems

Access to more data is widely talked about as the route to better insurance underwriting and creating an improved customer experience. Companies that have been created to offer new data and analytics have gained a lot of attention during the insurtech wave. There are certainly opportunities in this area, but the reality is that it’s been getting harder for new entrants to offer solutions that are meaningfully better than what already exists for underwriting and improving customer engagement in established markets such as property underwriting.

This rapidly growing demand for hard evidence of actions to address climate change is far more wide-ranging and urgent than what has been required for underwriting. Climate risk analytics is about to go mainstream.

Now that the starter’s gun has gone off in the race to find answers to what is fundamentally a different problem for insurers, many are going to be scrambling to address a pressing need for which there are currently few good tools available. The regulators have shifted the focus on climate change risk from being a problem of the future to a business critical need that must be addressed now.

This represents a major opportunity for both start-ups and established vendors.

Our InsTech London report reviewed the needs and profiled over 80 companies providing data and analytics related to property location, hazard, vulnerability and more. Many of the most successful data providers and modellers (see below) supporting insurers today pre-date the birth of insurtech by many years. They have well-validated offerings and continue to innovate. That's why it’s been hard for the new arrivals to break in. New entrants have had to discover or create new sources of data (e.g. HazardHub creating its fire hydrant database, Previsico offering higher resolution flood forecasting, Flyreel and e2Value with phone enabled property risk assessment); invested in major new technology (e.g. ICEYE with its constellation of mini-satellites) or have gained a market lead in providing proper, robust AI to add value to the data and technology developed by others (e.g CAPE Analytics, MIS, Archipelago, Safehub, Arturo). Other new arrivals have grown by offering a route to market as aggregators of data (WhenFresh, Addresscloud, LightBox).

A reboot for hazard data

The market for climate analytics and data is about to explode. Insurers and banks now need to access and use high-quality hazard and location information, and be able to assess future changes and impacts, in order to report on climate risk. This is going to turbocharge revenue opportunities for companies that are able to provide the necessary information and gain credibility with insurers and the regulators. But it’s by no means simply a “plug and play” opportunity. The current generation of models and data are predominantly backwards-looking. Understanding future climate risk properly requires these models to be “climate conditioned” which is rarely simple.

The Bank of England announcement is just one in a series of regulatory smoke signals. In the last year or so we have seen other strong signals of change including:

  • ESG (Environmental, Societal and Governance) reporting is already required for all public listed companies as part of their annual return. Investors and others are requiring ever more rigorous disclosures.
  • The Task Force on Climate Related Financial Disclosures (TCFD), supported by the G20 nations, started releasing recommendations in 2017 to help promote advancements in the availability and quality of climate-related disclosure by companies around the world. It is becoming increasingly influential.
  • In the US, President Joe Biden’s 2022 budget proposal includes more than $36 billion to fight global climate change. The president is committed to setting the US on a path to cut carbon emissions in half by 2030 and put the economy on a path to carbon neutrality by mid-century.
  • In March this year, The New York Department of Financial Services (DFS), which regulates the activities of nearly 1,800 insurance companies, became the first US financial regulator to issue climate-related guidance to require insurers to manage and report on the financial risks of climate change.
  • The French regulator – Autorité de contrôle prudentiel et de resolution, completed a stress test pilot in April 2021 which included 15 insurance companies. They expect that claims from natural disasters will increase from 2 – 5 times in some regions.

The previously mentioned Network of Central Banks and Supervisors for Greening the Financial System (NGFS) was launched in 2017. The organisation is supported by 91 regulators worldwide and so has a central role in defining the requirements for reporting by insurers on climate risk. The NGFS has already commissioned a number of organisations to develop the scenarios that it intends to be used to support requirements for stress tests around the world. Strangely, none of these organisations has experience in building the probabilistic models now widely used by insurers. 

The Bank of England is recommending the NGFS as the source of information to help insurers carry out the stress tests and the NGFS scenarios are available online. It’s clear that insurers, banks and regulators are going to be looking for quite a bit more help to meet the emerging regulatory demands. 

A short history of stress tests

In the UK, the Bank of England, usually through the Prudential Regulatory Authority (PRA), has been setting stress tests to assess the resilience of the UK financial system and individual institutions since the financial crisis in 2008. In addition to these annual solvency tests, every other year (“biennial”) the bank also runs exploratory scenarios to probe the resilience of the UK financial system to a wider range of risks.

Solvency tests are commonly used to monitor the current situation, but the Bank of England considers these biennial climate scenarios as a tool to;

“enhance participants’ strategic thinking on how to manage those (future) risks”.

When the PRA released its first climate risk stress tests in 2019 insurers were provided with the modelling scenarios and asked to plug in their data, run some simple tests and then submit the results. Unfortunately, such an approach gives little real insight given the complexity and uncertainty associated with modelling future climate risks. Responding to climate stress tests in a way that gives meaningful results requires using climate conditioned models and data. Granular assessments are needed for different classes of risk, and the type of analyses vary in different geographies and across a range of natural perils. With even the best models and experienced analysts, uncertainties in outcomes can be high.

For the 2021 tests, insurers are required to do a lot more of the modelling themeselves. They must also answer 98 qualitative questions about how they will be impacted by climate risk. The PRA wants to understand insurers’ exposure to the two primary types of climate risk: transition risk and physical risk.

  • Transition risk covers what happens when we move away from carbon-intensive activities and uses new or different technologies or fuel sources. For example, there are implications for investors, insurers and others when we all stop driving petrol or diesel cars and use electric vehicles.
  • Physical risk is how the climate impacts our assets. Questions for insurers include determining which properties will be further impacted by flooding as sea levels continue to rise, and what can be done to reduce the risk.

Detail around the requirements of the scenarios is defined on the BoE website.

RCPs - The good, the bad and ugly

The UK stress tests require companies to consider a framework of three climate scenarios. These scenarios are derived from the Representative Concentration Pathways (RCPs), originally proposed by the IPCC in 2018 following the Paris Agreement. We are going to be hearing a lot more about RCPs so it’s worth getting familiar with them. RCPs define different concentrations of CO2 in the atmosphere at future dates. The amount of CO2 directly determines the size of temperature increase over time. RCPs are widely used to assess climate change impacts.

The scenarios prescribed by the Bank of England are labelled the ‘early’, the ‘late’ and the ‘no policy’ action scenarios. More specifically these are defined as:

  • an orderly transition to 1.8C of warming by 2050. An orderly transaction means taking action and implementing plans now to move the world away from carbon.
  • a disorderly transition to 1.8C of warming. A disorderly transaction means taking insufficient action now, then in future having to take sudden and dramatic action in a very short space of time (for example, suddenly banning all petrol and diesel cars).
  • taking no action means doing nothing impactful and is assumed to result in a catastrophic 3.3C of warming.

The PRA has told banks and insurers that they must have “embedded” processes for understanding and managing climate-related risks by the end of this year. The PRA will then be reviewing how companies intend to adapt their business models over time. It is very keen to know the explicit actions management are taking now to managing climate risk.

Climate Change – no longer tomorrow’s problem

Stress tests, and other new regulatory requirements, are moving climate change beyond the well-intentioned, but somewhat optional investment in CSR (Corporate Social Responsibility) activities and the loosely defined ESG (Environmental, Societal, Governance) reporting. The pressure is now on for banks and insurers to define and take action in a 12-month period. At the same time pressure is coming from initiatives such as the Carbon Disclosure Project's campaign to ensure that data on climate change, deforestation and water usage are properly reported by publicly traded companies. This campaign is supported by 168 asset managers and financial institutions from 28 countries. Climate action is going to have to get tactical.

Innovation and novel modelling

The Bank of England is encouraging respondents to the stress tests to explore new analytics and..

“..use novel modelling to assess a detailed, bottom-up analysis of their largest counterparties”.

Insurers have been using models for pricing for decades, including as a core part of the reinsurance transaction. That's a strong head start over banks. Outside of insurance, most financial organisations use less sophisticated scenario modelling and rarely considere climate perils at a portfolio level.

Currently, the stress tests impact only a small number of major UK insurers. In general insurance Aviva, RSA and Direct Line are required to report across the whole group. AXA, Allianz and AIG are required to report for their UK entities only. The Corporation of Lloyd’s has selected 10 syndicates which it is asking to participate and will combine these to provide a Lloyd’s submission.

These companies represent 60% of the premium coming into the UK property and liability insurance market. The tests do not impact all insurance companies, but all insurers should be taking careful account of them. Kirsten Mitchell-Wallace, Head of Portfolio Risk Management at Lloyd’s, explained at our InsTech London event held with reask on 20 April that Lloyd’s is considering the role of climate change scenario tests for the entire Lloyd’s market and we can expect that more insurers will have to provide a response to the Bank of England in the next few years.

For banks, the stress test is focussed on their credit books. Insurers are being asked to assess risks to their investment assets, reinsurance recoverables and insurance liabilities (including inward reinsurance).

The perils to watch

Climate change will impact the frequency and intensity of many natural perils. Both hurricanes and flooding have a potentially larger geographical footprint from future climate change, along with the amplification effect caused by higher sea levels and bigger storms.

Wildfires can be expected to increase significantly around the world but losses will be more concentrated. A smaller number of companies may suffer disproportionately (such as happened to the Pacific Gas & Electricity company’s $13.5 billion pay-out after the 2018 California wildfires).

Climate Modelling 1.0

Climate modelling companies and data have been available to insurers for decades. RMS, Verisk and CoreLogic have been licensing modelling tools to insurers, reinsurers and brokers since the mid-nineties. New companies are emerging every year offering solutions to insurers. New models are becoming more accessible now that Nasdaq has been making it easier for smaller, niche modelling companies to release their products to the market through its Risk Modelling for Catastrophes product, linked to the open-source Oasis Loss Modelling Framework. 

Catastrophe modelling has become more sophisticated and reliable for underwriting risk selection and capital management. Until recently though, the main catastrophe model vendors have been cautious about offering forecasting capabilities to account for future changes in climate because the uncertainty has been so high and fundamental adjustments to models are required.

This is starting to change and we will be running events with many of the InsTech London corporate members and their clients and partners to discuss these new climate conditioned model releases.

In the UK, the increased impact from flooding is one of the most significant drivers of potential loss and disruption. Flooding is also one of the perils where climate change risk is (relatively) easier to model in the UK than elsewhere. Flood modellers such as Fathom, JBA, Verisk, and RMS have all announced enhancements to their flood models to add a more dynamic risk assessment that takes account of the climate change scenarios.

New companies are emerging with a strong specific focus on future climate risk. Recent entrant reask is offering global climate models that can be used alongside traditional catastrophe models. Jupiter Intelligence, backed by Liberty Mutual and OneConcern backed by Sompo have entered the market with substantial investment backing. Other companies, such as Swiss Re backed kWh Analytics, are helping assess and insure the risks of investing in renewable energy sources. Cambridge University has been helping insurers, banks and corporates for some years through its Centre for Risk Studies.

Understanding the climate hazard peril is only one side of the risk equation though. It’s equally important to understand the resulting physical and financial damage caused to properties from hurricane speed winds and flood. This is where established modellers still have the advantage.

Climate Modelling 2.0

Within insurance alone there will soon be hundreds of companies searching for the best tools to help measure and report on their “new” exposure to future climate risk. Banks and other financial companies will be required to carry out climate modelling many times more sophisticated than what they have had to in the past. Despite the availability of the NGFS scenarios many companies outside of insurance currently have little or no idea how to assess the risk to their assets from even “normal” historic wind and flood risk. Within a few years, the rest of the entire corporate world will be required to follow.

The NGFS acknowledges the challenges the industry faces:

“A major obstacle to undertaking this analysis has been the availability of detailed scenarios that analyse both the physical and transition risks from climate change and their economic impacts. The challenges
and costs of creating such scenarios are beyond most individual firms or institutions”.

The NGFS identifies two areas in particular that it is struggling to resolve:

  • lack of a systematic approach to quantifying uncertainty
  • research gaps on the nature and size of impacts from physical risks, including on exposure, vulnerability and adaptation.

The NGFS is working on a longer-term project to improve the scenarios and “expects to be engaging with the private sector”. It’s not clear yet how the NGFS will be tapping into the experience of the existing data and analytical companies servicing insurers. Let's hope it does.

Climate Modelling 3.0?

The additional risk of damage to properties from hurricanes and floods is not the only climate risk insurers face. Other recent examples of challenges for the industry include:

  • Liability modelling company Praedicat reports that 43 lawsuits have been filed against baby food manufacturers over elevated levels of arsenic and other heavy metals in their products arising from more frequent droughts.
  • The Financial Times has revealed that Legal and General Investment Management, the UK’s biggest institutional investor, has dropped AIG from some of its investment funds because of “insufficient policies in place over climate change risks”.
  • Climate change activists are targeting insurers that they believe are not moving fast enough to follow those insurers that have already committed to stop insuring and investing in coal-related industries.


Lloyd’s CEO John Neal described climate change as “the largest single underwriting and investment opportunity” that most re/insurers will see in their careers when he spoke at S&P’s virtual 37th Annual Insurance Conference this month. That’s a bold statement, but it reinforces the point that the role of the regulator is not just to say “no”.

The modelling of natural peril risk is today a successful but boutique niche market worth less than $5 billion and serviced by a handful of companies. If regulation promotes natural catastrophe modelling into a mainstream and critical function across multiple industries the market could be worth $50 - 100 billion or more.

If you are interested in continuing to learn more about the opportunities emerging from the Climate Change Stress Tests or would like to discuss how InsTech London can support your organisation in sharing your stories and improving your visibility amongst global insurers, technology companies, consultants and influencers, please contact me via LinkedIn or by emailing