Assessing climate risk for the insurance industry: How reliable are climate scenarios?

As the impacts of climate risks become more pronounced, regulators and the insurance industry are under increasing pressure to refine risk assessment frameworks. 

This article examines regulatory developments, a critical review of climate scenarios and limitations of models within the insurance sector, the implications of new regulatory developments, and the role of actuaries in establishing actionable frameworks to address the uncertainties of climate change.

You can access the detailed article below, which provides a comprehensive critical analysis of the literature on generating climate scenarios. 

Insurance regulatory developments for climate resilience

Recent years have seen heightened regulatory efforts to integrate climate risk into financial frameworks.

In response to the growing climate risks, EIOPA issued recommendations to integrate climate risk into the ORSA (Own Risk and Solvency Assessment) framework.

Initiatives such as France’s Autorité de Contrôle Prudentiel et de Résolution (ACPR) climate pilot exercises and the use of Network for Greening the Financial System (NGFS) scenarios have provided methodologies for assessing physical and transition risks, while also exposing the limitations of existing tools and models in capturing the full scope of potential impacts.

Building on this momentum, the European Central Bank (ECB) and the European Insurance and Occupational Pensions Authority (EIOPA) have added to the regulatory discourse with their December 2024 report, “Towards a European System for Natural Catastrophe Risk Management.” This report highlights the economic risks associated with the growing insurance protection gap for natural catastrophes and proposes EU-level measures to improve risk-sharing, modelling, and coordination between public and private entities.

These developments reflect the increasing demand for sophisticated tools to assess and mitigate climate risks effectively. As these regulatory efforts evolve, the role of climate scenarios becomes ever more crucial.

“The ECB and EIOPA’s latest report reinforce the drive for EU-wide solutions to address climate risks and the insurance protection gap, highlighting lessons such as the need to improve the availability, quality, and comparability of data on insured losses across EU countries.”

Eric Cloutier, Head of Global Financial Sector Regulatory Centre (RegCentre), Forvis Mazars UK

Limitations of integrated assessment models for generating climate scenarios

The generation of climate scenarios relies heavily on Integrated Assessment Models (IAMs), which link economic activity to climate outcomes. However, the process of creating these scenarios is fraught with challenges, including model limitations and uncertainties.

It has become clear that the impacts of climate risks are often shaped not by average patterns but by extreme scenarios. Integrated Assessment Models (IAMs), which commonly assume average deterministic responses to global warming, may underestimate these risks.

Nevertheless, actuaries and risk managers rely on climate scenarios to anticipate the impacts of global warming and the transition to a low-carbon economy. The critical literature on IAMs offers insights into addressing these challenges, helping to develop more consistent modelling approaches for climate risks and effective strategies for their management.

“Extreme climate scenarios are no longer abstract possibilities—they represent the critical risks that insurers must quantify and prepare for using advanced modelling and forward-looking strategies.”

Maxime Simoen, Partner, Forvis Mazars in France

Interpreting IAM results requires a good understanding of the uncertainties involved. Van Asselt & Rotmans (2002) suggest a taxonomy with multiple sources of uncertainty:

• Model structure uncertainty: this refers to the potential flaws in the structure of the model itself, including the assumptions made about how different variables interact.
• Model completeness uncertainty: Involves questioning whether the model captures all relevant processes. A model may exclude important climate or economic mechanisms, leading to incomplete or inaccurate predictions.
• Parameters, inputs, and initial states uncertainty: this refers to the inherent uncertainty in all models, arising from the challenge of accurately estimating parameters based on observational data.
• Model operation uncertainty: this relates to errors that might occur during the operation of the model, such as numerical inaccuracies, software bugs, or the accumulation of propagated uncertainties.

One of the first papers to tackle the problem of interpreting IAM results is Weitzman (2009). The aim of this article is to suggest a mathematically rigorous, though abstract, economic and statistical framework for high-impact and low-probability disasters. The key point of this article is that the impacts of global warming could be dominated by extreme scenarios. IAM, which generally assume an average deterministic response to global warming and its consequences, could therefore significantly underestimate the risk.

Yet, some research articles criticise IAM and address the various types of uncertainty mentioned above. The criticisms below are primarily drawn from two articles: Ackerman et al. (2009) and Stern et al. (2021).

It should be noted that some of the major limitations expressed in 2021 overlap with those mentioned in 2009, indicating a certain inertia in the most widely used models. These criticisms relate to several modelling aspects:

• Discounted utility expectation framework with a unique representative agent;
• Choice of discount rate;
• Unpredictability of key economic drivers;
• Substitutability of environmental goods;
• Social Cost of Carbon;
• Damage function;
• Mitigation costs.

Please read our full article below for details on these criticisms and the impact of these limitations on IAM results.

“Integrated Assessment Models lay the groundwork for climate risk analysis, but their limitations require innovation to fully capture the economic, social, and financial complexity of climate impacts.”

Alice Thou, Partner, Forvis Mazars France

Suggestions to address uncertainties and improve comparability

Actuaries and risk managers need climate scenarios to anticipate the impacts of global warming and the low-carbon transition. The critical literature on IAMs provides guidance, suggesting that a range of models and scenarios should be used to address uncertainties.

Proposals from leading experts

• Van Asselt & Rotmans (2002) advocate the use of a multitude of different models to reduce uncertainties.
• Ackerman et al (2009) and Stern et al (2021) suggest a more “insurance-like” use of these models. The idea would be to draw on the literature in the field of climatology to define a warming target not to be exceeded, and then to ask the IAMs to provide the “socially least costly” carbon tax trajectory to meet these targets.
• Kaufman and al (2020) suggest a similar approach, called “Near-Term Net Zero”. This approach can be broken down into four steps: set a date by which net CO2 emissions must be zero, choose an emissions reduction trajectory, use an IAM to estimate a CO2 price consistent with the chosen trajectory and periodically repeat the above steps to update the estimates.

Given the state of the art of IAMs, it is necessary not to rely solely on the climate scenarios produced by the NGFS. The results obtained must be nuanced and interpreted as a lower bound of the impact of transition scenarios on economic and financial systems. Actuaries therefore need to be trained in the limits of these models.

Supervisors can also share scenarios produced by models other than the NGFS to get a wider range in the risk estimations rather than a lower bound. Centralising the production of these scenarios via supervisors would also provide the significant advantage for the industry to work on a common basis. This would enable both supervisors and the industry to better compare the risks to which different companies are exposed.

“Actuaries are pivotal in translating complex climate scenarios into actionable insights, ensuring insurers can meet regulatory demands while preparing for the uncertainties of climate risks.”

Alice Thou, Partner, Forvis Mazars France

Preparing for the next wave of regulations

The regulatory landscape is poised to demand more precise, data-driven approaches to climate risk assessment and management. Insights from France’s ACPR climate exercises and NGFS scenarios, along with the ECB and EIOPA’s recent report, highlight the need for enhanced modelling, improved data sharing, and stronger public-private coordination. These initiatives aim to close the insurance protection gap and prepare the sector for the challenges posed by extreme climate scenarios.

For the EU insurance and actuarial industry, this presents both opportunities and challenges. Regulators are likely to mandate advanced scenario analyses, robust risk modelling, and actionable adaptation measures. Actuaries, in particular, must refine their skills in interpreting complex scenarios and their implications for financial stability. Insurers will also need to invest in resilience-building strategies to align with increasing regulatory expectations.

As the EU seeks to harmonise climate risk frameworks across Member States, it positions itself to lead globally in addressing the financial and societal challenges of climate change. By embracing these changes, the EU insurance sector can set a benchmark for innovation, collaboration, and resilience, strengthening its role as a global leader in the transition to a low-carbon, climate-resilient future.