20/12/2023

Climate scenario analysis for life insurers

This short article is an initial output from the IFoA Life Climate Scenarios Working Party (Assumptions and scenario modelling) – it provides a brief outline of how climate scenarios can be used by a life insurance company.

Climate scenarios are projections, either quantitative or qualitative, of how change in the world is driven or influenced by climate change. Scenarios may vary in terms of time horizon and granularity, ranging from global projections of how temperatures might change over the coming decades, through to granular analyses of the economic impacts of the energy transition. In all places these problems are complex and multidimensional.

Many companies have focused on net zero 2050 targets; often used interchangeably with the target of keeping warming below 1.5°C. Scenarios allow exploration of these assumptions. The latest projections show that a combination of a lack of action to reduce emissions and improved understanding of the climate system, demonstrate this equivalence no-longer holds. Recent academic literature estimates that the remaining budget for having a 50% chance of keeping warming below 1.5°C by 2050 could be as low as 75Gt of Carbon (P. Friedlingstein et al.: Global Carbon Budget 2023)[1] from the beginning of 2024, equivalent to seven years of current emissions.

When designing scenarios, we can add analysis and context, but we believe scenarios need to be “decision useful”. Therefore, before we start to develop scenarios, we need to define how they will be used. To illustrate this, we may consider four areas that a life insurance company might want to investigate (Table 1).

Table 1: Simple example of where climate scenarios may be employed to understand climate risk exposure

Use case	Example
Managing investment risk	Risks associated with long term buy-to-hold credit portfolio.
Business risk – impact on business volumes and customers for each line of business	Investigating the impact on lapse risk for a Unit-Linked pensions book resulting from the transition.
Physical risk to company assets	Risks of changing weather patterns and extreme weather events on an insurer’s own buildings.
Longevity risk	Understanding potential impacts on life expectancy that increased physical risk could have – both direct (e.g., more volatile weather patterns) and indirect (e.g., food insecurity, economic impacts)

Once a question has been defined, we need to move on to what we need a scenario to tell to support decision makers with. Is that a projection of economic returns, or a narrative exploring the risks and their drivers. For each of these scenarios we need to first determine the appropriate level of detail and depth of modelling, what information we might need, how we would go about obtaining this.

Whilst on one level generating these scenarios requires detailed and complex analysis, there are key messages that cut across all uses for these scenarios.

We know that the globe is already 1.2°C warmer than in pre-industrial times and even current CO₂ levels will contribute to ongoing warming for the rest of the century.
We know that emitting carbon heats the planet, and humanity is producing carbon at a rate that, unchecked, will lead to unacceptable levels of warming.
We know that if society wants to avoid an unliveable planet we will need to decarbonise. This means (at a minimum) stopping burning fossil fuels and electrifying everything. The timing, nature and outcomes of this transition are highly uncertain - and likely to define this century.

However, the timing, nature and outcomes of this transition are highly uncertain - and understanding this is key to developing decision useful scenarios.

As an example, looking at the credit risk analysis. Here there are two key features we might want to consider.

Whether the asset is likely to be exposed to transition risks?
What is the term of the lending and how does that interact with the refinancing risks?

The key risks for credit assets tend to be at maturity or re-financing. So often the key elements of scenarios might be understanding climate might have affected the world (and the financial markets) when an asset is scheduled to mature.

To take a simple and obvious example we will consider the debt of a coal mining company maturing in 2040. Here we might want to understand the likely demand for coal up to and at this point, how that demand might be met and whether it would be met by this miner. Table 2 provides more detail on the information that may be considered – this is not necessarily the gold standard, but what would be practical for actuaries/life insurers with limited resources.

Table 2: a summary of information required to begin to build climate scenarios for a fixed income investment in a coal miner

	Ideal information required	How to gather
Supporting a narrative	High level description of plausible future pathways of the world to 2040 and beyond. Descriptive, to foster creativity and imagination amongst stakeholders. Detailed, to build links between the scenario narrative and the economy. Societal impacts should also be considered in the narratives.	Internal stakeholders agree the range of scenarios useful to consider - either single risk or integrated – here we consider transition risk. Plausibility supported by the latest policy, technological development, literature or by: International Energy Association (IEA) forecasts Inevitable Policy Response (IPR) Nationally Determined Contributions (NDC) Informed by or consistent with internal views e.g. from risk committee.
Quantifying economic impacts	Once narratives are created with plausible policy actions and other features, economic and emission impacts need to be formulated (qualitatively) and/or quantified. Transition risk: Qualitative: describing specific drivers of coal demand and the factors that influence credit risk and refinancing. Quantitative: Projections of inputs and outputs, considering coal demand per scenario in the context of the economy. Systemic risk: Risks should be considered from a system wide lens to avoid missing important interconnections between economy, environment, society.	Modelling whole economies is not a trivial exercise and may be outside the remit of actuaries. Instead, some may take direct outputs from organisations such as the IEA. Other approaches may use simplifying assumptions or stylized facts as the basis for understanding future coal demand. Some risks to coal producers cannot be quantified e.g. immediate change to societal/political acceptance of coal use in light of increasing physical risk.
Estimating impact on holdings	Linkage between macro-economic drivers and credit and refinancing risk for the specific miner or type of miner. Allowing for corporate structure and other specific circumstances if the data is available.	Sources may include: In-house due diligence Corporate disclosures Industry bodies such as World Coal Association Peer benchmarking exercises Third-party data provider

We believe this approach could be used to guide the development of climate scenarios for a wide range of different questions and problems. Given the complexity of the systems we are looking at often narrative scenarios will be the best way to investigate the risks associated with climate change.

Different scenarios will be required for different problems, reflecting the specificity and granularity of information required to help support decisions. What is critical is that actuaries apply themselves to understand the climate scenarios being used to uncover the implicit assumptions to best inform present day decision making.

Following this initial blog, the Working Group’s goal is to flesh out in more detail the types of scenarios that would be required for some of the different areas that a life insurance company might want to investigate

[1] https://essd.copernicus.org/articles/15/5301/2023/essd-15-5301-2023.pdf