[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Environment: A change in the weather

The insurance industry is taking serious notice of climate change in all its activities — from product design and underwriting through management of its own risk and capital to developing risk management solutions for policyholders. In March 2008, a survey of insurance industry analysts placed climate change as the number one risk to the industry. By contrast, a recent survey of the Actuarial Profession by the Climate Change Working Party suggests that we are somewhat behind the curve in understanding the implications of climate change for our roles and for our employers. This article looks at the way in which climate change is influencing some of the areas of insurance activity most relevant to actuaries.

Risk management
Actuaries working for and advising insurance companies have a crucial role to play in the risk-management activities of these firms. It is impossible to manage the risks of any business without considering the immediate and longterm impact of climate change.

We will consider some aspects of modelling below; however, this is only one part of the risk and capital-management framework. Actuarial advice can be invaluable in framing risk appetite. One of the key challenges in this context lies in understanding how individual exposures are linked to one another.

While insurance is founded on the principle that a portfolio of risks diversifies, the unexpected aggregation of individual risks can cause enormous problems. This threat can be addressed from the bottom up and from the top down, and the two approaches can combine to give management some insight into the risks that the business is running. An example of bottom-up aggregation analysis is the mapping of property exposures to provide a picture of the aggregate risk presented in, say, one geographical area.

While such bottom-up analysis is an important, basic risk-management tool, it cannot be relied upon to provide a complete picture of aggregate risk. Although the 2007 floods in the UK may not be a manifestation of climate change, they provide a good example. The major household insurers now employ detailed and accurate digital flood mapping technology to help manage their risks. Nevertheless, the industry was surprised by the nature of the floods. No matter how well we think we understand risks, we must always consider the possibility that we have missed something. This can be addressed through consideration of scenarios — an exercise requiring more imagination than analysis, perhaps, but nevertheless one to which an actuary can add a great deal of insight.

As well as considering previously unobserved types of loss, scenarios can also help to make connections between risks that might previously have been regarded as uncorrelated. Remaining with the property theme, risks that are geographically separate could be exposed to the same event — for example, if climate change affects the paths taken by windstorms or increases their severity so that they can create longer and wider tracks of destruction. We must not be too quick to dismiss apparently implausible scenarios. Regardless of how rapidly climate science improves, it is clear that we are a long way from understanding the implications of climate change for future weather events.

Aggregation of risk is not limited to the insurance portfolio — it spans the entire risk spectrum of a business. Certain types of event can lead to losses in the asset portfolio as well as the insurance book and, if these are coupled with operational losses, the impact can be disastrous. Climate change is likely to have a strong but unpredictable effect on the frequency and severity of many events of the type that can lead to such aggregations.

Model behaviour
Actuarial modelling methods tend to rely on analysing the past and drawing inferences about the future. Climate science uses modelling techniques that are more forward-looking in their approach. Given the changing impact of the climate on insured risks, it is essential that insurance risk models are reconciled with the rapidly developing climate-modelling discipline.

Naturally, there are some actuaries who specialise in this area and we cannot all be at the cutting edge of this science. However, many of us use these models and have a responsibility to our stakeholders to ensure that the advice we give is based on the best available information. As users we should, therefore, be driving these modelling improvements when setting our requirements from the providers of our models.

Across the full range of insurance risks, climate change is influencing the variables we are trying to model. While driving trends in the expectation of losses, changes in our climate are also increasing the variability of many outcomes. In fact, in many cases the trends in variability are more extreme than the trends in the underlying expectation. This is often exacerbated as our attempts to reduce our uncertainty about loss outcomes through modelling improvements can themselves reveal new sources of uncertainty.

While the modelling problem is already sufficiently difficult, it is not helped by weaknesses in the data that is applied to the models. As information technology improves, the exposure data can, in theory, be stored at increasingly granular levels. However, there is still typically a gap between what can be stored and what is stored. Inaccurate or missing exposure data can be a significant cause of error when the risk profile of a portfolio is being modelled — for example, to define a suitable reinsurance programme.

One of the core skills of the actuarial profession is making use of collateral data — from graduation of mortality experience using standard tables to combining own claims data in rating models with third-party information about behavioural characteristics. There are various initiatives across the industry and academia to collect and pool relevant data that should help individual insurers understand their exposures better, in the context of relatively scant information about their own portfolios. Although some might be uncertain about the competitive implications of pooling data, the risk-management benefits to individual insurers — leaving them more free to seek competitive advantage through their core-business activities — and the benefits to society of financial stability should provide ample justification.

Product design and underwriting
Insurers are reacting to climate change in their product design. In a very positive sense, new products and features are being developed to exploit the perceived risk-reduction benefits associated with responsible behaviour. There are also opportunities for new insurance products as the very behaviour of policyholders changes in response to improving climate awareness. On a more defensive note, insurers are finding it necessary to review underwriting terms to address emerging understanding of risk. Let us take these three aspects in turn and look at the role that actuaries ought to play.

There appears to be a great deal of enthusiasm for insurers ‘rewarding’ responsible behaviour with lower premiums. For example, energy-efficient buildings are regarded as being more resilient to major perils, whether it is because they are built to stricter building codes or for more trivial reasons such as energy-efficient light bulbs being less of a fire hazard.

Pay-as-you-drive insurance products link the insurance exposure to the number of miles driven and they also give an incentive, albeit small, to drive fewer miles. Some insurers even believe that customers who tend to think responsibly — for example, those who drive hybrid cars — are also relatively better insurance risks. It is essential that these connections are founded on robust pricing and underwriting principles. The role of actuaries in pricing is never more important or challenging than when new rating factors are proposed, particularly if there is a strong likelihood of correlations with existing rating factors.

Even more challenging is the analysis of risks associated with new products, covering new perils in evolving circumstances. For example, how might one go about pricing a product that covers the failure of innovative and even experimental, renewable energy projects? This cannot be based on analysis of the past, but would have to begin with a clear understanding of the possible reasons for failure, the way that these are mitigated in the structure of the project and the residual risks. This understanding would be much more akin to the prospective climate models referred to above. As insurers increasingly build their own risk-management expertise into their insurance offerings for such projects, the actuarial analysis of possible loss scenarios can be made much more productive by feeding into the risk-mitigation features of the product.

As well as building new products and features in reaction to climate change, insurers are increasingly feeling the need to tighten underwriting terms as, in the extreme, some risks become uninsurable. This presents a challenge as historical claims experience becomes less relevant as a predictor of likely future experience. Actuaries involved in pricing and underwriting need to be able to react quickly to emerging experience and be ready to fine-tune the pricing adjustments that accompany variations of cover. There is also the risk that new sources of claims will emerge.

Many household insurers are already subsidising some of their policyholders in line with the Association of British Insurers’ Statement of Principles. Insurers can make active decisions to write some risks outside their normal risk appetite and even, in some cases, below the expected cost of claims. This can work perfectly well as long as the insurer’s portfolio is sufficiently large and diverse to subsidise these risks. From a risk management as well as a pricing perspective, there is a role for actuaries in monitoring the implications of these cross-subsidies, to help the insurer guard itself against adverse selection and maintain a balance across the portfolio.

The question of liability is potentially the area of greatest concern for the industry and its insurers. This demonstrates clearly that climate change is not a problem that we can ignore in the near term. Exposures being written now will be influenced by the way the climate changes in the longer term.

While we see the effects of climate change in aggregate, it can be very difficult to link this to the individual behaviour of corporations. However, recently reported developments in modelling techniques may help with the apportionment of blame. If this can be linked with the public nuisance aspects of the effects of climate change, for example, then the potential for litigation against corporations and their directors and officers could be huge. These risks must be factored into the pricing and underwriting of new contracts, as well as being allowed for when estimating provisions for historical exposures, written when awareness of these risks was more limited.

Be imaginative
The impact of climate change on the insurance industry will be fundamental. No matter what role you play in your organisation, you must not dismiss climate change as a factor that should affect your work. Be imaginative about the scenarios you consider, and do not be anchored on historical experience. Be imaginative about the sources of information you use and be demanding of the suppliers of the information and models you use.

Tony Brooke-Taylor is head of wholesale insurance at the Financial Services Authority (FSA), supervising the Lloyd’s and commercial insurance and reinsurance markets. He is also responsible for mitigation of the risks to the FSA’s statutory objectives presented by the impact of climate change on the insurance industry.