What's next in insurance modelling?
Models for risk assessment, capital allocation or projecting financial market trends are core to insurance business. Traditionally, models have been built relying mostly on historical data. In view of present- and future-day uncertainties introduced by factors such as climate change, record-low interest rates and low inflation, among many others, insurers need to debias their risk assessment models and make them also forward looking.
Article information and share options
Traditionally insurance models have relied on historical loss and exposure data, and sometimes reflect just the last one or two decades when it comes to financial markets. However, real-world risk dynamics are forever changing. To improve the basis of sustainable underwriting and investment decision-making, a modelled risk needs to reflect both past experience, and present and likely future developments. This means actively debiasing past data points for already-known trends and thinking of trending variables of the future.
Climate change, for instance, is impacting insurers’ assets and liabilities.1 Rising global temperatures are leading to increased intensity of severe storms and increasing losses when an extreme weather strikes areas of high population and economic value. By far the biggest risk driver remains the rapid increase of assets in exposed areas, mainly through urbanisation. The range of forward-looking variables to consider for insured risks is wide, and different by line of business. For instance, in long tail casualty, uncertainties exist not just around inflation and interest rates, but also the future trajectory of technological, legal, judicial and capital requirement developments.
Today, scientific understanding of how micro-climates and weather can shape potential losses is improving.2 And insurers can turn to professional scenario analysts and forecasters for a forward-looking view on a range of economic variables. Such knowledge can be incorporated as forward-looking risk assessment components in insurance modelling. There are several ways to do so, including counter-factual “what-if” analysis and consideration of potential crisis situations in the future, the latter important for better evaluation of tail risks in particular. As an example, systematic indices and heatmaps for local, industry and sector situations are enabling improved risk scenario analysis. Application of such future-oriented scenario assessments include mappings of the economic resilience of countries, estimations of climate change effects, say on health, and on biodiversity and ecosystem services.3 Forward-looking scenario analyses which combines historical experience and future trends and scenarios can help insurers anticipate and better price risk, and support underwriting performance.
1. An overview can be found e.g. in Swiss Re Institute, 13 July 2020, https://www.fintegral.com/storage/app/media/blog/Climate%20Risk%20Webcast/20200713_ Fintegral%20Webcast_Dr.%20Salomon%20Billeter_Climate%20stress%20testing%20for%20financial%20resilience.pdf
2. M. Bertogg, “Why hurricane risk modelling has to change,” Swiss Re, 6 January 2021, https://www.swissre.com/risk-knowledge/mitigating-climate-risk/why-hurricane- risk-modelling-has-to-change.html
3. For examples see: sigma Resilience Index 2020: global resilience put to the pandemic test, Swiss Re Institute, August 2020, https://www.swissre.com/institute/ research/sigma-research/2020-resilience-index.html, Swiss Re Institute’s Biodiversity and Ecosystem Services (BES) Index, September 2020, https://www.swissre. com/institute/research/topics-and-risk-dialogues/climate-and-natural-catastrophe-risk/expertise-publication-biodiversity-and-ecosystems-services.html, Swiss Re Climate Risk Score to assess the impact climate change will have on the wetness, dryness and sea level rise of any given location in the long-term (forthcoming).