Evaluating climate change impacts on mortality, life insurance, and annuities

Given the increasing frequency of extremely high temperature events due to global warming, life insurers need to incorporate temperature impacts in their mortality models and long-term liability valuations. Current actuarial literature, primarily using monthly or annual data, inadequately captures the non-linear and delayed effects of temperature on mortality. To address this, we examine the relationship between weekly deaths and daily mean temperatures using the distributed lag non-linear model (DLNM) from environmental epidemiology, where we extend to projecting mortality rates and considering age-specific temperature impacts. These extensions better accommodate actuarial needs for single-age mortality rates in pricing and valuations. To evaluate the impacts of climate change on future mortality, we propose a procedure combining pattern scaling from climate science and time series methods for projecting local temperatures under various future emission scenarios. Using data from Madrid and Centro of Spain, we find that extreme cold and hot temperatures significantly increase mortality for those aged 60 and above, with prolonged extreme temperatures having added effects. Global warming results in warmer summers and winters, which tends to shift the annual mortality peak from winter to summer by the end of the century, where overall annual mortality increases for scenarios with more warming. The resulting divergence in life insurance and annuity liabilities across different climate scenarios is visualised. We further apply our framework to East Austria to examine the geographic variability of temperature effects on human mortality under climate change.