Robust mortality forecasting in the presence of outliers
Stochastic mortality models are important for a variety of actuarial tasks, from best-estimate forecasting to assessment of risk-capital requirements. However, the mortality shock associated with the Covid-19 pandemic distort forecasts. In this paper and presentation, we look at the robustification of three broad model classes: univariate time indices, such as in the Lee-Carter model; multivariate time indices, such as in the Cairns-Blake-Dowd model family; and penalty projections, such as with the 2D P-spline model.
Speaker: Stephen Richards