Hao Wei Ching

Title: Incorporating cohort effects into a nonparametric Bühlmann credibility mortality mode
Date: Monday, August 24th 2026
Time: 3pm
Location: LIB 2020/ Zoom
Supervised by: Cary Tsai

Abstract:

Cohort effect is an important source in mortality forecasting as people born in different generations experience different mortality patterns. The nonparametric Bühlmann credibility mortality model proposed by Tsai and Lin (2017) provides a simple and effective approach to mortality forecasting and has been shown to produce accurate forecasts. However, the model only captures age-specific mortality improvement and does not account for cohort mortality information. This project extends the original age-based credibility mortality model by applying the Bühlmann-Straub credibility theory to yearly cohort mortality improvement patterns, allowing cohorts with different amounts of data being incorporated into the credibility framework. The proposed model then combines age-based and cohort-based forecasts using an optimal weighting approach. Empirical results show that the proposed model outperforms the age-based credibility, Lee-Carter, and Renshaw-Haberman mortality models. Overall, the proposed framework incorporates cohort effects into credibility-based mortality forecasting while maintaining simplicity and computational efficiency.