A multi-dimensional Bühlmann credibility approach to modeling multi-population mortality rates. Issue 5 (28th May 2019)

• Record Type: Journal Article
• Authors: Tsai, Cary Chi-Liang; Zhang, Ying
• Is Part Of: Scandinavian actuarial journal. Volume 2019:Issue 5(2019)
• Page Start: 406
• Page End: 431
• Publication Date: 2019-05-28
• Subjects: Bühlmann credibility theory -- lee-carter model -- CBD model -- mortality model; Insurance, Life -- Mathematics -- Periodicals; Insurance -- Mathematics -- Periodicals; 368.01
• DOI: 10.1080/03461238.2018.1563911
• Languages: English
• ISSNs: 0346-1238

Abstract

ABSTRACT: In this paper, we first propose a multi-dimensional Bühlmann credibility approach to forecasting mortality rates for multiple populations, and then compare forecasting performances among the proposed approach, the CBD model, the Lee-Carter model (LC), the joint- k (JoK-LC), the co-integrated (CoI-LC), and the augmented common factor (ACF-LC) Lee-Carter models for multiple populations. Mortality data from the Human Mortality Database are fitted to the underlying mortality models for both genders of three well-developed countries (the US, the UK, and Japan) and both genders of a developed country (France) and a developing country (Poland) with an age span 25–84 and a wide range of fitting year spans. Empirical illustrations show that the proposed multi-dimensional Bühlmann credibility approach contributes to more accurate forecast results, measured by AMAPE (average of mean absolute percentage errors over all fitting year spans), than the CBD, LC, JoK-LC, CoI-LC and ACF-LC models for three forecasting year spans 2004–2013 (10-year wide), 1994–2013 (20-year wide) and 1984–2013 (30-year wide).