Dependence modeling of multivariate longitudinal hybrid insurance data with dropout

作者：

Highlights：

• The cause–effect pattern in customer lapses and prices is usually bidirectional.

• A method to estimate the association between price and lapsing is presented.

• The approach is suitable for large longitudinal data sets and multivariate processes.

• An insurance case study shows the implications for assessing interrelated risks.

摘要

•The cause–effect pattern in customer lapses and prices is usually bidirectional.•A method to estimate the association between price and lapsing is presented.•The approach is suitable for large longitudinal data sets and multivariate processes.•An insurance case study shows the implications for assessing interrelated risks.

论文关键词：Copula,Generalized method of moments,Joint longitudinal and time-to-event,Insurance lapsation,Tweedie distribution

论文评审过程：Received 14 May 2020, Revised 4 June 2021, Accepted 1 July 2021, Available online 16 July 2021, Version of Record 20 July 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115552