Analysis of Life Insurance Contract Cancellations Using the Accelerated Failure Time Model

Vladimír Mucha, Patrícia Ďuďák Teplanová, Ján Gogola, Jana Špirková
Statistika, 106(1): 35–48
https://doi.org/10.54694/stat.2025.26

Abstract
The aim of this paper is to analyse the cancellation of life insurance contracts on death using an accelerated failure time (AFT) model. The study focuses on identifying risk factors that influence the time to cancellation, with the objective of determining which to identify those insureds who cancel their policies the fastest. The analysis revealed several notable findings regarding the impact of premium payment frequency on contract cancellation. Specifically, yearly premium payments were found to extend the time to cancellation by 27% compared with monthly payments, holding all other factors constant. For contracts with monthly premiums, 10% of clients cancel within approximately 376 days, whereas for yearly premiums, the corresponding period is 476 days. Additionally, the results indicate that clients who did not conclude their contracts through the tied agent distribution channel tend to cancel their policies sooner. The AFT model was constructed using established R packages for survival analysis.

Keywords
Survival analysis, accelerated failure time model, censoring, cancellation of an insurance contract, life insurance