Accelerated Failure Time Models:

An Application in Insurance Attrition

Abstract

Despite the seeming power of survival analysis over popular binary models in insurance attrition analysis, its consideration is now growing in the literature. Besides, studies have only considered the Kaplan-Meier estimator and the Cox proportional hazards model. To our knowledge, no single study has modeled insurance attrition using the accelerated failure time model. This study presents some parametric models in survival analysis, specifically, the accelerated failure time model. Furthermore, we investigate the applicability of this model in analyzing insurance attrition using life insurance data. We show for the first time that the accelerated failure time model offers an attractive alternative to the Kaplan-Meier estimator, and the Cox proportional hazards model in estimating insurance attrition. Based on the Akaike information criterion, the generalized gamma model provides the best fit for the data. This work will serve as the basis for the consideration of parametric survival models in estimating insurance attrition, deepen knowledge in retention analysis, and broaden the scope of survival analysis.