Risk-Based Premiums of Insurance Guarantee Schemes: A Machine-Learning Approach

Abstract

Introduction/Main Objectives: This study explores the application of machine-learning techniques to risk-based premium calculations for insurance guarantee schemes within the Indonesian insurance market. This study aims to develop a risk-based premium calculation model using machine-learning techniques in the Indonesian context. Background Problems: A gap exists in determining risk-based premiums for both the life and non-life insurance sectors within the Indonesian insurance market. Identifying and understanding the key variables that significantly influence risk-based capital (RBC) is important, and this research addresses this need. Novelty: This paper is the first to apply machine learning to calculate risk-based premiums in the context of the Indonesian insurance market. The distinction between the life and non-life insurance sectors in terms of the importance of its variables and itsselection of an optimal model further enrich its unique approach. Research Methods: We employed gradient-boosted and decision-tree models to identify key factors impacting risk-based capital. Furthermore, we leveraged clustering techniques to categorize companies into distinct risk tiers, aiming to enable more precise risk-based premium rate calculations. Finding/Results: The findings reveal significant differences between the life and non-life insurance sectors in terms of key variables that impact their risk-based capital. These insights lead to the categorization of insurance companies into distinct risk tiers whichhelps to more accurately calculate risk-based premiums. Conclusion: Machine learning can serve as a powerful tool in refining insurance risk management practices, ultimately benefiting insurers, policyholders, and regulators alike.