Integrating Behavioural Theory and Machine Learning to Predict Life Insurance Purchase Intention in India

Abstract

Life insurance penetration In India, remains significantly below global averages. This despite a young population and rising income levels. Traditional theories, with its emphasis of life insurance being a rational product, have not been able to explain this paradox. It is important, therefore, to look beyond rationality and look at psychological and financial factors to explain this phenomenon. This study extended Theory of planned behaviour (TPB). This research integrated the psychological determinants of TPB i.e. attitude, subjective norms and PBC with financial drivers i.e. financial literacy, saving motive and risk aversion. Primary data was used in the study. The SEM model found that all the TPB constructs significantly predict intention with attitude being the strongest predictor. Financial literacy and Saving motive also had a significant and positive effect on intention. Additionally, the findings indicate that saving motivation serves as a mediating mechanism linking financial literacy to behavioural intention. Risk aversion showed a weaker but significant effect without moderation effect. To validate behavioural findings two ML models, using Random Forest and XGBoost were used. Both the ML models demonstrated superior predictive accuracy confirming nonlinear behavioural dynamics. The findings extend existing theoretical frameworks and offer meaningful implications for practitioners and policymakers for insurance policy design, marketing communication, and consumer inclusion strategies.