DISCRIMINATION-FREE INSURANCE PRICING

Abstract We consider the following question: given information on individual policyholder characteristics, how can we ensure that insurance prices do not discriminate with respect to protected such as gender? address issues of direct and indirect discrimination, latter resulting from implicit learning characteristics nonprotected ones. provide rigorous mathematical definitions for introduce a simple formula discrimination-free pricing, avoids both discrimination. Our works in any statistical model. demonstrate its application health example, using state-of-the-art generalized linear model neural network regression An important conclusion is pricing general requires collection policyholders’ discriminatory posing potential challenges relation policyholder’s privacy concerns.