A Fair Price to Pay: Exploiting Causal Graphs for Fairness in Insurance

In many jurisdictions, insurance companies must not discriminate on some given policyholder characteristics. Omission of prohibited variables from models prevents direct discrimination, but fails to address proxy discrimination, a phenomenon especially prevalent when powerful predictive algorithms are fed with an abundance of acceptable covariates. The lack of formal definition for key fairness concepts, in particular indirect discrimination, hinders the fairness assessment of methodologies. We review causal inference notions and introduce a causal graph tailored for fairness in insurance. Exploiting these, we discuss potential sources of bias, formally define direct and indirect discrimination, and study the properties of fairness methodologies. A novel categorization of fair methodologies into five families (best-estimate, unaware, aware, hyperaware, and corrective) is constructed based on their expected fairness properties. A comprehensive pedagogical example illustrates the practical implications of our findings: the interplay between our fair score families, group fairness criteria, and sources of discrimination.

Please join the event.

About Marie-Pier Côté

Marie-Pier Côté is an Associate Professor at the School of actuarial sciences of Université Laval in Québec, Canada, where she holds the Chair in eductional leadership on big data analytics for actuarial sciences – Intact. She is a member of the Institute Intelligence and Data and the Big Data Research Center at Université Laval.

Her research interests are in dependence modelling, statistical learning, and fairness in actuarial sciences. She has ongoing research collaborations with insurance companies. With her co-authors, she won the North American Actuarial Journal 2021 Best Paper Award and the ARIA Shapiro-Brockett Award for her research on gradient boosting in insurance pricing. She is also dedicated to education through her implementation of high-quality inclusive teaching practices, and was recognized by Faculty teaching awards.

Marie-Pier completed a Master’s and a Doctoral degree in statistics at McGill University, under the supervision of Christian Genest. Prior to joining McGill, she studied actuarial sciences at the undergraduate level at Université Laval. She is a Fellow of the Society of Actuaries and an Associate of the Canadian Institute of Actuaries.