The Cohort Shapley value to measure fairness in financing small and medium enterprises in the UK

Abstract

Banks are relying on machine learning techniques to support their decisions in financing small and medium enterprises (SMEs). As regulators require that credit decisions are transparent, there is a need to develop methods to measure fairness. We propose a weighted average of the Cohort Shapley value, which removes impossible feature combinations, and a relative fairness score for assessing the fairness level within sub-populations. Based on our knowledge, this is the first paper that investigates the fairness of UK financial institutions in providing funding to SMEs. Our findings reveal discrimination against start-up, micro, women-led companies, and owners of Asian ethnic backgrounds.