Algorithmic Fairness and Economics

Abstract

We develop an economic perspective on algorithmic fairness and the surrounding empirical, theoretical and policy issues. Our perspective draws from clear parallels between algorithms and issues in economics of discrimination, crime, personnel and technological innovation; as well as more subtle connections to environmental economics, product safety regulation, behavioral economics and economics of information. We highlight the distinction between biased algorithmic predictions and biased algorithmic objectives. We conclude by discussing economic issues in policy policy and managerial practices around reducing algorithmic unfairness.

Keywords: economics, algorithms, bias, fairness

JEL Classification: J7, J16