Interventions Against Machine-Assisted Statistical Discrimination

TL;DR

This paper introduces a belief-contingent intervention called common identity that effectively eliminates statistical discrimination driven by verifiable beliefs, even with biased training data.

Contribution

It proposes a novel intervention strategy that constrains decision makers based on their beliefs, extending beyond traditional belief-free approaches like affirmative action.

Findings

Common identity intervention eliminates equilibrium statistical discrimination.

Effective even with biased training data.

Extends discrimination mitigation beyond belief-free methods.

Abstract

I study statistical discrimination driven by verifiable beliefs, such as those generated by machine learning, rather than by humans. When beliefs are verifiable, interventions against statistical discrimination can move beyond simple, belief-free designs like affirmative action, to more sophisticated ones, that constrain decision makers based on what they are thinking. I design a belief-contingent intervention I call common identity. I show that it is effective at eliminating equilibrium statistical discrimination, even when training data exhibit the various statistical biases that often plague algorithmic decision problems.