Microfoundation Inference for Strategic Prediction

TL;DR

This paper introduces a method to understand and estimate the long-term social impacts of predictive models, especially in strategic settings where predictions influence future outcomes, using optimal transport techniques.

Contribution

It proposes a novel approach to learn the distribution map of strategic responses, modeling agents' behavior as utility maximization and providing convergence guarantees.

Findings

Effective estimation of distributional impacts demonstrated on credit data

Method captures long-term strategic effects of predictive models

Provides theoretical convergence rates for the estimates

Abstract

Often in prediction tasks, the predictive model itself can influence the distribution of the target variable, a phenomenon termed performative prediction. Generally, this influence stems from strategic actions taken by stakeholders with a vested interest in predictive models. A key challenge that hinders the widespread adaptation of performative prediction in machine learning is that practitioners are generally unaware of the social impacts of their predictions. To address this gap, we propose a methodology for learning the distribution map that encapsulates the long-term impacts of predictive models on the population. Specifically, we model agents' responses as a cost-adjusted utility maximization problem and propose estimates for said cost. Our approach leverages optimal transport to align pre-model exposure (ex ante) and post-model exposure (ex post) distributions. We provide a rate of convergence for this proposed estimate and assess its quality through empirical demonstrations on a credit-scoring dataset.