# Protecting the Protected Group: Circumventing Harmful Fairness

arXiv: https://arxiv.org/abs/1905.10546 · HTML: https://mlsanity.com/paper/1905.10546/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 May 2019.

## Built on by (most cited first, 9 of 9)

- [Input-agnostic Certified Group Fairness via Gaussian Parameter Smoothing](https://mlsanity.com/paper/2206.11423/) (2022)
- [Generalized Reductions: Making any Hierarchical Clustering Fair and Balanced with Low Cost](https://mlsanity.com/paper/2205.14198/) (2022)
- [Distributive Justice as the Foundational Premise of Fair ML: Unification, Extension, and Interpretation of Group Fairness Metrics](https://mlsanity.com/paper/2206.02897/) (2022)
- [Fair Polylog-Approximate Low-Cost Hierarchical Clustering](https://mlsanity.com/paper/2311.12501/) (2023)
- [Fairness Through Counterfactual Utilities](https://mlsanity.com/paper/2108.05315/) (2021)
- [Improving Fairness in Credit Lending Models using Subgroup Threshold Optimization](https://mlsanity.com/paper/2403.10652/) (2024)
- [Achievement and Fragility of Long-term Equitability](https://mlsanity.com/paper/2206.12333/) (2022)
- [When fairness metrics fail: A utility-based perspective on varepsilon-fairness](https://mlsanity.com/paper/2405.09360/) (2024)
- [Fairness under Competition](https://mlsanity.com/paper/2505.16291/) (2025)

## Abstract (arXiv)

Machine Learning (ML) algorithms shape our lives. Banks use them to determine if we are good borrowers; IT companies delegate them recruitment decisions; police apply ML for crime-prediction, and judges base their verdicts on ML. However, real-world examples show that such automated decisions tend to discriminate against protected groups. This potential discrimination generated a huge hype both in media and in the research community. Quite a few formal notions of fairness were proposed, which take a form of constraints a "fair" algorithm must satisfy. We focus on scenarios where fairness is imposed on a self-interested party (e.g., a bank that maximizes its revenue). We find that the disadvantaged protected group can be worse off after imposing a fairness constraint. We introduce a family of textitWelfare-Equalizing fairness constraints that equalize per-capita welfare of protected groups, and include textitDemographic Parity and textitEqual Opportunity as particular cases. In this family, we characterize conditions under which the fairness constraint helps the disadvantaged group. We also characterize the structure of the optimal textitWelfare-Equalizing classifier for the self-interested party, and provide an algorithm to compute it. Overall, our textitWelfare-Equalizing fairness approach provides a unified framework for discussing fairness in classification in the presence of a self-interested party.