2000 character limit reached
De-Biasing Models of Biased Decisions: A Comparison of Methods Using Mortgage Application Data (2405.00910v1)
Published 1 May 2024 in cs.LG, cs.CY, and econ.EM
Abstract: Prediction models can improve efficiency by automating decisions such as the approval of loan applications. However, they may inherit bias against protected groups from the data they are trained on. This paper adds counterfactual (simulated) ethnic bias to real data on mortgage application decisions, and shows that this bias is replicated by a machine learning model (XGBoost) even when ethnicity is not used as a predictive variable. Next, several other de-biasing methods are compared: averaging over prohibited variables, taking the most favorable prediction over prohibited variables (a novel method), and jointly minimizing errors as well as the association between predictions and prohibited variables. De-biasing can recover some of the original decisions, but the results are sensitive to whether the bias is effected through a proxy.
- The effects of the 1930s HOLC redlining maps. American Economic Journal: Economic Policy, 13(4), 355–392.
- Racial segregation in housing markets and the erosion of black wealth. Review of Economics and Statistics, 1–45.
- Consumer-lending discrimination in the FinTech era. Journal of Financial Economics.
- Fairness in criminal justice risk assessments: The state of the art. Sociological Methods & Research, 50(1), 3–44.
- Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination. American Economic Review, 94(4), 991–1013.
- How much does racial bias affect mortgage lending? Evidence from human and algorithmic credit decisions..
- Racial discrimination in mortgage lending in Washington, DC: A mixed methods approach. The Review of Black Political Economy, 33(2), 9–30.
- How costly is noise? Data and disparities in consumer credit. arXiv preprint.
- Inaccurate statistical discrimination: An identification problem. National Bureau of Economic Research working paper.
- Building classifiers with independency constraints. In International Conference on Data Mining Workshops, pp. 13–18. IEEE.
- Three naive Bayes approaches for discrimination-free classification. Data Mining and Knowledge Discovery, 21(2), 277–292.
- Classification with fairness constraints: A meta-algorithm with provable guarantees. In Proceedings of the Conference on Fairness, Accountability, and Transparency, pp. 319–328.
- CFPB (2019). Consumer Credit Panel..
- Why is my classifier discriminatory?. Advances in Neural Information Processing Systems, 31.
- XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, pp. 785–794, New York, NY, USA. ACM.
- The racial wealth gap, 1860-2020. Manuscript, Princeton University and University of Bonn.
- Federal Financial Institutions Examination Council (2018). A guide to HMDA reporting: Getting It Right! 2018 edition..
- An observational implementation of the outcome test with an application to ethnic prejudice in pretrial detentions. Working paper.
- Proxy fairness. arXiv preprint.
- Bias mitigation for machine learning classifiers: A comprehensive survey. arXiv preprint.
- Policy statement on discrimination in lending. Federal Register, 59(23).
- Hull, P. (2021). What marginal outcome tests can tell us about racially biased decision-making. NBER Working Paper.
- Identifying and correcting label bias in machine learning. International Conference on Artificial Intelligence and Statistics, 702–712.
- Classifying without discriminating. In 2nd International Conference on Computer, Control and Communication, pp. 1–6. IEEE.
- Fairness-aware learning through regularization approach. In 11th International Conference on Data Mining, pp. 643–650. IEEE.
- Ladd, H. F. (1998). Evidence on discrimination in mortgage lending. Journal of Economic Perspectives, 12(2), 41–62.
- Racial and ethnic disparities in mortgage lending: New evidence from expanded HMDA data. OCC Working Paper.
- A survey on bias and fairness in machine learning. arXiv preprint.
- OCC (2023). Comptroller’s Handbook: Fair Lending. Office of the Comptroller of the Currency.
- Implementing anti-discrimination policies in statistical profiling models. American Economic Journal: Economic Policy, 3(3), 206–31.
- FairXGBoost: Fairness-aware classification in XGBoost. arXiv preprint.
- Color of credit: Mortgage discrimination, research methods, and fair lending enforcement..
- Tenev, N. (2020). Social connections and racial wage inequality. Working paper.
- Intergenerational economic mobility and the racial wealth gap. In AEA Papers and Proceedings, Vol. 111, pp. 206–10.
- Removing biased data to improve fairness and accuracy. arXiv preprint.
- Learning fair representations. In International Conference on Machine Learning, pp. 325–333. PMLR.
- Do lenders still discriminate? A robust approach for assessing differences in menus. National Bureau of Economic Research Working Paper.
- Zhang, Y. (2022). Machine learning, alternative data, and mortgage bias. SSRN working paper.