Equalized Odds

Equalized odds is a fairness metric that asks whether a model makes errors at the same rate across groups, given the true outcome. A loan model satisfying equalized odds would incorrectly approve and incorrectly reject applicants at similar rates regardless of demographic group. This metric cares about balancing both types of errors (false positives and false negatives) across protected groups. Mathematically, equalized odds requires that true positive rates and false positive rates be equal across groups, which is provably incompatible with predictive parity when base rates differ.