This article provides a new account of employers' advantages over employees in federal employment discrimination cases. We analyze the effects of judicial deference, in which judges use institutionalized employment structures to infer nondiscrimination without scrutinizing those structures in any meaningful way. Using logistic regression to analyze a representative sample of judicial opinions in federal EEO cases during the first thirty-five years after the passage of the 1964 Civil Rights Act, we find that when judges uncritically use the presence of organizational structures to reason about whether discrimination occurred, employers are much more likely to prevail. This pattern is especially pronounced in opinions written by liberal judges. In light of these findings, we offer recommendations for judges, lawyers, and policy makers-including legal academics-who seek to improve the accuracy and efficacy of employment discrimination adjudications.