Finding Law: Learning in the Judicial Hierarchy

Abstract

The U.S. Supreme Court supervises the disposition of hundreds of thousands of cases each year, and the law that governs these disputes is not static: new problems appear, new classes of disputes arise, and doctrine must adapt to govern their resolution. This dissertation explores how the Supreme Court learns to create and extend doctrine to adjudicate these cases. I argue the Supreme Court relies on the Courts of Appeals as laboratories of law, observing their decisions and reviewing those that best inform legal development.

I develop a formal-theoretic model in which the Supreme Court supervises two lower courts deciding similar cases. A tension arises because the Supreme Court can see only case outcomes from the lower courts---only upon costly review can the Supreme Court determine whether the arguments that support a lower court's decision are sound. But, because an unbiased judge only makes an extreme decision when there is an imbalance in the parties' arguments, the Supreme Court is able to draw some inferences without review. The Court thus leverages multiple Courts of Appeals decisions to identify which will be most informative to review. Upon review, it learns from the arguments presented in the instant case to develop legal doctrine.

The model generates a number of empirical predictions, which the subsequent chapters test. First, because moderate decisions are ambiguous and therefore more informative to the Supreme Court, the model predicts the Supreme Court should be more likely to review these. The data support this prediction. Second, consistent with the model's predictions, when the Supreme Court is resolving a conflict between lower courts, it is more likely to review cases on the side whose doctrine it ultimately rejects, especially when lower courts are heterogeneous. Finally, the model predicts the Supreme Court will reverse extreme decisions more than moderate decisions; empirically, the Court is more likely to reverse liberal decisions than moderate ones, but the same is not true for conservative decisions. The results shed light on how hierarchy eases the inherent difficulty and uncertainty of crafting law and on how the Supreme Court learns to create doctrine.