Ideal Point Estimation in Political Hierarchies: A Framework and an Application to the US Executive Branch

Abstract

Abstract This article develops a procedure for estimating the ideal points of actors in a political hierarchy, such as a public bureaucracy. The procedure is based on a spatial auditing model and is motivated by the idea that while agents within a political hierarchy are typically segregated in different policy fiefdoms, they are bound to a common principal that can scrutinize their policy proposals through selective reviews, or audits. The theoretical model shows how a principal’s decision to audit an agent’s proposal can reveal both actors’ spatial preferences, despite the strategic nature of the interaction. Empirical identification of the ideal points comes from leveraging settings where elections replace principals over time, but not agents. Although the procedure is quite general, I provide an illustration using data on federal regulatory policymaking in the United States and recover ideal point estimates for presidents and agencies across three administrations.