I study presidential power in the separation of powers system. My research asks what policy presidents pursue through unilateral actions, and how they calibrate the policy content of those actions under institutional constraint. To study this, I develop an anchored embedding scaling method to measure policy displacement of unilateral actions.
Substantively, I work on American political institutions, executive politics, legislative politics, and political communication. Methodologically, I use large-scale text analysis, large language models, and other computational tools to build new measures from political texts.
This dissertation examines how presidents calibrate the content of unilateral actions to manage institutional constraints and maximize political benefits. The first chapter proposes a new measurement framework to estimate the extent to which presidents move the policy status quo. The second chapter extends the unilateral action model, proposing a cost-constrained model. The third chapter studies whether presidents represent national constituencies through unilateral actions.
Chapter 1
Measurement
This paper develops a measure of policy displacement in presidential directives by placing unilateral actions in relation to the inherited statutory status quo. It provides the measurement foundation for studying how far presidents move policy through unilateral means.
Chapter 2
Theory
This paper develops a model in which presidents choose unilateral policy under institutional and political constraints. Rather than assuming binary costs of unilateral action, it argues that costs are continuous. Presidents stop short of their preferred policy when the political costs of displacement rise.
Chapter 3
Application
This paper examines whether presidents use unilateral actions to represent national constituencies or particular groups defined by race, class, and gender. Using the ideological locations of unilateral actions estimated in Chapter 1, it studies whether the content of directives is targeted toward specific constituencies rather than the national median voter.
My research brings together political institutions, political communication, and computational methods to study political actors' strategy and representation.
with Eunseong Oh
Under review at Political Analysis
Develops a diagnostic of vocabulary separability to guide model selection in political text classification, and explains when sparse representations outperform dense alternatives.
Text Analysis · NLP · Model Selection
Examines whether conspiracy beliefs shape campaign contributions in American politics, using the 2012 and 2016 ANES. The paper links conspiratorial belief to donor behavior across partisan contexts.
Campaign Finance · Conspiracy Theories · ANES
with Eunseong Oh
Presented at PolMeth 2025
Uses CLIP technology to analyze how political videos portray megadonors. Applies multimodal deep learning to understand the framing of political visual media.
Multimodal Analysis · CLIP · Political Communication
Teaching Assistant — UC Riverside (2023 – Present)
Vocabulary separability diagnostics for text classification.
Companion package for Oh and Yu, “When Sparse Beats Dense: Vocabulary Separability and Model Selection in Political Text Analysis.”
Interactive visualization of 20,961 presidential directives on the liberal–conservative scale using Anchored Embedding Scaling.
Preview available upon request.