Publication Type

PhD Dissertation

Version

publishedVersion

Publication Date

2-2026

Abstract

This dissertation investigates the complex behavioral responses to public policy shifts and institutional cues, quantifying their intended and unintended societal impacts. Across two distinct empirical contexts, this research examines how individuals adapt their risk profiles when signals from political or legal authorities change and how these adaptations generate measurable externalities.

The first chapter, "From Punishment to Treatment: Unintended Social Impacts of Changing Drug Policies," shifts the focus to the intersection of health policy and criminal justice. This chapter evaluates the sudden transition from punitive drug enforcement to health-oriented decriminalization, utilizing the 2021 State v. Blake decision in Washington as a natural experiment. Employing Synthetic Control and Difference-in-Differences methodologies, the analysis reveals that while the elimination of legal deterrence successfully reduced drug possession arrests, it inadvertently altered illicit market dynamics. The reduction in enforcement risk facilitated market expansion and drug substitution, evidenced by a significant increase in fentanyl seizures and a corresponding rise in specific violent offenses, notably robbery. Furthermore, the analysis uncovers significant racial and demographic heterogeneities in how these policy outcomes are distributed.

The second chapter, "Elite Cues and Cognitive Resistance: A Difference-in-Differences Analysis of Political Polarization and COVID-19," explores the morbidity cost of political influence during a public health crisis. Utilizing a Poisson Pseudo-Maximum Likelihood (PPML) Two-Way Fixed Effects (TWFE) design, it tests whether "cognitive resistance" to executive minimizing rhetoric—measured via continuous presidential disapproval ratings—altered state-level infection rates. While the preferred model suggests that high political disapproval generated an approximately 64 percent reduction in expected monthly cases relative to the baseline, rigorous econometric interrogation reveals these behavioral signals are highly sensitive to structural vulnerabilities. By demonstrating how unstandardized log-transformations artificially compress population variance and how missing pre-pandemic data triggers listwise deletion, this chapter provides a critical methodological critique. It concludes that structural epidemiology and urban density ultimately overpower political rhetoric when spatial panels are properly standardized.

Together, these essays highlight the fragility of behavioral compliance and the complexity of policy reform. By rigorously examining both the direct objectives and the unintended social externalities of institutional shifts, this dissertation underscores the necessity of comprehensive, methodologically sound approaches in the design and evaluation of public health and legal frameworks.

Keywords

Health Economics, COVID-19, Political Polarization, Causal Inference, Poisson Pseudo-Maximum Likelihood, Policy Evaluation

Degree Awarded

PhD in Economics

Discipline

Health Economics | Public Health

Supervisor(s)

ZHANG, Yichong

First Page

1

Last Page

58

Publisher

Singapore Management University

City or Country

Singapore

Copyright Owner and License

Author

Download

Share

COinS