UC Riverside

This dissertation provides innovative procedures for the estimation of the nonparametric triangular simultaneous equations model, extending beyond the traditional independent and identically distributed (i.i.d.) error structures in the reduced form and exploring its applications in empirical research.

Chapter 1 provides a comprehensive introduction to nonparametric simultaneous equations models, reviewing the various estimation methodologies discussed in the existing literature and limitation.

Chapter 2 advances the nonparametric estimation of structural equations by relaxing the i.i.d. error assumption in reduced form. I introduces a pre-whitening process designed to address heteroscedasticity in the nonparametric estimation of structural equations, followed by simulation analysis and asymptotic results.

Chapter 3 develops two novel methods for estimating structural equations under a nonparametric AR(1) error structure, supported by extensive simulation studies. The result demonstrates the consistency and efficiency of the proposed estimators, effectively managing both endogeneity and autocorrelation within the nonparametric estimation framework.

Chapter 4 highlights the practical application of the proposed estimators through an empirical analysis of cross-border air pollution spillover effects in the East Asian region. Using daily data from the recent COVID-19 period, this chapter employs a fully nonparametric approach to rigorously quantify and analyze the sources and magnitudes of air pollution and spillover effects. Additionally, I analyze the economic implications and policy recommendations derived from the findings.