Two essays on time series econometrics: Multiple time series analysis of the Asian crisis and estimation and inference of the nonstationary structural vector autoregressive model.

摘要

This dissertation is composed of three chapters. Chapter One outlines the motivation and the structure of this dissertation. It also summarizes some of the important findings in Chapters Two and Three. Chapter Two applies multiple time series analysis to the Asian currency crisis in 1997--98. In particular, it examines the role of interest rate in exchange rate stabilization and the real effect of capital inflows in the East Asia crisis countries. The estimation results indicate that an increase in the interest rate has the traditional impact of appreciating nominal exchange rate during the crisis, although the impact is small. It is also found that net capital inflows tends to appreciate the real exchange rates but has no impact on domestic real interest rates or real output. Chapter Three studies the estimation and statistical inference of a nonstationary structural vector autoregressive model. We focus on situations in which researchers have no knowledge of and are not interested in the number and/or location of the unit roots. We derive the asymptotic properties of the following estimators: the ordinary least squares estimator, the two-stage least squares estimator, the three-stage least squares estimator, fully modified two-stage least squares estimator and an alternative modified two-stage least squares estimator. Monte Carlo simulations are conducted to compare the finite sample performance of these estimators. Implications for hypothesis testing are discussed.