Applications of time-varying-parameter models to economics and finance

This dissertation focuses on applying time-varying-parameter models to the field of financial and monetary economics. The first two essays analyze the cross-sectional returns on the U.S. stock market by emphasizing the dynamics of risk loadings. The third essay studies the impact of a tight monetary policy on weak currencies during financial crises by examining the time-varying relationship between interest rates and exchange rates.; Motivated by the pricing errors found in small size and low book-to-market ratio portfolios in the Fama-French three-factor model, the first essay proposes a time-varying four-factor model. As small size and low book-to-market ratio firms are more sensitive to the risk related to innovations in the discount rate, the model incorporates a new risk factor to capture the information about the discount-rate risk for which the Fama-French three factors cannot fully account. In addition, the investors' learning process mimicked by the Kalman filter procedure is used to model the evolution of risk loadings. The results indicate that the model outperforms the Fama-French three-factor model in explaining the cross-sectional returns by substantially reducing pricing errors.; The second essay analyzes the risk-return relationship in a capital asset pricing model (CAPM) with a time-varying beta estimated by adaptive least squares (ALS) based on Kalman foundations. The results show the presence of a significant and positive risk-return relationship in the up market and the presence of a significant and negative risk-return relationship in the down market. In comparison with the model that assumes a constant beta, the CAMP with a time-varying beta reduces unexplained returns and improves the accuracy of the estimated risk-return relationship.; The third essay investigates the use of interest rates as a monetary instrument to stabilize exchange rates in the Asian financial crisis. Since previous studies suggest that the interest-exchange rate relationship may vary within, or across, regimes, a time-varying-parameter model with generalized autoregressive conditional heteroskedastic (GARCH) disturbances is used to estimate the impact of raising interest rates on exchange rates. The empirical evidence shows that an increase in interest rates leads to currency depreciation during certain periods of financial crises.