Essays on Bayesian Macroeconometrics

This dissertation consists of three chapters that study the determinants of macroeconomic fluctuations, with a particular emphasis on the roles of agents\u27 expectations and assessments of risks. In the first chapter, I study a business cycle model where the probability of transitioning to a downturn state characterized by low growth evolves over time. I call a change in the future probability of transitions to the downturn state a downturn risk shock. An increase in the risk of the downturn state leads to declines in consumption, investment, output, and hours. I take the model to the data using Bayesian methods. The fluctuations caused by expectations changes from the downturn risk shock account for substantial output variations at business cycle frequencies and hours fluctuations at medium run frequencies. The extracted time-varying probability process from the model matches well with the University of Michigan Index of Consumer Sentiments (ICS). Impulse response functions from downturn risk shocks produce the same comovement as those from ICS innovations in a structural vector autoregression. The second chapter, co-authored with Minchul Shin, proposes the multivariate stochastic volatility in vector autoregression model as a framework for studying the real effects of uncertainty shocks. We advance a new approach to structurally identify uncertainty shocks that does not rely on identification of the level structural shocks. We estimate the model using a Bayesian Markov chain Monte Carlo algorithm, and we show how to construct impulse response functions and variance decompositions that can be used to analyze identified uncertainty shocks. The third chapter, also co-authored with Minchul Shin, suggests using ``realized volatility\u27\u27 as a volatility proxy to aid in model-based multivariate bond yield density forecasting. To do so, we develop a general estimation approach to incorporate volatility proxy information into dynamic factor models with stochastic volatility. We study the density prediction performance on U.S. bond yields of including realized volatility into a dynamic Nelson-Siegel (DNS) model with stochastic volatility. The results clearly indicate that using realized volatility improves density forecasts relative to popular specifications in the DNS literature that neglect realized volatility