Applied Bayesian Time Series Methods

Catalog description

The course introduces students to Bayesian time series methods. Students will learn how to make likelihood-based inference about unobserved quantities, e.g. model parameters, policy impacts or future outcomes, conditional on the observed data. Applications include structural vector autoregressions, state space models and linearized dynamic stochastic general equilibrium macro models. Student will become familiar with numerical posterior simulation techniques such as Gibbs sampling and the Metropolis-Hasting algorithm. The course is useful for any students interested in empirical work that involves time series and/or structural likelihood-based estimation.