Applied Bayesian econometrics for central bankers - Centre for Central Banking Studies

This event is an introduction to some of the techniques in Bayesian econometrics which can be useful for modelling and forecasting in central banks. It will provide an overview of the theory and then focus on practical implementation through computer-based exercises. The seminar will be taught from the perspective of the practitioner with the aim of discussing techniques that can improve upon classical econometric methods, or are more convenient alternatives. The topics covered are likely to include: - introduction to Bayesian analysis and Gibbs sampling; - Gibbs sampling for linear regressions and vector autoregressions; - Gibbs sampling for state-space models including time-varying parameter and dynamic factor models; - non-linear regression models: the Metropolis-Hastings algorithm; - Bayesian model averaging; and - Bayesian estimation of dynamic stochastic general equilibrium models.