Applied Bayesian econometrics for central bankers

The aim of this handbook is to introduce key topics in Bayesian econometrics from an applied perspective. The handbook assumes that readers have a fair grasp of basic classical econometrics (e.g. maximum likelihood estimation).  It is recommended that readers familiarise themselves with Matlab© programming language to derive the maximum benefit from this handbook.  A basic guide to Matlab© is attached at the end of the handbook. The first chapter of the handbook introduces basic concepts of Bayesian analysis.  In particular, the chapter focuses on the technique of Gibbs sampling and applies it to a linear regression model.  The chapter shows how to code this algorithm via several practical examples.  The second chapter introduces Bayesian vector autoregressions (VARs) and discusses how Gibbs sampling can be used for these models.  The third chapter shows how Gibbs sampling can be applied to popular econometric models such as time-varying VARs and dynamic factor models.  The final chapter introduces the Metropolis Hastings algorithm.  We intend to introduce new topics in revised versions of this handbook on a regular basis. The handbook comes with a set of Matlab© codes that can be used to replicate the examples in each chapter.  The code (provided in code.zip) is organised by chapter.  For example, the folder 'Chapter 1' contains all the examples referred to in the first chapter of this handbook. The views expressed in this handbook are those of the authors, and not necessarily those of the Bank of England.  The reference material and computer codes are provided without any guarantee of accuracy.