Bayesian inference in dynamic econometric models. With a foreword by Jacques J. Drèze (Q2760416)

This textbook deals with dynamic econometric models using only Bayesian methods for estimation and inference. The book originates from different lecture notes of courses on Bayesian econometrics and contains the work of the last 25 years done at CORE in this research area. Given that comprehensive works on Bayesian econometrics are not numerous, this book fills the gap. Another advantage of the book is the thoughtful general introduction to Bayesian inference in econometrics with a comprehensive treatment of a class of models that have been the subject of extensive research over the past decades, namely dynamic models.NEWLINENEWLINENEWLINEDynamic econometric models have received sustained attention in recent years, both at the micro as well as at the macro level. The participation of Bayesian econometricians in this effort reflects the progress made with implementation of the Bayesian approach in general.NEWLINENEWLINENEWLINEThe book is divided into two major parts. Part one presents an introduction into Bayesian econometrics, which is enriched with a chapter on numerical integration. The second part of the book covers a wide field of econometrics. The dynamics may affect exogenous variables, endogenous variables or error processes; they may be linear or nonlinear; they may concern a single equation or a system of equations. The book fits most relevant cases into five solid and logically separated chapters.NEWLINENEWLINENEWLINEThe first part consists of four chapters starting with an introductory chapter on decision theory and Bayesian inference, followed by a chapter on Bayesian statistics and linear regression. Methods of numerical integration are presented in the third chapter and chapter four contains prior densities for the regression model. Part 2 starts with dynamic regression models as the starting point. In chapter six the authors discuss unit roots inference in the Bayesian context, followed by heteroscedasticity and ARCH in chapter seven. Chapters 6 and 7 are restricted to single equation models which are linear in the parameters. Extending the Bayesian approach, chapter 8 discusses nonlinear time series models and the final chapter addresses the issue of systems of equations.NEWLINENEWLINENEWLINETo summarize, the book presents a comprehensive survey of Bayesian econometrics which is unique in its presentation and which covers the most recent developments in this research area.