This book discusses the econometric foundations of structural vector autoregressive modeling, as used in empirical macroeconomics, finance, and related fields.

1. 1 Integrating results The empirical study of macroeconomic time series is interesting. It is also difficult and not immediately rewarding. Many statistical and economic issues are involved. The main problems is that these issues are so interrelated that it does not seem sensible to address them one at a time. As soon as one sets about the making of a model of macroeconomic time series one has to choose which problems one will try to tackle oneself and which problems one will leave unresolved or to be solved by others. From a theoretic point of view it can be fruitful to concentrate oneself on only one problem. If one follows this strategy in empirical application one runs a serious risk of making a seemingly interesting model, that is just a corollary of some important mistake in the handling of other problems. Two well known examples of statistical artifacts are the finding of Kuznets "pseudo-waves" of about 20 years in economic activity (Sargent (1979, p. 248)) and the "spurious regression" of macroeconomic time series described in Granger and Newbold (1986, §6. 4). The easiest way to get away with possible mistakes is to admit they may be there in the first place, but that time constraints and unfamiliarity with the solution do not allow the researcher to do something about them. This can be a viable argument.

This monograph is concerned with the statistical analysis of multivariate systems of non-stationary time series of type I. It applies the concepts of cointegration and common trends in the framework of the Gaussian vector autoregressive model.

This book contributes to re cent developments on the statistical analysis of multiple time series in the presence of regime shifts. Markov-switching models have become popular for modelling non-linearities and regime shifts, mainly, in univariate eco nomic time series. This study is intended to provide a systematic and operational ap proach to the econometric modelling of dynamic systems subject to shifts in regime, based on the Markov-switching vector autoregressive model. The study presents a comprehensive analysis of the theoretical properties of Markov-switching vector autoregressive processes and the related statistical methods. The statistical concepts are illustrated with applications to empirical business cyde research. This monograph is a revised version of my dissertation which has been accepted by the Economics Department of the Humboldt-University of Berlin in 1996. It con sists mainly of unpublished material which has been presented during the last years at conferences and in seminars. The major parts of this study were written while I was supported by the Deutsche Forschungsgemeinschajt (DFG), Berliner Graduier tenkolleg Angewandte Mikroökonomik and Sondeiforschungsbereich 373 at the Free University and Humboldt-University of Berlin. Work was finally completed in the project The Econometrics of Macroeconomic Forecasting founded by the Economic and Social Research Council (ESRC) at the Institute of Economies and Statistics, University of Oxford. It is a pleasure to record my thanks to these institutions for their support of my research embodied in this study.

Bayesian Multivariate Time Series Methods for Empirical Macroeconomics provides a survey of the Bayesian methods used in modern empirical macroeconomics. These models have been developed to address the fact that most questions of interest to empirical macroeconomists involve several variables and must be addressed using multivariate time series methods. Many different multivariate time series models have been used in macroeconomics, but Vector Autoregressive (VAR) models have been among the most popular. Bayesian Multivariate Time Series Methods for Empirical Macroeconomics reviews and extends the Bayesian literature on VARs, TVP-VARs and TVP-FAVARs with a focus on the practitioner. The authors go beyond simply defining each model, but specify how to use them in practice, discuss the advantages and disadvantages of each and offer tips on when and why each model can be used.

1. 1 Objective of the Study Vector autoregressive (VAR) models have become one of the dominant research tools in the analysis of macroeconomic time series during the last two decades. The great success of this modeling class started with Sims' (1980) critique of the traditional simultaneous equation models (SEM). Sims criticized the use of 'too many incredible restrictions' based on 'supposed a priori knowledge' in large scale macroeconometric models which were popular at that time. Therefore, he advo cated largely unrestricted reduced form multivariate time series models, unrestricted VAR models in particular. Ever since his influential paper these models have been employed extensively to characterize the underlying dynamics in systems of time series. In particular, tools to summarize the dynamic interaction between the system variables, such as impulse response analysis or forecast error variance decompo sitions, have been developed over the years. The econometrics of VAR models and related quantities is now well established and has found its way into various textbooks including inter alia Llitkepohl (1991), Hamilton (1994), Enders (1995), Hendry (1995) and Greene (2002). The unrestricted VAR model provides a general and very flexible framework that proved to be useful to summarize the data characteristics of economic time series. Unfortunately, the flexibility of these models causes severe problems: In an unrestricted VAR model, each variable is expressed as a linear function of lagged values of itself and all other variables in the system.