Empirical vector autoregressive modeling (Q1320556)

As the author puts it himself, the basic idea behind the study is to combine results out of the fields of econometrics, time series analysis and macroeconomics to develop, implement and apply a reasonably reliable method to uncover interesting linear time series relationships between macroeconomic variables. Going along this line, after the introductory chapter the author starts with an analysis of unrestricted vector autoregression and its components. Special emphasis is given to the simultaneous treatment of statistical unit root tests at different frequencies. This rather fresh approach is then applied to an analysis of aggregate investment in the final chapter of the book. The next step consists of a data analysis by vector autoregression. The two main topics addressed here are concerned with data-oriented measures of influence as well as with diagnostic checking. Two chapters follow presenting two special themes. The first one introduces the problem of seasonality into VAR-models and the second one deals with outliers. Here the author distinguishes between the effects of outliers on the results of a VAR-analysis and how to detect them. A derivation of the Lagrange multiplier statistic in the presence of outliers is also given. Under the heading ``Restrictions on the VAR'', the author draws attention to the issue of cointegration, the number of unit roots, and common trends. This chapter covers issues like estimating pushing trends and pulling equilibria as well as multivariate tests for unit roots. More technical details of the analysis as well as some brief applications are put in the appendices at the end of each chapter. In summary, it is a thoroughly written book about an up-to-date theme and it is recommended to all empirically oriented scientists in the fields of econometrics and macroeconomics.