New Directions in Econometric Practice: General to Specific Modelling, Cointegration, and Vector Autoregression

New Directions in Econometric Practice: General to Specific Modelling, Cointegration, and Vector Autoregression

New Directions in Econometric Practice: General to Specific Modelling, Cointegration, and Vector Autoregression

New Directions in Econometric Practice: General to Specific Modelling, Cointegration, and Vector Autoregression

Synopsis

Contents: Preface to Second Edition Preface 1. Traditional Methodology in Retrospect 2. Data Mining 3. Origins of a Modern Methodology: the DHSY Consumption Function 4. General to Specific Modelling 5. Cointegration Analysis 6. Vector Autoregression: Forecasting, Causality and Cointegration 7. Exogeneity and Structural Invariance 8. Non-nested Models, Encompassing and Model Selection References Index

Excerpt

The present volume represents an innovative approach to teaching econometrics. Most of the existing texts share two basic characteristics: (a) they concentrate on expositing econometric theory and techniques and (b) they cover 'standard' econometric techniques, such as (generalized) regression models, simultaneous equations models, limited dependent variable models, etc. The authors of the present text take a different approach, which practical researchers will find most welcome and which will answer the oft-heard question by students, 'Yes, yes, it's nice to know all these techniques, but if I do some research, how do I know what to do next?'

The volume exposits in detail the process by which empirical research is done. After a brief critique of the Cowles Foundation approach, it starts out with a most welcome chapter on data mining, a subject that is woefully neglected not only in the teaching of econometrics, but, one suspects, often in the practice of it. It then turns to the methodology introduced in the important 1978 paper of Davidson, Hendry, Srba and Yeo. It is a particularly interesting unifying theme of the book that the various techniques that are discussed chapter-by-chapter are nearly always illustrated with reference to the consumption model in that pioneering paper. The volume is thus one in which important techniques and methods are introduced and in which these methods are invariably applied to a concrete case that accompanies the reader from beginning to end.

An excellent discussion of error correction models is followed by a chapter that espouses the general philosophy of going from a general model to a more specific one and does this in the context of autoregressivedistributed lag models. Cointegration, vector autoregressions and Granger causality follow in neat order. The last two chapters deal with the various exogeneity concepts on the one hand and with nonnested models and encompassing on the other. The authors succeed entirely in reaching a happy balance between 'proving things' (although no arduous proofs of any kind are ever presented), providing rationales or explanations, and illustrating the material by working out the appropriate technique or procedure in the context of the consumption function model.

Many of these concepts that are systematically covered here are either completely omitted in standard textbooks or are given only the scantiest mention. It is a tour de force to write a 'book that makes all these concepts completely accessible to students on an intermediate level, i.e. those that have had no more than one course in statistics. It is safe to predict that teachers and students alike will relish the clarity, as well as the concrete orientation of this excellent volume.

Princeton University

RICHARD E. QUANDT

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.