Customer Reviews

Unit Roots, Cointegration, and Structural Change (Themes in Modern Econometrics)

5 star:		(3)
4 star:		(3)
3 star:		(0)
2 star:		(0)
1 star:		(1)

 

 

This is a book on specialized topics in econometric modeling. Like Franses recent book it deals with ARIMA models with unit roots and advances in the theory of cointegration. This book is somewhat advanced but is perfect for the right audience, the statisticians and econometricians that deal with time series modeling (univariate and multivariate ) and structural equation modeling.

The asymptotic theory is well covered but the unique feature of the book is that it points out that the asymptotics can give very poor approximations in small to moderate sample sizes. The authors provide alternatives including the use of the bootstrap for standard error estimates, confidence bounds and hypothesis testing (particularly tests for unit roots).

It is clear and covers the important literature. Much like Franses book it covers bootstrap and Bayesian methods and really does provide a current and useful approach to important problems and methodology in econometrics.

It could be used for a special topics graduate course or as a supplement to a graduate course in econometrics.

This book is extremely well written. It gives a good intuition behind unit roots and cointegration. It tells us to be critical when it comes to unit roots and cointegration; they are not very powerful in a statistical sense. The authors warn of using such techniques blindly. This book is a good start for an intuitive feel after a tough Time Series course where one is all entangled in sophisticated statistical techniques.It is also a good book for the professional economist not very knowlegeable in time series econometrics. I personally learned a lot from that book and it increased my critical capacity when it comes to econometrics.

I was required to read this book as a part of my field course in econometrics. The book is full of mistakes, especially on asymptotic distributions of estimators. Since the properties of these estimators of unit root tests depend on those asymptotic distributions, this kind of mistake is crucial. Also, the results they derive are sometimes misleading and other times even outright wrong! This book is a mere survey of papers other people wrote and not a good one at that. In summary, do not spend your money on this book!

This book gives very good high level semi-technical review and introduction on the many different techniques in the areas of unit root and cointegration. This is a good starting point if you want to get into this new area. It also lists quite complete references based on which you can start your development and research work.

This book is very helpful in case that your aim is to construct a basement for unit root and cointegration study. The language of the book is clear and easy to follow. Besides the introduction and explanation of different methods, it compares different methods by using recent empirical studies.

The book by Maddala and Kim is a very good and relativelynontechnical pointer to the time series literature on unit roots andcointegration. The book is meant to be used as a reference, not readas a text.

Maddala and Kim do a good job of discussing the strengths and weaknesses of the myriad unit root and cointegration tests that have appeared in recent years. If you want to know more about the asymptotic theory, then refer to _Time Series Analysis_ by Hamilton or read the original journal articles.

For a more balanced review of the Maddala and Kim book, one might also take a look at the book review written by Heather Anderson in the December 1999 issue of the Economic Record. END

Excellent book

This one of the best book about cointegration.