Customer Reviews

Econometric Analysis of Cross Section and Panel Data

5 star:		(9)
4 star:		(4)
3 star:		(2)
2 star:		(1)
1 star:		(1)

 

 

If you are interested in learning the latest microeconometric methods, this book will save you from long hours of sorting through the literature. Wooldridge has brought together in a well organized, clear and concise manner the state of the art techniques in microeconometrics. His book covers all of the core issues involving single equation and simultaneous equation models. The most important aspects of M-estimation, MLE, GMM and minimum distance estimation are carefully presented. Those interested in limited dependent and qualitative variables will find that this goes well beyond Maddala's classic book. In addition to new developments in logit, probit and tobit, Wooldridge explains sample selection, attrition and stratified sampling. He covers research by Heckman et al on estimating average treatment effects using instrumental variables. I found his material on negative binomial regression, binomial regression, exponential regression and fractional logit regression to be especially interesting. He concludes his book with a nice summary of research on duration analysis. Throughout the book Wooldridge shows how to handle panel data with the various techniques he covers. Anyone doing applied work with cross section or panel data runs the risk of being left behind if they fail to read this new classic of microeconometrics.

In my opinion, this is now the best introduction to cross-section econometrics. Wooldridge covers all the basics, he does it very well, with a lot of attention to empirical applications. Most of the empirical applications can be replicated using Stata datasets that you can download from his web page. One thing I love about this textbook is that robust versions of variance matrices are almost always provided, so one does not have to rely much on homoskedasticity. This is also one of the very very few textbooks to devote some space to important topics such as stratified sampling, clustering, weak instruments. The exposition is generally excellent, with intuition provided, and proofs rigorous enough for beginners, even if the most technical details are usually left out. Sometimes I find Wooldridge style a bit disorganized in the way he orders topics within a chapter, but overall I do think this is a great book, and the best introduction to the topic. I never liked Greene, which devotes too much space to irrelevant topics. Ruud is a good book too, but very technical, and probably one that you want to keep as a "backup". Hayashi is very nice too, but a bit unusual, with its strong emphasis on assumptions such as stationariy and ergodicity that are important in time-series (not covered AT ALL in Wooldridge). Amemiya is still a great reference, but it's too advanced for beginners. Overall, I highly recommend this book.

Here are the topics covered:
- Conditional Expectations and Related Concepts in Econometrics
- Basic Asymptotic Theory
- Linear Models
- The Single-Equation Linear Model and OLS Estimation
- Instrumental Variables Estimation of Single-Equation Linear Models
- Additional Single-Equation Topics
- Estimating Systems of Equations by OLS and GLS
- System Estimation by Instrumental Variables
- Simultaneous Equations Models
- Basic Linear Unobserved Effects Panel Data Models
- More Topics in Linear Unobserved Effects Models
- M-Estimation
- Maximum Likelihood Methods
- Generalized Method of Moments and Minimum Distance Estimation
- Discrete Response Models
- Corner Solution Outcomes and Censored Regression Models
- Sample Selection, Attrition, and Stratified Sampling
- Estimating Average Treatment Effects
- Count Data and Related Models
- Duration Analysis

If you are frustrated with the presentation in Greene and other econometric textbooks, then you may want to take a look at this book. It's the anti-Greene, full of clear, well-motivated presentations. What I love about this book is that the author clearly intended for it to be used by students -- even though you can find all the formal results that you would expect from an advanced econometrics text, there is so much intuition that can't be found anywhere else (at least in one book). If you plan on doing applied research, get this book!

Wooldridge fills an enormous gap in the market between applied manuals (which do not deliver enough theory for an academically serious audience), and theory books (that leave you scratching your head as to how one is supposed to implement their results). This is a wonderful book for economic practitioners who want enough theoretical knowledge to support their applied work, but who don't want the theory obscuring the practical issues. An econometric theoretician might find another work more satisfying.

This book does an excellent job in covering cross-sectional and microeconometric models (Stata codes for all examples in the book are available from UCLA's webside). However, the exposure to panel data is limited. For a theoretical overview of panel data econometrics, I'd recommend Hsiao and Baltagi. For applied work, Edward Frees wrote a good book, although his book is more from the social science perspective (SAS and Stata codes are provided).

It's difficult to decide how to rate this text because it receives very favorable reviews from most readers, but it was not the book I expected, nor is it a book that I find particularly useful. I have long been a fan or Wooldridge's basic econometrics text because of its accessibility and coverage of a broad range of useful topics. I had expected to have a similar experience with the Econometric Analysis of Cross Sectional and Panel Data: a lot of good material covered in accessible fashion making technically difficult topics seem almost easy. Sadly, I found this book much more difficult than Woolridge's basic text, making it much less valuable as a self-instructional tool.

In good part, no doubt, the higher level of difficulty reflects the increased complexity of the material. But it also seems clear that Woodridge, perhaps rightly, is addressing a very different audience than the one he expected to use his introductory text. In any case, his earlier efforts to make hard-to-understand material accessible to readers who do not work with econometrics every day seem to be absent from this book. A reasonable admonition, therefore, might be that just because a reader found Wooldrige's introductory econometrics text useful is absolutely no guarantee that he or she will find this more advanced effort similarly readable. Certainly, anyone not thoroughly comforatable with matrix algebra will struggle mightily and still miss a great deal of value.

All tolled, the book may very well merit more than a three star rating. My experience with it, however, suggests otherwise. If only there were an unambiguous way, short of actually procuring and trying to read a text, to make its level of difficulty and required mathematical and statistical background a good deal clearer. The need for this sort of information has become a good deal stronger now that econometric methods are being used in a much more varied range of disciplines.

As an alternative to Woolridge's text, anyone interested in learning to analyze panel data might give Singer and Willett's (2003) Applied Longitudinal Data Analysis a try. Happily, Singer and Willett cover a broad range of pertinent material, and they never lose sight of the fact that their audience contains a lot of folks with limited training in quantitative methods. Their material is framed in terms of multilevel analysis, which makes it even more valuable. The only pre-requisite is a course in applied regression analysis,

Wooldridge's "Econometric Analysis of Cross Section and Panel Data" is a gem. As a student of applied microeconomics, I found reading his undergraduate book "Introductory Econometrics" is a great complement of reading this graduate text -- especially part I of the UG book (especially the parts of interaction terms, dummy variables); part III (panel data) might be too chatty. I think both of his books worth every penny that I paid. Although, I think Hayashi text is great, maybe it is geared toward time series econometricians. For example, Wooldridge's example of omitted variable bias is clearer to me (using wage-education regression from labor economics and compute the asymptotic inconsistency of OLS estimator with omitted variable bias), compared to Hayashi's simultaneity bias. Both are nice illustration of endogenity, but I prefer Woolridge's text.

I agree that this textbook actually provides a truly thorough overview on microeconometric methods and stduies. Furthermore, it is immensely instructive in discussing the role of assumptions, on ways of setting up a good empirical investigation that actually answers the research question etc. Plenty (especially applied) problem sets will definitively further your skills, too.

However, all this comes at a cost. I personally like fully grasping the methods I use. That includes, for instance, at least having had a look at some of the details of how the particular estimator is derived. Wooldridge is rather silent on that, I was a lot more comfortable with Hsiao's monograph when dealing with panel data topics.

This book is a reasonably competent reference on panel data econometric methods. Its somewhat more clear and helpful than most metrics books, but not by a whole lot. It suffers from the same problem that plagues all econometrics texts. Like all metrics textbooks, it's long and overly detailed where it should be concise, and it's terse where it should be detailed. It also lacks a coherent, step by step structure. But that's essentially par for the course.

Statistics texts are almost universally either ultra-simple, hand waving outlines with no good mathematical insight, or they're a sprawling, useless mess of asymptotic derivations with no practical explanation, applicable intuition, or computational guidance. I'm hoping that one day, someone will write a really good econometrics textbook. This is a small step in the right direction, but it isn't the one we've all been waiting for.

I have bought this book to have a comprehensive intermediate-advanced text on panel data and other approaches. The book is good, but no tables, no datasets, no figures!
Compared to other books on the topic, it is a very understandable book, but only COMPARED to those books. You do not have to be a Mathematician to understand it, but you must be at least intermediate in econometrics (after Gujarati or an intermediate course).