Nonparametric Econometrics

Rating:
Summary: Great Book on Non-Parametrics
Review: I just started reading it, and I love the clear exposition of the book. Its a very fast-growing field, so don't expect this book to be the last word on the subject. Still, it's a must for an advanced graduate student in econometrics in need of a good introduction to non-parametric estimation.

Rating:
Summary: The best introduction to the field
Review: I think this is the best introduction to nonparametric and semiparametric estimation to date. It covers an impressive amount of material, and the focus is on density and regression estimation. The exposition is clear, a lot of crucial results are proved, and an immense quantity of others are sketched or at least mentioned. Most of the book is fairly or very advanced, but all topics are introduced in a neat, simple, and intuitive way, so even a beginner can benefit from several parts of this book (previous good knowledge of math & statistics are still necessary, of course). Densities, regressions, discrete dependent variable models, simultaneous equation models, selection models, it's all in here. The field is expanding, but this book really has almost all you need to know about what the field has done until the publication date.

There is also a nice and useful appendix for many of the asymptotic results used in the book. The only drawback (besides a few typos, but not so many to be annoying) is the scant presence of empirical applications, but this book is not supposed to be a guide for applied econometricians (at least, not mainly), so I don't think it's a serious shortcoming. If you are more interested in the applied side of np regression, but you still want a rigorous treatment, you may look at Yatchew's "Semiparametric regression for the applied econometrician", in the same Cambridge series. For an even simpler, shorter, and low-tech introduction to np esimation of densities and regression, I would suggest instead the last pages of Ch. 3 in the splendid "The analysis of household surveys", by Angus Deaton.

This book is really worth its price (which, by the way, is kept at a very decent level by the worthy Cambridge University Press. I wish Wiley or Chapman and Hall stopped with their policy of immoral prices...). Highly recommended.

Rating:
Summary: A comprehensive review of nonparametrics statistics
Review: Nonparametrics seems to be one of the most promising fields in econometrics. All econometricians should be aware of that and try to learn the basic tools. This book is a great beginning (perhaps you should read the chapter of nonparametrics in Johnston and Dinardo's "Econometric Methods" to get used to the very basic concepts). The manual contains practically all the stuff that has been done in the field. It begins pretty fast with the kernel estimation method and, by page 19, you will be face to nonparametric derivatives estimation equations. In the introduction there is a clear explanation of the difference between parametrics and nonparametrics; you will also learn the main basic methods and concepts, such as the nearest Neighborhood Estimator and the window's size problem. After that, you'll have to read about the statistical properties (finite sample and asymptotics) of the estimators. There is also a lot of stuff of semiparametric methods. You shouldn't expect an extremely easy-to-read manual, because nonparametrics is a pretty complex subject. The first 50 pages are easy and fun to read. You'll get excited by learning such interesting theory. But then, the hard topics begin and if you want to understand them all, you'll have to make a big effort. Not overwhelmingly complicated, neither elementary, this book is an excellent reference in the field, but I advice you to have two or three more books of the same subject (Hardle, for example) so you can understand faster some of the developments presented. A fairly good mathematical and probability knowledge is required.

Rating:
Summary: Up to date
Review: This is the most accessible and the most comprehensive text on nonparametric econometric methods I have seen. The field is highly technical, and there has been a need for the book that would combine ease-of-use with the scope. Moreover, the book is up to date and covers all econometric methods, instead of focusing on a specific branch. Recommended.