A Guide to Econometrics - 4th Edition

This book doesnt pretend to be a text. It answers the questions a text usually doesnt: like how does it all fit together? and what do we do when our assumptions dont hold? The author points out that econometrics texts are usually "cookbooks", and aim to tell you how to use their collection of receipes. This book explains the assumptions behind regression, and what happens when we violate them, and what we can do about it. It also gives a good overview of the topic. This book shows the connection between the receipes in a text. It doesnt pretend to teach econometrics, so the reader isnt innundated with equations. Instead, you get a discussion of why you would want to do (or not do) the things your text would teach you. Anyone who is taking an econometrics course (graduate or undergraduate) should get this book. After reading a chapter in your text, read the corresponding chapter in Kennedy to add some depth to your understanding.

I would also suggest this to graduate students who are facing preliminary or comprehensive exams in econometrics. This will help you to "bring it all together", and answer those vague, conceptual questions which seem to cause the most grief. The book has sufficient references to the literature so that you can easily follow up anything you want to explore in more depth, but its clear and self-contained enough that only an econometrician is likely to feel the need. Each chapter is organized in three parts: a discussion of the topic, an appendix of long footnotes which add details which would obstruct the flow of the discussion, and finally, a technical appendix, where you can find a few of those equations if you really want them. This makes it easy to read the book at the level you need.

For a book of this size, he covers a lot of territory. He covers the CLR model and hypothesis testing well, and discusses a few other things too. This guide is hardly encyclopedic. However, it covers the things economists need to know most.

Kennedy does more than just explain econometrics. He spells out the limits of econometric analysis. Texts often pay little attention to the con in econometrics. Not Kennedy. He discusses the limitations and defects in standard techniques, as well as their advantages.

The only thing wrong with this book is that it does not carry the reader along far enough. After reading this book, most readers will likely move on to a standard (i.e. badly written) econometrics textbook. In contrast, this book is written so well that it almost makes learning econometrics fun!