The first sentence of the first comment apparently confuses "mere" with "slight". Also, his judgement is too harsh. The book definitely deserves two stars.

The review posted by the anonymous reader from CT seems a little too harsh to me. The book offers a decent overview over many interesting areas of panel data. In particular, one and two way error components are discussed, one chapter is devoted to hypotheses testing, heteroscedasticity and serial correlation and dynamic panel data models respectively. However it is true, that the authour does not provide much help to understand the literature which is just reproduced (often word by word from the papers!). The book is not recommended to someone who wants to read about Panel Data but doesn't have a lot of knowledge about the subject yet.

To say that this book is scandalous would only be a mere exaggeration. Instead of clarifying the literature on Panel Data, the author just sums up some papers (often citing them word by word; why bother then buying a book? Just copy the articles!) without adding any value to them. Even worse, while summarizing, the essence of the papers often goes missing. The only value of the book consists of a quite complete reference list of the literature up to 1994. (However, the list is of course a little biased in favor of the author. He mentions every minuscule and unimportant publication of himself. One could almost get the impression that he contributed to Panel Data) Expecting a book that provides the main ideas and clarifies them together with a clear notation and proofs of the result, I was very disappointed with Badi Baltagi's book! Not recommended at all! (Unfortunately there is no alternative I know of!)