##
**Elements of econometrics.
Repr. of the 1986 2nd ed.**
*(English)*
Zbl 0935.62129

Ann Arbor, MI: University of Michigan Press. xiv, 786 p. (1997).

This book belongs to one of the standard textbooks in econometrics. In 1997 the book received its second edition. Following the author, the new edition is written in the spirit of “traditional econometrics” as it is taught in the classroom. In this sense several recent developments such as unit roots and cointegration are not considered as new issues but rather as special topics of traditional econometrics. Also the developments in financial econometrics are not considered explicitly. Here the argument is that many of the proposed procedures are based on statistical techniques which are not well founded in economic theory. Examples are trend removal procedures instead of trying to explain the trending behavior from an economic point of view. Furthermore, the author is not convinced that unit root techniques as well as financial econometrics are good substitutes for simultaneous equations models which in practical applications have lost in importance in the last decade. For these reasons, Kmenta still dedicates a complete chapter to estimation and testing in simultaneous equations systems. Another research area with growing importance since the seventies is related to the “longitudinal analysis of labor market data”. The omission of this important topic, as argued by the author, is due to the special character of this research field which is too specialized to be included into in a general textbook.

Compared with the first edition from 1971 (see the review Zbl 0272.62001), the author added all important new results in the established areas of econometrics. Here the most important changes are related to the chapters dealing with autocorrelated and/or heteroscedastic error terms. New results have also been added to the problem of multicollinearity, to specification errors, to distributed lag models as well as to restricted coefficients and nonlinear models. Finally, the chapter on pooling cross-section and time-series data has been considerably extended. Whenever possible, existing old material has been expanded and new material has been introduced.

To summarize, the book presents a concise and stringent survey of recent developments in traditional econometrics putting emphasis on the close relationships between economics and econometric techniques. In this sense the textbook should be highly recommended to beginners in applied economics or econometrics. But for those readers who will update their knowledge about most recent trends in time-series econometrics the content of the book is somewhat disappointing.

Compared with the first edition from 1971 (see the review Zbl 0272.62001), the author added all important new results in the established areas of econometrics. Here the most important changes are related to the chapters dealing with autocorrelated and/or heteroscedastic error terms. New results have also been added to the problem of multicollinearity, to specification errors, to distributed lag models as well as to restricted coefficients and nonlinear models. Finally, the chapter on pooling cross-section and time-series data has been considerably extended. Whenever possible, existing old material has been expanded and new material has been introduced.

To summarize, the book presents a concise and stringent survey of recent developments in traditional econometrics putting emphasis on the close relationships between economics and econometric techniques. In this sense the textbook should be highly recommended to beginners in applied economics or econometrics. But for those readers who will update their knowledge about most recent trends in time-series econometrics the content of the book is somewhat disappointing.

Reviewer: Herbert S.Buscher (Mannheim)

### MSC:

62P20 | Applications of statistics to economics |

62-01 | Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics |

91B84 | Economic time series analysis |