Mini models make a splash

This book is a welcome addition to the econometrician's library. The importance of econometrics in providing direction to economics has increased markedly over the past two decades. In part this has arisen from the extension of econometrics into areas away from its traditional applications in macroeconomics. Moreover, the increased availability of cheap computing power and microeconomic data sets has meant that a new range of techniques has developed that involves unfamiliar methodological issues and new modelling frameworks, techniques and problems.

Microeconometrics fills these needs in an admirable way: it is well organised and well written; it enables a reader who has progressed to William Greene's excellent intermediate-level textbook in econometrics, Econometric Analysis , to make real progress in understanding how and, more important, why, microeconometrics has developed into a recognisable area.

The text provides applications that serve to illustrate the econometric techniques, to motivate the reader in the flexibility of recent developments in econometrics and to stimulate new applications.

Naturally, some of the material is technical and requires application, but the exposition and coverage is second to none throughout its 1,000 or so pages. The book comprises chapters plus two appendices, so I will not attempt to list these. Rather, the organisation of these chapters is formed into six sections: preliminaries; core methods; simulation-based methods; models for cross-section data; models for panel data; and further topics.

There is a useful book outline provided by the authors, A. Colin Cameron and Pravin K. Trivedi, together with a suggested 20-lecture course.

The querulous might ask: but what is microeconometrics? In part it is defined by the data that are used and the problems that arise as a result.

The range of economic problems that can be tackled by microeconometrics seems almost unlimited, but the key is the individual rather than aggregate nature of the questions addressed and the data that are used.

Although not exclusively the case, micro-data often derive from surveys with large numbers of responses and involve problems such as discrete (rather than continuous) outcomes, truncated or censored outcomes, non-response, missing data and biased sampling.

To cite just one of the authors' examples, in chapter 24, data on ,700 individuals from the World Bank's Vietnam Living Standards Survey are used to estimate the income elasticity of the demand for healthcare.

The authors are to be congratulated on this sure-footed addition to the econometrics literature.

Kerry Patterson is professor of econometrics, Reading University.