Econometric methods for labour economics, by Stephen Bazen, New York, Oxford University Press, 2011, xi + 148 pp., £ 35.00 (hardback), ISBN 978-0-19-957679-1
This book is intended primarily for labour economists and aims to provide practical guidance on how to use those econometric or statistical methods and how to understand and use the results that are produced. It is interesting and well presented. It may be a useful reference for postgraduates and undergraduates, an aid for researchers conducting empirical research, and a guide for more experienced economists wishing to apply new tools in their research for the first time.
The book consists of six main chapters, with an introduction, a conclusion, a bibliography and an index. Chapter 1 introduces the use of linear regression in labour economics. It reviews estimation, hypothesis testing and other basic results, and covers specification issues in linear regression. It highlights the Mincer earnings equation and real-world data analysis, with dummy variables, quartic specification and instrumental variables and two-stage least squares. Chapter 2 addresses further regression issues. It covers decomposing of the differences between groups using the Oaxaca decomposition technique, quantile regression and earnings decomposition, regression with panel data and estimating standard errors. Chapter 3 explores dummy and ordinal-dependent variables. It starts with the linear regression model with least-squares estimation and progresses to logit and probit models with maximum likelihood estimation, the likelihood ratio test and McFadden's pseudo-R 2. It also covers how to interpret the output from logit and probit models, and (briefly) multinomial logit and ordered probit models. Chapter 4 discusses selectivity issues. It covers truncation bias and a pile-up of zeros. It also covers the sample selection bias of missing values and mentions the Roy Model, marginal effects and Oaxaca decompositions. Chapter 5 illustrates duration models. The chapter starts with the analysis of completed durations before moving on to explain censoring; complete and incomplete durations. It then covers modelling issues, including parametric specification and discrete data. Chapter 6 evaluates policy measures. It explains the experimental approach, the quasi-experimental approach and the role of selectivity in evaluating policies in a non-experimental context. Every chapter has a section dedicated to concluding remarks and suggests other material for further reading. Any technical support is placed in the appendices.
In summary, this book is easy to read and accessible to those with basic knowledge of econometrics, but is not a guide to cutting-edge research, nor is it an applied econometrics textbook. From an applied statistician's perspective, the book is very useful to those practitioners, researchers and students who appreciate the uses of regression models, real-world data analysis and duration analysis applied to labour economics and other areas, and may even adapt such methods in their own research and study.