The Analysis of Household Surveys (Paperback)

A Microeconometric Approach to Development Policy

By Angus Deaton

World Bank Publications, 9780801852541, 480pp.

Publication Date: August 1, 1997



Over the past 15 years, the availability of cheap and convenient microcomputers has changed the collection methods and analysis of household survey data in developing countries, making the data available within months, rather than years. Simultaneously, analysts have become more interested in exploring ways in which such data may be used to inform and improve the steps involved in policymaking. This book reviews the analysis of household survey data, including the construction of household surveys, the econometric tools that are the most useful for such analysis, and a range of problems in development policy for which the econometric analysis of household surveys is useful and informative. The author's approach remains close to the data, relying on transparent econometric and graphical techniques to present the data so that policy and academic debates are clearly informed. The author illustrates the analysis through examples of policy issues from several countries, including Cote d'Ivoire, India, Pakistan, Taiwan (China), and Thailand. Two introductory chapters discuss the content and construction of surveys, as well as the econometric methods that can be used with survey data. Survey design and sampling are covered in some detail, as are the effect of survey design on the calculation of statistics and the estimation of parameters. A chapter on econometrics explores issues such as heteroskedasticity, sample selection, and instrumental variables. Four chapters focus on poverty and inequality, demand analysis and intrahousehold allocation, price and tax reform, and savings. Each chapter presents a self-contained development of the topic, introduces the important empirical issues, and provides substantive results. The book also includes the computer code used to calculate the results so that readers can adapt the methods to yield similar results for other data.