Angus Deaton

Orley Ashenfelter

Gary Solon

This Working Paper reviews a number of aspects of the collection and use of panel data from households in developing countries. Sampling issues are discussed in Section 1. The authors conclude that there are likely to be real, if modest, benefits from incorporating some panel element into household survey data collection in developing countries. The recognition that panel data are likely to be subject to substantial errors of measurement does not invalidate this conclusion. Section 2 discusses the measurement of income dynamics, an issue that cannot be addressed without panel data. Recent research using U.S. data is reviewed to show that comparable work for developing countries would add an important dimension to discussions of poverty, inequality, and development. It is in the third area of review, that of econometric analysis, that the real benefits of panel data appear most fragile. While it is true that panel data offer the unique ability to deal with the contamination of econometric relationships by unobservable fixed effects, the presence of measurement error can compromise the quality of the estimates to the point where it is unclear whether cross-section or panel estimators are superior.

Gary Solon

Books

Your next favorite book is out there