Mostly Harmless Econometrics: An Empiricist's Companion

Comment from the Stata technical group

All graduate students and researchers should read Mostly Harmless Econometrics: An Empiricist’s Companion, by Joshua D. Angrist and Jörn-Steffen Pischke. This instructive and irreverent romp through microeconometrics is as much of a page turner as we are likely to see in a book about statistical methods.

Angrist and Pischke provide an excellent introduction to microeconometrics from the potential-outcomes approach. They advocate using the potential-outcomes approach as the method to identify and interpret causal effects, which can be estimated by the methods discussed in the book. The authors provide many examples of how this approach has been applied in recent empirical work and closely link these empirical papers to the methods discussed.

The book covers ordinary least squares, generalized least squares, instrumental variables, panel-data estimators, regression-discontinuity methods, quantile regression, and standard-error estimation. Maximum-likelihood methods are not discussed in detail; in fact, Angrist and Pischke argue that simple regression methods frequently provide more robust estimators of the causal effects of interest in cases where maximum-likelihood methods are traditionally used.

This book will be required reading; methods experts and practitioners are already discussing its pros and cons. Fortunately, Angrist and Pischke have made this book as entertaining as it is informative.

Table of contents

View table of contents >>