Mostly Harmless Econometrics

Mostly Harmless Econometrics. It is a highly influential textbook in applied econometrics authored by Joshua Angrist and Jörn-Steffen Pischke, published by Princeton University Press in 2009. The book advocates for a focus on credible causal inference using research designs that mimic randomized controlled trials, positioning itself as a practical guide for empirical researchers over traditional theoretical econometrics. Its accessible style and emphasis on real-world application made it a standard reference in economics graduate programs and across social sciences like political science and sociology.

Overview and Purpose

The book emerged during the credibility revolution in applied economics, a movement emphasizing robust identification strategies. Joshua Angrist and Jörn-Steffen Pischke wrote it to bridge the gap between advanced econometric theory and the practical needs of researchers conducting empirical work in fields like labor economics and public finance. Its purpose is to demystify core methods for estimating causal effects, arguing that with clever research design, sophisticated techniques can be rendered "mostly harmless." The text is deliberately concise, avoiding the extensive matrix algebra found in classics like William Greene's *Econometric Analysis*, and instead focuses on intuition and application, drawing heavily on influential studies from the National Bureau of Economic Research.

Key Methodological Approaches

A central tenet is the prioritization of research design over complex statistical correction. The authors champion the use of natural experiments and quasi-experimental variation to approximate the conditions of a randomized controlled trial. Key approaches include the instrumental variable method, where an external factor provides random variation, and regression discontinuity design, which exploits arbitrary cutoffs in program eligibility. They also emphasize the use of difference-in-differences estimators to control for unobserved confounders by comparing changes over time between treatment and control groups, a method with roots in work by Orley Ashenfelter and David Card.

Core Econometric Techniques

The book provides detailed guidance on implementing specific techniques. It covers the mechanics of two-stage least squares estimation for instrumental variables and discusses the properties of limited dependent variable models. A significant portion is devoted to panel data methods, including fixed effects and random effects models, which control for unobserved heterogeneity. It also treats matching estimators and propensity score methods, though often with a critical eye toward their assumptions. The technical discussion is consistently grounded in the potential outcomes framework, associated with Donald Rubin and Guido Imbens, for defining causal effects.

Applications and Empirical Examples

The text is renowned for illustrating methods with landmark empirical studies. For instance, it discusses Joshua Angrist and Alan Krueger's use of quarter-of-birth as an instrument to estimate the return to education, and David Card's analysis of the Mariel boatlift on the Miami labor market. Examples from health economics, such as the RAND Health Insurance Experiment, and from development economics, like evaluating the PROGRESA program in Mexico, are used to show how research designs can be applied. These cases, often drawn from the American Economic Review or the Quarterly Journal of Economics, serve as practical blueprints for students and researchers.

Influence and Reception

*Mostly Harmless Econometrics* has had a profound impact on empirical practice in the social sciences. It solidified the pedagogical shift toward causal inference in graduate programs at institutions like the Massachusetts Institute of Technology and the London School of Economics. The book received praise in publications like the Journal of Economic Literature for its clarity and practicality. However, it also sparked debate, with some econometricians like Edward Leamer and James Heckman critiquing its perceived narrow focus on internal validity at the expense of structural economic modeling. Despite this, it remains a seminal work, influencing a generation of researchers in political science, sociology, and public policy.

Category:Econometrics textbooks Category:Princeton University Press books Category:2009 non-fiction books