Mostly Harmless Econometrics: An Empiricist's Companion [Tapa blanda]

The first thing I want to say is this: If you plan on doing regression analysis in your research, stop what you are doing, and read this book first. I think this book represents THE current statement on how we should use regression. For Angrist and Pischke, regression is a technology for summarizing data. If regression is to be used for causal inference, then there is nothing in the specification of the model or the choice of estimator that can ultimately make the causal story persuasive. That is, you don't identify causal effects simply by including "control" variables in your regression. The identification comes from elsewhere---either a real or "quasi" experiment---and the regression is what you use to clean up the imperfections of the experiment and measure effects. Angrist and Pischke have done an enormous service to social science by writing a regression textbook that nonetheless emphasizes the primacy of design. This is a terrific corrective for the "101 flavors of regression" approach of textbooks to date.

Even with this emphasis on design, Angrist and Pischke show us that are a lot of nuances to the way that regressions measure such effects---e.g., in the presence of effect heterogeneity---and that's what this book explores in exquisite detail. It's a hugely important book and a very serious and rigorous treatment, despite it's apparently causal style. They make some claims that may strike some as outrageous---e.g., always using OLS, even for limited dependent variables---but the rigor of their presentation means that the onus is on those who disagree to think harder about why, exactly, they would prefer, say, a more parametric approach.

Nonetheless, it isn't a "5 star" book. It often feels a bit rough-draft-like. The presentation of technical material skips important steps rather haphazardly. I wonder if this was due to bad editing? Hopefully there will be a second edition that cleans up these rough edges, in which case it would be the ideal textbook on regression analysis.

This is a wonderful book. Despite having taken many courses and read many statistics and econometrics books, I'm sometimes stuck in my own applied problems. Of course, these courses and books taught me how to derive asymptotic theorems or be careful when maximum likelihood fails, but this is not what I really need to know to solve my problems. Now with Angrist and Pischke, I have a book that is truly applied in focus -- one that explains why and how certain empirical strategies are convincing and one that is up-to-date with the latest examples of recent research employing these strategies. Real econometrics applied to real problems. This book should be on every applied economists bookshelf.

In their introduction Angrist and Pischke say "empirical research is most valuable when it uses data to answer specific causal questions, as if in a randomized clinical trial". The book aims to teach this approach to students of applied econometrics.

The emphasis on natural experiments and quasi-experiments which the authors espouse has become influential in some sub-areas of econometrics and the authors, particularly Angrist, have played a leading role in this development. However, this approach is not uncontroversial. The Journal of Economic Perspectives has an entire issue (Spring 2010, Vol. 24, No. 2, full text online for public) discussing the pros and cons and you may want to glance through it before buying this book.

Taken on its own terms, the book intermingles four levels of discussion: the philosophical and methodological issues around causality, tips and tricks on how the apply the workhorse models of (micro)econometrics, case studies, and the mathematical properties of models and estimators. This intermingling may be useful for a practitioner trying to see the big picture, but it makes things hard for a beginning student. The problem is compounded by the sketchiness of the mathematical derivations.

If you are starting out in econometrics you may be better served by traditional textbooks with more detailed and linear presentations such as Wooldridge's Introductory Econometrics and Econometric Analysis of Cross Section and Panel Data.

If your interest is primarily in causal inference, book-length treatments that focus only on that aspect can be found in Judea Pearl's Causality: Models, Reasoning and Inference or Morgan and Winship's Counterfactuals and Causal Inference.