Argument versus Super-Duper Regression

Randall Morck and Bernie Yeung recently published a thoughtful essay entitled “Economics, History, and Causation.” They argue that in the quest for causal identification, economists have come to rely too much on econometric techniques such as instrumental variables regression. (A friend once called these techniques “super-duper regression.” There are a couple of biting jokes embedded in the term “super-duper.”) Morck and Leung argue that econometrics, while obviously valuable, should be complemented with careful historical argumentation. That is, when nature does not offer the analyst a research design that can be given some kind of treatment effect interpretation, it is often still possible through judicious use of historical data and a close attention to historical context to produce quantitative evidence that has a causal interpretation.

Another way to put this is, learning how to construct a good argument is just as valuable as learning how to execute a super-duper regression. Both are indispensable parts of our methodological toolbox.

This really is one of those essays that I wish I had thought to write myself, and it’s well worth a read for anyone interested in how political science can remain creative and thoughtful when confronting the credibility revolution.