Practical Advice on Applied Econometrics

The Ten Commandments of Applied Econometrics

From using common sense to not confusing statistical significance with substance: Peter Kennedy’s dissatisfaction with the way empirical researchers think about, apply, and interpret econometric methods led him to writing the ten commandments of applied econometrics. The paper is available here.

Estimated regression model has the “wrong” sign

Peter Kennedy (again!) provides a long list of ways in which a “wrong” sign in empirical work can occur and how it might be corrected. The papers discusses issues varying from faulty economic theory to data problems to classic econometric problems (e.g., omitted variables, non-stationary). The paper is available here.

Good and Bad Control Variables

Students and researchers often find themselves needing to decide whether or not the addition of a variable to a regression equation helps controlling for potential confounding factors to get estimates closer to the parameter of interest. This paper (available here) discusses the problem of bad controls.

On the (Mis)Use of the Fixed Effects Estimator

Adding more rounds of data is not necessarily better if the goal is to identify a causal relationship because the amount of stuff that remains constant over time decreases as the data grows to cover a longer time period. Daniel Millimet and Marc Bellamare propose alternatives to the fixed effects estimator. The paper is available here.

What are we weighting for?

When estimating descriptive statistics, weighting might be needed to make the analysis sample representative of the population. When estimating causal effects, weighting has different motives. Gary Solon, Steven Haider and Jeffrey Wooldridge discuss how motive sometimes does not apply in situations where practitioners often assume it does. The paper is available here.