Obtaining Externally Valid Heterogeneous Treatment Effects
16 Pages Posted: 30 Sep 2020
Date Written: August 28, 2020
In this research note I will explain a method to obtain heterogeneous treatment effects that are externally valid across the dimension of time. While it is interesting to see how treatment effects vary across the distribution of observed heterogeneity in the study period, the observed heterogeneity in the study period may not be entirely representative of this dimension over time. An example of a heterogeneity dimension where interannual variation is important is rainfall. I employ an example based around rainfall as the guiding case. I will show that one can employ a historical administrative data record to recover estimates of heterogeneous treatment effects in the average year and a distribution of treatment effects. I have provided code for this algorithm to be performed in Stata.
Keywords: research methods, heterogeneity, treatment effects, rainfall
Suggested Citation: Suggested Citation