Code for My Papers
| Paper | Description | Language |
|---|---|---|
| Quasi-Experimental Shift-Share Designs | Replication archive for the application to Autor et al. (2013) and ssaggregate command that helps implement shift-share (Bartik) designs (link) (updated Aug 2018) | STATA |
| Revisiting Event Study Designs | Implements estimators, graphical, and statistical tests developed in the paper (by request) | STATA |
Other useful Commands
Below are some STATA commands that I wrote and found useful in data work. Hope you will find them useful, too. More detailed usage instructions are provided in each file.
| Module | Description |
|---|---|
| reclink4 (by request) | Updates the reclink2 package by Wasi and Flaaen for exact and fuzzy merging of firm names and addresses. Substantially improves performance by minimizing loops. |
| MostFreq.ado |
Reports the list of most frequent values of a set of variables. For example, in a dataset of counties, it can print 20 largest states (by population), their population shares in the country, and the average income per capita in all counties of each state. |
| binscatter2.ado |
Adds more functionality to the binscatter command: variable labels, robust standard errors, automatic footnotes, etc. |
| Demean.ado |
Demeans a list of variables (optionally within cells), with replace or generating new variables. |
| recover.ado | Fills in empty variables after running fillin or tsfill. |
| share.ado |
Reports the share of observations (or weights) which satisfy some condition, within a subsample defined by another condition. |