Academic Resources

I am currently collecting some (what I think) useful resources for early-stage PhD students.

Econometric Resources

I highly recommend the empirical methods course by Paul Goldsmith-Pinkham. Check out his website or his Applied Methods PhD course github Page.

Difference-in-differences: I recommend the book project of Clément de Chaisemartin and Xavier D'Haultfœuille. Link.  The project is ongoing, it already includes binary, staggered, and heterogeneous adoption.

Also see the website of Asjad Naqvi Link which includes Stata, R, and Julia code, literature, teaching resources, and much more. 

Regression Discontinuity

Matias Cattaneo (and coauthors) provide several resources concerning regression discontinuity. Link.  I recommend the foundations as well as the extensions of the "A Practical Introduction to Regression Discontinuity Designs" articles.

The Journal of Econometrics has a special section for "How To" papers. See here. Topics include clustering standard errors, etc.

Datasets and Code

www.icpsr.umich.edu - Contains all ICPSR projects. Search for (data) publications including detailed descriptions, datasets, and code. 

www.researchdatagov.org - A platform to conveniently file FOIA requests. Groups more than 15 federal agencies together. Great overview of available data.

Artificial Intelligence Tools

I use elicit.org for initial literature searches. You provide some context or a research question and can find some related papers. Highly recommend.

ChatGPT - Provide a prompt with as much (or as precise) info as possible. I use AI for critiquing my writing and for coding. More use cases (here).

Replication files. Every Journal of Finance article should contain code and date from Jan 1st 20xx onwards. Sometimes authors only provide cleaned anonymized datasets and the code only contains regression commands. Sometimes authors include the code that was used to clean the raw data. This is a great resource and often saves considerable time. If more researchers would do this, it would also lead to more streamlining, making comparisons across samples easier, etc.