Paul Goldsmith-Pinkham
@paulgp
Yale SOM professor & Bulls fan. I study consumer finance, and econometrics is a big part of my research identity. He/him/his
jamesbrandecon.github.io/blog/posts_h... So after having GPT-5.4 agents review 4,800 proofs and GPT-5.5 give a second review to a random subset of them, what did we find? To my absolute surprise, if anything the rate of errors in proofs has increased(!) in 2025 relative to previous years.
If you look hard, you can see exactly when my children are born in my MBA evaluations...
In a new post, I pull data from four different sources --- arXiv, NBER working papers, SSRN, and the AEA --- to look at how much working paper growth and submissions have changed with the growth of AI. paulgp.substack.com/p/the-number...
If you want to incorporate full costs, including board, we can subset to those schools that report the overall COA
2023 data for annual STICKER price at U.S. Colleges. Upper-bound on cost paid by students (because it does not consider aid and grants)
Well this is a bit insane but in case anyone wants to see us live:
Python implementation of `rdrobust` that is significantly faster! github.com/leostimpfle/... These runtimes are in logscale!!
Hi @ulrichatz.org , yeah the dashboard figures are different than some of the ones in the paper (that needs to be rectified). The little ? next to the word denotes how it's derived -- in this case it's a crazy long string:
Papers now also indicate my analysis of the data accessibility in their data repos. Shows up on the paper listings, and on the exported paper CSV. So, easy exercise is to search for "regression discontinuity" and filter the CSV by "full data" to get a bunch of examples.
Redid my leniency design IV slides based on the paper for JEP that Peter Hull, Michal Kolesar and I wrote: github.com/paulgp/appli... arxiv.org/abs/2511.03572
Version 2, with more explanations and clean-up. paulgp.com/replication-...
First pass at a website that keeps track of the data availability status of various published papers: paulgp.com/replication-... Very much in Alpha mode, so data will be updated soon.
Come for my thoughts on AI writing, stay for Dostoevsky caricature as an econometrician: