🗣️ Published today in @aeajournals.bsky.social. Ethan Lewis and I use a visa lottery, as a nationwide randomized experiment, to test the effects of foreign labor on US firms and workers. doi.org/10.1257/app.... A concise summary from @nber.org —> www.nber.org/digest/20221...

🚀 Announcing T v0.53.0 "L'Initiation"! 🚀 This release takes Nix-native data science pipelines to the next level: 📊 Dynamic Branching: Map, cross, slice, head/tail, and sample pipeline nodes automatically 🖥️ Atelier TUI IDE: Tmux-based interactive environment with live REPL variable & plot sync

I continue to be impressed by the Rousseeuw Prize for Statistics, focusing on important innovations, rather than always trying to tease out individuals. Here it is given to the R project. Particular shout out to Brian Ripley whose contribution to the statistical commons was massive.

Been waiting for Census to OK our disclosure of new results on a project with the Longitudinal Business Database. RDC administrator just informed us, no disclosure because of new rule. This is gonna mess up a lot of ongoing research & especially job-market candidates. #EconSky

Are you on this @aaronsojourner.org @justinwolfers.bsky.social @mclem.org @dynarski.bsky.social ?

Gribouille 0.3.0: 'guides(x: none)' hides axis ticks, 'guides(default: none)' hides all legends at once. 'compose()' gains 'theme:', 'geom-area()' stacks by default, 'annotate()' gets 'clip: false'. See the release announcement: mickael.canouil.fr/posts/2026-0... #Typst #DataViz #OpenSource

This whole article is a banger:

'In the words of a wag whose name escapes me: “Econometric theory is an industry with a very large output but no sales.”' www.tandfonline.com/doi/full/10....

two property wrongs don't make a property right

university admin when faced with any problem

💔 Now, being cared for by full-time assistants, confined to her apartment, unable to feed or bathe herself, robbed of her charismatic personality by her struggles to speak, I asked what life is like for her. “It’s not so scary,” Nancy replied. www.nytimes.com/2026/06/11/s...

Oh NOW they make group chats

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.

Found a cool result I checked my code Now I have correct code

Today's cool young researcher #econtwitter #econsky is @fabalbav @WakeForest#econtwitter#econskyho works on topics related to education and health, particularly in Latin America

Lots of people with strong opinions about use of AI detectors to filter what papers we review (or what stories we consider for awards). I wrote up a toy model of using AI detection to infer author type to get at some implict assumptions we make when we argue that AI detection is or is not useful

Great thought exercise for how to think about detection

Gelman absolutely on fire on the blog statmodeling.substack.com/p/13-aspects...

Examining weight stigma shows that the "obese" cutoff increases patient mortality. Analyses of diagnostic effort and clinical documentation point to stigma as a mechanism, from Manasvini Singh www.nber.org/papers/w35277

Still riding the high from this

I mean this was just yesterday. There’s really a pattern where actual data analysis does not support the impression of radical disruption you would get from journalism / social media.

If you look hard, you can see exactly when my children are born in my MBA evaluations...

I haven't dug into the details of this but this is really satisfying to my priors re: AI & activity across many domains. As individual acceleration and experimentation hits the relational, vetting and hand-off layer, initial spikes hit like waves smacking into a seawall.

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...

Lumpy infrequent price adjustment

Thrilled to see this paper in print. What started as a Medicaid managed care natural experiment produced some of the best evidence we have on coverage restriction effects. www.aeaweb.org/articles?id=...

The Sloan Foundation is searching for a new Director of its Economics Program. At a critical time when federal research funding is limited, this is a critical position. Please circulate this opportunity to shape and support economics and economists widely! sloan.org/about/careers

I watched John Wick and I discovered that labour is capital’s dead wife. open.substack.com/pub/rottenan...

New working paper out today for the real UI heads, from me and David Wasser: ⬆️ UI generosity - ⬆️ wages - for UE, J2J, and EE workers - especially at the bottom - ⬇️ hiring from U - ⬇️ LF exit - maybe ⬇️ job postings - ⬆️ posted wages - ⬆️ new entrant hiring - and more!

Can someone who (1) has a PhD; (2) is an economist; (3) understands the competing risks identification problem; and (4) is an expert in causal inference please write a paper on how badly-designed most studies are of the impact of AI on task completion? I only meet two of those criteria.

Culmination of my weekend experimenting with local models: Haven't hooked it up to an image model to see if it works yet, but here's the PoC of what Qwen 3.5 9B put together in ~2 hours today. Will probably rename the repo since I moved toward a computer vision approach. github.com/ajnafa/TabFe...

New working paper with Amol Singh Raswan and Chris Udry: "The Sisyphean Pursuit of Evidence for Poverty Traps." A central idea in development economics is that poverty can trap people. We went looking for the cleanest evidence. Here's what we found – and didn't.

More Markus Academy things: paulgp.substack.com/p/skills-spe...

This song worth a listen or five www.youtube.com/watch?v=wz3p...

NOT TWITTER IS INSANE erickrouss.github.io/not-twitter/...

You write an epitaph that's completely biblical -- I insist the causal graph is purely cyclical better

Wrote a great qualifying exam question that I'm really excited about? Being a professor is weird

This is lovely An Angrist family tree

I will take this time to beat my competing risk identification problem drum and note that the highlighted passage from the article below by @paulgp.com highlights an issue I have with AI benchmarks: They implicitly assume that tasks in the real world get completed instead of abandoned or repurposed.

michael kremer for nobel prize twice! please!! (another banger from paul)

Really enjoyed contributing to this write-up on @causalinf.bsky.social Substack — an NBER panel on AI in research production w/ Kosali Simon, David Bradford, Coady Wing + commentary from me, Jason Fletcher & @paulgp.com open.substack.com/pub/causalin...

Good thread. Also worth reading is this post by @paulgp.com paulgp.substack.com/p/the-median...

It's official. Our FIRST academic publication is officially live @aeajournals.bsky.social! ✨ TLDR: In an online experiment, @emmarackstraw.bsky.social and I find evidence that evaluators assigned to Black supervisors perceive Black workers as more productive on average. #econsky #blacksky

wow — the preprint host, arxiv, is banning authors for a year if they submit papers with hallucinated citations 🤖

In a worker-firm effects (AKM) model, the worker+firm effects are only separately identified within mobility groups - the subgraphs of workers+firms connected by workers moving across firms. Something I just learned: the mobility groups for OLS vs. Poisson estimators are not necessarily the same.

Japan is one of the most R&D intensive non-US countries, and RIKEN is its most prestigious (non-university) research institute. If you isekai’d it into the US, it only has the fourth biggest R&D budget in the state of North Carolina, which is like the twelfth biggest state bsky.app/profile/awma...

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)