Examining the financial consequences of the 2007 California Fair Pricing Law, which places a price ceiling on hospital bills for financially vulnerable individuals, from Yaa Akosa Antwi, Marion Aouad, and @nathanblascak.bsky.social www.nber.org/papers/w35080
Twice the normal amount of good news this morning! Crossed off an academic bucket list item with my first NBER working paper (w/ Yaa Akosa Antwi and Marion Aouad) going live this morning:
www.nber.org/papers/w35080
More importantly, it was just accepted at the Journal of Health Economics!
NBER Behavioral Finance conference starts now! Program and link to livestream here: conference.nber.org/c/BFs26
Update: I'm 30+ runs in now. Battled Hades x3 times, but not close to beating him. Gameplay loop is great and writing is excellent.
Claude Code learning definitely lagging behind.
Week five of AI in Finance: Fraud detection and compliance
• How embeddings, LLMs, and graph methods can improve fraud detection
• Where AI addresses the compliance and regulatory burden
• The arms race of AI advancing cybersecurity risks and defenses
arpitrage.substack.com/p/finding-ne...
As in previous years, the @AtlantaFed is running a visiting program for PhD student. Come and join us for an exciting summer full of economics and fried chicken! Deadline soon: April 4th! rb.wd5.myworkdayjobs.com/FRS/job/Atla...
New paper! Expanding public insurance to parents benefits parents and generates spillovers to children, improving their health care utilization. The estimates highlight intergenerational effects relevant for welfare analysis. www.journals.uchicago.edu/doi/10.1086/733364 @michellemmarcus.bsky.social
Hope on pop
Accurate representation of lying on the rug with twins
Do you have a paper on financial decision-making?
Want to present research & hike?
Does Bozeman in early August sound good?
The Financial Decision-making Conference is Aug 6-7!
Submission deadline April 20.
The call for papers is here:
fdmconference.pages.dev#cfp
🧵A thread on "AI one-shot papers".
This has not been a focus on Bluesky, but on Twitter/X and LinkedIn, there has been a large discussion of the ability to "one-shot" policy evaluation, which David Yanagizawa-Drott has been pursuing in an interesting way: ape.socialcatalystlab.org
Probably should be learning more about Claude Code, but Hades is on Game Pass. Thanks @paulgp.com and @economeager.bsky.social.
My attempt at articulating why I think LLMs are a clear net positive for all research, and it's a mistake to ignore it despite private concerns:
paulgp.com/2026/03/16/r...
Please submit, health economists, and share widely!
Has anyone yet run a head-to-head comparison of the different advice that competing AI tools offer for expecting parents and for parents of newborns, infants, and toddlers?
tldr: nature of previous fraud exposure matters for take-up of new precautionary actions. More serious past experience ➡️ higher take-up. This updated version has new results examining what happens to consumer credit outcomes after freezing their credit report.
Updated working paper with my KC Fed Colleague Ying Lei Toh examine how past experiences with fraud affect take-up of new precautionary actions (e.g., credit report freezes) in consumer credit markets: www.philadelphiafed.org/-/media/FRBP...
I’m always for more data. But, be careful using the new Medicaid spending data released by the Trump administration. It’s missing a lot of important details and context.
www.kff.org/medicaid/wha...
The extremely loosely goosey nature of data governance in AI land means that use of some types of confidential data may end up being signals of true human effort. The AI assisted papers that have been demonstrated only work because data is on the open internet and already cleaned/aggregated
I'm teaching a new course at Stern on AI in Finance and opening it up!
Syllabus/slides on Github: github.com/arpitrage/ai...
Weekly summaries on Substack
arpitrage.substack.com/p/1-three-ru...
First post is on Amdahl's Law, Jevons' Paradox, and why finance was slow to learn the bitter lesson
Then again, the twins sometimes just want to eat bread and butter and nothing else 🤷♂️
We also do instant mashed potatoes where I boil water in the electric kettle.
Got it! If I have time the night before, I prep something for the slow cooker for the next day (e.g , pasta sauce, pork shoulder, pot roast). Rice in a proper rice cooker is also super helpful since you can leave the rice in there for most of the day without it going bad.
For the entire family? Or just for the kids?
We study how radiologists use AI to diagnose pulmonary embolism (PE), tracking over 100,000 scans interpreted by nearly 400 radiologists during the staggered rollout of an FDA-approved diagnostic platform. When AI flags PE, radiologists agree 84% of the time; when AI predicts no PE, they agree 97%. Disagreement evolves substantially: radiologists initially reject AI-positive PEs in 30% of cases, dropping to 12% by year two. Despite a 16% increase in scan volume, diagnostic speed remains stable while per-radiologist monthly volumes nearly double, with no change in patient mortality—suggesting AI improves workflow without compromising outcomes. We document significant heterogeneity in AI collaboration: some radiologists reject AI-flagged PEs half the time while others accept nearly always; female radiologists are 6 percentage points less likely to override AI than male radiologists. Moderate AI engagement is associated with the highest agreement, whereas both low and high engagement show more disagreement. Follow-up imaging reveals that when radiologists override AI to diagnose PE, 54% of subsequent scans show both agreeing on no PE within 30 days.
Posted a very early stage draft with rock star collaborators.
Key question: when we actually roll out AI tools, how do people use them? Do they just defer completely? Does it improve productivity and ability?
We look in the medical setting of pulmonary embolisms
paulgp.com/papers/Radio...
Submission deadline for the Philadelphia Fed's Mortgage Market Research Conference is January 20th. Details here: bit.ly/4mT9Ocz
There’s a paradox around remote work. Jamie Dimon says it will kill productivity, while startups are hiring for remote roles. Who is right?
With Abhinav Gupta and Elena Simintzi we try to resolve this dispute
Paper: papers.ssrn.com/sol3/papers....
Substack: arpitrage.substack.com/p/remote-wor...
More on dementia & finances: similar to credit outcomes, we see declines on asset side start about 6 years before dementia onset, patterns are consistent with mistakes driving declines, not healthcare spending or spenddown
Recently published at JPAM: "Can Machine Learning Target Health Care Fraud? Evidence From Medicare Hospitalizations" by Shubhranshu Shekhar, Jetson Leder-Luis, & Leman Akoglu.
doi.org/10.1002/pam....
Best video games I played in 2025:
1. Finally played Outer Wilds. 10/10 no notes, go play it without spoiling yourself. It’s like a fun Myst with sweet space physics