iiinova — where your digital self keeps evolving

Peter Hull @instrumenthull.bsky.social · 24/03/2026

Very excited for this Friday's Chamberlain Online Seminar, where we'll muse about "The Future of Econometrics." Check out this panelist lineup! Friday 3/27 at Noon ET. Come see! t.co/pE1gZCMI7a

24411

Peter Hull @instrumenthull.bsky.social · 13/10/2025

Thanks BeNA (and Berlin) for a great few days!

170

Peter Hull @instrumenthull.bsky.social · 26/09/2025

My coauthors and I are hiring a project coordinator for new research on an innovative jail-based rehabilitation paper (IGNITE), building on our recent QJE Apply at careersearch.stanford.edu/jobs/project...

01010

Peter Hull @instrumenthull.bsky.social · 26/09/2025

Come hear what Larry Katz, Xavier D'Haultfoeuille, Kei Hirano, Francesca Molinari, and Andres Santos think about publishing econometrics in general interest journals! Tomorrow's Chamberlain Seminar: 12pm ET

1186

Peter Hull @instrumenthull.bsky.social · 08/09/2025

The Chamberlain Seminar continues this semester with two exciting online sessions on 9/26 & 10/31! Check out the details below and click below to join our mailing list! mailman.stanford.edu/mailman/list...

273

Peter Hull @instrumenthull.bsky.social · 03/09/2025

New 'Metrics Note! "One weird trick" to characterize compliers / the effective population underlying design-based OLS & IV specifications Check it out: www.dropbox.com/scl/fi/lgba6...

44712

Peter Hull @instrumenthull.bsky.social · 23/07/2025

Tomorrow (Thu) at 1:45pm ET: Check out @borusyak.bsky.social's presentation of work w/ Mauricio Caceres Bravo on demand estimation with recentered IVs @ the Frontier Econometric Methods NBER SI. We're honored to have Steve Berry as a discussant! Should be livestreamed at youtube.com/@nbervideos

0185

Peter Hull @instrumenthull.bsky.social · 15/07/2025

If you wanna learn more about this, check out @borusyak.bsky.social's presentation next Thursday at the new @nber.org Frontier Econometric Methods day nber.org/conferences/...

093

Peter Hull @instrumenthull.bsky.social · 05/06/2025

Trump vs Elon spat was just a warmup today

1250

Peter Hull @instrumenthull.bsky.social · 02/06/2025

I'll be leading a Shift-Share Mixtape workshop on July 29th & 30th Sign up here! www.eventbrite.com/e/shift-shar...

2203

Peter Hull @instrumenthull.bsky.social · 28/05/2025

Thank you Tinbergen Institute! tinbergen.nl/news/1045/ti...

290

Peter Hull @instrumenthull.bsky.social · 14/04/2025

cool to see our IGNITE work highlighted in this. Go @marcellaalsan.bsky.social !

1424

Peter Hull @instrumenthull.bsky.social · 03/04/2025

Not too much goin' on today, so thought I'd tell you about a new paper out in @nejm.org w/Angrist, Gao, and Yeh We explain & demonstrate why instrumental variable methods are essential for trials of medical interventions that fail to play out as intended Check it: www.dropbox.com/scl/fi/lsnmm...

2627

Peter Hull @instrumenthull.bsky.social · 28/03/2025

📢New WP w/ @borusyak.bsky.social We derive the most powerful recentered IVs for formula treatments & propose an algorithm for approximating them This approach yields *huge* power gains, relative to conventional simulated IV, when estimating Medicaid crowdout effects Check it out! t.co/UPg2XQBefN

16312

Peter Hull @instrumenthull.bsky.social · 07/03/2025

Coming soon to Zoom rooms near you! We're excited to announce a reboot of the Chamberlain Seminar, five years after its 2020 debut We'll kick things off on 3/21 with a tribute to Gary. More info and future seminars here: chamberlainseminar.org Register here: stanford.zoom.us/meeting/regi...

1368

Peter Hull @instrumenthull.bsky.social · 29/01/2025

TIL Abraham Wald, who famously worked on selection bias in assessing aircraft damage, died in a plane crash :(

44311

Peter Hull @instrumenthull.bsky.social · 19/01/2025

The latter is probably the most realistic case. See p 15-16 of the draft for a discussion of how the choice of evaluation metric can be captured by our definition of systemic discrimination

020

Peter Hull @instrumenthull.bsky.social · 08/01/2025

Can't wait till Friday am when I get to read what Joe Biden thinks of recentered IV

4492

Peter Hull @instrumenthull.bsky.social · 23/12/2024

Some things I've learned in 2024: 6. Design-based regression and IV specifications do not suffer from the same "negative weight" issues popping up in difference-in-difference (More: www.dropbox.com/scl/fi/c6hd5...)

1171

Peter Hull @instrumenthull.bsky.social · 23/12/2024

Some things I've learned in 2024: 5. Algorithmic discrimination can be measured and reduced in high-stakes decisions by leveraging quasi-experimental variation (More: www.dropbox.com/scl/fi/2ji7z...)

1101

Peter Hull @instrumenthull.bsky.social · 23/12/2024

Some things I've learned in 2024: 4. IV analysis of the effects of revascularization on quality of life show they're are larger and more persistent than we'd previously thought (More: www.dropbox.com/scl/fi/dgs74...)

140

Peter Hull @instrumenthull.bsky.social · 23/12/2024

Some things I've learned in 2024: 3. Systemic discrimination in real-world decisions can be measured with "Iterated Audits," that capture the interactions of decisions across interconnected nodes (More: www.dropbox.com/scl/fi/ab5yx...)

1102

Peter Hull @instrumenthull.bsky.social · 23/12/2024

Some things I've learned in 2024: 2. Racial discrimination in foster care placement decisions is widespread across the US, but declining (More: www.dropbox.com/scl/fi/l6nft...)

170

Peter Hull @instrumenthull.bsky.social · 23/12/2024

Some things I've learned in 2024: 1. A jail-based education program in Flint MI which emphasizes rehabilitation and improved jail culture dramatically reduced misconduct and recidivism (More: www.dropbox.com/scl/fi/ep5xy...)

16123

Peter Hull @instrumenthull.bsky.social · 20/12/2024

I have a new paper with a bunch of great folks, including Josh, on estimating dynamic effects in RCTs with imperfect compliance Using a new assumption, we find that the medical benefits of revascularization are larger + more persistent than previously thought www.dropbox.com/scl/fi/dgs74...

2303

Peter Hull @instrumenthull.bsky.social · 19/12/2024

I'm teaching an online two-day workshop on advanced IV topics this February 10th and 12th, through @causalinf.bsky.social's Mixtape platform. All are welcome. Sign up here! www.mixtapesessions.io/session/iv_f...

13611

Peter Hull @instrumenthull.bsky.social · 18/12/2024

At least as it's been conventionally defined (in e.g. Aigner and Cain '77) statistical discrimination is "direct", capturing the effect of perceived race (etc) on actions conditional on signals. Agreed that informally the term may be used more broadly; that's why formalizing systemic is useful, IMO

120

Peter Hull @instrumenthull.bsky.social · 15/12/2024

Here's the punchline -- a "checklist" for researchers hoping to either apply our exogenous-shocks approach or the exogenous-shares approach studied by @paulgp.com and coauthors We pair this with discussion of several real-world papers and how they may or may not fit into the two frameworks

610923

Peter Hull @instrumenthull.bsky.social · 15/12/2024

There's much more new in the paper, including an expanded discussion of how our framework relates to earlier notions of indirect discrimination in economics (e.g. "pre-market" discrimination) and other fields like sociology and computer science Comments welcome, as always! Thanks for reading

2130

Peter Hull @instrumenthull.bsky.social · 15/12/2024

The second shows how gendered language in recommendation letters can drive the vast majority of discrimination in STEM hiring and wage decisions Here the experimental IA approach is key, because of the high-dimensional "signal" of rec letter language which systemic discrimination operates through

180

Peter Hull @instrumenthull.bsky.social · 15/12/2024

The biggest addition, however, is two new field experiments using these IA strategies to measure systemic discrimination in two labor markets The first combines a standard correspondence study with other data on market thickness to show how racial discrimination becomes embedded in job histories

190

Peter Hull @instrumenthull.bsky.social · 15/12/2024

Second, we've revised the section on identification to detail the new Iterated Audit (IA) approach to measuring systemic discrimination We outline two IA strategies -- constructive and experimental -- each appropriate for estimating systemic discrimination in different data environments

140

Peter Hull @instrumenthull.bsky.social · 15/12/2024

First, we've overhauled the theoretical framework to better highlight how systemic discrimination arises from direct discrimination at other points of an interconnected system Researchers first chose which systemic forces to study in the system, which sets the target discrimination estimand

170

Peter Hull @instrumenthull.bsky.social · 15/12/2024

We've significantly updated our paper on modeling + measuring systemic discrimination! Check it out: www.dropbox.com/scl/fi/ab5yx... (cc @aleximas.bsky.social + @aislinnbohren.bsky.social!) A short 🧵 on what's new...

915744

Peter Hull @instrumenthull.bsky.social · 10/12/2024

Luigi Mangione, have I got a paper for you (screenshot c/o @alvinchristian.bsky.social)

3100

Peter Hull @instrumenthull.bsky.social · 09/12/2024

new WP y'all www.nber.org/papers/w33236

927958

Peter Hull @instrumenthull.bsky.social · 04/12/2024

alright alright you've convinced me I guess

3360

Peter Hull @instrumenthull.bsky.social · 02/12/2024

Interesting-looking JMP on using quasi-experiments to measure discrimination in high-stakes settings nikh-rao.github.io/site_files/d...

1286

Peter Hull @instrumenthull.bsky.social · 19/11/2024

Last but not least: Luca Rizzotti (www.lucarizzotti.com), criminal justice data + institutions expert! Luca does applied micro, focusing on the economics of crime, political economy, and labor. His JMP develops + applies a new test for discrimination against crime *victims*, via clearance rates!

2122

Peter Hull @instrumenthull.bsky.social · 19/11/2024

Next up: Pedro Picchetti (sites.google.com/view/pedropi...), master of causal inference! Pedro came to us from Insper in Brazil, and he works on topics in both theoretical + applied econometrics. His JMP develops a neat new method for estimating dynamic treatment effects with IV!

1121