iiinova — where your digital self keeps evolving

David A. Jaeger (econ) @davidajaeger.bsky.social · 20/08/2026

The ones that stick are the ones they use organically (online searching, word processing). AI will be one of those

David A. Jaeger (econ) @davidajaeger.bsky.social · 20/08/2026

Wow, what a privilege to welcome the first Master of Public Policy cohort to the @uniofstandrews.bsky.social! So gratifying to see this talented group of 29 here!

2291

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

For simple estimators with a few tens of thousands of observations, estimating 8 or so specifications takes about 5 minutes of computer time, more complicated estimators, large numbers of specifications, take longer. Thanks in advance!

020

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

There is a worked example for both the Stata and R version for generating the bootstrap estimates. Stata repo is here: github.com/davidajaeger... R repo is here: github.com/davidajaeger... Paper is here: www.djaeger.org/research/rob...

130

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

I'd be most grateful for feedback on the packages. They are thoroughly tested (the estimates in the paper use them), but it would be great to have feedback on how they function "in the wild".

110

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

For my most recent paper, there are Stata and R packages to calculate a suite of robustness statistics. The statistics are calculated from B bootstrap estimates of the relevant parameters, where B is the number of bootstrap samples, chosen such that (B+1)\alpha is an integer.

1112

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

I'd be most grateful for feedback on the packages. They are thoroughly tested (the estimates in the paper use them), but it would be great to have feedback on how they function "in the wild".

000

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

That might actually be a good thing for academia…

130

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/08/2026

But also not pay social insurance and retire with 65% of your last year’s salary forever, if you’re verbeamtet.

120

David A. Jaeger (econ) @davidajaeger.bsky.social · 13/08/2026

Robustness? in @iza.org clothing, Discussion Paper 18851! docs.iza.org/dp18851.pdf

030

Reposted by David A. Jaeger (econ)

Count Binface @countbinface.bsky.social · 13/08/2026

Checkmate, Nige.

14678211992

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

I suggest that 𝑅^*_{.95} and 𝑝_𝑅 be reported whenever multiple specifications are presented as evidence of robustness.

010

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

In a pre-registered survey of CEPR and NBER affiliates, expert judgments align with the framework in obvious cases but diverge in intermediate cases, where the verdict depends on the joint sampling distribution rather than on the visible spread of point estimates alone.

120

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

Simulations show approximately correct size and demonstrate that common visual rules often provide false comfort. Applications to five prominent papers validate some conventional robustness claims while revealing cases in which apparent agreement reflects imprecision rather than stability.

110

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

Because the bootstrap resamples the same observations and re-estimates all specifications in each replication, it captures dependence across estimates without requiring inversion of a potentially near-singular variance-covariance matrix.

120

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

The range-based equality 𝑝-value, 𝑝_𝑅, tests whether the estimates are statistically distinguishable. Together they separate two ideas that informal robustness checks often conflate: failure to detect differences and affirmative evidence of agreement.

110

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

I formalize this claim with two bootstrap statistics. The minimum equivalence bound, 𝑅^*_{1−𝛼}, is the smallest tolerance within which the estimates can be judged equivalent.

110

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

This common practice makes an implicit inferential claim about the range of estimates, but the estimates are usually generated from the same data and have a joint sampling distribution that published tables do not show.

140

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

Applied economists routinely compare estimates across specifications, observe that they are “similar,” and conclude that their results are “robust.”

110

David A. Jaeger (econ) @davidajaeger.bsky.social · 29/07/2026

Ever estimated a bunch of specifications, glanced across the coefficients and said “eh, looks robust to me!” Who *hasn’t*?? Then my new working paper is for you! www.djaeger.org/research/rob.... Be the first on your block to stop eyeballing…. Bonus R and Stata packages, just ‘cause I like you.

djaeger.org

Robustness? Range Tests for Equivalence and Equality Across Multiple Specifications – David A. Jaeger

1224

Reposted by David A. Jaeger (econ)

Per Engzell @pengzell.bsky.social · 27/07/2026

Has anyone tried this with influencers and autocrats

1287

David A. Jaeger (econ) @davidajaeger.bsky.social · 21/07/2026

Gives “Bonus slides” a new meaning!

010

David A. Jaeger (econ) @davidajaeger.bsky.social · 01/07/2026

Now it’s down to rooting for “Anyone but England”.

100

David A. Jaeger (econ) @davidajaeger.bsky.social · 01/07/2026

So when is Fable actually going to be available again??

020

Reposted by David A. Jaeger (econ)

Jonathan Portes @jdportes.bsky.social · 27/06/2026

Times argues that Labour sticking to its plans fro remigration and deportation is the only way to stave off Farage - because that's worked so well so far. There is no appeasing these people, and no point in trying and the sooner govt realises that the better. www.thetimes.com/comment/the-...

ut Ms Mahmood believes the changes are essential, both for maintaining social cohesion and limiting the burden on public services. As things stand, some 1.6 million people will be eligible for ILR status by 2030, granting them greater access to the welfare state. She also believes changing the system is the only way of seeing off the political threat from Nigel Farage. She is right on all counts.

25431127

Reposted by David A. Jaeger (econ)

Shelly Lundberg @shellyjlundberg.bsky.social · 25/06/2026

I'm so grateful to the colleagues who have been willing to read my forthcoming book How Economics Discovered Women in advance (#1).

59921

David A. Jaeger (econ) @davidajaeger.bsky.social · 22/06/2026

Very glad we took that decision. It was time.

1120

David A. Jaeger (econ) @davidajaeger.bsky.social · 22/06/2026

I’ve always found that a bit cringe.

000

David A. Jaeger (econ) @davidajaeger.bsky.social · 19/06/2026

But hasn’t Burnham reiterated Starmer’s non-sensical immigration policy?

140

David A. Jaeger (econ) @davidajaeger.bsky.social · 17/06/2026

Ok, but hasn’t the Census Bureau considered this for a long time (and abandoned after pressure)? The access issue is super important and separate from the statistical noise issue I think.

000

Reposted by David A. Jaeger (econ)

Eleonora Alabrese @eleala.bsky.social · 17/06/2026

🧵 New working paper out today: "Measuring the Regional Economic Cost of Brexit: Evidence as of 2026" with @jacobedenhofer.bsky.social @trfetzer.com & Shizhuo Wang Here's what the data say about who paid the price in these 10 years from the referendum 1/…

6172119