Matthew Gudgeon

Tufts University is thrilled to host the North East Development Consortium (NEUDC) Conference on November 8-9, 2025!

📝 Call for Papers: Submissions start June 16, and end August 17. Submit your papers here:

editorialexpress.com/cgi-bin/conf...

More details to come!

#NEUDC #NEUDC2025

NEUDC 2025

The North East Universities Development Consortium (NEUDC) Conference will be held at Tufts Nov 8-9, 2025. Submissions are now open. Deadline to apply is Oct 17. neudc2025.com

Please see the link if interested in a full-time, pre-doc RA position working with labor, public, and education economists in OEMA and the Economics Program at West Point —-
an excellent opportunity to develop skills and work with a great team. otie.hrmdirect.com/employment/j...

A recent paper by Eoin McGuirk and @nathannunn.bsky.social highlights how agricultural development projects can backfire in traditionally pastoral areas of Africa, fueling conflict instead of fostering development.

#EconSky

www.nber.org/papers/w33191

Check out this paper! Gerard is also on the market :) scoop him up! 📈📉

Econometrics Thread (#EconSky)

Today, I will talk very briefly about a few recent methodological papers that I think are super useful to applied researchers. Basically, below, you will find some new tools that may help you to answer relevant empirical questions.

+

Still possible to sign @imbernomics.bsky.social @adamaltmejd.bsky.social

Yup, just drop me an email!

IZA fellows and affiliates can still sign if they choose to! Reach out to any of the signatories for a link. @davidajaeger.bsky.social shared it with me and i signed. #econsky 📈📉

Grateful for the last years at IZA and the new spirit to labor economics and the policy discourse Simon Jäger brought with him.
The inclusive and cooperative leadership approach has been contagious and maybe this year’s summer events gave a glimpse of what path IZA was on to.

What could have been…

If you're a senior male economist and would be interested in discussing the drafting of an open letter re the newly announced IZA leadership change, e-mail me.

Female econs are rightly tired of the burden of such efforts falling on them.

Exactly! Simon is such a star. Getting him to head a German research institute was a coup unlike any I have seen in that space before, and his contributions to the debate so valuable. I am a bit ... lost.

Adding #econsky 📉📈

Oh. My. God.

www.iza.org/wc/files/pre...

The Deutsche Post Foundation is undergoing a period of change and has decided to reorganize its research institutes. This involves changes in the strategic direction of both the IZA Institute of Labor Economics and the briq Institute on Behavior & Inequality as well as in the funding of the research institutes. As part of this transformation, the briq Institute will be integrated into IZA, and the Berlin location of IZA will be closed down. As a consequence of these developments, the current CEO of IZA, Simon Jäger, has decided to depart from his leadership role at IZA by the end of 2023 in mutual agreement with the Deutsche Post Foundation. Both IZA and the Deutsche Post Foundation regret this decision and would like to express their gratitude to Simon Jäger for his exceptional work and his contributions to the further development of IZA. Effective January 1, 2024, Armin Falk, the current CEO of the briq Institute, will take over the leadership of IZA.

😳

#Econsky 📈📉

I have a new working paper out with Pablo Guzman, Johannes Schmieder, Simon Trenkle, and Han Ye: "When Institutions Interact: How The Effects of Unemployment Insurance are Shaped by Retirement Policy. www.nber.org/papers/w31807

See replies for a short thread.

#Econsky 📈📉

12. In summary, the same UI change can have diff. effects on the unemp. rates of older workers under diff. retirement institutions.  The same likely goes for other inst’s, as well, like different UI or welfare benefits.  We hope this helps inform when to expect differences and how large they can be.

Panel a) shows that the simulated unemployment rates of older workers would be substantially smaller had PBD remained at 12 months instead of increasing to 32 months. There is less of an effect on slightly younger workers. Panel b) shows that the unemployment rates of older workers would have been much lower had retirement via UI not been allowed at age 60 (and instead only allowed starting at age 63).

11. This matters:  Based on estimates for younger workers or more recent workers, we wouldn’t have expected historical PBD extensions to explain the distinctive 10pp rise in the unemp rate of older workers in the 90s. Instead, due to ret. policies, they explain over 50% of the rise.

10. We use the model to show that, in 2014, the effect of increasing PBD by 1 year on the unemp. rates of those 56-59 is a 0.87pp increase, whereas had retirement institutions remained at their historically more generous levels, this same PBD extension would have increased unemp. by  2.85pp

These show that the model simulations for both employment-to-unemployment transitions for the 1935 birth cohort and for non-employment duration fit their empirical counterparts well, matching the inflow spike and other broad patterns.

This figure plots the empirical and model-simulated unemployment rates of workers aged 52-55 and 56-59 separately. The model broadly captures the empirical trends, matching the large differential spike in the unemployment rate of relatively older workers in the 1990s.

9. Our model fits the key patterns in the data as well as the key trends in the unemployment rate for workers aged 52-55 and 56-59.

8. In order to make statements about the total effect of PBD extensions for older workers in varying contexts, we take a structural approach and specify a dynamic model that allows workers to transition between Emp., UI, and being out of the labor force.

7. Where we can, we also estimate RDs at older ages at which PBDs are discontinuously extended that yield the effects of an additional month of PBD conditional on UI entry.  These are similar to younger workers, perhaps a little larger.

6. These inflow responses are large: At certain points in time, between 10 and 25% of the ENTIRE (sampled) birth-year cohort was receiving UI at age 59. In the mid 1990s over 50% of total UI payments in any given month accrued to workers aged 55+.

These two figures show that for the 1924 birth cohort, UI entries spike at age 59, and for the 1935 cohort, UI entries spike at age 57 and 4 months.

5.First we present reduced form evidence showing that older workers UI inflows spike at precisely the bridge-to-retirement age, i.e. the age that allows workers to transfer directly from UI to pension without any uncovered period.  Inflows adjust as expected as PBDs increase.

This shows the unemployment rate for workers in different age groups in both the USA and Germany over time. In the US. unemployment rates largely co-vary for the different groups, in Germany, the unemployment rates of workers aged 55 to 59 diverge from the rest, reaching 15% in the mid 1990s (other groups stay at 8% or so).

4. We show this in Germany, where reforms over four decades provide essential policy variation on both UI maximum potential benefit duration (PBD) and retirement rules.  And where historical unemp. rates of older workers have exhibited a distinctive and hard-to-explain pattern.

3. Our paper shows that the non-employment effects of UI for older workers depends in a first-order way on the structure of retirement policies.  The effect of the same UI extension on same-aged workers can be significantly larger under one set of pension policies than another.

2. For e.g., the non-employment effects of UI extensions may be larger if welfare options following UI worsen. Or, for older workers in particular, they might depend on retirement policies. Are these interactions quantitatively important in practice?

#EconSky:
The effects of key policy changes -- such as the non-employment effects of unemployment insurance (UI) extensions – are often estimated in isolation, holding other policies fixed.  But theory tells us that the effects of such policies can vary with the broader policy context/institutions.