Jesse Rothstein

American Economic Association AREUEA Poster Session Poster Session Hosted By: American Real Estate and Urban Economics Association Chair: Lu Han, University of Wisconsin-Madison AEA Poster Session Poster Session Hosted By: American Economic Association AFA PhD Student Poster Session...

@assameeting.bsky.social attendees, join the Institute, LERA, & @neaecon.bsky.social in celebrating the many contributions of Bill Spriggs to econ profession & policy at a luncheon on Jan 4 and two paper sessions. Search “Spriggs" at https://bit.ly/4ohwkg4 for details 

This article is well worth reading.

I love this languages map! Would love to see one for LA also.

Is the SAT serving the mission of public universities?

CSHE Senior Associate Saul Geiser tackles this question in an op-ed for Inside Higher Ed. Drawing on his ROPS paper, Geiser argues that "treating the test as a neutral measure of merit is an illusion."

www.insidehighered.com/opinion/view...

The California Policy Lab provides independent, practical research that helps our government make better decisions and improve more lives. In this interview, Prof. Till von Wachter explains CPL's research showing the LA wildfires impacted people both inside and outside the directly impacted areas.

Tune in to @kqedforum.bsky.social at 9a tmrw to hear me talk rising student loan delinquencies

@jrothst.bsky.social makes excellent points & language is important. I'm estimating "full employment" ~given~ cruel immigration policy. To measure slack, Fed compares actual labor mkt to estimate of current "full emp" (also dicey language as that estimate generally means UR is high for some groups).

Final note: Of course, this all presumes we know what employment growth has been. We don’t. The Republican failure to pass a budget means that the data aren’t being collected or reported. That makes it much harder to ensure that policy matches to conditions. The Fed can only go on vibes. 7/7

So yes, the Fed should take account of this, and not interpret job growth in the low 5 digits as a sign of a catastrophic collapse in demand. But it is nevertheless bad, not good, news and a sign of policy that is not only illegal and inhumane but also bad economics.

That is, the slowdown in job growth consistent with full employment is direct evidence that immigrants take jobs that otherwise wouldn’t exist in this country, not jobs that natives would otherwise get. And that our (lawless, inhumane) immigration policy is making us all poorer.

To the extent there is an economic justification for the immigration crackdown, it has to be that immigrants take jobs that would otherwise be held by native workers. The evidence has long made clear that isn’t true. But if it were, a reduction in immigration wouldn’t reduce employment growth.

But the low employment growth is not good - it is bad. It means that our economy isn’t growing as fast as it otherwise would, and we aren’t getting wealthier the way we otherwise would. It is a sign that the administration’s (lawless, inhumane) immigration policies are making us all poorer.

@wendyedelberg.bsky.social is (of course) right that we need to recalibrate expectations of what employment growth we can expect in an economy where demand matches supply. If the Fed tried to maintain previous employment growth via monetary policy, it would just create inflation.

This is a good thread. But I want to take exception to the naming of low unemployment and low employment growth as “good.” A thread of my own: 1/7

Sign says "We're HIRING"

The California Policy Lab is hiring a Researcher/Senior Researcher, focused on health & safety net in California: capolicylab.org/careers/rese...

Infograph has a "We accept EBT" sign. And text above it reads: "330,000 low-income college students use CalFresh to help pay for their groceries. But, because of the government shutdown, these college students may not receive their November CalFresh benefits."

More than 330,000 low-income college students in California use #CalFresh to help pay for their groceries. But they may not receive their benefits in November because of the government shutdown. capolicylab.org/news/analysi... #CaLeg

I'd caution against over-relying on that skepticism. That (people find jobs elsewhere) is usually what happens in response to technical change, so I think we should require evidence before concluding that it isn't happening this time.

Interesting - thanks. That does seem like it is cutting things pretty finely. And on top of that, if COVID had any impact on time-to-degree, that would also show up in EPOPs at those ages.

They do have one graph, their Figure 4, showing total headcounts. But it isn't normalized for population (baby booms and busts do funny things to headcounts by narrow age group), and is based only on a subset of ADP data. I'd like to see more evidence in support. Is there any?

This is what I am referring to in my second paragraph: They show evidence for specific occupations. That I can believe. But people seem to be jumping from that to "it is a bad time to be a young worker looking for work," and that I'm not seeing.

Part of what? My point is that there isn't much evidence that hiring has been held off on, regardless of whether AI is the cause or not. Or at least I haven't seen it.

The unemployment rate over time for college graduates, age 20-24 and 25-34, from the FRED site.

The third graph as an image:

The unemployment rate over time for 16-24 year olds and for age 20+, from the FRED site.

The second graph as an image:

Graph of unemployment rate for 16-24 year olds over time, from the FRED site.

The graph as an image:

So excited to be working with @jrothst.bsky.social and the amazing program committee to make this year's SoLE annual meeting the best ever-don't forget to submit your paper so you can join us there!

www.sole-jole.org/2026-program...

What evidence am I missing that AI is having broader impacts on young people's access to jobs - or even that *anything* unusual is impacting that?

This graph focuses on college grads (only back to 2000). Again, the young people series is always higher. The gap always grows in (a) recessions and (b) summer. The most recent data point is Aug. 2025 (thanks, Republican shutdown!), and seems right on trend.
fred.stlouisfed.org/graph/fredgr...

This adds the UR for the full labor force, age 20+.
fred.stlouisfed.org/graph/fredgr...
The increase in the relative unemployment rate of the younger group is small compared to the permanent age difference. And the gap always grows in recessions. Was AI affecting the labor market in 1955?

Here's a graph that was shared with me, showing a small rise in unemployment among 16-24 year olds.
fred.stlouisfed.org/graph/fredgr...

Is there anything more than vibes to the idea that unemployment is up among the young, due to AI?

Yes, I am aware that there evidence of impacts in specific occupations like software development. I'm interested in the claim that there's been a change in aggregates.

#econsky