Is each American generation doing better?
We construct a posttax, posttransfer income measure from 1963 to 2023 based on the Current Population Survey Annual Social and Economic Supplement that allows us to consistently compare the economic well-being of five generations of Americans at ages 36–40. We find that Millennials had a real median household income that was 20% higher than that of the previous generation, a slowdown from the growth rate of the Silent Generation (36%) and Baby Boomers (26%), but similar to that of Generation X (16%). The slowdown for younger generations largely resulted from stalled growth in work hours among women. Progress for Millennials younger than 30 has also remained robust, though largely due to greater reliance on their parents. Additionally, lifetime income gains for younger generations far outweigh their higher educational costs.
That is from Kevin Corrinth and Jeff Larrimore in Demography. Via the excellent Kevin Lewis.
Those old factory sector jobs
As AI sweeps into white-collar workplaces, old-timey hands-on jobs are getting a new look—and some of those professions even have shortages.
Consider tailors. Sewing is a vanishing skill, much like lacemaking and watchmaking, putting tailors in short supply when big retailers like Nordstrom and Men’s Wearhouse, as well as fashion designers and local dry cleaners, say they need more of them.
The job, which can take years to master, can be a tough sell to younger generations more accustomed to instant gratification. But apprenticeships that offer pay to learn on the job and new training programs are helping entice more people…
For the first semester of its program, which concluded in December, FIT received more than 190 applications for 15 spots. The nine-week course requires prior sewing experience. Nordstrom hired seven students from the inaugural class.
“It’s increasingly becoming more challenging to find people to fill these alterations jobs,” said Marco Esquivel, the director of alterations and aftercare services at Nordstrom, which employs about 1,500 tailors. Similar to other high-end retailers, Nordstrom offers free basic tailoring for garments purchased at the department-store chain and charges a fee for those bought elsewhere.
Tailored Brands, which employs about 1,300 tailors at its Men’s Wearhouse, Jos. A. Bank and other chains, is updating its apprenticeship program to include more self-guided videos with the goal of moving people through the training faster.
Here is more from Suzanne Kapner at the WSJ. Via LJ Fenkell.
The exposed counties (from my email)
Professor Cowen,
Built a county-level AI displacement model across all 3,204 US counties. Top 5 most exposed counties are all in the DC metro, not the Rust Belt.
https://jakeprokopets.substack.com/p/why-the-most-ai-exposed-counties
18, built it in three days.
Jake Prokopets
Wednesday assorted links
1. On health care price transparency.
2. Interview with Sindarov’s trainer.
3. Tariff increases are contractionary.
5. U.S. manufacturing capacity has been growing for sixteen consecutive quarters.
6. Dean Ball book on AI is coming.
7. DEI statement requirements in academic hiring have more than halved within a year.
Ending the Occupational Licensing Racket
VinNews: The Rockland County Legislature approved amendments to the Home Improvement Law, dissolving the existing Home Improvement Licensing Board and shifting primary licensing authority to the Legislature itself…Under the new rules, the former licensing board will be reduced to an advisory role, losing its power to issue or revoke licenses. Licensing responsibilities will now fall under the Rockland County Legislature…
This is an interesting change and worth studying. In the Licensing Racket, which I reviewed for the WSJ, Rebecca Haw Allensworth emphasizes that occupational licensing boards put the fox in charge of the chickens:
Governments enact occupational-licensing laws but rarely handle regulation directly—there’s no Bureau of Hair Braiding. Instead, interpretation and enforcement are delegated to licensing boards, typically dominated by members of the profession. Occupational licensing is self-regulation. The outcome is predictable: Driven by self-interest, professional identity and culture, these boards consistently favor their own members over consumers.
Ms. Allensworth conducted exhaustive research for “The Licensing Racket,” spending hundreds of hours attending board meetings—often as the only nonboard member present. At the Tennessee board of alarm-system contractors, most of the complaints come from consumers who report the sort of issues that licensing is meant to prevent: poor installation, code violations, high-pressure sales tactics and exploitation of the elderly. But the board dismisses most of these complaints against its own members, and is far more aggressive in disciplining unlicensed handymen who occasionally install alarm systems. As Ms. Allensworth notes, “the board was ten times more likely to take action in a case alleging unlicensed practice than one complaining about service quality or safety.”
Moving regulation out of the hands of the regulated could be an improvement but there are also advantages to self-regulation. See my review for other reform possibilities.
Hat tip: Heshy.
The empirically inscrutable climate-economy relationship
From Finbar Curtin and Matthew G. Burgess, here is the paper. Here is the thread, worth a read. Important stuff, I hope to hear more about this. The whole climate to gdp transmission thing does not seem to be working very well?
Technological unemployment in Victorian Britain
We do not know whether technological unemployment swept across England in the wake of the British Industrial Revolution. In this paper, I propose an approach to quantify jobs lost to, and created by, creative destruction in the 19th century. Using over 170 million individual records from the full-count British census (1851–1911), I generate sub-industry “task” level occupational data. I apply this to the English bootmaking industry as it mechanized. The new data reveal sharp structural changes: 152,000 artisanal jobs disappeared as skills became obsolete, while 144,000 new jobs emerged. However, incumbent bootmakers were rarely displaced. Instead, the decline was driven by young men no longer entering the artisanal trade. These findings challenge assumptions about displacement, showing how slow adoption and persistent demand can shield existing workers, while opportunities vanish for new entrants.
That is a recent paper by Hillary Vipond, a recent PhD from LSE. Via Lukas Freund. Here are other papers by Hillary, some of them on what we can learn about automation from economic history. Here is Hillary on Twitter.
Imagegen 2.0
Created by Alex T., and of course GPT as well.
Tuesday assorted links
1. Desmond Morris, RIP (NYT).
2. A long NYT feature on how to be cultured.
5. People laid off from USAID (NYT).
6. Ariel Rubinstein tells his story. And his home page more generally.
The Average is Over generation?
The result is that even though today’s young adults, and graduates in particular, are over-represented in the top quartile of the earnings distribution, they are also far more likely to be at the bottom than the top for earnings relative to reasonable expectations. In both the UK and US, even though only 10 percent of graduates are in the lowest earnings quartile, one in three is in the bottom bracket for earnings relative to expectations.
The Luddites Were the First to Attack AI
Everyone knows the Luddites smashed looms. What is less appreciated is that the loom was the first serious programmable device — the direct ancestor of the computer. Thus, the Luddites weren’t just the first to resist automation. They were in some ways the first to attack AI.
The Jacquard loom, introduced in France circa 1805, used a chain of punched cards to control which threads were raised for each pass of the shuttle. The ability to change the pattern of the loom’s weave by simply changing cards was an important conceptual precursor to computer programming. Babbage borrowed the idea directly for the Analytical Engine in the 1830s.
The Luddites lost–they were violently suppressed by the UK military–but more generally they lost because programmable looms brought patterned clothes to the masses.
Prior to its invention, the creation of complex patterns required skilled and labour-intensive manual labour, often involving large teams of weavers. With the Jacquard loom, a single operator could control the machine and produce intricate designs with relative ease.
This innovation greatly increased the speed and efficiency of textile production. It also opened up new possibilities for creativity and design, as the loom enabled the production of intricate patterns that were previously unattainable. The Jacquard loom contributed to the democratization of textile manufacturing, making intricate fabrics accessible to a wider audience
By the time Jacquard died in 1834, thousands of his looms were operating in Manchester, an epi-center of the Luddites riots. Moreover, just over 100 years later, Manchester birthed the Manchester Baby and the Manchester Mark 1, the first electronic stored-program computer. And who was hired to program the latter? None other than Alan Turing.
Ada Lovelace had foretold it all beautifully: “the Analytical Engine weaves algebraical patterns just as the Jacquard-loom weaves flowers and leaves.”
Addendum: I thank Claude for assistance on this post.
A Comparison of Agentic AI Systems and Human Economists
This paper compares agentic AI systems and human economists performing the same causal inference tasks. AI systems and humans generally obtain similar median causal effect estimates. While there is substantial dispersion of estimates across model instances, the human distributions of estimates have wider tails. Using AI models as reviewers to compare and rank “submissions,” the following ranking emerges regardless of reviewer model: (1) Codex GPT-5.4, (2) Codex GPT-5.3-Codex, (3) Claude Code Opus 4.6, and (4) Human Researchers. These findings suggest that agentic AI systems will allow us to scale empirical research in economics.
I enjoy the name of the author, namely Serafin Grundl. Here is the paper, via Ethan Mollick. You could interpret these results as showing the AIs have fewer hallucinations. And just to reiterate a key point from the paper:
The second part of this paper is an AI review tournament in which “submissions” (codes and write-ups) from humans and the AI models are compared and ranked against each other. The reviewers are the following AI models: Gemini 3.1 Pro Preview, Opus 4.6 and GPT-5.4. For each review the reviewer is asked to write a report comparing four submissions (human, Opus 4.6, GPT-5.3-Codex, GPT-5.4). Each reviewer model writes comparison reports for the same 300 comparison groups. The average rankings are strikingly similar across reviewer models: (1) Codex GPT-5.4, (2) Codex GPT-5.3-Codex, (3) Claude Code Opus 4.6, and 2(4) Human Researchers.
Who comes in last? Hi people!
Zimbabwe facts of the day
Zimbabwe, often considered an economic basket-case because of its history of farm seizures and hyperinflation, is enjoying an idiosyncratic boom. High prices for the metal and other commodities have led to a surge of cash through its highly informal economy. They have made it easier for authorities to stop printing money and meddling in currency markets; inflation is at its lowest in about 30 years. The IMF has repeatedly revised upwards estimates for economic growth, most recently to at least 7.5% for 2025, almost double the African average…
Gold is not the only source of growth. The current tobacco crop will be the largest on record. Lithium, chrome and platinum miners, many of them Chinese, have raised production. Zimbabwe’s diaspora, mainly in South Africa, sent back $2.5bn last year. So overall demand is higher than ever, says a banker.
Here is more from The Economist. We are told that the private vault sector is booming too.
Monday assorted links
1. Sindarov profile. And: “I know this GM who made 2600 at 19 without reading a chess book in his life.” And Magnus on Sindarov vs. Gukesh.
2. Inflation-adjusted book prices over time.
3. On the Amanat Iran book and its excellence.
4. More on the wet market hypothesis. We should all be uncertain, but it is mood affiliation (with conspiracy theorizing, for one thing) to be convinced of Lab Leak. It is contributing to negative emotional contagion.
5. Review of the new Knausgaard series. By Max Norman: “(I’d rather read Knausgaard on defecation than predestination, let alone whether machines can think or trees can feel.)”‘
6. AI and the arts, a short Instagram video.
7. AI and the pancreatic vaccine. More testing is needed, but there is a reasonable chance that we have a good treatment for pancreatic cancer, and AI was instrumental in that. It is mRNA as well, so a double burn on the haters.
Eight Rules to Regain Public Trust in Academia
The Yale Report was quite good but for concision I prefer Kevin Bryan’s Eight Rules:
1. Produce and Teach Useful Knowledge
Universities exist to generate and teach useful knowledge. This knowledge is grounded in skeptical inquiry, empirical evidence, and logical deduction. “Useful” includes not only practical applications but also fundamental discoveries that expand our understanding of the world, even if their benefits are long-term.
2. Be Useful to All of Society
Universities are subsidized only if society at large finds them valuable. Research may take time to bear fruit, but its insights should ultimately serve the public good, communicated openly and accessibly, and presented with epistemic humility. Teaching should be done with care and draw on up-to-date research.
3. Attract Talent from All of Society
Useful knowledge can be created by people from any social or economic background. Do not waste talent. Do not select talent based on who knows “how to play the game”. Avoid insular language or norms that deter people from entering research.
4. Neutral, Objective Research Produces Useful Knowledge
Research must be neutral and objective. It is true that everyone has their individual background and preferences; nonetheless, unbiased research is still possible. Tradition, folk knowledge, and storytelling all play an important roles in society, but they are not the purpose of universities. There is no “Western science” or culturally-determined “ways of knowing”. Rather, research is open to all and can be performed identically regardless of background.
5. Hire, Promote, and Cite Based on Knowledge Contribution
Hiring, promotion, and citation must be based on an individual’s contribution to knowledge. Nepotism, group preferences, and adherence to specific “schools of thought” corrupt this process. When advancement is not based on merit, the public rightly questions our integrity and the objectivity of our findings.
6. Keep Personal Views Out of Research and Teaching
A scholar’s personal politics should be invisible in their research and teaching. If a finding is predictable based on the author’s identity or known views, the process has failed. Objectivity is the hallmark of credible science. Academics may hold private beliefs like anyone else, but their academic work must stand apart from them.
7. Research Fraud is Unacceptable
Fraud destroys trust. Misrepresentation of results, selective reporting, or methods designed to publish rather than to discover are also harmful. Proven fraud must bring immediate dismissal, as it violates the core purpose of academia.
8. Scientific Institutions Should Be Apolitical
Universities, journals, and scientific societies must remain non-partisan. Their public statements must be rare, restricted to issues of direct expert consensus, and made only when silence would be a greater threat to their integrity than speaking. Activism sacrifices credibility for influence – or worse yet, sacrifices credibility and influence alike.
I would add 9) Grades must be objective and useful discriminators of talent.