What would you ask Simon Johnson — Nobel laureate in Economics, MIT Sloan professor, and co-director of the Stone Center — if you had the chance?

The Stone Center will be recording an “ask me anything” video with Simon, and we want to hear from you.

Post your question in the comments below, or email shapingwork@mitblueprintlabs.org by September 9th with the subject line “AMA”.

Leave a comment

Over the past seven decades, birth rates have fallen substantially worldwide, fueling widespread concern among economists and policymakers that these declines will slow economic growth. In a new paper, Daron Acemoglu, David Autor, Keelan Beirne, and Andrew Scott study the effects of these past demographic changes on economic performance, comparing across countries and within the US.1 Contrary to standard economic models, they find that lower birth rates have actually raised GDP per working-age adult, with no negative impacts on aggregate GDP or earnings.

Birth rates have fallen substantially worldwide

In 1950, there were 3.78 births each year for every 100 people globally. By 2025, this number had more than halved to 1.71. As Figure 1 shows, birth rates declined both in industrialized countries and across the developing world. The United Nations projects that by the second half of this century, the global population will see its first sustained decline since the bubonic plague of the 1300s.

Figure 1: Global Trends in Birth Rates

Economists and policymakers often express concern that declining birth rates will slow economic growth. Older workers are assumed to be less productive, innovative, entrepreneurial, and technologically adept than younger workers. Shrinking, aging populations are expected to generate less consumer spending and investment and fewer workers, while simultaneously increasing pension and healthcare costs.

While these concerns are often forward-looking, evidence from the past 70 years of falling birth rates can inform our predictions. New findings from Acemoglu and co-authors challenge conventional wisdom: lower birth rates have not hampered economic growth, and in fact have increased productivity per worker.

Finding #1: Lower birth rates increased productivity per worker, without hurting economic output

Lower birth rates, and the aging and shrinking populations that follow, have raised GDP per worker since 1950, not lowered it. Across countries, a one percentage point lower birth rate is associated with an approximately 22% higher GDP per worker2 40 years later, and 29% higher 60 years later. Within the US, regions with lower birth rates have experienced significantly faster average wage growth.3

These gains in productivity per worker were large enough to fully offset the reduction in total output resulting from fewer workers. Total economic output and overall earnings were unaffected. This evidence goes against standard models of economic growth, which predict that lower birth rates will cause an overall decline in GDP.

Finding #2: Labor scarcity causes firms to invest in technology, increasing productivity

Why do declining birth rates result in faster economic growth? The main driver is that when labor becomes more scarce, companies develop and adopt labor-saving technology. These investments in technology make each worker more productive and increase wages on average.

Countries with lower birth rates shift exports toward high-tech industries, and see significantly faster growth in total factor productivity (a measure of how efficiently an economy uses its labor and capital to produce goods). Similarly, in the US, areas with lower birth rates shift employment away from labor-intensive industries and toward R&D-intensive industries. These areas also see an increase in the share of patents that reduce the need for workers, such as IT and automation technology.

The data do not support other conventional potential explanations for the relationship between lower birth rates and higher productivity per worker. For example, there is no evidence that lower birth rates bring more women into the workforce, nor that fewer births leads to more investment per child through higher-quality childrearing or more resources per student in schools. The hypothesis that lower birth rates spur transitions from agriculture to manufacturing, increasing productivity per worker, is also unsupported.

Finding #3: Scarcity of younger workers – rather than scarcity of population – drives these results

Are these increases in productivity driven by the fact that the population is smaller, or that the workforce is older? To separate these two factors, the researchers analyze the impacts of civilian versus military deaths during WWII across countries. Because the war affected countries differently for reasons unrelated to their economies, it provides a sobering natural experiment. Civilian deaths reduced population size without fundamentally changing the age structure, whereas military deaths were concentrated among young men, shrinking the population and raising the average age of the workforce.

Across countries, more military deaths are associated with higher future GDP per capita, while civilian deaths correspond with lower GDP per capita. This suggests that it is specifically a scarcity of younger workers that prompts companies to adopt labor-saving technologies. These technological shifts then increase productivity per worker.

Conclusion: Smaller, older populations do not inevitably shrink the economy

These findings challenge traditional economic models, which predict that lower birth rates lead to declines in overall GDP, and that increases in productivity per worker do not persist. However, many countries are on the cusp of even more significant (and faster) population changes. For example, China’s working-age population is expected to fall from its peak of 1 billion to around 300 million by 2100. China leads the world in the development and adoption of robotics, but will these investments be enough to counteract the economic impacts of a declining population in the coming decades?

The future is far from pre-ordained. Life expectancy is rising along with these demographic shifts, which could bring about other institutional changes, policies, and investments. And technology and investment responses in the age of artificial intelligence may differ from decades past. The past 70 years make clear that the future of global economies will depend not only on demographic changes, but on our technology and policy choices in an era of rapid technological transformation.

Share

Source: Daron Acemoglu, David Autor, Keelan Beirne, and Andrew Scott, “Baby Busts and Growth Booms: Demographic Change and the Macroeconomy,” NBER Working Paper 35401 (2026), https://doi.org/10.3386/w35401

One key intervention is to rebalance the tilted investment playing field in the tax code, which incentivizes firms to invest in machines and algorithms over hiring and retaining workers. Simon explains:

“I think there’s two issues in the tax code with regard to capital and labor and how they’re treated. One is this idea we’ve had since the 1980s that capital income should be taxed less than labor income. Now, if it’s the case that AI is going to divert money from labor to capital, that means you’re moving it to a less taxed sphere. So you may want to help workers, you may have good ideas for helping workers, but where’s the revenue going to come from?”

Second, employing workers comes with additional costs to firms such as Social Security and Medicare contributions, raising the relative cost of labor.

“This I think is going to be a very active and intense debate as we head into next year, which is can we find reasonable and responsible ways to level the playing field? Nobody wants to subsidize the displacement of labor. But when you have a differential tax system, seeing them as inputs, that is unfortunately what’s happening.”

Book reviews

In The Washington Post, Julia Cartwright describes:

The MIT economist and Nobel laureate best known for his 2012 title “Why Nations Fail” has written the most ambitious entry yet in the crowded liberalism-in-crisis genre. Where others blame social media, demagogues or racism, he makes a much more incriminating argument: Liberalism is dying primarily of self-inflicted wounds. And unlike most diagnosticians, he offers an alternative: what he calls “working-class liberalism.”

The book was also reviewed by Steven Poole in The Guardian and Martin Wolf in the Financial Times.

Wolf writes,

Acemoglu’s most recent book is timely and thought-provoking. It is also an important call to arms. It is so for good reasons: hard-won and precious freedoms are now at stake. What Happened to Liberal Democracy? is also no mere polemic. It is factual.

Above all, the book addresses fundamental questions. What is liberalism? Why is it so precious? Why is it internally conflicted? What has made it endangered? Above all, how are we to save it?

Subscribe now

Interviews

Daron appeared on The Remnant podcast with Jonah Goldberg, Pitchfork Economics with Nick Hanauer, Conversations with Tyler with Tyler Cowen, Stay Tuned with Preet Bharara, and The Good Fight with Yascha Mounk.

On Tyler Cowen’s podcast, Daron explains,

Liberalism’s crisis is related to the rise of a post-industrial economy in which what you just called the “educated elite” has become both much more powerful and more numerous, and a philosophical shift that gave more and more priority and power to the educated elite. Whether you approach that from the left liberal point of view or from more of a right liberal point of view, that’s problematic in my mind.

We cannot have true freedom and true flourishing of a population when there is a clear hierarchy, whether it is by education, it is by race, it is by income. Just empowering the elite in that way is not going to cut it.

Share

Daron describes the motivation and central thesis of the book:

The crisis of liberal democracy is the defining challenge of our age. Our lives would be transformed for the worse without democracy and the basic freedoms that liberal democratic institutions enshrine.

The irony is that liberal democracy was once very popular because it delivered what people wanted: shared prosperity, true voice in public affairs, and high quality public services. This crisis is rooted in liberal democracy, in theory and in practice, turning its back on these founding aspirations.

My book, What Happened to Liberal Democracy?, argues that we need a new governing philosophy for democracy and proposes “working-class liberalism” as the bedrock of this philosophy. It prioritizes shared prosperity and issues that matter to the whole population, not just to a narrow elite.

Liberal democracy, renewed this way, would be the right framework for addressing new challenges, from AI to climate change to shifting global balances.

Learn more and pre-order What Happened to Liberal Democracy? here

Insights from the Stone Center’s co-directors on AI, aging populations and economic growth were featured across media outlets this month.

The Atlantic: Automation Led to Economic Misery. AI Doesn’t Have To.

The advent of digital technology over the past 40 years increased inequality and polarized the labor market. In The Atlantic, Daron Acemoglu argues that AI does not have to do the same – but only if we redirect the technology away from excessive automation and toward “pro-worker AI.”

“At stake is a core tenet of the liberal economic order: that jobs and wage growth remain available to workers of different skills and backgrounds. The past 40 years have jeopardized this promise. But they have also produced new, capable technologies—most important AI—that can help workers rather than sideline them. First we have to decide that automation won’t be the north star of the AI age.”

VoxEU: Baby busts and GDP booms: Demographic change and the macroeconomy

A recent VoxEU column from the Centre for Economic Policy Research (CEPR) summarizes new research by Daron Acemoglu, David Autor, Keelan Beirne, and Andrew Scott on declining birth rates and economic growth.

Despite widespread economic pessimism, historical cross-country and US commuting zone data point to no impact of ageing populations on aggregate GDP or earnings growth. Countries with lower birth rates have higher GDP growth per working-age adult, and US commuting zones with lower birth rates have faster wage growth. The growth-reducing potential of demographic slowdowns is countered by faster technological change. While future demographic changes will occur at greater speed and scale, the economic record suggests that current pessimism is premature.

The Wall Street Journal: Is an AI Jobs Apocalypse Coming? Three Economists Square Off

David Autor was interviewed in The Wall Street Journal alongside Martha Gimbel and Anton Korinek about how AI will impact jobs.

“The view that everything exposed to AI will shrivel up and die makes no sense. We’ve been using computers for decades and that has not eliminated all work. There are now a quarter million data scientists in the U.S., and that’s a job that didn’t exist without computers. Instead of thinking primarily about running out of jobs, we should think how we will use AI to meet vast unmet needs in healthcare, better education and other tough, societal problems.”

We Must Act Now: A Statement on AI’s Transformation of the Economy

The Stone Center’s co-directors signed a statement on AI’s transformation of the economy, along with hundreds of other economists, AI researchers and tech leaders. The statement was covered by The New York Times, Axios, Fortune, and The Associated Press, among other outlets.

1. AI may become radically more powerful over the next 10 years.

2. This could drive an unprecedented transformation of our economy, larger than the Industrial Revolution, but unfolding over a vastly shorter time frame. It could bring risks, including large-scale job displacement, as well as opportunities such as major gains in living standards.

3. Economists, policymakers and technology leaders must act now to understand the economics of transformative AI and to build the incentives, guardrails, and institutions needed to steer AI in a direction that complements humans and benefits society.

The promise of agentic AI models is to provide context-specific information and recommendations to human decision-makers. While this may benefit individuals in the short run, it discourages the formation of collective knowledge by disincentivizing learning effort.

In a recent paper, Daron Acemoglu, Dingwen Kong, and Asuman Ozdaglar develop a theoretical model for how agentic AI interacts with human learning incentives and long-run collective knowledge. While powerful AI models can improve the efficiency and quality of decision-making today, they replace human effort in a way that threatens to erode long-run community-level knowledge.

Individual human learning contributes to general knowledge

Humans require two types of knowledge to make well-informed decisions: general knowledge and individual, context-specific information. For example, to make investment decisions, people need general knowledge about different financial instruments (e.g., stocks versus treasury bonds), how stock markets and economies are performing, and macroeconomic risks. But they also need information specific to their individual context, such as their personal risk tolerance, planning horizon, and other existing sources of income.

Human decision-makers typically acquire general and context-specific knowledge at the same time when learning about a topic. As humans learn, they contribute to growing the stock of general knowledge. And when the quality and availability of general knowledge is higher, individuals find it more valuable to invest in their context-specific learning. Consider, again, the example of investment decisions: the more you understand markets in general, the more useful it is to figure out your own risk tolerance.

While individuals benefit from the collective knowledge that already exists on a topic, because this collective knowledge is the result of the contributions of hundreds of thousands of people, each individual’s direct benefit from their own contribution is small. For example, when a software engineer diagnoses a rare bug in their code, they benefit mainly from fixing their own system (a context-specific fix). If they write up a summary of the problem and its solution on a public forum like Stack Overflow, this will be mostly for the benefit of others in the future.

Importantly, general knowledge is not static. It is dynamically updated over time as the frontier of knowledge changes – for example, as environmental changes cause new diseases and medical advances lead to new treatments.

AI disincentivizes learning effort

The promise of agentic AI is that it provides context-specific, individualized information to human decision-makers. Returning to the example of investing, existing textbooks and online resources allow people to teach themselves about finance and gain general knowledge. But future agentic AI might recommend a specific investment choice based on an individual’s context (or even trade autonomously on their behalf). Since humans can now receive this information without investing in their own learning on a topic, the AI is a substitute for their effort. The better the recommendations from the AI, the less value humans will gain from putting effort into their own learning.

Agentic AI thereby disincentivizes individual learning efforts. And since those individual learning efforts build up the stock of collective knowledge, AI indirectly erodes public knowledge.

At first glance, it is unclear whether these combined effects will make people better or worse off. On the one hand, individuals benefit personally from high-quality information provided by AI. On the other hand, the erosion of general knowledge makes people (and society) worse off.

The researchers’ model suggests that there is a “sweet spot” of AI accuracy from a welfare perspective. Modestly accurate AI can have a net benefit. But when agentic AI is very accurate, the negative effect dominates.

These dynamics can create a tipping point. Because general knowledge is not static, ongoing human learning effort is required to sustain it. If a community’s stock of general knowledge stays above a critical threshold, continued learning will maintain it. But if the stock dips below that threshold, learning incentives and knowledge can fall together, creating a downward spiral.

AI could trigger the collapse of general knowledge

All together, these forces could reduce collective knowledge in society or even lead to a state of knowledge collapse, in which society’s stock of useful general knowledge gradually disappears over time. AI may still provide context-specific information to individuals. But because good decisions require combining general and context-specific knowledge, personalized AI advice becomes useless once the underlying general knowledge has eroded. Individuals can’t apply tailored information well without understanding the broader context.

How likely is this knowledge-collapse scenario? A key factor is how well the knowledge individuals generate gets pooled and made widely available. Well-aggregated general knowledge – for example, well-functioning public forums, professional networks, open repositories, and Internet search – reduces the risk of knowledge collapse and unambiguously makes people better off. Larger and better-connected communities not only spread existing knowledge, but also motivate people to invest in their learning by making learning more worthwhile.

On the flip side, as agentic AI improves, the erosion of general knowledge could paradoxically become more likely. The risk increases when people respond to high-quality AI by sharply reducing their own learning. Where learning is more ‘sticky’ (e.g., in professions with mandatory training), AI erodes but doesn’t eliminate collective knowledge.

If AI itself adds to the stock of available knowledge or produces novel insights, then knowledge is less prone to collapse – but unless these channels are very strong, public knowledge will still erode.

Policy should consider the optimal level of AI accuracy

There is already suggestive evidence that generative AI has hindered collective knowledge-building, including declining activity and human engagement on knowledge-sharing platforms like Stack Overflow and Wikipedia. To combat these effects, policy could attempt to reduce the extent to which individuals rely heavily on AI recommendations.

Because highly accurate agentic AI is what discourages human learning, one policy lever could be to manage its accuracy directly. This approach would limit the precision of AI recommendations in order to target the ‘sweet spot’ of AI accuracy for preserving human learning and knowledge formation. In general, the analysis shows that a two-phase policy is necessary. Such a policy would start with a temporary moratorium on agentic AI to rebuild the stock of public knowledge above the ‘tipping point’ for knowledge collapse. The second phase would set a permanent cap on the AI’s accuracy at an optimal level.

Other types of policy may also be effective to reduce individual reliance on AI recommendations. As agentic AI models become more powerful, policymakers should carefully consider how AI could degrade collective knowledge in the long run – and not be blinded by its short-term benefits.

Source: Daron Acemoglu, Dingwen Kong, and Asuman Ozdaglar, “AI, Human Cognition and Knowledge Collapse,” NBER Working Paper 34910 (2026), nber.org/papers/w34910

Automation isn’t inherently bad. But it is critical to also invest in innovation and technology development that will increase the demand for human skills.

Without these complementary investments, AI is likely to displace middle-skill workers, pushing them into lower wage jobs and further polarizing the labor market — just as earlier waves of automation did for workers in manufacturing and clerical jobs.

To learn more about the labor market effects of automation from the 1980s to the mid-2010s, check out this paper from Daron Acemoglu and Pascual Restrepo and this paper from David Autor.

Looking for something new to read or listen to? Here are some articles, books, and academic papers that our faculty co-directors have been enjoying.

David Autor:

• This New York Times article, which describes how Schneider Electric is deploying AI to make workers more productive, rather than to replace them.

• Project Maven: A Marine Colonel, His Team, and the Dawn of AI Warfare by Katrina Manson. This book, released in March 2026, details the Pentagon’s decade-long campaign to make AI warfare a reality.

• Furious Minds: The Making of the MAGA New Right by political theorist Laura K. Fields, which explores the intellectual underpinnings of the New Right in American politics.

Daron Acemoglu:

• “The Generative AI Learning Penalty: Evidence from Chinese Secondary Education” by David Strömberg, Victor Lei, and Yanhui Wu. This paper, released in June 2026, studies how generative AI affects homework productivity and learning among students in China (grades 7-12). The researchers find that using AI enables students to complete their homework faster and more accurately, but lowers their exam scores.

• The God Test: Artificial Intelligence and Our Coming Cosmic Reckoning by Robert Wright. This book examines our evolutionary history as a species to understand how AI could bring about dramatic social transformation.

The healthcare and education sectors comprise about 25% of US GDP, and a significant share of their funding comes from the public sector. This offers an opportunity for the government to set targets to deploy AI in these sectors in ways that complement workers.

For example, AI could help teachers to assign and grade work more effectively and to provide customized resources for students — not making teachers less necessary, but making them more effective with better tools.

For more policy recommendations, see the report on “Building pro-worker AI,” authored by the Stone Center’s co-directors.

Insights from Stone Center co-directors and staff on AI, the future of work, and the global economy were featured across media outlets this month.

The Wall Street Journal: Economists Weigh In on the Future of Work and AI

Daron Acemoglu and David Autor were among the 16 top economists who shared their perspectives and predictions on AI and work in this Wall Street Journal article.

What does history suggest will happen to workers displaced by AI?

David Autor: “If it works out well, we will be vastly more productive and hence potentially wealthier. But whether those gains are broadly shared, and whether the workers whose careers are destroyed make successful transitions, these things do not depend on the technology. They depend on the societal institutions and policies we build to share the gains and redress the costs. The United States does not have a glorious history of doing either. We are currently unprepared, and most signs from Washington and Silicon Valley say: let it rip and damn the consequences.”

New York Times Magazine: Who Will Actually Thrive in the Hybrid A.I.-Human Work Force

Daron Acemoglu was interviewed in the New York Times Magazine as part of a panel discussing how job seekers should prepare for the future of work in the age of AI.

“The job of electricians, HVAC technicians, all of these are going to become more complex. We already have a huge shortage of these trades. Again, A.I. can be massively helpful. A novice electrician with the right A.I. tool could be 10 times as productive as they are at the moment. Using A.I. for training electricians with new equipment could be hugely useful. But that’s not where the center of gravity of the investments in A.I. are going.”

Project Syndicate: Why is Europe’s Economy Falling Short?

Simon Johnson interviewed fellow Nobel laureate Philippe Aghion about the European economy, creative destruction, AI, regulation, competition, and global markets. They discuss how China escaped the “middle income trap,” and whether the AI race changes the picture.

“Competition may exist in the Chinese economy, but there’s much less creative destruction in the political sphere.

I wonder whether China may be automating too much. Wages are lower than they are in Europe or the US, and China already has a high degree of automation in manufacturing and is increasingly automating the gig economy as well. That may generate productivity gains, but what happens to the displaced workers? Because China’s model is so top-down and emphasizes capital accumulation, infrastructure, and computing power, I’m not sure it can create enough good jobs to keep people happy and sustain economic growth.”

MIT Sloan: Pro-worker AI, explained

A recent article from the MIT Sloan School of Management features insights from the paper “Building Pro-Worker Artificial Intelligence,” which was co-authored by Daron Acemoglu, David Autor, and Simon Johnson and released in February.

Not every productivity gain is pro-worker: A technology can make a company more efficient while reducing the need for expertise, narrowing a job role, or shifting more value from labor to capital. The authors define pro-worker AI as AI that moves in the opposite direction: It expands what workers can do and makes their expertise more important to the work.

Inclusion Geeks: AI Policy and Work Inequality with Julia Regier

The Stone Center’s Policy and Research Manager, Julia Regier, was interviewed on the Inclusion Geeks podcast. The discussion included the latest research on wage inequality for workers with and without a college degree, the pitfalls of “so-so automation,” and what it would take to redirect AI to be pro-worker.

“When it comes to AI, the vision is really being set by a small handful of people who are leading the big tech companies, and they have been focused for years on achieving artificial general intelligence, or AGI.

And so I wouldn’t even necessarily say that firms are being so logical and so rational and calculating an ROI for replacing workers. I think some of them are. I think there’s a lot of uncertainty in those calculations … [but] some are just hopping on the bandwagon and worried about being left behind. And I think too often we all are just taking for granted what that vision for AI is without acknowledging that there could be another way.”

First, companies tend to innovate within their existing business models and product lines. In the tech industry, the monetization strategy of the few dominant players has been overwhelmingly focused on digital advertising.

There is also a coordination problem: If tech companies don’t think firms will adopt pro-worker tools, they won’t create them, and if firms don’t think these tools are coming, they won’t invest in changing their strategy or organizational structures to adopt them.

But there is also an ideological dimension: the tech industry has been dominated for decades by a vision of artificial general intelligence. AI developers are focused on building technology that mimics humans, instead of considering more expansive possibilities.

For more insights on firms and automation, check out this post:

At any given time, technology does two things to employment: It replaces traditional jobs, and it creates new lines of work. Machines replace farmers, but enable, say, aeronautical engineers to exist. So, if tech creates new jobs, who gets them? How well do they pay? How long do new jobs remain new, before they become just another common task any worker can do?

A new study of U.S. employment led by MIT labor economist David Autor sheds light on all these matters. In the postwar U.S., as Autor and his colleagues show in granular detail, new forms of work have tended to benefit college graduates under 30 more than anyone else.

“We had never before seen exactly who is doing new work,” Autor says. “It’s done more by young and educated people, in urban settings.”

The study also contains a powerful large-scale insight: A lot of innovation-based new work is driven by demand. Government-backed expansion of research and manufacturing in the 1940s, in response to World War II, accounted for a huge amount of new work, and new forms of expertise.

“This says that wherever we make new investments, we end up getting new specializations,” Autor says. “If you create a large-scale activity, there’s always going to be an opportunity for new specialized knowledge that’s relevant for it. We thought that was exciting to see.”

The paper, “What Makes New Work Different from More Work?” is forthcoming in the Annual Review of Economics. The authors are Autor; Caroline Chin, a doctoral student in MIT’s Department of Economics; Anna M. Salomons, a professor at Tilburg University’s Department of Economics and Utrecht University’s School of Economics; and Bryan Seegmiller PhD ’22, an assistant professor at Northwestern University’s Kellogg School of Management.

And yes, learning about new work, and the kinds of workers who obtain it, might be relevant to the spread of artificial intelligence — although, in Autor’s estimation, it is too soon to tell just how AI will affect the workplace.

“People are really worried that AI-based automation is going to erode specific tasks more rapidly,” Autor observes. “Eroding tasks is not the same thing as eroding jobs, since many jobs involve a lot of tasks. But we’re all saying: Where is the new work going to come from? It’s so important, and we know little about it. We don’t know what it will be, what it will look like, and who will be able to do it.”

“If everyone is an expert, then no one is an expert”

The four co-authors also collaborated on a previous major study of new work, published in 2024, which found that about six out of 10 jobs in the U.S. from 1940 to 2018 were in new specialties that had only developed broadly since 1940. The new study extends that line of research by looking more precisely at who fills the new lines of work.

To do that, the researchers used U.S. Census Bureau data from 1940 through 1950, as well as the Census Bureau’s American Community Survey (ACS) data from 2011 to 2023. In the first case, because Census Bureau records become wholly public after about 70 years, the scholars could examine individual-level data about occupations, salaries, and more, and could track the same workers as they changed jobs between the 1940 and 1950 Census enumerations.

Through a collaborative research arrangement with the U.S. Census Bureau, the authors also gained secure access to person-level ACS records. These data allowed them to analyze the earnings, education, and other demographic characteristics of workers in new occupational specialties — and to compare them with workers in longstanding ones.

New work, Autor observes, is always tied to new forms of expertise. At first, this expertise is scarce; over time, it may become more common. In any case, expertise is often linked to new forms of technology.

“It requires mastering some capability,” Autor says. “What makes labor valuable is not simply the ability to do stuff, but specialized knowledge. And that often differentiates high-paid work from low-paid work.” Moreover, he adds, “It has to be scarce. If everyone is an expert, then no one is an expert.”

By examining the census data, the scholars found that back in 1950, about 7 percent of employees had jobs in types of work that had emerged since 1930. More recently, about 18 percent of workers in the 2011-2023 period were in lines of work introduced since 1970. (That happens to be roughly the same portion of new jobs per decade, although Autor does not think this is a hard-and-fast trend.)

In these time periods, new work has emerged more often in urban areas, with people under 30 benefitting more than any other age category. Getting a job in a line of new work seems to have a lasting effect: People employed in new work in 1940 were 2.5 times as likely to be in new work in 1950, compared to the general population. College graduates were 2.9 percentage points more likely than high school graduates to be engaged in new work.

New work also has a wage premium, that is, better salaries on aggregate than in already-existing forms of work. Yet as the study shows, that wage premium also fades over time, as the particular expertise in many forms of new work becomes much more widely grasped.

“The scarcity value erodes,” Autor says. “It becomes common knowledge. It itself gets automated. New work gets old.”

After all, Autor points out, driving a car was once a scarce form of expertise. For that matter, so was being able to use word-processing programs such as WordPerfect or Microsoft Word, well into the 1990s. After a while, though, being able to handle word-processing tools became the most elementary part of using a computer.

Back to AI for a minute

Studying who gets new jobs led the scholars to striking conclusions about how new work is created. Examining county-level data from the World War II era, when the federal government was backing new manufacturing in public-private partnerships throughout the U.S., the study shows that counties with new factories had more new work, and that 85 to 90 percent of new work from 1940 to 1950 was technology-driven.

In this sense there was a great deal of demand-driven innovation at the time. Today, public discourse about innovation often focuses on the supply side, namely, the innovators and entrepreneurs trying to create new products. But the study shows that the demand side can significantly influence innovative activity.

“Technology is not like, ‘Eureka!’ where it just happens,” Autor says. “Innovation is a purposive activity. And innovation is cumulative. If you get far enough, it will have its own momentum. But if you don’t, it’ll never get there.”

Which brings us back to AI, the topic so many people are focused on in 2026. Will AI create good new jobs, or will it take work away? Well, it likely depends how we implement it, Autor thinks. Consider the massive health care sector, where there could be a lot of types of tech-driven new work, if people are interested in creating jobs.

“There are different ways we could use AI in health care,” Autor says. “One is just to automate people’s jobs away. The other is to allow people with different levels of expertise to do different tasks. I would say the latter is more socially beneficial. But it’s not clear that is where the market will go.”

On the other hand, maybe with government-driven demand in various forms, AI could get applied in ways that end up boosting health care-sector productivity, creating new jobs as a result.

“More than half the dollars in health care in the U.S. are public dollars,” Autor observes. “We have a lot of leverage there, we can push things in that direction. There are different ways to use this.”

This research was supported, in part, by the Hewlett Foundation, the Google Technology and Society Visiting Fellows Program, the NOMIS Foundation, the Schmidt Sciences AI2050 Fellowship, the Smith Richardson Foundation, the James M. and Cathleen D. Stone Foundation, and Instituut Gak.

Reprinted with permission of MIT News

When power looms automated weaving in early 19th-century Britain, profits and output in the cotton industry boomed – but weavers’ wages collapsed and stayed low for decades. New technologies do not automatically benefit workers, even when they increase total productivity.

Conventional wisdom in economics says that all technological advances which increase total productivity will ultimately benefit workers, at least on average. Technology – even automating technology – is often understood as a rising tide that lifts all boats.

However, research by Daron Acemoglu and Simon Johnson, published in the Annual Review of Economics, studies the case of weavers in the early Industrial Revolution to show that this assumption does not always hold.

In early industrial Britain, productivity surged in the cotton industry first with the mechanization of spinning and then with the invention of the power loom. Weavers benefited when the price of spun cotton declined, but when factory-based weaving machines became dominant, the wages and working conditions for handloom weavers deteriorated. The influential early 19th-century economist David Ricardo initially believed machines would necessarily benefit workers, but changed his mind after witnessing firsthand the collapse of weavers’ wages and living standards.

The British economy boomed in the early Industrial Revolution, but workers did not reap the benefits

The British cotton industry grew massively between 1780-1820, initially driven by the mechanization of spinning, which turned raw cotton into yarn. By 1813, cotton made up at least 7% of Britain’s GDP, compared with 1% in the early 1780s. Total British exports more than tripled during this period, with cotton accounting for 42% of the value of total exports in the mid-1810s. The cotton industry was one of the first to widely introduce machines in factories.

The ability to spin cotton more efficiently into yarn increased demand for weavers, who turned the yarn into fabric. In the late 1700s and early 1800s, fabric was produced by handloom weavers who worked out of their homes or small workshops. The profession was relatively well-paid, and workers had independence and autonomy over their schedules.

But between 1800 and 1820, the power loom spread throughout Britain, automating the work of weaving. As a result, wages for handloom weavers collapsed. Figure 1 shows the steep drop in handloom weaver wages in the early 1800s. Handloom weavers had previously earned twice as much as factory workers. By 1820, they earned 25% less.

Figure 1: Wages of Handloom Weavers vs. Cotton Factory Workers, 1806-1846

The real wages of handloom weavers – wages adjusted for inflation and purchasing power – dropped to about a quarter of their peak level, and then continued to decline. Despite this collapse in earnings, hundreds of thousands of handloom weavers remained in the profession and struggled to survive.

Why didn’t they move into other occupations? There is evidence to suggest that handloom weavers resisted better-paying factory jobs because of the extremely poor working conditions and lack of autonomy. In addition, for at least the first two decades of the 19th century, cotton factories did not grow fast enough to employ many handloom weavers. There was no other booming sector in the British economy in which to find work, especially in Lancashire, where many handloom weavers lived. The main complementary activity to weaving – spinning – was already mechanized. Some new jobs were created (e.g., repairing the power looms), but not enough to compensate for the handloom weavers who were displaced by machines.

In 1820, there were still double the number of handloom weavers in Britain as there were factory workers. As Figure 2 shows, it wasn’t until other types of factory work expanded in the 1820s that factory employment took off.

Figure 2: Employment of Handloom Weavers vs. Cotton Factory Workers, 1806-1846

Real wages for all workers in the British economy largely remained flat from the late 1700s into the 1820s. In fact, real wages declined sharply in the sectors with the most notable productivity gains, including the cotton industry. While entrepreneurs and factory owners made huge profits from the mechanization of weaving, workers did not share in the gains.

In addition, working conditions and quality of life for workers deteriorated with the arrival of the power loom and the growth of factories. Factory workers not only had little autonomy, but were forced to work in unsafe conditions and live in unsanitary housing. Crucially, workers did not have the legal right to bargain collectively or the political voice to advocate for higher wages and better working conditions. The historian E.P. Thompson argued that the factory system fundamentally shifted power away from workers, and that this power imbalance was critical to determining who benefited from new technology.

New technologies can increase workers’ wages, but only when they create new tasks or new demand in complementary sectors – and when workers have sufficient power

As the case of handloom weavers illustrates, workers do not automatically benefit from new technology. Technologies that automate work can reduce the demand for labor. Unless the technology also creates new tasks for workers to perform, or spurs growth in related industries, wages are likely to decline. Even if these conditions are met, workers are unlikely to benefit from new technology unless they have enough power and political voice to push for their share of the productivity gains.

Today, AI may increase average productivity, but there is no guarantee that these gains will be shared with workers. Previous waves of automation since 1980 have brought down real wages for affected workers. If used mainly to automate work, AI could replace many workers and worsen job quality for others (e.g., through surveillance). However, if AI is used to create new tasks for workers and complement their capabilities, it is more likely that workers will share in the benefits. Even in this latter scenario, history teaches us that workers will need adequate political and bargaining power to secure their share of the gains from AI.

Source: Daron Acemoglu and Simon Johnson, “Learning From Ricardo and Thompson: Machinery and Labor in the Early Industrial Revolution and in the Age of Artificial Intelligence,” Annual Review of Economics (2024), https://doi.org/10.1146/annurev-economics-091823-025129

Insights from Stone Center co-directors Daron Acemoglu, David Autor, and Simon Johnson on technology and the future of work were featured across media outlets this month.

Slate Money: The AI Job Apocalypse is Avoidable

On the Money Talks podcast, Daron Acemoglu discussed what it would take to make the vision of pro-worker AI a reality.

“We are moving in the direction of AI replacing people because that’s what the tech companies are thinking about, that’s what they’re targeting, that’s what their business model is, that’s what their efforts are seeking to achieve. Our point is that there’s an alternative direction, but it’s not one that’s going to happen by itself.”

WSJ: What Is a ‘Compute Tax’ and Why Is the Idea Gaining Traction?

Simon Johnson was quoted in the Wall Street Journal about a possible tax on AI processing.

A compute tax is a sensible policy lever to consider in order to slow down automation, says Simon Johnson, a Nobel laureate and professor of entrepreneurship at the MIT Sloan School of Management. The tax could make it less appealing to fire tens of thousands of workers and build data centers, he adds. As AI advances, there should also be new things for humans to do. “You want to create new tasks, you want to boost the demand for labor and that’s where the tech sector is kind of falling down,” he says.

The Economics of Work: David Autor – How Technology Affects Work

David Autor discussed how new technologies transform work with Ben Zweig, CEO of Revelio Labs, in the inaugural episode of the Economics of Work podcast.

“If my supporting tasks are automated, great. If my expert tasks are automated, my labor is all of a sudden commodified. Both things can go on, and I think people very intuitively understand this. They don’t think all automation is bad; they don’t think all of it is good. They know, when does automation threaten your livelihood? When it makes your specialized labor no longer scarce.”

MIT Technology Review: Three things in AI to watch, according to a Nobel-winning economist

Daron Acemoglu shared his views on some of the latest developments in AI, including AI agents, AI apps, and tech companies’ recruitment of economists, in the MIT Technology Review.

Whether or not agents will supercharge AI’s impact on jobs will come down to whether they can eventually handle the orchestration between tasks that humans do naturally. AI companies are in heated competition to prove that their AI agents can work independently for ever longer periods without making mistakes, sometimes exaggerating the results—but Acemoglu says many jobs will be spared from an AI takeover if agents can’t fluidly switch between tasks.

When we hear about automation and artificial intelligence replacing jobs, it may seem like a tsunami of technology is going to wipe out workers broadly, in the name of greater efficiency. But a study co-authored by an MIT economist shows markedly different dynamics in the U.S. since 1980. 

Rather than implement automation in pursuit of maximal productivity, firms have often used automation to replace employees who specifically receive a “wage premium,” earning higher salaries than other comparable workers. In practice, that means automation has frequently reduced the earnings of non-college-educated workers who had obtained better salaries than most employees with similar qualifications. 

This finding has at least two big implications. For one thing, automation has affected the growth in U.S. income inequality even more than many observers realize. At the same time, automation has yielded a mediocre productivity boost, plausibly due to the focus of firms on controlling wages rather than finding more tech-driven ways to enhance efficiency and long-term growth.

“There has been an inefficient targeting of automation,” says MIT’s Daron Acemoglu, co-author of a published paper detailing the study’s results. “The higher the wage of the worker in a particular industry or occupation or task, the more attractive automation becomes to firms.” In theory, he notes, firms could automate efficiently. But they have not, by emphasizing it as a tool for shedding salaries, which helps their own internal short-term numbers without building an optimal path for growth.

The study estimates that automation is responsible for 52 percent of the growth in income inequality from 1980 to 2016, and that about 10 percentage points derive specifically from firms replacing workers who had been earning a wage premium. This inefficient targeting of certain employees has offset 60-90 percent of the productivity gains from automation during the time period.

“It’s one of the possible reasons productivity improvements have been relatively muted in the U.S., despite the fact that we’ve had an amazing number of new patents, and an amazing number of new technologies,” Acemoglu says. “Then you look at the productivity statistics, and they are fairly pitiful.”

The paper, “Automation and Rent Dissipation: Implications for Wages, Inequality, and Productivity,” appears in the May print issue of the Quarterly Journal of Economics. The authors are Acemoglu, who is an Institute Professor at MIT; and Pascual Restrepo, an associate professor of economics at Yale University.

Inequality implications

Dating back to the 2010s, Acemoglu and Restrepo have combined to conduct many studies about automation and its effects on employment, wages, productivity, and firm growth. In general, their findings have suggested that the effects of automation on the workforce after 1980 are more significant than many other scholars have believed. 

To conduct the current study, the researchers used data from many sources, including U.S. Census Bureau statistics, data from the bureau’s American Community Survey, industry numbers, and more. Acemoglu and Restrepo analyzed 500 detailed demographic groups, sorted by five levels of education, as well as gender, age, and ethnic background. The study links this information to an analysis of changes in 49 U.S. industries, for a granular look at the way automation affected the workforce. 

Ultimately, the analysis allowed the scholars to estimate not just the overall amount of jobs erased due to automation, but how much of that consisted of firms very specifically trying to remove the wage premium accruing to some of their workers. 

Among other findings, the study shows that within groups of workers affected by automation, the biggest effects occur for workers in the 70th-95th percentile of the salary range, indicating that higher-earning employees bear much of the brunt of this process. 

And as the analysis indicates, about one-fifth of the overall growth in income inequality is attributable to this sole factor.

“I think that is a big number,” says Acemoglu, who shared the 2024 Nobel Prize in economic sciences with his longtime collaborators Simon Johnson of MIT and James Robinson of the University of Chicago.

He adds: “Automation, of course, is an engine of economic growth and we’re going to use it, but it does create very large inequalities between capital and labor, and between different labor groups, and hence it may have been a much bigger contributor to the increase in inequality in the United States over the last several decades.” 

The productivity puzzle

The study also illuminates a basic choice for firm managers, but one that gets overlooked. Imagine a type of automation — call-center technology, for instance — that might actually be inefficient for a business. Even so, firm managers have incentive to adopt it, reduce wages, and oversee a less productive business with increased net profits.

Writ large, some version of this seems to have been happening to the U.S. economy since 1980: Greater profitability is not the same as increased productivity.

“Those two things are different,” says Acemoglu. “You can reduce costs while reducing productivity.” 

Indeed, the current study by Acemoglu and Restrepo calls to mind an observation by the late MIT economist Robert M. Solow, who in 1987 wrote, “You can see the computer age everywhere but in the productivity statistics.” 

In that vein, Acemoglu observes, “If managers can reduce productivity by 1 percent but increase profits, many of them might be happy with that. It depends on their priorities and values. So the other important implication of our paper is that good automation at the margins is being bundled with not-so-good automation.” 

To be clear, the study does not necessarily imply that less automation is always better. Certain types of automation can boost productivity and feed a virtuous cycle in which a firm makes more money and hires more workers. 

But currently, Acemoglu believes, the complexities of automation are not yet recognized clearly enough. Perhaps seeing the broad historical pattern of U.S. automation, since 1980, will help people better grasp the tradeoffs involved — and not just economists, but firm managers, workers, and technologists. 

“The important thing is whether it becomes incorporated into people’s thinking and where we land in terms of the overall holistic assessment of automation, in terms of inequality, productivity and labor market effects,” Acemoglu says. “So we hope this study moves the dial there.”

Or, as he concludes, “We could be missing out on potentially even better productivity gains by calibrating the type and extent of automation more carefully, and in a more productivity-enhancing way. It’s all a choice, 100 percent.”

Reprinted with permission of MIT News

Looking for something new to read or listen to? Here are some articles, books, podcasts, and academic papers that our faculty co-directors have been enjoying.

David Autor:

• The Substack Labor Matters from Gad Levanon, Chief Economist at The Burning Glass Institute. The newsletter offers practical economic insights on labor market trends.

• The Dwarkesh Podcast with Dwarkesh Patel, which features well-researched interviews with prominent tech leaders.

• This New York Times article, which tells the story of 85-year old Jan Worrell and ElliQ, an AI robot designed to be a companion for seniors.

Simon Johnson:

• The Sparrow by Mary Doria Russell. This visionary science fiction novel, published in 1996, explores themes of knowledge, language, AI, the Catholic Church, and aliens — and the implications for jobs.

If you haven’t listened already, you may also enjoy co-director Simon Johnson’s new podcast with former SEC chair Gary Gensler, “Power and Consequences.” The first ten episodes cover topics ranging from tariffs to AI policy to wealth taxes, synthesizing historical perspectives with policy analysis.

A new premium may emerge, favoring those with the skills or the luck to benefit from AI, and those who have an ownership stake in AI.

More insights on AI and work from the Stone Center

Policy paper: “Building Pro-Worker AI” (February 2026) by Daron Acemoglu, David Autor, and Simon Johnson: https://www.hamiltonproject.org/publication/paper/building-pro-worker-ai/

Panel discussion at the launch of the Stone Center: “Pro-Worker AI: What Is It, and How Do We Do It?” (November 2025), featuring Daron Acemoglu, Zana Buçinca, and Ethan Mollick:

Podcast episode: The Weekly Show with Jon Stewart (April 2026) featuring Daron Acemoglu and David Autor on AI and the future of work:

Podcast episode: Power and Consequences with Simon Johnson and Gary Gensler (April 2026) on AI policy in the US:

Power and Consequences

Episode Seven: Artificial Intelligence

In the latest version of our Power and Consequences podcast, we (Gary Gensler and Simon Johnson) talk about Artificial Intelligence, with a particular focus on what is and should be policy for AI in the United States…

Listen now

5 months ago

New work – novel job tasks that demand human labor – is a significant component of total employment. Between 2011-2023, 18% of US workers were employed in jobs introduced since 1970.

In a new paper, forthcoming in the Annual Review of Economics, David Autor, Caroline Chin, Anna Salomons, and Bryan Seegmiller show that this new work differs systematically from more jobs in existing work.

Understanding the drivers and characteristics of new work is important in an era of increasing automation. Since the early 2000s, the “labor share” – the share of a country’s national income that goes to workers, rather than to owners of machines and capital – has declined 10% in the US. A similar pattern holds in many industrialized countries. New work helps counteract the effects of automation, which displaces workers and reduces the labor share.

The researchers use granular data from the linked US Census of Populations from 1940 and 1950 and the American Community Survey (ACS) for 2011 through 2023, including write-in responses from individuals who report their occupations. These data allow them to analyze the characteristics of new work over time, and who is employed in this new work.

Finding #1: Workers in new work are younger and more educated

Who performs new work? Younger and more educated workers are more likely to hold new job titles. This may be because it is easier for these workers to obtain the new skills and expertise needed in new work.

Across time periods, workers below age 30 are the most likely to be employed in new work, followed by workers ages 30-54. Workers ages 55-64 are the least likely to be employed in new work.

More educated workers are consistently more likely to be employed in new work, both overall and within occupational categories. Between 2011-2023, workers with advanced degrees were 2.9 percentage points more likely to be employed in new work than high school graduates.

The types of new work that have emerged recently differ from those that emerged several decades ago. In 1950, a large portion of new work was in production, clerical, and administrative occupations. In 2011-2023, the largest share of new work was in professional and managerial occupations.

Figure 1: Distribution of New Work by Broad Occupation

These professional and managerial positions are more likely to be held by workers with a college degree. In recent years, new work for workers without a college degree is more likely to be in health services and personal services. These jobs tend to be lower-paying and less specialized than earlier work in production, clerical, and administration occupations.

Workers in new work in 1940 were disproportionately likely to be employed in new work in 1950, including in different new work. This suggests that workers gain specialized expertise and skills in new work that continue to be valuable in the labor market.

Finding #2: Workers earn higher wages in new work, reflecting the value of their scarce, specialized expertise

Across time periods, workers in new work earn a “wage premium” – higher wages than workers with similar characteristics in existing occupations. The wage premium is higher for “newer” new work and declines over time.

These findings suggest that new work requires valuable expertise (capability, skills, human capital, and/or know-how). They also indicate that the wage premium for new work reflects temporary scarcity in the skills required, rather than being a permanent feature of new occupations. As more workers gain these skills and the expertise becomes less scarce over time, the wage premium declines. The novel human expertise that new work requires fundamentally distinguishes new work from pre-existing work.

Figure 2: New Work Wage Premium by Vintage, 2011-2023

The wage premium is four times larger for new work associated with technological change than for other types of new work. Other types of new work generally reflect changing tastes, income levels, and demographics – for example, the new job titles of “mental health aide” in 1970 and “conference planner” in 1990. The technology-linked wage premium is also more persistent over time. This likely reflects the fact that new work associated with technology often requires more specialized and scarcer expertise.

Finding #3: Public policy can drive the emergence of new work

Other research has shown how technological change and demand shifts can drive the creation of new work. This new research offers a novel example in which public policy seems to have directly contributed to the emergence of new work.

In the 1940s, the US government commissioned large manufacturing plants for production during World War II. Garin and Rothbaum (2025) show that these investments led to increased employment and earnings for men born in the counties where plants were built, primarily by bringing more high-wage, blue-collar jobs to these locations.

Autor and co-authors find that locations where plants were built generated not only more work, but more new work. The share of employment in new work was 0.35 percentage points higher in 1950 in counties where plants were built, compared to other counties with similar characteristics. The difference was roughly twice as large in the manufacturing sector. Workers living in these counties in 1940 were more likely to be employed in new work in 1950.

There was no difference in the prevalence of new work between the same places in 1940, suggesting that the policy had a causal effect. The construction of plants during WWII seems to have helped propel local workers into higher-paying, more expertise-intensive work.

Conceptual schematic: New work, expertise scarcity, and wage premia

Conclusion: New work counteracts the job-displacing effects of automation

Not only does new work create new tasks for workers, but it also generates new demand for scarce human expertise. Because this expertise is scarce, workers performing new work command a wage premium. Though new work emerges from many sources, the United States post-WWII history demonstrates that public policy can drive the emergence of new work. New work is ultimately a central mechanism for maintaining and extending the value of human expertise in the face of automation.

Source: David Autor, Caroline Chin, Anna M. Salomons, and Bryan Seegmiller, “What Makes New Work Different from More Work?,” NBER Working Paper 34986 (2026), https://doi.org/10.3386/w34986.