✨ Daron Acemoglu: too much faith in government, too little in markets (and AI)

The new Nobel laureate should offer more Up Wing optimism based on economic history and current technological progress

MIT economist and freshly minted Nobel laureate Daron Acemoglu (with whom I podcasted in 2017) has outlined three emerging “epochal” challenges for the American economy. If you tried to guess them, you would probably nail all three: global population aging, artificial intelligence, and deglobalization. As Acemoglu concedes in his New York Times essay earlier this week, “There should be little surprise in this, since all these are evolving slowly in plain sight.”

Obvious doesn’t necessarily mean unimportant, of course. Acemoglu is correct to be worried about the economic impact of all three trends on workers, including labor shortages, job market disruption, and an unprepared workforce for changing manufacturing and trade patterns. His main macro-solution is as predictable as his economic diagnosis: investing in workers' skills and adaptability. 

Again, obvious but also important! The great policy ideas are worth repeating and evangelizing — especially ones with broad consensus among economists. For example:

• My AEI colleague Michael Strain says America should “invest in all workers” rather than trying to restore the manufacturing status quo of mid-20th Century America. This involves increasing earned-income subsidies, investing in skills training to boost wages, and removing barriers from social policies and anticompetitive labor practices, including occupation licensing schemes.

• In the recent study “Technological Disruption in the US Labor Market,” economists David Deming, Christopher Ong, and Lawrence H. Summers argue that in the near term, AI will likely raise expectations for knowledge workers rather than replace them. This shift requires increased public investment in STEM education and training. Reskilling “will be necessary to help workers adapt to and effectively use new technologies.

Great stuff. A good chunk of the current presidential election should concern ideas like these. That said, I have deep reservations about the following concept, as put forward by Acemoglu:

We need a broad national strategy so that A.I. doesn't only automate work and sideline workers, but creates new tasks and competencies for them. This isn't just because of the inequality that rapid A.I.-based automation could create or the fear of tech elites that the resulting joblessness will bring out the pitchforks. Evidence suggests that new technologies increase productivity much more consistently when they work with workers, enabling them to perform their jobs better and allowing them to expand into new, more sophisticated tasks.

Acemoglu is talking about a lot more than worker training here. He wants a broad federal effort to encourage worker-augmenting technologies that enhance human capabilities across industries. It’s the same idea expressed in Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity that he coauthored with fellow Nobel winner Simon Johnson (with whom I podcasted in 2023 about the book). The two economists want government to “redirect” and “rechart” technological progress and innovation so that it makes workers more productive in what they do and creates new things for them to do. That, rather than “just automating work, making workers redundant, or intensifying surveillance.”

In addition to worker training programs, they suggest “subsidies and support for more worker-friendly technologies, tax reform … data-ownership and data-protection schemes, breaking up of tech giants, and digital advertisement taxes.” Similarly in the NYT essay, Acemoglu calls for establishing a “new federal agency tasked with identifying and funding the types of A.I. that can increase worker productivity and help us deal with looming labor shortages.”

Here are my main reservations: