Generated by Codex with GPT 5.6 Sol XHigh
Daron Acemoglu occupies an unusual position in economics: he is both one of the field’s most accomplished scholars and a target of mounting unease among his peers. The Massachusetts Institute of Technology professor shared the 2024 Nobel prize for research on how institutions shape prosperity, has produced hundreds of papers and routinely influences public debate. The article’s concern is not that his work is incompetent. It is that prestige, productivity and technical sophistication may have given several contestable ideas more authority than their evidence deserves.
A landmark result with fragile foundations
Acemoglu’s reputation rests heavily on a 2001 paper written with Simon Johnson and James Robinson. It argues that colonial conditions helped determine the institutions countries inherited. Where European settlers faced high mortality, colonial powers tended to build extractive systems designed to concentrate wealth and remove resources. Where settlement was safer, they created more inclusive institutions, leaving a legacy that helps explain why some countries are richer today.
The paper changed how economists think about institutions, but its celebrated empirical result has not emerged unscathed. Some mortality figures were borrowed from other countries, and critics have found that revising the data can make the estimates unreliable. Acemoglu counters that prominent reanalyses omit important countries and make statistical choices that introduce their own problems. He says he would now change parts of the estimation, though not the underlying mortality data.
A newer paper on falling birth rates raises a related concern. It finds that lower fertility has historically been associated with faster growth in output per working-age person. The authors acknowledge that past relationships may be poor guides to countries now facing unprecedented demographic decline. Yet that caveat can disappear beneath the paper’s mathematical sophistication, leaving a stronger impression than the evidence warrants.
When a grand theory explains too much
The article’s deeper criticism is that Acemoglu’s central idea can become difficult to test. Saying that nations prosper with good institutions sounds persuasive, but “institutions” can encompass laws, norms, enforcement and culture. Explaining institutional change through earlier institutions, critical junctures and gradual drift risks a circular story that works best in hindsight. China’s rapid growth under a political system often described as extractive also complicates a simple division between inclusive and extractive regimes.
Acemoglu replies that understanding change requires tracing specific institutions and historical events, not invoking an empty tautology. His framework remains valuable because it directs attention to power and the rules governing economic life. The question is whether it predicts enough, rather than merely organizing history after the fact.
His account of technology is similarly skeptical. In Power and Progress, written with Johnson, innovation often appears as a story of elite capture and harmful side effects. The cotton gin strengthened slavery; the Haber-Bosch process enabled new weapons. The article argues that this emphasis understates the more basic fact that technological progress has made ordinary people vastly richer than they were before industrialization.
Why the AI debate raises the stakes
The dispute matters most in Acemoglu’s work on artificial intelligence. He has projected only modest productivity gains from AI over a decade, partly because his analysis excludes the transformative potential of products that do not yet exist. He concedes that he did not anticipate agentic AI and may have underestimated the technology’s value for scientific discovery. Even so, he defends his method as a corrective to extravagant forecasts.
At the same time, Acemoglu warns that AI could damage jobs and democracy. His preferred alternative is “pro-worker AI” that makes human labor more valuable instead of replacing it. That goal is attractive, but the article finds it underspecified. Who should steer development, and how should anyone decide whether a system complements workers or displaces them? A call signed by many prominent economists repeats the aspiration without resolving those questions.
The closing lesson is about intellectual power. Acemoglu’s scholarship has opened productive lines of inquiry, and even critics describe him as brilliant and generous. But reputation should invite more scrutiny, not less. When one economist’s framework begins to shape how governments and researchers approach institutions, technology and AI, elegant models and worthy intentions are not enough; the assumptions, evidence and policy mechanisms must remain open to challenge.