The Great Decoupling
The Great Decoupling describes the historical break, documented rigorously by Brynjolfsson and Andrew McAfee in The Second Machine Age (2014), between productivity and median compensation in the American economy. For roughly three decades after World War II, these two curves moved together — when the economy became more productive, median workers saw their wages rise proportionally. Beginning in the late 1970s, the curves diverged. Between 1973 and 2016, American productivity roughly doubled. Median household income, adjusted for inflation, grew by only about twenty percent. The gap represented enormous economic value generated by the economy's expanding capacity but not reaching the median worker — flowing instead upward to executives, capital owners, and the technologically elite. The AI transition, if its gains distributed along the same lines as previous digital technology gains, threatened to accelerate the divergence to a point that threatened social cohesion.
In the AI Story
The decoupling was not caused solely by technology, but technology was a primary driver. Digital systems disproportionately augmented high-skilled workers while displacing or deskilling middle-skilled occupations — the pattern labor economists call job polarization. The factory worker monitoring a machine was replaced by the machine monitoring itself. The office clerk processing information was replaced by software. Globalization, institutional changes, and union decline contributed, but the technology-mediated mechanisms were the most directly relevant to understanding what AI would mean.
Brynjolfsson and McAfee organized their analysis around two concepts: bounty (the total economic value created) and spread (the distribution of that value). The two were not opponents in a zero-sum contest. They were twin dimensions of a single phenomenon: technological progress that expanded the total pie while distributing slices with increasing unevenness. The bounty from digital technologies was enormous and growing. The spread was widening with each passing year.
The AI transition arrived into this already-decoupled landscape. The argument for further decoupling was straightforward: AI amplified the most skilled, so the premium for being at the top of the skill distribution would grow. The argument against was more nuanced: AI also expanded who could participate, lowering barriers to entry for work that had previously required specialized training. Brynjolfsson's empirical work — particularly Generative AI at Work — showed AI helping the least skilled the most, but separate data showed entry-level positions in AI-exposed occupations disappearing. Both dynamics were true simultaneously.
The historical parallel Brynjolfsson invoked repeatedly was the mid-twentieth-century response to earlier technology transitions. The industrial economy had generated wealth on an unprecedented scale, but the institutional infrastructure that translated that wealth into broadly shared prosperity — labor legislation, the eight-hour day, universal public education, social insurance — had been built over generations at enormous political cost. The AI transition was compressing the timeline, leaving less room for the slow institutional learning that had historically managed distributional consequences.
Origin
The term Great Decoupling appears throughout Brynjolfsson and McAfee's work, most fully developed in The Second Machine Age (2014) and its sequel Machine, Platform, Crowd (2017). The empirical documentation drew on Bureau of Labor Statistics data showing the divergence between labor productivity and median worker compensation beginning in the 1970s, a pattern documented by many economists but integrated by Brynjolfsson and McAfee into a specific analysis of digital technology's distributional consequences.
The intellectual lineage runs through David Autor's work on job polarization, Daron Acemoglu and Pascual Restrepo's framework on displacement versus reinstatement effects, and Claudia Goldin and Lawrence Katz's race-between-education-and-technology framework.
Key Ideas
The coupling broke in the late 1970s. Productivity continued rising; median wages stopped keeping pace, with the gap widening each decade.
Bounty and spread are distinct. Total economic value (bounty) expanded while distribution (spread) concentrated — both dimensions required separate analysis.
Technology is a primary mechanism. Digital systems disproportionately augment high-skilled workers while displacing middle-skill occupations, producing job polarization.
AI threatens acceleration. Without deliberate distributional intervention, AI gains will likely concentrate along the same lines as previous digital technology gains, at faster speed.
Distributional infrastructure is built, not automatic. The institutions that translated previous technology gains into broadly shared prosperity were political achievements, not natural consequences.
Debates & Critiques
The empirical magnitude of the decoupling is debated — some economists argue measurement issues (especially around fringe benefits and consumer surplus) overstate the gap. More consequentially, economists disagree about the relative weight of technology versus other causes (globalization, institutional decline, policy choices). The AI-specific debate concerns whether the tool will compress or widen the skill distribution, with Brynjolfsson's own empirical work producing evidence for both directions depending on which population segment and time horizon is examined.
Appears in the Orange Pill Cycle
Further reading
- Brynjolfsson, Erik and Andrew McAfee. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. W.W. Norton, 2014.
- Brynjolfsson, Erik and Andrew McAfee. Machine, Platform, Crowd: Harnessing Our Digital Future. W.W. Norton, 2017.
- Autor, David. The China Shock: Learning from Labor-Market Adjustment to Large Changes in Trade. Annual Review of Economics, 2016.
- Goldin, Claudia and Lawrence Katz. The Race Between Education and Technology. Harvard University Press, 2008.