Occupational complexity explains four decades of U.S. wage inequality
A Federal Reserve Board paper reveals that occupational problem complexity drives U.S. wage inequality growth since 1980. Researchers Colin Caines, Florian Hoffmann, and Gueorgui Kambourov show that technological forces shifted from capital automation to task augmentation around 2000.
Two decades of shifting technology
The study combines Census, ACS, and O*Net data across 317 occupations to evaluate wage dynamics from 1980 to 2020.
The wage premium for high-complexity occupations grew by over 60 log points, while employment shares shifted by only 0.5 percentage points.
Structural estimation identifies two distinct periods of technological change.
Prior to 2000, falling equipment prices and capital-skill complementarity drove inequality, with capital-labor substitution elasticities ranging from 0.5 for complex roles to 1.6 for simple tasks.
Post-2000, growth stemmed from supply-side task augmentation, as occupations became 1.5 times more efficient at utilizing worker problem-solving skills.
Bridging task automation and skills
The paper unifies two major economic frameworks: skill-biased technological change and task-based automation.
By defining complexity through psychological problem-solving requirements rather than automation risk, the authors isolate skill comparative advantage from capital substitutability.
Counterfactual simulations reveal that without capital-skill complementarity, the wage gap between high- and low-complexity roles would have increased by only half its observed value.
Additionally, average capital-labor substitution is estimated at 0.80 to 0.85 across the economy.
A unifying model with empirical bite
This paper successfully unifies capital-skill complementarity and task automation into one coherent macroeconomic framework.
Yet, relying on static O*Net task descriptors potentially masks ongoing intra-occupational skill shifts.
Nevertheless, the structural insight offers a valuable template for assessing future artificial intelligence shocks.