Rising complexity premia explain four decades of US wage inequality
A Federal Reserve Board study shows that occupational problem-solving complexity strongly predicts US wage growth since 1980. While capital-skill complementarity drove rising skill premia until 2000, post-2000 wage patterns reflect supply-side task augmentation within occupations.
The race between technology and skills
Analyzing 317 occupations across five decades, the authors establish a continuous measure of occupational problem complexity using O*Net data.
The quantitative general equilibrium model shows that wage premia for complex problem solving grew monotonically since 1980, while overall occupational employment shifts remained modest.
Capital-skill complementarity plays a central role: the estimated elasticity of substitution between equipment capital and labor ranges from 0.55 in the most complex occupations to 1.6 in the least complex, aggregating to an economy-wide average of 0.80 to 0.85. Cheaper equipment capital induced firms to substitute away from low-complexity labor, driving wage inequality prior to 2000.
A structural break at the turn of the century
The study uncovers a structural break around 2000 in the drivers of wage inequality.
Before 2000, selection-corrected skill prices outgrew raw occupational wages due to falling equipment capital costs.
After 2000, aggregate skill price growth stalled, yet raw wage gaps continued to widen.
The researchers attribute this post-2000 divergence to supply-side technological change, as occupations became more efficient at utilizing worker skills for complex problem solving.
Bridging two paradigms of labor economics
The study neatly bridges skill-biased technological change and task automation within a single framework.
Yet, abstracting from search frictions and regional barriers oversimplifies short-run labor market adjustments.
Nonetheless, it provides a vital quantitative foundation for understanding how AI may reshape skill returns.