Task-based labor market
This is a calibrated production function in which every parameter is sourced and can be swapped out, so you can see which assumption drives which result.
The problem
There are now seven major measures of AI exposure, and they disagree. The measures built before LLMs are negatively correlated with the ones built after. Meanwhile, claims about AI's effect on wages and employment have gotten ahead of the evidence. Another estimate would not settle this. What helps is a framework where every parameter is disclosed, sourced and swappable, so that when results differ you can see which assumption made them differ.
How it works
01The economy as tasks
The model has three nested levels. An industry combines tasks in stable expenditure shares, each task is produced from capital and labor with its own substitution elasticity, and the labor input is a bundle of skills. Tasks that resist automation and tasks that are exposed differ in a single parameter.
02Calibrated from published sources
No elasticity is estimated from market data and no equilibrium is solved. Every parameter is calibrated to a published source, such as staffing matrices, task-importance profiles, output weights and skill proximities, and records where it came from. The model can be defended because every number can be traced.
03Exposure as a pluggable adapter
The AI-exposure parameters come from interchangeable adapters that share one interface, with several published measures available. Results are reported for each measure as a standard output, so the disagreement between measures is shown directly.
04Validated against observed shifts
The headline test is predictive: does calibrated task-level exposure anticipate observed wage shifts better than flat per-occupation scores and a zero-information null? Validation runs on healthcare occupations, where within-industry variation in substitutability is largest.
The structure nests Cobb-Douglas over tasks, capital against labor within each task, and skills within labor, and the two engine outputs follow directly from it.
None of the parameters are estimated structurally. Each is traceable to a published source.
Tools used
Final state
The engine is built and its outputs are cached at scale: skill exposure, role resilience, and a task atlas of several hundred thousand calibrated rows, each carrying its adapter and calibration vintage. The framework is also used for research, and a companion paper on measuring labor market resilience and AI exposure is in progress. The equations above are the ones the production system uses. Worked magnitudes are not shown here.