Tail Risk Report
This is the full stress test prepared for policy planning. It is not a base-case forecast. It is a sizing tool for severe but plausible AI labor-market disruption through 2040.
Scenario framing
The report sizes the policy-relevant support population, not just official unemployment. That means it counts AI-linked unemployed, underemployed, and detached workers together when considering program demand.
2035 planning metrics
- Severe tail support population: 17.6 million
- Extreme plausible support population: 31.8 million
- Severe official unemployment: 7.4 percent
- Extreme official unemployment: 10.1 percent
- Severe labor-force participation: 58.3 percent
- Extreme labor-force participation: 55.2 percent
Wage pressure
The report models cumulative wage-level gaps versus a no-tail-shock reference path. By 2035 in the extreme plausible case, lower-barrier human tasks are about 21 percent below the reference path, mid-barrier tasks about 15 percent below, and higher-barrier tasks about 6 percent below.
Program sizing implication
The report argues that agencies should size for a broad support population rather than for the headline unemployment rate alone, and should consider multi-lane responses such as unemployment support, wage insurance, retraining with hiring channels, and support for detached workers.