Monthly state-level full-employment unemployment rates for 2001–2025 reveal persistent regional differences. Using national rather than state benchmarks changes whether a state is classified as slack, near full employment, or tight in 15.5 percent of state-months.
The paper lets the unemployment-gap coefficient in Taylor-type rules vary with cross-state dispersion. In a pre-COVID VARX shock replay, dispersion-based attenuation lowers stabilization loss by 4 to 15 percent relative to matched benchmarks and outperforms constant attenuation. The national gap is also less informative about inflation when dispersion is high.
This paper demonstrates that the extent to which unemployed human capital is heterogeneous and specific, rather than homogeneous and general, plays a key role in the labor market, accounting for a significant share of joblessness in the pre-COVID era.