Beyond U-3: A County-Level Comprehensive Labor Underutilization Index

The Bureau of Labor Statistics U-3 unemployment rate captures a single dimension of labor market health. This paper constructs a County-Level Comprehensive Labor Underutilization Index (CLUI) using Fay-Herriot small-area estimation on 3,082 U.S. counties. The index combines three layers of distress: official unemployment, labor force non-participation, and poverty-wage employment. Nationally, poverty-wage employment accounts for 37 percent of the CLUI, the participation gap for 25 percent, and unemployment for just 15 percent. The Fay-Herriot EBLUP achieves 78 percent variance reduction over direct survey estimates. Five contributions follow. First, the CLUI predicts county premature mortality 3.6 times better than U-3 (R-squared = 0.417 versus 0.116) and disability non-participation 3.1 times better (R-squared = 0.419 versus 0.134); neither outcome enters the CLUI construction. Under spatial buffered cross-validation, the CLUI reduces premature death prediction error by 39.3 percent. Second, Vuong (1989) and Williams (1959) tests confirm CLUI superiority in 12 of 20 pairwise comparisons against five competing indices. Third, spatial autocorrelation is substantial (Moran’s I = 0.603); a spatial Fay-Herriot extension preserves baseline rankings (Spearman rho = 0.987), and the Chen (2026) diagnostic indicates that standard shrinkage understates distress in small rural counties. Fourth, 292 hidden-distress counties representing 6.6 million residents have official unemployment below 4 percent but severe comprehensive underutilization. Fifth, these counties shifted toward the Republican ticket in 2024, though this association does not survive spatial HAC correction. Results are stable across ten robustness checks.