Job-Finding Rates Drop in Unusual Pattern for Expansion

Those two patterns are anomalies for this stage of a cycle, the authors note. The weakening pipeline for nonparticipants may reflect a normalization after the unusually strong post-pandemic labor demand. The drop for skilled workers points to something else: technology-related downsizing, managerial restructuring, government contractor reductions, or early effects of AI-related displacement in professional occupations.

The demographic split also differs from the last sustained decline in job finding. During the 2007–09 recession, the drop was broad-based across education levels. The unemployment-to-employment rate fell about 13 percentage points, a 45% decrease, and the nonparticipation-to-employment rate dropped 1.3 percentage points, a 24% decrease. Both were more than double the current declines. This time, the pain is split by education: lower-educated nonparticipants and higher-educated unemployed workers.

The authors used a shift-share analysis to separate compositional changes in the pool of unemployed people from actual rate changes by demographic group. They found the decline from unemployment reflects mostly rate declines by age and education, not a shift toward groups with typically lower rates. The nonparticipation decline reflects rate decline plus some contribution from an aging population.

Nonparticipants matter more than their low job-finding rate suggests. Even though only 4–5% of them find work in a given month, they make up two-thirds of all people who transition from nonemployment to employment each month, because the pool is far larger than the unemployed population.

The authors list contributing factors for future research: immigration-driven changes, sector-specific slowdowns in technology and government contracting, policy uncertainty, and early signals of broader labor market deterioration. The unemployment rate has stayed relatively low even as these rates fell, which the authors call an unusual development for a stable economy.