IZA DP No. 16373: Labor Force Transition Dynamics: Unemployment Rate or Job Posting Counts?
Job posting counts (JPCs) are increasingly being used as indicators of employment dynamics, but they have not received sufficient research attention to establish their value as a metric of these dynamics. This study aims to assess the efficacy of the traditional survey-based unemployment rate versus big-data-based JPCs in capturing labor market transitions in the United States. Using the Current Population Survey, our comparison focuses on the ability of these two types of metrics to predict individuals' transitions between employment and unemployment. Unlike with the unemployment rate, we not only examine the raw national JPCs but also consider four additional versions of JPCs that measure labor demand at various disaggregated levels. Our findings suggest that JPCs and the unemployment rate provide comparable predictive power for labor market transitions, with each capturing different aspects of the variation in these transitions. The estimated coefficients of both types of metrics remain statistically significant when considered together. Notably, the correlation between the unemployment rate and labor market transitions switches signs when year fixed effects are added, but a similar phenomenon is not observed when JPCs are examined. Among the various versions of JPCs, the most refined measure—JPCs by state, occupation, and industry—demonstrates the strongest predictive capabilities, outperforming other JPC measures and the unemployment rate.