Jobless recoveries and time variation in labor markets

Abstract

We study the nature of changes in the relationship between the labor market and output by using a broader set of labor market variables and a flexible time-varying-parameter vector autoregression model with stochastic volatility (TVP-SV). Overall, the TVP-SV model fits the labor market data well and has good predictive performance. We find heterogeneity both in the timing and the type of changes in the relationship between labor market variables and output. We find evidence that the relationship between output and employment growth, hours, vacancies, and wages has changed over time, with employment, wages, and vacancies being less responsive over time. In contrast, the responses of hours and part-time employment to output have become stronger over time, indicating a shift towards utilization of the intensive margin and towards utilizing more flexible labor inputs.