Search Frictions and Evolving Labour Market Dynamics

This paper puts search frictions models under novel empirical scrutiny and tests their ability to match empirical observations. To capture changing dynamics we fit an extended Bayesian time-varying parameter VAR to US labour market data from 1962–2016. Our results indicate that these models are unable to match the responses of key variables to identified structural shocks. We document substantial changes in the transmission, and economic importance, of shocks throughout time that this framework is unable to explain. In particular, although search frictions models use productivity and job separations shocks to explain the labour market, these shocks only explain at best 50% of fluctuations in key labour market variables.