Unemployment Fluctuations, Match Quality, and the Wage Cyclicality of New Hires

Recent papers interpret micro-level findings of greater cyclicality in the wages of job-changers as evidence for flexible wages of new hires, thus concluding that wage rigidity is not an empirically plausible mechanism for resolving the unemployment volatility puzzle. We analyze data from the Survey of Income and Program Participation (SIPP) to argue that greater cyclicality of wages for job changers reflects cyclicality in the composition of match quality across new hires. After introducing controls for the cyclicality of job-to-job flows, we find no evidence for greater wage flexibility among new hires. In light of our empirical findings, we study an equilibrium model of unemployment with staggered Nash bargaining, heterogeneous match quality, and on-the-job search. Workers in bad matches vary their search intensity according to the probability of finding a better match, generating cyclicality in the contribution of bad-to-good transitions to total job-to-job flows. Using simulated data from our model, we compute measures of new hire wage cyclicality analogous to those found in the literature and show that cyclical match composition in our model generates spurious evidence of new hire wage flexibility of comparable in magnitude to what we estimate from the SIPP. The model is also successful in accounting for the cyclicality of job creation and the dynamics of aggregate unemployment.