This paper develops a general equilibrium model of nonsequential employer search with recruiting selection and heterogeneous workers, and characterizes its equilibrium. I depart from the standard search model by allowing firms to simultaneously meet several applicants and choose the best candidate. Recruiting selection is important: firms interview a median of 5 applicants per vacancy and spend 2.5% of their total labor cost-about US$4200 per recruit – in these activities. The model provides an endogenous matching process with heterogeneous workers in which the hazard rate out of unemployment increases in productivity. The model also accounts for the empirical evidence of negative duration dependence of both hazard rates and re-employment wages. Under recruiting selection, lifetime inequality increases relative to the sequential search benchmark because low wage workers go through longer and more volatile unemployment spells, and have less valuable outside options to bargain with firms. I also show that stronger recruiting selection worsens the productivity of the unemployed and may not generate a more efficient job assignment at the aggregate level. Search frictions coupled with recruiting selection generate new kinds of externalities that affect not only transition probabilities, but also the expected productivity of recruited workers. The calibrated model can replicate moments of the distribution of wages and unemployment durations in CPS data. Using this parametrization, I also show that an increase of screening costs reduces inequality and productive efficiency, and decreases negative externalities on other employers.
Keywords: Duration dependence, Employer search, Hazard rate heterogeneity, Inequality, Nonsequential search, Recruiting selection, Search externalities.