The Diamond-Mortensen-Pissarides search and matching model is the workhorse of labor
macro, but it has difficulty in simultaneously matching the cyclical behavior of job loss and
vacancies when taken to the data. By completely ignoring frictions in job creation and
focusing instead on firm-level heterogeneity, one can match the cyclical behavior of job flows
and vacancies relatively well. In particular, one can generate a Beveridge Curve which looks
much like the real Beveridge Curve, and one can replicate the approximately equal
contributions of job creation and destruction to the cycle. Focusing on heterogeneity rather
than on hiring costs seems to give an improved picture of hiring activity over the cycle.