Aggregate employment dynamics: Building from microeconomic evidence

This paper studies quarterly employment flows of approximately 10,000 U.S. manufacturing establishments. We use establishments' hours-week to construct measures of the deviation between desired and actual employment and use these as the establishments' main state variables. Our main findings are: (i) microeconomic adjustment functions are nonlinear, with plants adjusting disproportionately to large shortages; (ii) adjustments are often either large or nil, suggesting the presence of nonconvexities in the adjustment cost technologies: (iii) the bulk of average employment fluctuations is accounted for by aggregate, rather than reallocation, shocks; and (iv) microeconomic nonlinearities amplify the impact of large aggregate shocks.