Firm Dynamics at the Zero Lower Bound

Can the large decline in aggregate employment during the US’s Great Recession be explained by the unusually large decline in employment of young firms? In this paper I develop a quanti- tative heterogeneous-firm model with both financial and nominal frictions and use it to analyse the recent recession through the lens of firm-level data. I first use BDS data for the US to show that 40% of the decline in aggregate employment during the recession can be accounted for by declining employment in young firms (age 0-5). While this supports the role of financial frictions in impeding employment, it still leaves 60% of the decline to be explained by old firms (age 6+). Decomposing this further into changes in average employment within firms, and the number of firms, I find that most of the total decline in employment can be accounted for by (1) a decline in the number of young firms by over 25%, and (2) a decline in employment within old firms of 7.5%. I then extend the heterogeneous firm model of Khan and Thomas (2013) to include nominally denominated firm debt, downwards nominal wage rigidity (DNWR), and the zero lower bound (ZLB). Absent nominal features, a financial shock is only able to match the response of young firms during the crisis, because declines in real factor prices reallocate resources from young to old firms. However, adding DNWR and the ZLB inhibits this fall in wages and interest rates, leading the financial crisis to spill over to old firms. This allows the model to better match the data across the whole firm age distribution. Additionally, rather than leading to the misalloca- tion of resources across firms, the crisis instead manifests as a drastic decline in employment, as occurred in the US. Finally, I show how the power of fiscal policy at the ZLB depends on the firm distribution. The government spending multiplier is larger when more firms are financially constrained, and fiscal transfers to young firms are more powerful than transfers to older firms.