639 - Measuring trends and persistence in capital and labor misallocation

DNB Working Papers
Date 7 June 2019
We analyze trends and persistence in the misallocation of labor and capital using firm-level panel data for the Netherlands in the period 2001-2017. We use the dispersion in marginal revenue products of labor and capital to measure the extent of misallocation. Compared to a counterfactual efficient allocation we find that misallocation has had a sizable negative impact on aggregate productivity. Especially capital misallocation has increased over time. The relatively high and rising capital misallocation is caused by a combination of small, highly productive firms facing relatively high capital wedges and large and unproductive firms facing relatively low capital wedges. Exploiting a panel data error components model we find that capital misallocation has a much more permanent character than labor misallocation. Moreover, it is the permanent component of capital misallocation that has increased over time. Finally, we show that in our sample the measurement of misallocation is largely insensitive to capital adjustment costs and alternative specifications of the production function. The contribution of heterogeneous markups to observed misallocation, however, is non-negligible.
 
Keywords: misallocation, panel data, persistence, productivity.
JEL-Classifications: C23, D24, O47.