The importance of reallocations in cyclical productivity and returns to scale: evidence from plant-level data

Procyclical productivity plays an important role in many models of aggregate fluctuations. However, recent studies using aggregate data to directly measure technology shocks in the Solow residual find that technology shocks are not procyclical. This paper provides new evidence that, due to countercyclical composition changes between producers, the procyclicality of productivity observed in aggregate data may be understated. Using plant-level microdata, this paper finds that the reallocation of output shares across continuing plants, as well as the entry and exit of plants, creates a countercyclical component in aggregate productivity. This paper shows that such composition changes may cause a downward bias in industry-level estimates of returns to scale. The findings of this paper suggest that, without correcting for the countercyclical effects of reallocations, estimates based on aggregate data may not reflect the true cyclicality of technology shocks, which a representative agent faces over the business cycle.