Aggregate productivity and the productivity of aggregates

Explanations of procyclical productivity playa key role in a variety of business-cycle models. Most of these models, however, explain this procyclicaIity within a representative-firm paradigm. This procedure is misleading. We decompose aggregate productivity changes into several terms, each of which has an economic interpretation. However, many of these tenus measure composition effects such as reallocations of inputs across productive units. We apply this decomposition to U.S. data by aggregating from roughly the two-digit level to the private economy. We find that the compositional terms are significantly procyclicaI. Controlling for these terms virtually eliminates the evidence for increasing returns to scale, and implies that input growth is uncorrelated with technology change.