Should Capital Input Data Receive a Utilization Adjustment?

Abstract

A common procedure in productivity research is to use estimates of the stocks of physical capital as proxies for service flows. A number of authors propose cyclical adjustments for capital input which, if large enough, will eliminate findings of procyclicality in the behavior of TFP. This paper argues that for the preponderance of assets in the fixed capital stock, fluctuations in utilization have little effect on user costs. In the aggregate, adjustments to capital input data for utilization should consequently be small, much smaller, for example, than those suggested by Solow (1957), Griliches and Jorgenson (1966), Tatom (1980), Shapiro (1993), or Basu and Fernald. (2000).