The Dynamics of Plant-level Productivity in U.S. Manufacturing

Recent work in I.O. has emphasized the importance of firm- and plant-level heterogeneity in total factor productivity. Jensen and McGuckin (1996) argue that the major empirical regularity in studies of firm or establishment level productivity is heterogeneity within sectors and across plant characteristics. In a seminal paper, Olley and Pakes (1996) presented an econometric method for estimating establishment-level productivity taking into account selection issues. This paper extends Olley and Pakes in two directions. First, instead of focusing on the single industry, this paper applies the Levinsohn-Petrin (1999) modification of the Olley-Pakes methodology to estimate plant-level productivity across many industries. This is made possible by use of a unique plant-level dataset that covers all industries in the U.S. manufacturing sector. Second, while Olley and Pakes were interested in the dynamics of productivity, they focused on measuring year-by-year or multi-year average productivity within an industry. This paper augments their approach by estimating an establishment-level idiosyncratic productivity process.