IT Adoption Costs and Productivity: A Reply to Diego Comin

Diego Comin (2002, Review of Economic Dynamics 5 (2)), critiques my paper "Technology adoption costs and productivity growth: The transition to information technology" (2002, Review of Economic Dynamics 5 (2)), concluding that all of my major results are unfounded. I contend, to the contrary, that my results hold up against his criticism. My paper analyzes the effect of investment adjustment costs on subsequent productivity growth for manufacturing industries. I find that adjustment costs increased substantially during the late 70's and early 80's and that this increase is associated with increased investment in information technology. I also estimate the effect of these adjustment costs on BLS productivity growth estimates. This analysis uses two datasets, the NBER–CES Manufacturing Industry Database (Bartelsman and Gray, 1996. NBER Working Paper 205) and the BLS Manufacturing Productivity release of April 2001. For the former I calculate productivity growth using a standard Solow residual; for the latter, I use BLS productivity growth estimates. Comin's critique, however, uses a different measure entirely. He introduces a "corrected TFP" measure that I believe is flawed. Most of his critique uses this measure, rather than the standard productivity growth measures I use. As a consequence, most of his criticism is not actually about my paper, but is, instead, about a paper Comin might have written using his measure. And because this measure is flawed, so is this alternative. In addition, he raises criticism of my evidence of a link between information technology (IT) and adjustment costs and of my corrections to BLS productivity estimates. I present additional evidence to address these concerns and to examine the robustness of my results. (Copyright: Elsevier)