Abstract

I would like to preface my remarks by saying that I enjoyed reading this paper, which deals with a subject I knew too little about, but on which I now feel reasonably informed. The author has succeeded in writing down a model which to some degree captures the essence of three of the leading explanations of variation in research and development (R&D) activity across firms and industries. While the paper has some theoretical appeal, it falls short of its goal to distinguish these hypotheses empirically. As a result, the empirical aspects of the paper are the focus of my comments.

I begin with an introductory review of the goals of the paper, the methods applied to achieve those goals, and its conclusions. This is followed by some remarks regarding the model's ability to capture the hypotheses posited in the literature on R&D. I then discuss the empirical content of the model: specifically the mapping from model parameters to predictions for observable variables. On this basis I proceed to a criticism of the author's empirical work: what it tells us, what it cannot tell us, and what it ought to tell us about R&D.