Research and Productivity Growth : Theory and Evidence from Patent Data

I use the aggregate behavior of three indicators of technology (employment of research and scientists and enginerrs, patented inventions, and total factor productivity) to identify a model of endogenous technological change. In the United States and several other developed countries, research employment and total factor productivity have both grown substantially while the rate of patenting has remained relatively flat. One interpretation of these facts is that: (i) patentable inventions are becoming increasingly difficult to discover as the quality of techniques in use increases, (ii) inventions that are patented represent percentage improvements on techniques currently in use, and (iii) the size of the economy is growing, making patents increasingly valuable and justifying increased research efforst devoted to discovering them. I develop a general-equilibrium search-theoretic model of invention to formalize this view. I then fit the model to data for the U.S. economy. The calibrated model implies that as much as 25% of U.S. productivity growth can be attributed to industrial research activity and that the equilibrium fraction of resources devoted to research is less than two-thirds of the level that a social planner would choose