Patents, Learning by Doing, and Market Structure in the Chemical Processing Industries

This paper analyzes propensity to patent and the effect of patented process innovations on output prices for a sample of 24 chemical products. Econometric count data models (Poisson and negative binomial) are used given the discrete nature of the patent data. The results show that process patent activity was stimulated by growth in industry cumulative output (“learning by doing”) rather than by expenditures on R and D. Market concentration had a significant negative effect on patenting. Reductions in output prices were closely linked to growth in cumulative output, but only weakly to the rate of patenting. These results differ from those obtained in patent studies based on more aggregate data.