This study estimates the patent-R&D relationship using count panel data. The data is an original panel of 318 firms making R&D investments and applying for patents during the period from 1984 to 1993. A negative binomial model with fixed effects is estimated, taking into account both the discrete nature of the count dependent variable and firm-specific unobserved heterogeneity as well as overdispersion in the data. Firm-level R&D capital, concentration ratios, and various firm size proxies are used as independent variables. Analysis of the data fails to reveal support for the basic tenets of the Schumpeterian Hypothesis. In particular, firm size has a significant negative impact on innovation while industry concentration is statistically insignificant.
James J. Jozefowicz. "A Count Panel Data Study Of The Schumpeterian Hypothesis." New York Economic Review. vol. 34, Fall 2003, p. 21-37
BibTeX entry download