Research and development, competition and innovation - Pseudo-maximum likelihood and simulated maximum likelihood methods applied to count data models with heterogeneity

This paper focuses on the relationship between the research and development (R&D) expenditures undertaken by firms and the number of patents claimed by them. The econometric methods we use incorporate two important features of the data. First, we introduce a heterogeneity term accounting for the unobservable factors affecting the production of innovations. Second, we allow for a large proportion of the sample not to apply for a patent. Depending on the specification, estimation proceeds by maximum likelihood, pseudo-maximum likelihood or simulated maximum likelihood methods, Our main findings are the following: (i) the returns to scale in the innovation technology may be constant on average and (ii) competitors' R&D may have a negative impact on own innovation, thus revealing a rivalry effect of intellectual property.