This paper uses an extended version of the well-established Crépon, Duguet and Mairesse model (1998, CDM hereafter) to model empirically the innovation-productivity relationship using data on patent applications by Dutch firms to the European Patent Office.
We use an extended version of the well-established Crepon, Duguet and Mairesse model (1998, CDM hereafter) to model empirically the innovation-productivity relationship using data for the 2000-2006 period on patent applications by Dutch firms to the European Patent Office. The CDM model disentangles the impact of R&D expenditures on patents and the impact of patents on productivity. A multiple-equation dynamic panel data model of R&D, patent applications or citations and multi-factor productivity (MFP) growth is estimated that suits multiple data distribution properties. We explicitly take into account the role of dynamics and firm-level unobserved heterogeneity in each of the innovation processes and productivity. We find evidence that the output innovation affects productivity positively, which seems to be robust across specifications. We also find that the strong presence of random effects for individual heterogeneity in explaining the R&D patents relationship is an important driver to innovation. While the estimates of R&D and dynamics depend on whether these unobserved characteristics are taken into account, we find robust evidence on the role of firm size in explaining patent
and citation counts.