Fraunhofer ISI discussion papers innovation systems and policy analysis 24
We evaluate the impact of cooperation with public research institutions on firms' inno-vative activities in France and Germany, using data from the fourth Community Innova-tion Survey (CIS4). We propose an original econometric methodology, which explicitly takes into account potential estimation biases arising from self-selection and endoge-neity, and apply it to both process and product innovation. We find a positive effect of cooperation on both types of innovation. This effect is significant in both countries, but much higher in Germany than in France. Drawing on a comparison of the institutional context of cooperation across both countries, we interpret this difference as a conse-quence of the more diffusion-oriented German science policy. Finally, our robustness checks confirm the importance of controlling for selection and endogeneity. We show that these problems can be serious, and may lead to inconsistent estimates if ne-glected.
Public/private research partnerships University/industry linkages Innova-tiveness Heckit procedure with endogenous regressors