ناهمگونی در استراتژی های همکاری تحقیق و توسعه

We explore heterogeneities in the determinants of innovating firms' decisions to engage in R&D cooperation, differentiating between four types of cooperation partners: competitors, suppliers, customers, and universities and research institutes (institutional cooperation). We use two matched waves of the Dutch Community Innovation Survey (in 1996 and 1998) and apply system probit estimation. We find that determinants of R&D cooperation differ significantly across cooperation types. The positive impact of firm size, R&D intensity, and incoming source-specific spillovers is weaker for competitor cooperation, reflecting greater appropriability concerns. Institutional spillovers are more generic in nature and positively impact all cooperation types. The results appear robust to potential simultaneity bias.

The growing role of R&D collaboration in firms' innovative activities (Hagedoorn, 2002) has spurred research into the determinants of R&D cooperation and the effects of cooperative R&D. Two major strands of theoretical literature can be distinguished. The Industrial Organization (IO) literature has extensively examined the incentives and welfare effects of R&D cooperation among competing firms, focusing on the role of R&D investments and R&D spillovers. Theoretical contributions in the management literature have stressed that R&D collaboration aims at minimizing transaction costs and exploiting complementary know-how between partner firms (e.g. Kogut, 1988 and Das and Teng, 2000). Empirical work on R&D cooperation has utilized micro-level survey data from the European Community Innovation Surveys (CIS), and has focused mainly on the impact of firm size and R&D intensity as determinants of cooperation (Becker and Dietz, 2002, Leiponen, 2001, Kaiser, 2002 and Veugelers, 1997).

This paper has explored the heterogeneity in the determinants of firms' decisions to engage in vertical (suppliers, customers), horizontal (competitors) and research institutional (universities and research labs) R&D cooperation. We took into account a broad set of determinants but paid particular attention to the effects of different types of spillovers, a central focus in the industrial organization literature on R&D cooperation. We limited potential problems of simultaneity bias between cooperation and its determinants (notably R&D intensity and incoming spillovers) by utilizing a two-period dataset on innovating firms, which allowed us to employ lagged variables. In addition, we considered a sample of firms that had not cooperated in the first period, further minimizing potential simultaneity problems. We used a multivariate probit model to reflect that firms consider simultaneously the decisions to cooperate with various partners. We found significantly positive correlations between the equations, which might indicate that the various cooperation decisions tend to be viewed by the firms as complementary rather than substitutes, but could also be due unobserved firm heterogeneity.