关键词: 创新网络, 开放度, 边际合作效应递减, 创新绩效

Abstract: In order to decrease the uncertainty of S&T innovation at the market, enterprises need not only cooperate actively with external organizations, but also seek complementary internal and external resources. Open innovation is an important way for enterprises to obtain the key resources needed by technology innovation. However, due to transaction cost, information-searching cost, technology disclosure and other problems, open innovation may have a negative impact on innovation performance for enterprises. Therefore, the relationship between the degree of openness of the external organization and innovation performance of enterprises is complicated.
Cooperation openness is an indicator to describe the degree of cooperation between enterprises and external organizations, including the breadth and depth of cooperation. The breadth of cooperation refers to the extent to which an enterprise establishes cooperative relations with other organizations. The more extensive the enterprise cooperation is, the more beneficial it is to widen the breadth of the knowledge base, to increase the type of knowledge source, and to promote the innovation performance. The depth of cooperation refers to the average number of cooperation between enterprises and partners. It means that enterprises absorb the depth of external information and knowledge from external sources of knowledge. Industry-University-Research institution through deep-level cooperation can strengthen the degree of trust, establish lasting trust relationship, weaken knowledge transfer risk and cost, reduce information asymmetry and improve innovation ability.
To explore the relationship between the openness and the innovation performance has also become the focus of researchers. But little attention has been paid on the impact for openness of cross-organizational cooperation on innovation performance from the perspective of networks and dynamic level. To make up for the lack of existing research, the Industry-University-Research institution (I-U-R) cooperation networks have been created with three-year windows, based on co-applied patents data from the State Intellectual Property Office (SIPO) in the Chinese Information and Communication Technology (ICT) industry during 1999-2010. With a longitudinal sample of 2,603 Chinese firms we analyze the impact openness on firm innovation performance using Negative Binomial Regression.
The results show that taken Patentsit+1, Patentsi t+2 as the explained variables, the primary term coefficients of cooperation breadth were significantly negative at the significant level of 5% and 1%, and the Quadratic term coefficients of cooperation breadth were significantly positive at the significant level of 1% and 10%. It means that the cooperation breadth has a U-shaped impact on innovation performance of the enterprise. It shows that the breadth of Industry-University-Research institution (I-U-R) cooperation lies in the U-shaped region of the learning curves with marginal diminishing cooperation effects, not in the inverted U-shaped region of the learning curves.
Taken Patentsit+1 as the explained variables, the primary term coefficients of cooperation depth were significantly negative at the significant level of 1%, and the Quadratic term coefficients of cooperation depth were significantly positive at the significant level of 1%. It means that the cooperation depth has a U-shaped impact on 1-year lagged innovation performance of the enterprises. It shows that the depth of Industry-University-Research institution (I-U-R) cooperation lies in the U-shaped region of the learning curves with marginal diminishing cooperation effects, not in the inverted U-shaped region of the learning curves.
Taken Patentsit+2 as the explained variables, the primary term coefficients of cooperation depth were significantly negative at the significant level of 1%, and the Quadratic term coefficients of cooperation depth has not passed the significant test. It means that the cooperation depth has a negative impact on 2-year lagged innovation performance of the enterprise. The possible explanation is that appropriate cooperation depth can improve the innovation performance of enterprises, but if enterprises focus only on in-depth cooperation with a few organizations, and ignore the cooperation breadth, it would hider the improvement of innovation performance of the enterprise in the future.
Previous research has indicated that there is a best openness point between openness and innovation performance of enterprises, that is, the relationship between them is inverted U-shaped. But the conclusions of the paper are quite different. It shows that the learning curves of U shape with marginal diminishing cooperation Effects appear between openness and innovation performance. From dynamic capabilities point, there is a dynamic cyclical trend in the influence of openness on innovation performance, namely wave-like curves of learning effects.
The conclusions of the study provide a theoretical basis for the design of the mode of in-depth integration of industry-university-research in China. At the same time, it will also help to further improve the theory of open innovation. Policy suggestions are as follows. First, enterprises should evaluate their own growth stages and their comprehensive cooperation capabilities, so as to choose appropriate openness, rather than blindly expand the scope of cooperation. Second, On the basis of selecting potential high-quality R&D partners, we should further deepen R&D cooperation with partners, accumulate cooperation experience, and promote the formation of cooperation practices and cooperation capabilities to enhance innovation performance. Finally, open innovation is both an opportunity and a challenge. The effect of openness on innovation performance may show a spiraling upward and cyclical trend, which requires enterprises to continuously learn in cooperation and maintain the dynamics of their cooperation capabilities.
As for how to determine the “appropriate openness” of the enterprise, that is, the inflection point of the U-shaped region of the learning effect curve with diminishing marginal cooperation effect, how the cooperation breadth and the cooperation depth interact with each other are the directions for further research in the future. In addition, the research object of this paper is ICT enterprises, which are technology-driven enterprises. Whether the conclusion can be extended to an experience-driven enterprise still needs to be tested.

Key words: innovation networks, openness, marginal diminishing cooperation effects, innovation performance