Varieties of systems of innovation: A survey of their evolution in growth theory and economic geography

The systems of innovation (SI) approach has been established and extended during the last two decades. Although elementary goals and progress have been reached through seminal contributions by Freeman (1987), Lundvall (1992) or Nelson (1993), in designing a generic approach, displaying the dynamics of collaboration, networking and interactive learning, criticism has been raised that systems of innovation are still 'undertheorized'. The objective of this paper is to describe briefly the historical evolution of the SI concept within the academic literature and the policy sphere. This review primarily attempts to highlight some of the most important contributions that strongly assisted to the framework, by providing more consistency and a more theory- oriented perspective. Consequently, the system concept itself seems to be a kind of 'boundary object'. Within both, the academic and the policy field, different levels of conceptualization have been challenged and advanced in the course of time. These conceptualizations basically differ in their scale of analysis, taking geographical perspectives, technologies or sectoral classifications as foci for theorizing and empirical research. Despite these substantial levels of research, the SI framework is increasingly challenged, analyzed and extended in the context of globalization. As a result, regarding the openness and flexibility of the SI approach, this paper particularly tries to focus on the difficulties of contemporary research in defining functional and spatial boundaries in theory and empirical research. Agglomeration tendencies, knowledge externalities and localized learning are primarily based upon the concepts of knowledge diffusion, tacit knowledge and proximity. In spite of that, ICT and global business linkages foster inter-regional and trans-border knowledge flows. Thus, knowledge diffusion is also related to international and global 'pipelines' that could support, strengthen and reinforce localized learning.