Learning networks of academic spin-offs - A spatial perspective

The importance of new knowledge in innovative activities of firms and the impact of these activities on economic development of regions have been acknowledged in many studies. In particular, universities and firms that are established on university knowledge, spin-off firms, function as nodes and channels through which new knowledge is diffused into the wider (regional) economy. New knowledge is a strategic resource of competitive advantage for young high-tech firms. Of course, many of these firms are based on new technical knowledge but they may lack market and managerial knowledge and skills. An important way of learning on these different aspects is through social networks and business networks, with a local (national) and/or international coverage. Many studies have attempted to understand the characteristics of networks of young high-tech firms, but a detailed picture and understanding of the time and space dimension of models of learning relationships are rare. What may be true is that a well-developed local learning network performs as an important condition for establishing international learning relationships, indicating a stepwise model. The theory of 'born globals' indicates, however, international learning from the start of the firm. This paper explores the learning models of university spin-off firms, including some aspects of absorptive capacity, with a focus on various combinations of local and global knowledge networks and changes in these combinations by age of the firms. The analysis draws on a sample of 100 spin-offs from two universities: TU Delft University in the Netherlands and Norwegian University of Science and Technology (NTNU) in Trondheim in Norway. Resource based views and organizational learning theory will be applied to design an analytical model of young high-tech firms shaping of learning connections and improving of innovation performance. The empirical part of the paper will include descriptive and explanatory results, the latter derived from correlation analysis.