1. IntroductionThe concept of national innovation systems (Freeman, 1988; Nelson, 1993) is the starting point of the present investigation. This is a complex institutional context that characterizes modern capitalist economies, involving a diversity of actors – companies with and without R&D departments, universities, research institutes, government, financial agencies, legal framework – and a division of labor between them. Within such an institutional context, the relationships necessary to exchange and combine knowledge and experiences between actors are a crucial factor that differentiates developed and undeveloped NISs ( Lundvall, 2003 ). Among these, those between businesses, universities and research institutes stand out as they work in a complementary way, bringing benefits to both academia and industry (Mowery and Sampat, 2005). From the point of view of more developed industrial countries, they represent a key and dynamic component of the NIS, especially in the form of bidirectional linkages (Narin, 1997; Pavitt, 1991; Rosenberg, 1990). They are also described as complex and multifaceted in the most advanced form – the case of the United States (Klevorick et al, 1995; Cohen et al, 2002). These studies also show that the channels through which knowledge flows between companies and public research organizations (PROs), as well as the relationships between the channels used and the benefits obtained. However, they have strong historical roots, with very specific patterns of change over time and space, and are difficult to construct (Rosenberg, 1982; Mokyr, 1990). On the one hand, they thrive in academic environments that valorise the so-called “entrepreneurial role” of universities beyond the traditional focus on higher education, the production of scientific knowledge and the provision of...... ...the creation of scientific institutions “ahead of demand”. This has been the case in information and communication technologies, information technology and computer engineering. On the other hand, despite recent progress, Brazil still lags behind in its National Innovation System (NSI). Unlike South Korea, the Brazilian path is slow and shows a relatively small trade-off between scientific and technological production, as Ribeiro et al (2009) demonstrate. The scientific side of the system has improved significantly, but the imbalance with production technology is representative of a less developed national socioeconomic formation. Furthermore, it is understood that an NIS constitutes a precondition for overcoming underdevelopment, as it produces solutions to the problems faced by the economy and society (Suzigan, Rapini and Albuquerque (2009).