Content area
Full Text
INTRODUCTION
Knowledge flows are economically important because they increase the efficiency of the innovation process. The recombination of knowledge drives innovation; thus wider access to knowledge facilitates more efficient innovation by reducing the need to re-create what already exists elsewhere. In fact, contemporary economic theory focuses on knowledge spillovers - knowledge flows that occur outside market mechanisms - as the central determinant of economic growth (Romer, 1986, 1990).
Surprisingly, given the acknowledged importance of knowledge flows, we know very little about how they move through the economy and the mechanisms that influence flow patterns. However, we are reasonably certain about one feature of knowledge flow patterns: prior research shows, with reasonably conclusive empirical evidence, that such flows stay geographically localized (Agrawal & Cockburn, 2003; Almeida & Kogut, 1999; Audretsch & Feldman, 1996; Jaffe, Trajtenberg, & Henderson, 1993; Thompson & Fox-Kean, 2005).
The localization finding is important for a number of reasons. First, it provides insight into general flow patterns; knowledge does not flow uniformly across geographic space. Second, the finding implies that knowledge does not flow freely across the marketplace; public policy and firm strategy may influence flow patterns in self-serving ways. Finally, this finding offers insight into the mechanisms that cause knowledge to flow the way it does; despite its apparent costless-to-disseminate properties, knowledge flows faster locally. Some may not find this intuitive, since researchers often place knowledge in the public domain by way of widely accessible monographs, journals, patents, etc. Moreover, some types of inventor - particularly scientists - have strong incentives to disseminate their findings as fast as possible (Dasgupta & David, 1987, 1994). Why does knowledge flow this way?
Knowledge includes both codified and non-codified components. Even inventors able to codify knowledge frequently do not do so, because of a lack of incentives (Agrawal, 2006). In order for inventors to apply knowledge, they often need access to both the codified and the non-codified components. The non-codified components of knowledge are likely to contribute to geographic stickiness, as non-codified knowledge often requires direct interaction with the inventor for effective transfer.
This need to interact with the inventor may explain why knowledge frequently flows locally. Prior research shows that scientists and engineers impart knowledge among their peers, or "invisible college", particularly if they...