Geography Reference
In-Depth Information
in determining the degree to which an industry disperses is the extent to which existing
social connections within the industry prove advantageous to entrepreneurs. In this
chapter, originally published in
Research Policy
, we argue that the complexity of the
underlying knowledge importantly inl uences when a connection to the source of the
knowledge proves most valuable. In particular, relationships of er the greatest advantage
in accessing knowledge of intermediate interdependence.
The logical implication of this idea for the geographic evolution of industries is that
those industries that rely on technology and processes of intermediate complexity should
remain most concentrated in space. Those based on less complex knowledge, meanwhile,
should dif use more readily to new locations. Though from the theory one might also
expect greater geographic dispersion in industries dependent on highly complex technol-
ogy and processes, it seems quite likely that such industries fail to emerge, as the transfer
of the knowledge proves so dii cult as to preclude its replication across i rms. Though
this proposition has yet to receive rigorous testing, Sorenson (2004) shows that it is con-
sistent with the cross-sectional variation in the geographic concentration of industries in
the United States in the 1990s.
Though clearly of relevance to an evolutionary perspective on economic geogra-
phy, the processes described in the chapter also have implications beyond geography.
Given the range of potential applications, we chose to frame the original study more
broadly.
2.
Complexity, networks and knowledge l ow
Original introduction
The l ow of knowledge plays a central role in a wide variety of i elds (for a review, see
Rogers, 1995). Sociologists began investigating dif usion processes - and the impor-
tance of social structure to those processes - to understand the adoption patterns of
agricultural and medical innovations (Coleman et al., 1957; Ryan and Gross, 1943).
To students of technology management, knowledge l ow i rst arises as an important
issue in the context of technology transfers within the i rm (Allen, 1977; Teece, 1977),
but questions of dif usion also arise when technology scholars ask whether incumbent
i rms or upstarts i rst develop and commercialize new inventions (Reinganum, 1981;
Tushman and Anderson, 1986). Both students of organizational learning (for a review,
see Argote, 1999) and industrial economists (Griliches, 1957; Irwin and Klenow, 1994;
Zimmerman, 1982) study how knowledge moves through i rms and how it spills over
to other i rms. In short, a diverse array of scholars shares an interest in knowledge dif-
fusion processes.
The normative interpretation given to dif usion, however, dif ers dramatically across
i elds. Economists and sociologists tend to focus on the societal benei ts of spillovers (i.e.
the l ow of knowledge across actors, usually i rms). The generation of new knowledge
often requires substantial investment in research and development, but the repeated
application of this knowledge, once produced, entails little if any incremental cost
(Arrow, 1962). Knowledge dif usion, therefore, engenders scale economies and stimu-
lates economic development by allowing several i rms to benei t from the R&D activities
undertaken by a single i rm (Marshall, 1890; Romer, 1987; Scherer, 1984). Management
scholars, by contrast, note that when knowledge escapes to competing i rms the returns
