Geography Reference
In-Depth Information
We analyze patent data to test our thesis empirically. Citation patterns across patents
of er something of a fossil record for the l ow of knowledge - providing a lasting rel ec-
tion of ephemeral interactions. Using this record, we estimate the ef ect of knowledge
complexity on the likelihood of future citations as a function of the social proximity
of future inventors to the inventor of the original piece of knowledge, comparing those
socially close to and far from the source. To assess social proximity, we calculate the geo-
desic length between patents' inventors in a collaboration network. We also supplement
this metric with indicators of geographic proximity and employment within the same
organization. To gauge complexity, we develop a measure that rel ects the historical
interdependence of a patent's subcomponents with other subcomponents. The i ndings
provide strong support for our core hypothesis: the higher likelihood of citation among
proximate inventors peaks for knowledge of an intermediate level of complexity (inter-
dependence).
This work contributes to the literature in several ways. First, from the perspective of
social networks, it identii es one condition under which social proximity should prove
especially important to knowledge l ow: for knowledge of intermediate complexity.
Though social scientists have usefully demonstrated that networks matter for the dif-
fusion of knowledge, relatively little research considers precisely when those networks
should matter most (Baker and Faulkner, 2004; Strang and Soule, 1998). By synthe-
sizing the social network perspective with work on conceptions of knowledge receipt
as search, we identify scope conditions on the relevance of social connections to the
dif usion process. Second, with respect to evolutionary economics, our work highlights
social connections as an important channel through which 'insiders' gain superior access
to knowledge. Extant work asserts that insiders - dei ned usually as those within the
same i rm as the source - have better access to an original success, which serves as a
template in ef orts to transfer and extend that knowledge (Nelson and Winter, 1982,
p. 119; Rivkin, 2001). Yet this work fails to establish the source of this preferential
access. Does it come from incentives that reward transfer, from the coni dentiality
agreements that employees sign, or from some other source? Our research points to
direct social connections as a critical factor dif erentiating these internal parties from
those outside the i rm.
3. Thel ow of complex knowledge
Our discussion begins with the most common i nding of classic dif usion studies: the
S-shaped cumulative adoption curve (Griliches, 1957; Ryan and Gross, 1943; and
Rogers, 1995, provides an excellent review). Researchers consistently i nd that the
adoption of an innovation over time follows a common pattern: growing slowly at
i rst, then accelerating rapidly, and i nally slowing to reach some asymptotic satura-
tion level. These dynamics resemble that of an epidemic spreading through a popu-
lation; the innovation i rst 'infects' those most at risk of exposure - actors closest
to the original source (Hägerstrand, 1953 [1967]) - and those most susceptible to
infection - those most prepared to accept the uncertainty associated with an untested
technology (Mansi eld, 1968) or whose idiosyncratic characteristics make the innova-
tion appear most attractive (Griliches, 1957). Over time, awareness of the innovation
spreads, uncertainty ebbs, and the economics of the invention become favorable to
a larger share of the population. Dif usion then takes of . In this classic perspective,
