Geography Reference
In-Depth Information
Organizational proximity
Social networks also concentrate within foci, such as organi-
zations (Feld, 1981). On a daily basis, most fully employed individuals spend more
waking hours engaged in work than in any other activity. Employees regularly meet
other employees through work to cooperate on projects, to confer on decisions, to trans-
fer information, and to socialize. Hence, we use employment at the same patent assignee
as another indicator of social proximity:
Organizational proximity
ij
= 1 if the same organization owns both patents in a dyad, 0
otherwise.
We expect common ownership to boost citation likelihood, especially for focal patents
of moderate interdependence.
We test our hypothesis by regressing
Cite
ij
on the indicators of social proximity
directly, the indicators interacted with
k
, and the indicators interacted with
k
2
. We expect
social proximity to boost citation probability directly. The core test of our hypothesis
resides not in the direct ef ects but in the interaction terms: the impact of proximity on
citation probability should have an inverted-U relationship with respect to interdepend-
ence
k
.
13
In light of our empirical context, patent citations, it is useful to elaborate our expec-
tations about the direct ef ect of
k
on citation likelihood. Our hypothesis describes the
impact of interdependence on the
gap
between near and distant actors' success in receiv-
ing and building on knowledge. We examine this gap by examining
interactions
of
k
and
k
2
with social distance. In developing the hypothesis, however, we also paint a picture of
the
direct
impact of
k
on knowledge reproduction: we suggest that greater interdepend-
ence increases the dii culty for a party of receiving and building on prior knowledge,
regardless of the party's distance from the source. This argument concerns an actor's
success in receiving and building on knowledge
conditional on an attempt to do so being
undertaken
. Patent citation data nevertheless rel ect not only success conditional on an
attempt being undertaken, but also the sheer number of attempts being undertaken.
We have reason to believe that the number of attempts may rise with interdependence,
simply because interdependence increases the fertility that comes from mixing and
matching components (Fleming and Sorenson, 2001). Accordingly, we of er no hypoth-
esis about the direct ef ects of
k
on citation rates. Instead, we focus on the
gap
between
near and distant actors' citation rates, which should have a robust inverted-U relation to
interdependence. (See the Appendix for a more detailed treatment of this point.)
Controls
The non-monotonic interactions between interdependence and proximity that we predict
- if found in the data - lend themselves to few alternative interpretations. The models
nevertheless include as controls several of the most important variables used in prior
patent studies (e.g. Lanjouw and Schankerman, 2004).
Activity control
The activity control accounts for the typical number of citations
received by a patent in the same technological areas as the focal patent. In a i rst stage,
we calculated the average number of citations that each patent in a particular USPTO
class received from patents granted between January of 1985 and June of 1990 (equation
