Geography Reference
In-Depth Information
suggest that the ability of future inventors to receive and build on the original knowledge
varies as a function of their social proximity. Second, analyzing citations at the level of
the citing-patent/cited-patent dyad avoids the potential for aggregation bias inherent in
count models.
Interdependence
Following Fleming and Sorenson (2001), we measure the complexity of the knowledge in
a patent by observing the historical dii culty of recombining the elements that constitute
it. Though it involves intensive calculation, the intuition behind the metric is straightfor-
ward: a technology whose components have, in the past, been mixed and matched readily
with a wide variety of other components has exhibited few sensitive interdependencies.
The measure considers the subclasses identii ed in a patent as proxies for the underly-
ing components. Though in many cases subclasses correspond to identii able physical
components (such as in the example below), they do not always align so closely. Our
measure, however, requires only that these subclasses dei ne pieces of knowledge rather
than physical components. Combining some pieces that interact sensitively to each other
proves more dii cult than connecting relatively independent chunks of knowledge.
We calculate the measure of interdependence,
k
,
in two stages.
8
Equation 15.1 details
our measurement of the ease of recombination - the inverse of interdependence - for sub-
class
i
used in patent
j
. We i rst identii ed every use of the subclass
i
in previous patents
from 1980 to 1990.
9
The sum of the number of prior uses provided the denominator. For
the numerator, we counted the number of dif erent subclasses appearing with subclass
i
on previous patents. Hence, our measure increases as a particular subclass combines
with a wider variety of technologies, controlling for the total number of applications,
and captures the ease of combining a particular technology. To create our measure of
interdependence
for an entire patent, we averaged the inverted ease of recombination
scores for the subclasses to which it belongs (equation 15.2).
Ease of recombination of subclass
E
i
5
Count of subclasses previously combined with subclass
i
Count of previous patents in subclass
i
i
(15.1)
;
k
j
5
Count of subclasses on patent
j
i
[
j
Interdependence of patent
j
(15.2)
;
E
i
Intuitively, the measure operates as follows. Suppose a patent embodies subclasses
that have been combined with a wide variety of subclasses, even in a handful of previous
patents. This indicates that the patent's components do not have delicate interdepend-
encies that prevent widespread recombination and the components can be mixed and
matched independently. Such a patent receives a low value of
k
. Suppose instead that
a patent embodies subclasses that have been combined, again and again, with the same
small set of other subclasses. We presume those subclasses to be highly interdependent;
their repeated joint appearance in patents suggests that the presence of one requires the
appearance of the others. Hence the patent's
k
is high.
In addition to the measure's face validity, it has been validated externally via a survey
