Geography Reference
In-Depth Information
16 The geography of knowledge spillovers: the role of
inventors' mobility across i rms and in space
Stefano Breschi, Camilla Lenzi, Francesco Lissoni and
Andrea Vezzulli
1. Introduction
In the past 20 years, research on the geography of innovation has revolved largely
around the concept of 'localized knowledge spillovers' (hereafter LKSs). LKSs are
'pure externalities' (Griliches, 1992): they exist insofar as scientii c and technological
knowledge may escape its producer's control, and yet dif use only locally. LKSs may
explain why innovation activities are often found to be spatially clustered (Feldman,
1999).
For long supported only by circumstantial evidence, the LKS hypothesis was i rst
tested by Jaf e, Trajtenberg and Henderson (1993; hereafter JTH). The three authors
argued that knowledge spillovers may be measured by the 'citations to prior art' con-
tained in most patent documents, and produced a statistical experiment showing that
such citations come disproportionately from the same geographical area of the cited
patents. The experiment requires matching each citing patent to a control one, with the
same application date and technological classii cation, in order to compare their location
in space.
The JTH experiment has become a classical reference for most empirical work on
the geography of innovation, both within mainstream economics and for unorthodox
approaches, such as evolutionary and institutional ones. However, its interpretation as
proof of the existence of LKSs relies on the twin assumptions that scientii c and techno-
logical knowledge is largely tacit, so that face-to-face contacts are the necessary vehicle
for its dif usion, and that geographical proximity is a necessary condition for those
contacts to take place. As a result, JTH's work treats
geographical proximity
as a proxy
for
social proximity
, which inventors may derive from professional collaboration or
common ai liation to companies, technical and scientii c societies, or former institution
of higher education. A limitation of this strategy is that it makes it dii cult to distinguish
between dif erent social ties, according either to their nature (for example, professional
vs. friendly) or strength (such as older vs. more recent ties).
Various attempts have been made to overcome this limitation. Agrawal et al. (2006)
have tested whether inventors who move from one company to another and across dif-
ferent locations still pass on knowledge to former colleagues active in the cities they
have left. Breschi and Lissoni (2009) and Singh (2005) have resorted to a more direct
approach, which consists of measuring professional ties between inventors resulting
from co-invention data, and in applying to the resulting data some standard tools of
social network analysis.
In this chapter we put together the two strands of research. In particular, we explore
the social network of inventors who are mobile in space, and test whether its geographical
353
