Geography Reference
In-Depth Information
what drives the formation of sales partnerships in the stock photography sector in
Germany. Based on a network survey of i rms, Glückler explains the likelihood of estab-
lishing a sales partnership between two agents. Borrowing network measures from social
network analysis, his i ndings suggest that the likelihood of forming a sales partnership
in this sector is primarily shaped by multi-connectivity, rather than geographical prox-
imity and homophily, for instance. In Chapter 15, Olav Sorenson, Jan Rivkin and Lee
Fleming analyse US patent data and citation rates across inventors to determine which
forms of proximity are important for knowledge l ows, and relate these to the nature of
knowledge (whether complex or not). Synthesising a social network view with a perspec-
tive of knowledge transfer as a search process, they claim that the advantages of being
proximate to some knowledge source depend crucially on the nature of the knowledge
at hand. Their i ndings show that simple knowledge l ows equally to actors near and far,
while complex knowledge is unlikely to dif use, no matter how proximate actors are.
With knowledge of moderate complexity, however, the outcomes show that closer actors
are in a better position to benei t from knowledge dif usion, in contrast to more distant
recipients. Sorenson et al. conclude that an interesting line of research would be to inves-
tigate how the tendency of industries to agglomerate might depend on the complexity of
knowledge.
Regional networks may be formed through the movement of labour. Labour mobility
has the potential to provide ef ective channels of knowledge dif usion across countries
and regions (Saxenian, 2006). Agrawal et al. (2006) found that knowledge is transferred
between i rms across large geographical distances when their respective employees are
socially linked because of a shared past in the same school or same company. In other
words, knowledge networks are formed through social proximity between agents, irre-
spective of their geographical distance, although social networks are often geographi-
cally bounded (Breschi and Lissoni, 2003). In Chapter 16, Stefano Breschi, Camilla
Lenzi, Francesco Lissoni and Andrea Vezzulli explore a large set of patent applications
by US inventors registered at the European Patent Oi ce in three high-technology i elds
in the period 1991 to 1999. They demonstrate that inventors who patent across i rms do
not dif use their knowledge that much across space, i rst, because inter-regional mobility
of inventors is rather limited, and second, because inventors create social networks at
the regional level, not across regions. The few inventors who do move between regions,
however, tend to maintain their ties with former co-inventors, providing a channel of
knowledge dif usion to their prior location. In the latter case, knowledge dif uses across
space through the professional networks of inventors.
Another issue is the ef ect of (regional) networks on economic performance. Studies
often report a positive relationship (Ozman, 2009). However, this is not necessarily the
case. This may depend on, among other things, the degree of proximity between the net-
works partners. While a high degree of any form of proximity might be considered a pre-
requisite to make agents interact, proximity between agents does not necessarily increase
their innovative performance, and may possibly even harm it (Boschma, 2005; Broekel
and Meder, 2008; Cantner and Meder, 2007; Grabher and Stark, 1997; Nooteboom et
al., 2007). This is in line with what Gilsing et al. (2007) found when they assessed the
impact of technological distance in high-tech alliance networks on the innovative per-
formance of the partner i rms. As expected, they found an inverse U-shaped function
between technological distance and exploration (i.e. real breakthroughs), meaning that
