Geography Reference
In-Depth Information
space. Consequently, entrepreneurs are most likely to be tied to the region where they
have useful social relations, even if another region is otherwise more attractive (Sorenson
and Stuart, 2001). In general, we may expect that clusters of new i rms in a particular
industry continuously evolve in regions where human and other resources are abun-
dantly present and where entrepreneurs are 'produced' on a large scale in the incumbent
i rms (Sorenson and Audia, 2000). Geographical immobility of entrepreneurs would
then make the existing structural base of a region a dominant source of clusters and
create an agglomeration ef ect similar to the ones discussed by Arthur (1990, 1994).
Jane Jacobs (1969) described how clusters often are driven by employees leaving
incumbents to start new i rms. She described how an employee learns a craft by working
in an organisation and being taught by a master. Later, that employee may leave the i rm
and set up his own shop, employ new people, and teach them the craft. Jacobs argued
that especially breakaways of able workers would be a source of regional development,
since the entrepreneur would rarely leave the area, but stay in the place where he has the
social connections. Jacobs's ideas are consistent with recent i ndings (some of which are
reviewed below) that spinof i rms are an important engine of growth in the early phases
of clusters. Experienced workers that leave a i rm to start up their own i rm in the same
industry are a well-known mechanism for dif usion of knowledge. Historically, there
have been many attempts to limit spinof s and geographical mobility of employees in, for
example, craft-based industries. Veblen (1912) colourfully described how entrepreneurs
faced the death penalty if they left the area to start up somewhere else than in the cluster
of glassmakers in Venice.
The transfer of routines and experience between a new i rm and the previous employer
of its founder has been analysed by organisational sociologists. Blueprints of the incum-
bent are passed on through the new i rm's founders (Brittain and Freeman, 1986; Carroll,
1984; Hannan and Freeman, 1986). The relation between performance of the incumbent
and the spinof has also received attention in economics and management (e.g. Agarwal
et al., 2004; Klepper, 2001, 2002a). Klepper (2001) proposes a model that combines the
ideas of reproduction and inheritance with the notion of organisational routines. This
notion was originally developed by Nelson and Winter (1982), assuming that i rms to a
large extent are governed by routines. A i rm has separate routines for each of the dif-
ferent functions (R&D, marketing, management, etc.) and products. When a new i rm
is born, organisations will reproduce, because founders rely on routines they are already
familiar with from their previous employment experience. The quality of these routines
will determine the future success and performance of the new i rm. Spinof s may inherit
more suitable routines than any other kind of start-up. This may on average enable
spinof s to outperform other start-ups.
Market growth as well as the industry and technological life cycles inl uence the emer-
gence and evolution of clusters in a particular industry (Dalum et al., 2005; Klepper,
2002a). Market growth creates opportunities for additional i rms to enter, and evolution
of the industry opens new opportunities for market and application diversii cation for
existing i rms and spinof s in a cluster.
Entrepreneurs may be motivated to found competing i rms when the growth of their
current employer indicates a substantial demand for its products and services. High
growth can indicate the presence of unmet demand and great market opportunities
(Romanelli and Schoonhoven, 2001). In these cases, employee learning becomes impor-
