RKSi =γ10+γ11πie+γ12MPOSi+γ13RBASEi+γ14RISj+γ15ITECHk+ε1
SCKSi =γ20+γ21πie +γ22MPOSi +γ23RBASEi +γ24RIS +γ25ITECHVk+ε2 (1)
i 20 21 i 22 i 23 i 24 j 25 k 2
XSCKSi =γ30+γ22πie+γ32MPOSi+γ33RBASEi+γ34RIS +γ35ITECHk+ε3
i 30 22 i 32 i 33 i 34 j 35 k 3
where πei is the expected level of post innovation returns, MPOSi is a group of
variables representing the market position of the firm, RBASEi is a group of variables
reflecting the strength of the firm's internal resource base, RISj is group of variables
reflecting the strength of the regional innovation system within which the firm is
located, and ITECHk is a series of indicators reflecting the character of the industry in
which the firm is operating.
The inclusion of variables to represent the market position of the firm (i.e. MPOS) is
intended to reflect issues of appropriation and potential Schumpeterian or monopoly
effects related to plant size. CDM, for example, in their model for R&D investment,
include measures of firms' market share and diversification. Indicators of business
size (and size squared) are also included by CDM, LH and LR to reflect potential
scale effects. The characteristics of the internal resource-base of the business are less
well represented in CDM and LH, although both include a measure of the quality of
firms' workforce. In their knowledge sourcing equations, LR also include measures
relating to the nature of firms' production activities and the organisation of any R&D
being undertaken in-house.
Other factors included in equation (1) reflect the industrial and regional context in
which the firm is operating. In terms of the technological characteristics of the
industry, for example, CDM include dummy variables to indicate whether
technological change in each sector is characterised by demand pull or technology
push and a series of industry dummy variables2. LH (2000) on the other hand use a
technological classification (their INTE dummy) indicating whether the firm is
operating in knowledge-intensive, labour intensive or capital-intensive industries. LR
adopt a more explicitly empirical approach focusing on the organisation of production
activity within the industry (i.e. concentration ratios, MES estimates) and the general
level of knowledge sourcing activity in the industry. Other recent studies (e.g. LR
Roper et al., 2001) have also suggested that the spatial context within which a firm is
operating - i.e. the regional innovation system (e.g. Braczyk et al., 1998) may also
have an important influence on its innovative capacity. CDM and LH (2000, 2001)
use no spatially distinct variables in their analysis. LR, however, do indicate the
importance of a range of spatially distinct agglomeration (e.g. population density)
industrial composition (proportion of employment in high-tech sectors, size structure
of local firms) and R&D investment variables (e.g. government and private R&D
spending) for firms' knowledge sourcing activities.
Measuring the intensity of firms' knowledge sourcing through R&D is relatively
straightforward with standard indicators (used by CD, LH and LR) measuring R&D
2. These are derived by CDM from the 1990 French Innovation Survey and express
whether in the opinions of the firms surveyed demand and technology factors had a
'weak', 'moderate', or 'strong' influence on its innovative activities over the
preceding five years (CDM, p. 121).