You are on page 1of 16

Innovation and exports: New evidence

on persistence and complementarities

among Spanish manufacturing firms

Dragana Radicic, Lincoln International Business School

Sara Aylon, University of Girona
Research questions

Innovation and exports are considered to be vital driving

forces for enterprises to gain a competitive advantage.
Two research questions:
1. Is there persistence in technological (product and process)
innovations and exports?
True state dependence or spurious?
2. Is there complementarity between technological
innovations and exports?
True vs spurious state dependence
 Genuine or true state dependence
 Decision of a firm to innovate in one period in itself increases the
likelihood of innovation in a subsequent period.
 Decision of a firm to export in one period in itself increases the likelihood
of exporting in a subsequent period.
 Related to path dependence
 Spurious state dependence
 Observed and unobserved firm characteristics positively affect firms’
persistent activities.
 Past dependence suggests that a process is shaped only by the initial
conditions and by observed and unobserved heterogeneity.
 Past dependence is in line with the resource-based view of the firm,
whereby innovation persistence is a result of intrinsic characteristics of a
Persistence in product innovation

R&D sunk costs

‘Success breeds success’ hypothesis
Dynamic increasing returns and the cumulative
nature of knowledge
Persistence in process innovation

R&D sunk costs

Process innovation is likely to be less persistent than product
Industry characteristics
It often occurs in mature industries, in which the importance of
process innovation is reflected in a firm’s efforts to implement
more efficient production methods, rather than new products.
Persistence in exports

 The source of true state dependence is related to sunk costs incurred

when firms enter new markets.
 The accumulation of export experience plays a significant role as
 The dynamic nature of exports is reflected in the path-dependence
view of export behaviour, which is consistent with the resource-
based view of the firm and in particular with the dynamic-
capabilities approach.
 If, however, observed and unobserved heterogeneity and initial
conditions shape a firm’s current propensity to export, then the
observed persistence is spurious.
Complementarity between product and process
 Complementary nature of technological innovations is one of the crucial
factors in identifying their different patterns of persistence.
 Implementing new products or upgrading existing ones might render process
innovation necessary (e.g. new production lines).
 Process innovation might allow firms to enhance the quality of their existing
products or to manufacture new products.
 Most previous empirical studies support this theoretical argument.
Complementarity between innovation and
 In relation to the interdependence between innovation and exports, the
cognitive theory of the firm (i.e. the competence-based view of the firm)
posits that these strategies are complementary.
 The premise of the approach is that firms, as learning organizations, adopt
both strategies (innovation and exports) as mechanisms for accumulating
knowledge, enhancing their capabilities and contributing to their
competitive advantage.
 Most empirical studies exploring the interdependence between
technological innovations and exports report their mutual

Longitudinal data from the Spanish Business Strategies

Survey (SBSS)
All manufacturing firms in Spain
Period 2001–2014
Model specification
 Dependent variables
 Three binary variables measuring product and process innovations and exports.
 Independent variables
 Four proxies for absorptive capacity: Internal R&D intensity, External R&D
intensity, R&D department and R&D staff
 Firm characteristics: firm size, firm age, labour productivity, belonging to a group
and market share.
 Industry heterogeneity: inclusion of dummy variables for high-technology
industries, medium high-tech, medium low-tech and low-technology industries.
 Time dummy variables

𝑝𝑖𝑡∗ = 𝛼0 + 𝛼1 𝑃𝑅𝑖𝑡 + 𝛼2 𝐸𝑖𝑡 + 𝛼3 𝑃𝑖𝑡−1 + 𝛼4 𝑃𝑅𝑖𝑡−1 + 𝛼5 𝐸𝑖𝑡−1 + 𝛼6 𝑍𝑖𝑡′ + 𝑎𝑖 + 𝑑𝑖𝑡

𝑝𝑟𝑖𝑡∗ = 𝛽0 + 𝛽1 𝐸𝑖𝑡 + 𝛽2 𝑃𝑖𝑡−1 + 𝛽3 𝑃𝑅𝑖𝑡−1 + 𝛽4 𝐸𝑖𝑡 −1 + 𝛽5 𝑍𝑖𝑡′ + 𝑏𝑖 + 𝑒𝑖𝑡

𝑒𝑖𝑡∗ = 𝛾0 + 𝛾1 𝑃𝑖𝑡−1 + 𝛾2 𝑃𝑅𝑖𝑡−1 + 𝛾3 𝐸𝑖𝑡−1 + 𝛾4 𝑍𝑖𝑡′ + 𝑐𝑖 + 𝑓𝑖𝑡

Empirical strategy
 A joint dynamic random effects probit model
 jointly models product innovation, process innovation and exports
 takes into account true state dependence and feedback effects
 controls for observed and unobserved characteristics, as well as for
initial conditions
 Feedback effects between innovation and exports render them
 The recursive structure of the model
 process innovation and exports measured at time t enter as
explanatory variables in the first equation for product innovation.
Strong evidence of persistence in technological innovations
and exports (both genuine and spurious).
Evidence of complementarity between product and process
No complementarity between product innovation and
Marginally significant complementarity between process
innovation and exports.
Policy implications
 Current policy effects in promoting technological innovation and
exports have not only contemporaneous effects, but a rather long-
lasting influence on the decisions of firms to innovate and export.
 A potential synergistic effect of public support for innovation and
Coordination between public agencies supporting innovation and
those supporting exports
 Significant effects of observed and unobserved firm characteristics
suggest that policy makers should also focus on these firm-specific
factors in promoting technological innovations and exports.
Limitations and future research

Comparison with other countries, in particular, those

at a similar level of innovation performance.
SMEs and large firms are not analysed separately.
The sample does not include the service sector.