Technological Collaboration: Bridging The Innovation Gap Between Small And Large Firms
ABSTRACT The economics of recombinant knowledge is a promising field of investigation. New technological systems emerge when strong cores of complementary knowledge consolidate and feed an array of coherent applications and implementations. However, diminishing returns to recombination eventually emerge, and the rates of growth of technological systems gradually decline. Empirical evidence based on analysis of the co-occurrence of technological classes within two or more patent applications, allows the identification and measurement of the dynamics of knowledge recombination. Our analysis focus on patent applications to the European Patent Office, in the period 1981-2003, and provides empirical evidence on the emergence of the new technological system based upon information and communication technologies (ICTs) and their wide scope of applications as the result of a process of knowledge recombination. The empirical investigation confirms that the recombination process has been more effective in countries characterized by higher levels of coherence and specialization of their knowledge space. Countries better able to master the recombinant generation of new technological knowledge have experienced higher rates of increase of national multifactor productivity growth.
- [Show abstract] [Hide abstract]
ABSTRACT: There are several studies in entrepreneurship investigating determinants of innovation outcomes in SMEs. Although entrepreneurs’ entrepreneurial creativity is often seen as a prerequisite, previous research indicates it is not an exclusive determinant of innovation. We use theoretical logics of social cognitive theory and innovation theory to develop a conceptual model of entrepreneur’s creativity, self-efficacy, and innovation outcomes. The model is then tested on a large sample of small and medium firms from two distinct economies: the United States and Slovenia. Empirical findings partially support the proposed moderation effects of entrepreneurial self-efficacy, but with the same variations between countries. The implications of these results in relation to entrepreneurship theory and practice are discussed.Small Business Economics 06/2014; 43(1). · 1.55 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: This research uses confirmatory factor analysis and structural equation modelling to examine how organizational size -made up of four dimensions -control, resources, trust and complexity -impacts on utilization of industry-led supply chain innovation capacity in a traditional agribusiness industry, the Australian beef industry. It confirms small business rather than larger business accords greater importance to exploiting supply chain dynamic capabilities, particularly in relation to utilizing industry –led supply chain innovation capacity. For small business in Australian beef supply chains, being agile and able to adapt and align their business practices with supply chain partners is integral to ensuring these businesses remain relevant and competitive in this market. In theoretical terms this is supported by authors in the dynamic capabilities literature as they argue these types of capabilities enable organizations to innovate faster (or better), often leading to the creation of newer sources of competitive advantage.
- [Show abstract] [Hide abstract]
ABSTRACT: This study investigates product, process, and management innovation among a sample of Spanish manufacturing small and medium-sized enterprises (SMEs) during the current economic downturn and a period of economic growth. Tobit analyses examined whether the level of process, product, and management innovation changed during the recessionary period relative to the expansionary period; and MCO estimations were used to show whether the recession affected the relationship between SME innovation and performance. Three main findings are: (1) innovation among Spanish manufacturing SMEs declined during the recent economic crisis; (2) the type of innovation at Spanish manufacturing SMEs changed during different economic conditions; and (3) innovation was positively associated with firm performance during the economic expansion and recession years. The results demonstrate the importance of adopting innovation into SME strategy over the business cycle. The findings have relevance to SMEs, government policymakers, and providers of consulting services.Journal of Small Business Management 01/2014; 51(4):578-601. · 1.39 Impact Factor
Working Paper 06-66
Business Economics Series 20
TECHNOLOGICAL COLLABORATION: BRIDGING THE
INNOVATION GAP BETWEEN SMALL AND LARGE FIRMS
María Jesús Nieto 1 and Lluís Santamaria 2
Universidad Carlos III de Madrid
Calle Madrid, 126
28903 Getafe (Spain)
Fax (34-91) 6249607
This paper analyses technological collaboration as an input to the innovation
processes of SMEs. Technological collaboration may be a useful mechanism to
offset some of the weaknesses in SMEs’ resource endowments and bring their
innovation capabilities closer to that of their large counterparts. The results, based
on a large longitudinal sample of Spanish manufacturing firms, show that
technological collaboration is a critical factor in improving the capabilities and
innovativeness of SMEs. While a general bridging of the gap between the
innovativeness of SMEs and large firms was observed, the most significant advance
was in product rather than process innovations.
Keywords: SMEs, technological collaboration, product innovation, process innovation
JEL code: O31; O32; L24
1 Sección de Organización de Empresas. Tel: +34-91-6245826.; fax: + 34-91-6245707. E-mail address:
2 Departamento de Economía de la Empresa. Tel.: +34-91-6248643; fax: + 34-91-6249607. E-mail address:
Most current economies are largely composed of small and medium-sized enterprises
(SMEs). In European Union, for instance, SMEs make up 99% of industry and
account for more than 70% of employment1. Their innovative capability is a key
driver of sustainable competitive advantage in today’s rapidly changing markets. This
situation has fuelled growing concern among managers and policy makers and has led
to a strong commitment to use policy initiatives to support innovation within SMEs
(Hoffman et al., 1998; Jones and Tilley, 2003).
This concern is also apparent in academic circles. In the economics of innovation and
technological change literature, the relationship between firm size and innovation
activity has received a good deal of attention (see Cohen, 1995; for an overview). The
Schumpeterian debate over which firms – large or small – are more able and more
likely to innovate is one of the oldest in political economics (Harrison, 1994), and has
lost none of its relevance for today’s world. Numerous studies have attempted to help
boost the innovative capacity of smaller firms by trying to explain differences in
innovation activity and pin down the key success factors (Acs and Audretsch, 1988a,
b, 1990; Kleinknecht and Reijnen, 1991; Nooteboom, 1994; Rothwell and Dodgson,
1994; Bougrain and Haudeville, 2002; Narula, 2004; Hewitt-Dundas, 2006; among
In general we can argue that SMEs have behavioural advantages over large firms,
which in turn have material advantages over SMEs (Rothwell, 1989; Rothwell and
Dodgson, 1994). SMEs, however, tend to be less innovative than large firms and to
dedicate less resources to the acquisition of external technologies (Bougrain and
Haudeville, 2002). This is clearly the case in Spain. In our representative sample of
Spanish SMEs only 36% claim to have introduced innovations compared to 65% of
the large firms.
Despite the increasing attention being given to the role of SMEs and innovation,
however, this is still an area that is under-researched (Edwards et al., 2005; O’Regan
et al., 2006). Many of the weaknesses of small firms when innovating can still be
observed and many questions remain unanswered. The research of Hewitt-Dundas
(2006) took up this issue and examined the resources and capabilities that firms
perceive to be constraining their innovation activity. One of the most conclusive
findings of this research was that the lack of external partners was an important
barrier to undertaking product innovation for small firms – this is a major difference
between small and large firms. Although there is research that stresses the role of
strategic alliances and collaboration as alternatives to undertaking innovation
activities (Love and Roper, 1999; Rogers, 2004; Nieto and Santamaria, 2007), we
need to delve more deeply into the specific role that innovation networking plays as a
possible determining factor in developing the innovation capacity of SMEs (Edwards
et al., 2005).
This study sets out to discover if collaboration, specifically technological
collaboration, enables SMEs to overcome their lack of resources and capabilities, and
thus boost their innovativeness. We consider that small firms have weaknesses that
put them at a disadvantage compared to large ones when it comes to innovation and
1 Based on data from the “Informe 2003/7 del Observatorio de la PYME Europea” elaborated by
that collaboration can go some way to levelling the playing field. We shall then
proceed to analyse the potential effect of collaboration on innovation outcomes (in
terms of process and product), and lastly go on to reveal if innovation networking can
close the gap between SMEs and large firms.
In the next section we place our research in the context of the classic debate over firm
size and innovation, paying particular attention to the innovation strengths and
weaknesses of SMEs. After this we review the existing literature on technological
collaboration, innovation and firm size. In the section on methodology we describe
the data, variables and statistical techniques used. We then go on to analyse our
results. The final section contains a discussion of the results and our conclusions.
2. Conceptual Foundations
2.1. Innovation strengths and weaknesses of SMEs
The large body of theoretical and empirical research on firm size and innovation
reveals the interest there is in this relationship (see, for example: Acs and Audrestch,
1988a, b, 1990; Cohen, 1995). The contradictory nature of both the conceptual and
empirical findings, however, does not provide clear guidance on what to expect in
general (Stock et al., 2002: pp. 541). Although it has not been possible to establish a
strong relationship between firm size and innovation per se, some empirical research
has suggested that small and large firms have different determinants of innovation
efforts (Van Dijk et al., 1997; Rogers, 2004) and do not pursue the same types of
Dirección General de la Pyme (Spanish Government). Firms with fewer than 200 employees are
classified as SMEs.
innovation (Nooteboom, 1994; Vossen, 1999; Cohen and Klepper, 1996; Fritsch and
These differences in innovation activity and innovation results can be explained by
the various advantages traditionally ascribed to large and small firms. The main
relative strengths of SMEs lie in behavioural advantages, while those of large firms
reside in their resource advantages (Rothwell, 1989; Rothwell and Dogson, 1994).
Smaller firms generally enjoy internal conditions that encourage innovativeness, such
as entrepreneurship, flexibility and rapid response (Schumpeter, 1942; Penrose, 1959;
Lewin and Massini, 2003). Shorter and more informal lines of communication allow
SMEs to take decisions more quickly than large firms (Nooteboom, 1994; Narula
2004). Small firms may also find it easier to adjust employee incentives to provide
optimal innovative effort thanks to their advantages in resolving agency problems
(Holmstrom, 1989; Zenger, 1994). In addition, the closeness of SMEs to the market
makes them faster at recognising opportunities (Rogers, 2004).
On the other hand, the relative weaknesses of small firms compared to large ones lie
in the constraints they face on gaining access to critical resources and capabilities for
innovation (Hewitt-Dundas, 2006). The advantages of scale and scope provided by
the size of large firms make them better equipped for innovations that require large
and specialised teams or sophisticated equipment (Cohen and Klepper, 1992). When
licensing methods are not available, then, large firms have a greater incentive to
pursue all types of innovation (especially process innovation), as their higher sales
volumes allow them to spread the fixed costs of innovation over a larger sales base
(Cohen and Klepper, 1996).
Large firms, in addition to their advantages of scale and scope, are more likely to
possess the experience and financial resources required for capability development
than small firms (Woo and Cooper, 1981). Studies of SME growth and innovation
consistently stress that a lack of finance is the most important constraint on innovation
(Vossen, 1998). Therefore, because of their lack of financial resources, SMEs are
often disadvantaged in their ability to gather technological resources (Bougrain and
SMEs are also usually at a disadvantage when it comes to intangible resources, as
they have access to a smaller range of knowledge and human capital skills than large
firms (Rogers, 2004). While it is true that SMEs may be more efficient at retaining
and motivating specialists through performance-contingent contracts (Zenger, 1994),
they face greater problems to recruit highly skilled staff (Barber et al., 1989) and tend
to invest less in on-going employee training than larger firms (Brown et al., 1990).
In summary, then, one would expect SMEs to face more constraints on their resource
endowments and – despite their behavioural advantages – more obstacles to
innovation than large firms. SMEs, however, have alternatives to internal
development that may enable them to bridge the resource gap that exists with large
firms. Searching for complementary resources outside of the firm through network
relationships can be an efficient way to build up resource endowments (Choudhury
and Xia, 1999; Gulati et al., 2000).
2.2. Improving innovation capabilities through technological collaboration
According to Duysters et al. (1999), alliances have shifted from being regarded as a
peripheral aspect to a cornerstone of the firm’s technological strategy. In fact, over the
last two decades there has been a tremendous increase in the use of external networks
by firms of all sizes (Hagedoorn, 1996, 2002). Their usefulness has been examined in
various studies that have included networking to capture a contextual determinant of
innovation. Becheikh et al. (2006) found that in studies including a variable to capture
this networking effect2, the relationship between innovation outcomes and networks
(or external interactions with customers, suppliers, universities, research centres and
other actors of a firm’s environment) was either positive or insignificant, but never
A possible explanation for this increase, apart from the declining costs of monitoring
and exploiting networks, is the growing need for firms to possess multiple
technological competences (Granstrand et al., 1997). In this way, technological
collaboration is seen as a strategic mechanism to achieve several objectives
(Cassiman, 1999; Hagedoorn et al., 2000; Bayona et al., 2001; Caloghirou et al.,
2003; among others): 1) to increase the technological capabilities of the firm; 2) to
gain access to new markets and to exploit new business opportunities; 3) to have
access to public funding; and 4) to complete the innovation process.
Regarding small firms, external networks specifically create unique benefits and
challenges, as a growing number of studies suggest (Powell et al., 1996; Zahra et al.,
2000; Sarkar et al., 2001). Most recently, Hewitt-Dundas (2006) found that while a
lack of partners for innovation had a negative impact on the ability of small firms to
undertake innovation, it did not have a significant effect on the probability of
innovating in larger firms. Her resource-based interpretation of this finding is that the
external resources and capabilities that small firms can access through external
2 In their survey of empirical research on innovation from 1993 to 2003.
innovation partnerships may provide them with the stimulus and capability to
innovate that they would not otherwise have.
SMEs, then, have to use alliances astutely to overcome barriers to growth imposed by
absolute limits to resources (Ahern, 1993; van Dijk et al., 1997). Networks allow
SMEs to receive and decode flows of information. They reinforce SMEs’
competitiveness by enabling them to access new knowledge, sources of technical
assistance, expertise, sophisticated technology and market requirements; they also
strategically reduce the irreversibility costs of the innovation process (Bougrain and
Houdeville, 2002; Freel, 2005).
Rogers (2004) points out that SMEs may rely more heavily on external knowledge
networks as an input to innovation than do large firms. Given that small firms seem to
have potentially more to gain from innovative partnerships than larger firms, the very
success of SMEs vis-à-vis their larger competitors may be due to their ability to use
external networks more efficiently (Nooteboom, 1994; Rothwell and Dodgson, 1994).
Based on the preceding discussion, we conclude that technological collaboration may
be a good way of strengthening and complementing the resource endowments and
capabilities of SMEs and of improving their innovativeness. In the following sections
we shall empirically examine how collaboration – along with other factors – helps to
bridge the innovation gap between SMEs and large firms.
3.1 Sample and Data
The source for our empirical analysis is the Spanish Business Strategies Survey
(SBSS). This is a firm-level longitudinal survey compiled by the Spanish Ministry of
Science and Technology and the Public Enterprise Foundation (Fundación Empresa
Pública - FUNEP) from 1991 to 2002. The SBSS covers a wide range of Spanish
manufacturing firms operating in all industry sectors; approximately 1,800
observations are available for each year. The sample is representative of the
population of Spanish manufacturing firms; it is random and stratified according to
firm size and industry sector (for more information on the sample see Huergo and
Jaumandreu, 2004). The 1998 survey was the first to give information on firms
engaged in technological collaboration, including partner specifications (i.e., type of
partners). Consequently, our study is based on data for the period from 1998 to 2002.
Our final sample contains 6,500 observations from 1,300 firms that have remained in
the survey during the five-year period.
3.2 Model specification
We have taken the difficulties SMEs face to obtain the same innovation outputs as
their larger counterparts as our starting point to specify the different models.
Furthermore, technological collaboration – along with internal resources – is a key
input to firms’ innovation processes. Therefore, we specified two series of models to
measure the impact of collaboration as an input to innovation processes on SMEs and
(1) The first series of models examine the relationship between SMEs and innovation
outputs, in terms of process and product innovations. This relationship is
analysed first by considering the category of SME and then by introducing the
interaction effect between being a SME and to be engaged in technological
agreements. These models will enable us to recognise how being a SME affects
the probability of innovating, and if this a priori negative effect varies among the
firms that collaborated.
(2) The second series of models look at how technological collaboration acts as an
input to improving the innovation capacity of the firm. In this case, different
models will be specified for a sub-sample of SMEs and a sub-sample of large
Model 2A (Sub sample of SMEs):
Model 2B (Sub sample of large firms):
Industryresources Intion Collaboratt improvemenness Innovative
We considered innovation output indicators to estimate the first series of models.
These were measured using two variables:
(1) Product Innovation (PRODUCT-INN). This is a dichotomous variable that takes
the value 1 when the firm declares at least one product innovation in the survey
year; otherwise its value is 0. A firm was considered to have achieved a product
innovation if it replied positively to at least one of the following items: i) the
product incorporated new functions; ii) the product had a new design and
appearance; iii) the product incorporated new materials and/or iv) the product
incorporated new components.
(2) Process Innovation (PROCESS-INN). This is a dichotomous variable that takes
the value 1 if the firm has achieved at least one process innovation in the survey
year; otherwise its value is 0. A firm was considered to have achieved process
innovations when: i) new machinery had been introduced; ii) new methods of
organising production had been introduced; or iii) both of the above occurred.
To estimate the second series of models we had to construct an indicator that captured
some type of improvement in innovation capacity. This improvement was measured
via two variables (see the Appendix for a detailed description of their construction):
(1) Improvement in product innovation capability (IMPROV-PD). It is a dichotomous
variable that takes the value 1 when the firm ‘improves’, going from being a non-
innovating firm in t-1 to an innovating one in t. It also takes the value 1 when the
firm ‘keeps’ its condition of innovating firm. It takes value 0 if the firm’s situation
has not changed and it continues to have no product innovations in year t. It also
takes the value 0 when the firm “gets worse” going from being an innovating firm
in t-1 to a non-innovating one in t.
(2) Improvement in process innovation capability (IMPROV-PC). This dichotomous
variable is constructed in the same way than previous one, but considering process
innovations instead of product innovations.
Independent and control variables
Given the objectives of our research, we tried to measure the potential innovative
behaviour of SMEs, especially how this behaviour was affected by technological
collaboration. To do this, we sorted the firms by size using a dichotomous variable
(SME) that takes value 1 when the firm has fewer than 200 employees, and value 0 if
there are 200 employees or more.
To take account of technological collaboration, we constructed a dichotomous
variable that indicates whether firms have collaborated technologically with other
firms or research organisations (COLLABORATION). This is included in the models
as a lagged variable.
We also included a series of variables to measure the firm-level characteristics related
to resource endowments: technological resources, financial resources, ownership,
commercial resources, organisational resources and relative scale. These are
important as they may have an effect on innovativeness.
Technological resources. To accurately measure the effect that technological
collaboration has on innovativeness (and the improvement in innovation capabilities),
we needed to take into account firms’ internal resources, particularly their
technological resources. For this reason we constructed a variable for R&D intensity,
measured as the ratio of R&D expenditure to total sales (R&D). This is included in the
models as a lagged variable.
Financial resources. Following Galende and Suarez (1999), we captured firms’
financial autonomy and resources by introducing the level of debt – calculated by
taking the ratio of Total Debts to Total Liability (LEVERAGE).
Ownership. Numerous studies have recognised the effect of ownership structure on
innovation, capturing its influence by focusing on foreign ownership (see Becheikh et
al., 2006). For this reason we included the percentage of foreign equity in a firm’s
Commercial resources. A good indicator of reputation and commercial activities is
the possibility that the firm exports part of its production to other markets (Galende
and Suarez, 1999). Therefore, we used export intensity – calculated as the ratio of the
firm’s sales in foreign markets to total sales (EXPORT).
Organisational resources. The age of the firm is a possible measure of its
organisational resources (AGE). It is a variable commonly used to measure the
experience and the learning of the firms in empirical studies of innovation (Kumar
and Saqib, 1996).
Relative scale. In addition to classifying firms as SMEs or large firms based on
number of employees, we felt that it was prudent to include sales figures (SALES) to
control for the relative scale of their activities. We also included these figures squared
to capture possible non-linear effects of the variable (SALES2).
Industry characteristics. Industry effects are critical control variables for innovation.
Because of size, firms can be more or less productive depending on the industrial
sector they are operating in (Nooteboom, 1994; Acs and Audrestsch, 1988b); the
knowledge requirements will also vary from sector to sector (Rothwell, 1991). The
classification proposed by Pavitt (1984) makes it possible to capture the impact of the
industrial sector as well as the purely technological effects. This classification
includes four dummy variables: 1) supplier-dominated sectors (SUPPLIER); 2) scale-
intensive sectors (SCALE); 3) sectors with specialised suppliers (SPECIALISED); and
4) science-based sectors (SCIENCE).
Lastly, year dummy variables were included in the models (YEAR). Table I contains
the descriptive statistics and correlations of the independent and control variables
used in this study.
[Table I about here]
3.3 Estimation techniques and data analysis
First, we performed a series of Tests of difference between means; the results of these
Tests are presented in Table II. The results reveal that small firms are much less likely
to collaborate than large ones: less than one out of five small firms decides to
collaborate, compared to almost three out of four large firms. Significant differences
in the measures of innovation outcomes can also be observed in both process and
[Table II about here]
The differences in these variables for the sub-sample of collaborating firms are far
smaller for the two measures of innovation outcome; in fact, for product innovations
they are not significant. The percentages of innovating firms, both for the full sample
and the sub-sample of collaborating firms, are presented in Figure I.
[Figure I about here]
To test the robustness of these differences between SMEs and large firms, and
observe the effect of collaboration on both types of firm, we specified a series of
models. As we have already mentioned, the first series of models analyse the impact
of collaboration on different types of innovation (product and process), while the
second series of models explore the effect of collaboration on improvement in
innovativeness. As our dependent variables are dichotomous, estimation models such
as logit or probit (Aldrich and Nelson, 1984; Greene, 2000) would normally be
appropriate. Given that product and process innovations may be related to each other
(Martínez-Ros, 2000; Fristch and Meschede, 2001), the error terms of the two models
are likely to be correlated. Thus, an extension of probit known as bivariate probit
(Greene, 2000) is usually a more appropriate estimator. The bivariate probit model
has the following specification (Breen, 1996):
iy = si
iy = si
iy = si
iy = si
This model produces estimates of the coefficient vectors β1 and β2 for the two
equations, of ρ (the correlation between the error terms εij of the equations), and of
the standard errors for these parameters. We can then test if the correlation between
the equations is statistically significant and decide whether the bivariate estimator is
the most appropriate model3. The bivariate probit model was estimated using the
Stata 8 routine, based on the method of simulated maximum likelihood.
4. Empirical findings
Table III gives the results of the models measuring the impact of a series of variables
on the likelihood of process and product innovations. The Wald test indicates high
joint significance of the variables for both models. The ρ parameter is highly
significant in both models, signalling that the error structures of the equations are
correlated. This suggests that product and process innovation are not independent and
the bivariate model is the correct specification.
The models corroborate the results of the Tests of difference between means. The
negative effect that being a small firm has on the likelihood of achieving both types of
3 If the correlation is not significant, separate (univariate) probit estimation of the equations is
preferable as bivariate probit is less efficient than estimating separate models when the error terms are
not correlated (Greene, 2000: 853-4).