Journal of Cleaner Production xxx (2015) 1e10

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Does the development of environmental innovation require different

resources? Evidence from Spanish manufacturing firms
Giulio Cainelli a, Valentina De Marchi b, *, Roberto Grandinetti b
a
University of Padova and CERIS-CNR, Milan, Italy
b
University of Padova, Via del Santo 33, 35123 Padova, Italy

a r t i c l e i n f o a b s t r a c t

Article history: In recent years, environmental innovations are getting increasingly at the center both of the scientific
Received 9 May 2014 debate and of the policy-maker agenda. Adopting the resource-based view, this paper provides a
Received in revised form comprehensive framework for understanding the specific role of e internal, external and hybrid e re-
2 February 2015
sources in the environmental innovation development. Using a dataset of 4829 Spanish manufacturing
Accepted 3 February 2015
Available online xxx
firms and estimating probit models, it is found that internal resources have greater importance for
environmental innovations. Moreover, green innovators seem to be characterized by more intensive
external relationships. For hybrid resources, acquisition of equipment but not of patents is more relevant.
Keywords:
New environmental products and processes
Finally, evidence suggest the relevance of complementing the analysis of the external factors triggering
Innovation resources environmental innovations with those of the internal resources the firm has access to in order to fully
R&D understand and support the development of environmental innovations.
Cooperation © 2015 Elsevier Ltd. All rights reserved.
Spain

1. Introduction development is a key issue for both policy and research. The first
determinants or drivers of EIs to be investigated have been those
Firms are increasingly challenged to include environmental exogenous to the firm, namely environmental regulations (e.g.
concerns in their business activities. While in the past firms were Brunnermeier and Cohen, 2003) or the role of stakeholders in
considered to be mainly the source of the pollution problem, more spurring firms to adopt greener practices (e.g. Kammerer, 2009).
recently they are increasingly seen as a possible solution, largely More recently, a conspicuous stream of literature has highlighted
thanks to their innovative activity. According to a widespread the importance of complementing the analysis of such external
definition, environmental innovation (EI) consists of ‘the produc- drivers with firm-level factors, both specific to EIs e such as the
tion, application or exploitation of a good, service, production presence of environmental management systems like EMAS and
process, organizational structure or management or business ISO14000 (e.g. Rennings et al., 2006; Wagner, 2007) e and common
method that is novel to the firm or user and which results, to all innovations, including R&D investment, the purchase of
throughout its life cycle, in a reduction of environmental risk, patents and cooperation aimed at innovation (e.g. De Marchi, 2012;
pollution and the negative impacts of resource use compared to Horbach et al., 2012). Regarding these second type of variables, few
relevant alternatives’ (Kemp and Pearson, 2008, p. 7). This defini- papers (five out of 17 in a recent review by Del Río et al., 2013)
tion is purposefully broad and includes all the changes in the measure the possible differences between green and non-green
product portfolio or in the production processes, whether radical or innovations, and their empirical focus does not allow for a full
incremental and whether initially intended or not, which tackle understanding of key elements, such as the role of resources and
sustainability targets, such as waste management, eco-efficiency, capabilities in environmental innovations.
reduction of emissions, recycling and eco-design (Rennings, Against this literature on firm-level determinants of EI, the aim
2000; Markusson, 2010). of this paper is to provide theory-driven evidence that the
Considering the relevance of EIs from an economic and social resources employed by a firm in developing (and not merely
perspective, the understanding of the specific determinants of their adopting) green innovations differ from non-green innovations.
More specifically, in this paper it is examined whether the two
categories of innovators (green vs. non-green) make use of different
* Corresponding author. Tel.: þ39 049 827 3848. types of resources. For this purpose, the distinction traditionally
E-mail address: [email protected] (V. De Marchi).

http://dx.doi.org/10.1016/j.jclepro.2015.02.008
0959-6526/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
2 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10

adopted in the (general) innovation literature and based on the Cohen and Levinthal (1990) e corresponds to the firm's ability to
resource-based view e distinguishing between internal and monitor, value, assimilate and exploit external knowledge. Such
external resources e is used together with an additional ‘hybrid’ ability is twofold: it allows the firm to identify which specific
category, added because of its relevance in environmental inno- external knowledge is needed for its innovation needs and to
vation development. assimilate the knowledge and employ it effectively in its innovation
In addition to providing a comprehensive framework for un- process.
derstanding the specific role of resources in EI development,
another important contribution of this research regards the 2.1. Different resources for green innovations?
method adopted. Existing firm-level econometric studies on EIs are
cross-sectional (e.g. Kammerer, 2009; Mazzanti and Zoboli, 2009; While the resource-based view has been widely adopted in the
Horbach et al., 2012). However, this paper employs a panel anal- general innovation literature, this has not been the case for EIs. In
ysis that illuminates which causal relationships are in place while the following, the framework discussed before has been adopted to
measuring the differential role of the variables for green vs. non- understand EI development. In particular, this paper argues that, in
green innovators.1 order to develop EIs, all three categories of resources mentioned
above play a more vital role than has been previously suggested by
the literature on general innovations.
2. Resources for innovation Especially but not exclusively when it comes to cleaner tech-
nologies, EIs are ‘complex and sophisticated’ (Del Rio et al., 2011, p.
Since the 1980s, the resource-based view has provided a 1172). A first argument in support of the peculiarity of EIs for firms
powerful theoretical framework for understanding the role of re- is that they are, on average, more complex than non-green in-
sources in supporting a firm's competitive advantage and innova- novations (Andersen, 1999; De Marchi, 2012). Such trait is not
tion processes. The initial focus of theorists has been on distinctive necessarily related to the process or product technology (De Marchi
resources and competences being internal to the firm (Wernerfelt, and Grandinetti, 2013) but is linked to other dimensions. Using
1984; Barney, 1991). The empirical literature has measured inter- concepts developed by innovation management scholars when
nal resources mostly in terms of a firm's R&D structure, but several assessing the complexity of an innovation, it is possible to assert
studies suggested that training (internal and external to the firm) is that EIs are, on average, characterized by higher levels of novelty,
another important factor influencing a firm's innovation perfor- uncertainty and variety with respect to the traditional technolog-
mance (e.g. Gupta et al., 1993; Laursen and Foss, 2003). ical or market domain the firm usually competes within. As will be
In later years, researchers pointed to the importance of external discussed in the following, novelty spurs the need for greater
resources as well, leveraging the awareness that it is not convenient internal innovation resources, uncertainty the need for external
of even possible for firms to internally develop all the resources resources and variety the need for both external and hybrid
needed to compete, innovate and grow in their competitive envi- resources.
ronments. Thus, firms seek access to the necessary resources Secondly, eco-innovation scholars hold that EIs are a peculiar
through inter-organizational relationships, alliances and networks subset of innovations because they are characterized by the so-
(Das and Teng, 2000; Ireland et al., 2002; Lavie, 2006). This is called ‘double externality problem’ (see e.g. Rennings, 2000;
particularly evident when it comes to innovation, as Pittaway et al. Horbach et al., 2013), which highlights the key role of existing
(2004) show in their review of research linking firms' networking environmental laws and other regulatory factors (regulatory push
behaviour with their innovative capacity. Even firms having strong effect) and consumers' requests (demand pull effect) to overcome
internal R&D activities and investing significantly in training hu- the low incentive to eco-innovate. Considering the increasing reg-
man resources for innovation often rely on cooperation to diversify ulatory and demand pressures that have characterized EU countries
risks or gain access to competences that would be too expensive or in the years following the initial work by Rennings (2000), firms are
time-intensive to develop internally. therefore likely to have a higher need for resources in order to be
Beyond the internal and external resources traditionally ana- able to respond to such inputs while remaining competitive. Both
lysed in the literature, following the rationale of the resource-based arguments for EIs' diversity play a role in explaining the importance
view and considering their importance to innovation, the category of the three categories of resources analysed.
of hybrid resources was included. Hybrid resources encompass the
knowledge embodied in patents, R&D services or the machineries 2.2. Green innovations and internal resources
that firms acquire from external suppliers. As explained by Edith
Penrose (1959, p. 79), ‘many developments in technological The innovations developed by firms can be distinguished based
knowledge become available to firms not simply as new knowl- on the degree of novelty or newness e that is, the degree of
edge, but physically embodied in the form of the capital equipment familiarity with the given technology (Tatikonda and Rosenthal,
they buy’. This third type of innovation resources can be considered 2000). Developing internal resources is a key determinant of a
a hybrid one in the sense that it consists of knowledge developed firm's capability to innovate any of its products or production
outside the company that then becomes an integral part of the methods but is even more important when it comes to innovations
company's stock of internal resources. Moreover, in order for the with a high degree of novelty. As supported by the resource-based
knowledge acquired to be effectively useful in the innovation pro- view, to make such innovations, the gap that a firm must fill to
cess of the firm, the firm itself has to be endowed with an adequate achieve a good fit between the requirements for the innovation
absorptive capacity that e according to the seminal contribution by project and the firm's resources and capabilities is large (Danneels
and Kleinschmidt, 2001).
EIs are, on average, characterized by a higher degree of novelty
1
Contributions that have studied EIs through panel data exist, but they have as compared to other innovations. In fact, EIs represent a techno-
either i) used industry as the level of analysis (e.g. Jaffe and Palmer, 1997; Del Río logical frontier in which firms, users and policy makers are often
et al., 2011), ii) had a different analytical focus, such as policy impacts (e.g. Kerr
and Newell, 2001) or iii) had a similar focus to the present study but failed to
inexperienced or ignorant, to quote the words used by Porter and
fully exploit the panel approach, instead just using some variables from previous van der Linde (1995a) in their seminal contribution on EI and
waves (e.g. Mazzanti and Zoboli, 2005; Horbach, 2008). competitiveness. Two decades later, their statement about the

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10 3

relative higher novelty of EI is still valid. In fact, the awareness of usually address such uncertainty by leveraging on external
environmental problems by firms and other relevant stakeholders resources, in particular co-innovating with other partners.
has grown, but so have environmental problems and technological EIs represent a frontier of innovation that is still very open in
requirements. The causes of this peculiarity with respect to other terms of technologies, production processes, products, business
innovations are multiple; Carrillo-Hermosilla et al. (2010) suggest models and consumer behaviours (De Medeiros et al., 2014). In
they are linked with the design of innovation, the user behaviours, such a context, both exploration and exploitation are trickier
the relationship between the product and the service, the config- because of the high uncertainty, which spurs firms to look for
uration of the value network and the governance of innovation. partners with whom to share the risk. On the one hand, EI-oriented
Having an adequate internal base of knowledge and skills to firms are facing technological uncertainties, especially considering
address environmental concerns is therefore crucial in enabling that in several industries there are no widely accepted standards,
firms to realize environmentally sound products, processes or either in terms of specific technological solutions or measures for
business models (Van Dijken et al., 1999). evaluating the environmental performance of products and pro-
Additionally, because of the ‘double externality problem’ cesses (Rennings, 2000; Sierzchula et al., 2012). On the other hand,
mentioned above and the subsequently increased pressure on firms, especially smaller ones, are often unaware of the market
firms, many have been convinced of the necessity to invest in potential of green products or the commercial payoffs of environ-
internal resources for EIs. Such an argument is consistent with the mental practices (Brammer et al., 2012). In addition, the role of the
winewin theory proposed by Porter and van der Linde (1995a, regulatory pushedemand pull effect in spurring EIs supports the
1995b), which suggests that firms investing in R&D in response need for higher external resources (Horbach et al., 2013). Following
to well-designed regulatory pressure contribute to environmental this discussion, the higher uncertainty for EIs suggests a high
improvements via EIs. Precisely, this is the so-called ‘weak’ version propensity to rely on external resources.
of the Porter hypothesis, which has been tested and confirmed by Indeed, the extensive empirical literature suggests that external
several empirical studies, as testified by the recent review by resources play a pivotal role in green innovations. An empirical
Ambec et al. (2013). study on Spanish manufacturing firms by De Marchi (2012)
Few contributions have empirically studied the distinctive role indicates that formal cooperation with external partners is even
of R&D in the introduction of EIs vs. other innovations. Despite more important for EIs than for other types of innovations. Horbach
there are contributions pointing to the fact that it may not be (2008) and Del Río et al. (2013) report a similar result for German
differentially important across the two types of innovation and Spanish firms. The contrary is true in the case of the Spanish
(Horbach, 2008; Cuerva et al., 2014), the majority of the empirical food and beverage sector analysed by Cuerva et al. (2014), which
literature supports the view that having a structured organization may point to the existence of important differences across
working on innovation is an important distinctive driver of EI (e.g. industries.
De Marchi, 2012; Del Río et al., 2013) and that such a difference in Such evidence has also been corroborated by studies focussing
regard to non-green innovation increases with the relevance of on specific external partners. In particular, cooperation with sup-
environmental concerns for the firm's innovation strategy (De pliers is important to ensure the supply of inputs or components
Marchi and Grandinetti, 2013).2 As far as personnel training is with eco-friendly features e which may not be readily available on
considered, it has been claimed to be particularly useful for EI, in the market e and to verify whether they fulfil the requirements
that it serves both as a competence-enhancing and motivating (Seuring and Müller, 2008; Albino et al., 2009). Studies on
factor (Sarkis et al., 2010). This is especially the case in small and knowledge-intensive business services (KIBS) have highlighted the
medium-sized enterprises (SMEs), where such training can suc- important role of such service firms in supporting their clients'
cessfully complement lower levels of other internal resources (see innovation in general, and green innovations specifically (Klewitz
Del Brío and Junquera, 2003). If there is evidence that the higher et al., 2012; De Marchi and Grandinetti, 2013). Cooperation with
the share of trained employees, the higher the probability a firm universities may be needed to introduce more complex forms of
will introduce EIs (e.g. Cainelli et al., 2012), no study so far has environmental innovation, whether they are products, processes or
empirically evaluated whether training plays a larger role in green organizational EIs (Wagner, 2007; Triguero et al., 2013). Moreover,
innovations than in non-green innovations. green strategies require manufacturers to interact with clients
Following the discussion presented so far, the first hypothesis is (consumers, retailers or downstream manufacturers) in order to
as follows: apply a life-cycle approach and ensure the recyclability of their
products (Handfield et al., 1997; Darnall et al., 2010; Thun and
Hypothesis 1. Supporting innovation via internal resources (in
Müller, 2010). The exploratory analysis by De Marchi and
terms of R&D investments or employee training) is a main driver for
Grandinetti (2013) takes this discussion a step further, suggesting
environmental innovators with respect to non-environmental
that the number of partners a firm collaborates with in the pursuit
innovators.
of innovation activities is higher for green innovators than non-
green innovators, and it increases with the number of environ-
mental concerns the firms address through their innovative efforts.
2.3. Green innovations and external resources Hence, the second hypothesis follows:

The more an innovation is novel, the more the uncertainty and Hypothesis 2. Cooperating with external partners for innovation is
risk associated with it increases (Daft and Lengel, 1984; Tatikonda more relevant for environmental innovators than non-environmental
and Rosenthal, 2000). Starting from the seminal contribution by innovators.
Teece (1986), innovation scholars agree in supporting that firms
2.4. Green innovations and hybrid resources
2
Such mixed results may be explained by the different empirical setting being Other than uncertainty, another dimension of innovation
considered (both in terms of sectors and countries analysed) or by the different
methods adopted to differentially measure green innovation and firms' resources
complexity to be considered is what the authors of this contribution
and capabilities, which often do not control for the potential bias emerging from refer to as ‘variety’, which plays a role when innovation develop-
the exclusion of non-innovators from the analysis. ment requires firms to combine a wide variety of resources and

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
4 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10

capabilities and look beyond their own boundaries (Srivastava and comparison between green and non-green innovators. The fact that
Gnyawali, 2011). As the resource-based view suggests, different the questionnaire was not designed specifically to investigate EIs
partners contribute different resources and technological capabil- allows it to avoid possible respondent bias in terms of over-
ities that improve and complement the firm's resources for inno- reporting the environmental performance of a firm to portray a
vation (Becker and Dietz, 2004; Nieto and Santamaria, 2007). The better social image. Second, being a panel, the dataset allows us to
more diverse the knowledge and competences needed to develop control for potential endogeneity problems stemming from the
an innovation process, the more the firm needs to rely on external omission of unobservable, firm-specific characteristics.3
resources to succeed, either by collaborating with other organiza- This paper focuses solely on manufacturing firms and identifies
tions or by acquiring them (e.g. via patents). green innovation based on self-reported data on the object of the
Indeed, environmental innovative solutions often stem from the innovations introduced, utilizing a question of the PITEC that asks,
original combination of knowledge and competences endowed by on a four-point scale, about the importance of the objective
different organizations (Halila and Rundquist, 2011). Variety “reduced environmental impacts” in driving innovation activities,4
supports the greater need for resources for EIs compared to other an approach already used in other studies (e.g. Horbach, 2008; De
innovations, both external (corroborating what has already been Marchi, 2012; Del Río et al., 2013).5 Unlike patent analyses (e.g.
stated previously in hypothesis 2) and hybrid ones. Very few papers Jaffe and Palmer, 1997; Berrone et al., 2013), this approach allows
have empirically tested the peculiar role of hybrid resources for EIs for the consideration of both radical and incremental, new-to-
and have never studied them in a comprehensive manner, allowing market and new-to-firm EIs and reflects the definition of EI
an analysis of them as a whole. This is the case, for example, of the adopted in the paper, which is widely accepted in the literature. The
acquisition of patents or of other external knowledge e often dependent variable, ENVINN, has a value of 1 if the company re-
studied together e that received very little attention from scholars ported that the importance of this object for the innovations,
and for which mixed results have been developed: evidence that introduced between 2008 and 2010, has been high or medium, and
they might not be significantly more important for EIs than for 0 otherwise. According to the PITEC dataset, 46.3% of the
other innovations (e.g. De Marchi, 2012; Horbach et al., 2012; Del manufacturing firms in Spain can be considered environmental
Río et al., 2013) co-exists with indications supporting the oppo- innovators based on this definition (see Table 1), a percentage that
site (De Marchi and Grandinetti, 2013). Empirical evidence is also is in line with data characterizing other European countries (see e.g.
scant as far as machinery and equipment acquisition is concerned. Chiarvesio et al., 2014 or Horbach et al., 2012 for the Italian and
Demirel and Kesidou (2011) find that this investment is recurrent German contexts, respectively).6
and enables important energy and material savings, both when it Given that the dependent variable is a dummy, the econometric
comes to end-of-pipe and integrated solutions. Horbach et al. specifications are estimated adopting a binary outcome model,
(2012) add that the role of machinery acquisition is peculiar to controlling for possible selection bias arising from the exclusion of
the propensity to introduce EIs. non-innovative firms from the analysis (maximum-likelihood
Despite the scant empirical evidence, following the theoretical probit model with sample selection or Heckman probit). In the first
argument on the higher variety that characterizes EI, a further stage, the so-called selection equation, a probit model is estimated
hypothesis is defined: that calculates the probability that a firm will become a techno-
logical innovator, using TECHINN as the dependent variable
Hypothesis 3. Hybrid resources are more relevant for environmental
(a dummy that takes on a value of 1 if the firm has introduced a
innovators than non-environmental innovators.
product or process innovation in the period 2008e2010 and
0 otherwise). In this analysis, the same regressors of the structural
3. Data, variables and econometric modelling equation are included, plus the dummy EXTRA_UE-MKT, capturing
whether the firm sells into extra-UE markets.7 This is because the
The data for the analysis are drawn from merging two different selection equation must contain at least one variable that is not in
waves of the Spanish Technological Innovation Panel (PITEC), the structural equation.
which include the periods from 2008 to 2010 and 2007e2005.
Whereas the majority of comparative analyses on EIs so far have
been performed in the German context, the choice of this empirical 3
Unfortunately, because of changes in the sample and the questionnaire and the
setting will allow us to verify the relevance of innovation resources low number of waves available, it is possible to perform panel data models on only
in the context of a moderately innovating country with a relatively two waves. More information on the dataset is available at the website http://icono.
fecyt.es/PITEC/Paginas/descarga_bbdd.aspx.
low level of environmental regulation stringency (Del Río et al., 4
The dependent variable is based on the question E.6 of the 2010 PITEC ques-
2013), where, therefore, it is even more important to understand tionnaire, asking to report how important have each of the objectives listed in the
at the firm level which innovation characteristics and strategies survey been in orienting the innovation activity of the firm. The survey listed 16
play a major role. objects, divided into four categories: product-related objects (5 objects), process-
PITEC is based on the Community Innovation Survey (CIS), one related objects (5), employmenterelated objects (3) and other objects (3),
including ‘reduced environmental impacts’, which has been used to build the
of the most used datasets in innovation studies (e.g. Laursen and
dependent variable.
Salter, 2006), which has been employed in studying environ- 5
The question used in Horbach et al. (2013) to disentangle green innovators from
mental innovations (e.g. Horbach, 2008). Administered in the other innovators asked whether the innovation led to significant improvements to
current form since 2003, it includes both SMEs and large firms. the environment or health conditions; that used by Del Río et al. (2013) referred
Initially, it included a sample of firms with 200 employees or more only to a high (and not also medium) positive impact on the environment.
6
In order to comply with the aim of the paper, that is to investigate innovation
and a sample of firms with intramural R&D expenditures. Starting
development rather than innovation adoption, firms that, in the period considered,
in 2004, however, it also included two samples of firms with fewer introduced innovations developed mainly by external partners rather than by the
than 200 but more than 10 employees (one of the firms reporting firms themselves were dropped from the analysis, reducing the initial sample by
external but no intramural R&D expenditures and one with no 6.45%.
7
innovation expenditure). The use of this dataset allows for impor- Several contributions in the literature suggest that export is not a factor when
explaining the decision to innovate green or not green (see e.g. Horbach, 2008; De
tant improvements to the knowledge regarding EIs. First, since it Marchi, 2012; Cainelli et al., 2012), whereas it has been demonstrated that being
contains information on innovative strategies and the structural active in international markets supports the decision to innovate (Kafouros et al.,
characteristics of all technological innovators, it allows for a 2008).

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10 5

Table 1 external firms or institutions, including supply chain partners,

Descriptive statistics: dependent and explanatory variables. knowledge providers such as KIBS, universities and research cen-
Mean Std. Dev. Min Max tres and competitors. Moreover, it has been tested whether firms' EI
ENVINN 0.463 0.498 0 1
activities benefited from the acquisition of external knowledge e
R&D 0.522 0.499 0 1 the variables EXT-R&D and EXT-KN indicate whether firms bought
TRAINING 0.117 0.322 0 1 extramural R&D or acquired patents or other forms of knowledge,
SKILL 19.16 28.71 1 100 respectively e or from the acquisition of new technologies and
INTRO 0.066 0.249 0 1
machineries incorporating the needed knowledge (EQUIPMENT).
COOPERATION 0.232 0.422 0 1
EXT-R&D 0.251 0.433 0 1 In order to correctly estimate the factors correlated with EI
EXT-KN 0.027 0.162 0 1 propensity, several controls were included. First, firm size was
EQUIPMENT 0.159 0.366 0 1 controlled through the inclusion of the variable SIZE, measured as
SIZE 4.011 1.752 0 10.586
the natural logarithm of the number of employees. In fact, the
REACTIVE 6.444 1.650 2 8
literature suggests that if SMEs engage in product, process and
organizational innovations (Klewitz and Hansen, 2014), they face
higher constraints in terms of financial and technological resources
The second stage e the estimate of the structural equation e than larger firms and may have a lower incentive to go green,
focuses on firms that have introduced technological innovations considering the lower scrutiny from stakeholders (Brammer and
but controls for possible sample selection bias by incorporating an Millington, 2006; Del Río, 2009). Moreover, the count variable
additional explanatory variable, the so-called Inverse Mill's Ratio REACTIVE was included, which was used to control for a firm's
(IMR). In addition, the panel nature of the dataset has been attitude toward innovation: ranging from 2 to 8, this variable
exploited by including in the econometric specifications the re- captures to what degree the perceived lack of need to innovate or
gressors as lagged variables (i.e., referring to the period the absence of a demand for innovation hampered the innovative
2007e2005). The inclusion of these lagged variables should miti- effort of the firm. Finally, variables capturing factors related to the
gate endogeneity and simultaneity problems.8 firms' location (four dummies considering whether a firm is based
For all specifications, we found that the Inverse Mill’s Ratios and in Madrid, Catalonia, Andalusia, or other Spanish regions, respec-
the correlations between the two error terms (see the Wald tests) tively) and industry specialization (20 industry dummies) were
were not significant. For this reason, we re-estimated our structural inserted, which may account for differences in regulatory pressure
equations using simple probit models finding very similar results across sectors. The inclusion of this set of covariates should mitigate
(available upon request). The only difference is related to the fact the problem of unobserved heterogeneity. Table 1 reports the
that the STATA heckprobit procedure estimates both equations descriptive statistics of the variables described, and Table 2 displays
using a maximum likelihood estimator. Anyway, we prefer to pre- the simple correlations among the regressors.
sent our results in this way not only to control for potential sample
selection, but also to show the impact of our main explanatory
variables on the general innovation variable (even if in this case an 4. Main results and discussion
‘additional’ regressor is taken into account). In fact, most of our
explanatory variables do not refer explicitly to environmental, but Results for the probit model with sample selection are reported
to product or process innovations in general. in Table 3, including both the selection equation, capturing the
To test the role of internal resources in enabling EI adoption with factors explaining the introduction of product or process in-
respect to other innovations, the dummy R&D was included e novations (green or not), and the structural equation, reporting
measuring whether the firm performed intramural R&D activities factors correlated with environmental innovation propensity as
during 2005e2007, which was adopted as a measure of the stock of compared to other innovators.
technological competences e and TRAINING, which indicates The results indicate that the internal innovative effort positively
whether the firms' personnel attended training specifically aimed affects green innovation propensity, and this factor is even more
at developing technological innovations. Furthermore, the variable important than for the introduction of other innovations, providing
SKILL was included, which captures the presence of highly skilled support for hypothesis 1. More precisely, having internal R&D
personnel within R&D offices, measured as the incidence of capacity (R&D) is positively and highly significantly correlated to
employees graduated in relation to the total number of R&D em- the introduction of EIs. This result is similar to the findings of other
ployees. The dummy variable INTRO captures whether the firm has studies in Spanish and Italian settings (e.g. Mazzanti and Zoboli,
made other investments to support the introduction of the inno- 2009; Del Río et al., 2013; Cuerva et al., 2014) but contrasts with
vation into the market. what has been found, for example, in Germany and France
In order to understand whether green innovators rely to a (Horbach et al., 2013). Moreover, training personnel specifically for
greater extent on external innovation resources, either by cooper- innovation activities (TRAINING) positively supports the introduc-
ating with network partners or acquiring such resources, a set of tion of EIs, as seen in Cainelli et al. (2012). In contrast with Horbach
variables was used. The dummy variable COOPERATION indicates (2008), the coefficient of the variable SKILL is not significant, which
whether a firm reported to have cooperated on innovation with may be explained by the differences in industry composition and
institutional structure between the two countries. This result sug-
gests that the education level of R&D workers does not affect the
8 introduction of EIs differently from other innovations. In this
An alternative econometric approach could be to estimate an ordered logit
model. In fact, the main variable, the ‘importance of reduced environmental im- regard, this variable is also not significantly correlated with the
pacts’, is measured on the basis of a four-point scale. However, since the share of propensity to become an innovator, whether green or not. The joint
censored observations (i.e. firms which have not introduced technological in- analysis of the variables SKILL and TRAINING may be interpreted as
novations) is equal to 18%, it was preferred to consider the problem of potential evidence that contextual, applied knowledge is the most relevant
sample selection explicitly, estimating the main structural equation as a second
stage of a probit model with the sample selection. Moreover, this approach allow
for developing (green) innovation or that, at least in the Spanish
greater comparison with the existing literature, where a large part of contributions context, the knowledge provided by universities is not relevant to
used a binary dependent variable approach (logit or probit). the innovation needs of firms.

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
6 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10

Table 2
Correlation matrix.

[1.] [2.] [3.] [4.] [5.] [6.] [7.] [8.] [9.] [10.] [11.]

[1.] 1.000
[2.] 0.138 1.000
[3.] 0.184 0.065 1.000
[4.] 0.154 0.084 0.316 1.000
[5.] 0.069 0.071 0.022 0.069 1.000
[6.] 0.050 0.027 0.035 0.043 0.131 1.000
[7.] 0.029 0.093 0.020 0.076 0.148 0.162 1.000
[8.] 0.110 0.099 0.150 0.176 0.206 0.274 0.176 1.000
[9.] 0.127 0.066 0.262 0.350 0.067 0.082 0.106 0.165 1.000
[10.] 0.037 0.047 0.233 0.138 0.023 0.045 0.037 0.071 0.120 1.000
[11.] 0.102 0.018 0.557 0.204 0.001 0.000 0.013 0.122 0.205 0.144 1.000

[1.] e ENVINN.
[2.] e SIZE.
[3.] e R&D.
[4.] e EXT-R&D.
[5.] e EQUIPMENT.
[6.] e INTRO.
[7.] e EXT-KN.
[8.] e TRAINING.
[9.] e COOPERATION.
[10.] e REACTIVE.
[11.] e SKILL.

In terms of external resources, the analysis suggests that green innovation studies but has had little application in the study of EIs
innovators fuel their innovation effort through interorganizational (an exception being Chiarvesio et al., 2014). The count variable
relationships (COOPERATION) more intensively than other COOPBREADTH was therefore included in the analysis, which takes
innovators (hypothesis 2), which is in line with the existing on values from 0 e if the firm cooperates with none of the external
empirical literature (e.g. De Marchi, 2012; Triguero et al., 2013). As partners listed in the PITEC survey e to 7 if it cooperated with all of
supported by several studies, the impact of networking activities on them to develop its technological innovations, being: i) customers,
innovation performance varies according to the range of external ii) suppliers, iii) competitors, iv) universities, v) private research
partners involved (see e.g. Miotti and Sachwald, 2003; Laursen and centres and consultants, vi) public research centres and vii) tech-
Salter, 2006; Vega-Jurado et al., 2009). Leveraging the results ach- nological centres. In this sense, such a variable can be considered to
ieved in studies focussing on general innovation e without be an alternative proxy for measuring hypothesis 2, being highly
considering its impact on the environment ethe validity of correlated with COOPERATION, but also entailing additional infor-
hypothesis 2 is tested considering the variety of the external mation regarding the nature of the network involved. As emerges
innovation resources activated by the firm on its EI propensity. To from Table 5, reporting the same analysis in Table 4 but substituting
this end, the methods developed by Laursen and Salter (2006) in COOPERATION with this count variable, COOPBREADTH is posi-
their seminal paper investigating the impact of openness in inno- tively and significantly correlated with the probability of becoming
vation performance was adopted, which is extensively used in

Table 4
Table 3 Probit model, with sample selection, explaining the factors correlated to the pro-
Probit model, with sample selection, explaining the factors correlated to the pro- pensity to develop EIs considering different specifications of the cooperation
pensity to develop EIs. variable.

Structural Eq. Selection Eq. Structural Eq. Selection Eq.

Dep. Var.: ENVINN Dep. Var.: TECHINN Dep. Var.: ENVINN Dep. Var.: TECHINN

Coeff. t-values Coeff. t-values Coeff. t-values Coeff. t-values

R&Dt-1 0.424*** 4.13 1.297*** 16.60 R&Dt-1 0.381*** 3.48 1.310*** 16.65
TRAININGt-1 0.315*** 4.80 0.525*** 2.64 TRAININGt-1 0.306*** 4.65 0.535*** 2.65
SKILLt-1 0.0005 0.69 0.0002 0.15 SKILLt-1 0.0006 0.72 0.0003 0.18
INTROt-1 0.053 0.72 0.484*** 2.92 INTROt-1 0.039 0.52 0.495*** 2.98
COOPERATIONt-1 0.154*** 3.09 0.405*** 4.33 COOPBREADTH t-1 0.037*** 3.12 0.089* 1.73
EXT-R&Dt-1 0.120** 2.50 0.246*** 2.81 EXT-R&Dt-1 0.123** 2.54 0.302*** 3.37
EXT-KNt-1 0.059 0.52 0.355 1.08 EXT-KNt-1 0.063 0.55 0.361 1.10
EQUIPMENTt-1 0.173*** 3.15 0.560*** 5.36 EQUIPMENTt-1 0.158*** 2.84 0.560*** 5.36
SIZEt-1 0.156*** 8.73 0.162*** 7.53 SIZEt-1 0.146*** 7.85 0.158*** 7.32
REACTIVEt-1 0.026* 1.83 0.061*** 4.11 REACTIVEt-1 0.028** 2.00 0.062*** 4.16
EXTRA_UE-MKTt-1 … … 0.256*** 4.46 EXTRA_UE-MKTt-1 … … 0.262*** 4.57
Geographic dummies Yes Yes Yes Yes Geographic dummies Yes Yes Yes Yes
Industry dummies Yes Yes Yes Yes Industry dummies Yes Yes Yes Yes
N. Obs. 4829 N. Obs. 4829
Censored Obs. 849 Censored Obs. 849
Uncensored Obs. 3980 Uncensored Obs. 3980
Wald test (p-values) 0.577 Wald test (p-values) 0.967

***Significant at 1%; **Significant at 5%; *Significant at 10%; standard errors are ***Significant at 1%; **Significant at 5%; *Significant at 10%; standard errors are
robust. robust.

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10 7

Table 5 Table 6
Probit model, with sample selection, explaining the factors correlated to the pro- Average marginal effects of probit models, with sample selection, explaining the
pensity to develop EIs considering interaction effects. factors correlated to the propensity to develop EIs.

Structural Eq. Selection Eq. Base model Model with different Model with
proxy for cooperation interaction
Dep. Var.: ENVINN Dep. Var.: TECHINN
dy/dx t-values dy/dx t-values dy/dx t-values
Coeff. t-values Coeff. t-values
R&D 0.151*** 4.47 0.137*** 3.70 0.176*** 4.49
R&Dt-1 0.498*** 4.08 1.385*** 16.75
TRAINING 0.112*** 4.84 0.110*** 4.69 0.113*** 4.91
TRAININGt-1 0.319*** 4.86 0.527*** 2.70
SKILL 0.0002 0.69 0.0002 0.72 0.0002 0.66
SKILLt-1 0.0005 0.66 0.0004 0.23
INTRO 0.019 0.72 0.014 0.52 0.020 0.79
INTROt-1 0.059 0.79 0.467*** 2.80
COOPERATION 0.054*** 3.14 0.120** 2.57
COOPERATIONt-1 0.339** 2.48 0.769*** 5.66
COOPBREADTH 0.013*** 3.12
COOP  R&D 0.217 1.49 0.664*** 3.85
COOPxR&D 0.077 1.52
EXT-R&Dt-1 0.126*** 2.61 0.265*** 3.03
EXT-R&D 0.043** 2.52 0.044** 2.57 0.044*** 2.64
EXT-KNt-1 0.055 0.48 0.351 1.07
EXT-KN 0.021 0.52 0.023 0.55 0.019 0.48
EQUIPMENTt-1 0.180*** 3.24 0.552*** 5.31
EQUIPMENT 0.062*** 3.22 0.057*** 2.90 0.063*** 3.34
SIZEt-1 0.160*** 8.84 0.164*** 7.63
SIZE 0.055*** 9.61 0.053*** 8.61 0.056*** 9.90
REACTIVEt-1 0.025* 1.75 0.060*** 4.00
REACTIVE 0.009* 1.81 0.010** 1.98 0.008* 1.73
EXTRA_UE-MKTt-1 … … 0.252*** 4.39
Geographic dummies Yes Yes Yes Yes ***Significant at 1%; **significant at 5%; *significant at 10%; standard errors are
Industry dummies Yes Yes Yes Yes robust.
N. Obs. 4829
Censored Obs. 849
Uncensored Obs. 3980 being considered: eco-efficiency innovations, for example, which
Wald test (p-values) 0.394
are incremental or incorporated in existing machinery and tech-
***Significant at 1%; **significant at 5%; *significant at 10%; standard errors are nology, may require a lower R&D effort, whereas the contrary is
robust.
true for innovations entailing design for recyclability.
The average marginal effects, reported in Table 6, support the
a green innovator, confirming the higher relevance of the cooper- emerging analysis and allow for the discussion of the magnitude of
ative attitude of green innovators. The more types of partners a firm the relations identified above. Interestingly, the variable with the
cooperates with, the more likely it is to effectively develop a new highest impact refers to internal resources: R&D and TRAINING.
product or process with eco-friendly features. However, the sig- Cooperation on green innovation activities with other firms or in-
nificance of this variable is lower than when using COOPERATION stitutions (COOPERATION) and the acquisition of external knowl-
as a regressor. edge in various forms (EXT-R&D and EQUIPMENT) have a marginal
Results on hybrid forms of innovation resources, which occur effect that is more than half, contrary to what was suggested, for
through the acquisition of knowledge developed by external example, by Borghesi et al. (2015) in the Italian context. When
partners, are less straightforward: support for hypothesis 3 is including the interaction effect COOPxR&D (see the third model in
mixed. The coefficients of EQUIPMENT and EXT-R&D are positive Table 6), the marginal effect of both its composing variables
and significant, whereas that of EXT-KN e measuring whether the increases, but such growth is even larger in the case of COOPERA-
firms acquired patents or non-patented inventions, know-how or TION, which almost doubles it. The impact of cooperation, however,
other types of knowledge developed by external partners e is is much lower when measured through the variable capturing the
negative and non-significant, similar to the results found in France extension of the cooperation network (COOPBREADTH) (second
by Horbach et al. (2013). In other words, the acquisition of ma- model in Table 6), suggesting that the difference across the two
chinery or R&D services is more relevant in spurring the intro- types of innovations largely concerns the propensity to get involved
duction of green innovations than other innovations, whereas the in innovation with external partners, and to a lesser extent, the
mere acquisition of patents or licenses is not differentially impor- variety of resources to which they give access.
tant. The fact that the acquisition of the first two forms of inno- Finally, the sign and significance of the coefficients of the control
vation resources entails e to a much higher degree than the third variables introduced are coherent along all the models presented.
one e an internal effort on the part of a firm to effectively assimilate In particular, the results suggest that the larger a firm, the more
and use embedded knowledge may explain such a difference, but likely it is to become a green innovator e a result consistent with all
further examination is needed to assess the degree to which this the empirical evidence reviewed except Cainelli et al. (2012), whose
applies. empirical focus, however, is mostly on micro and small firms.
An open question emerging from the results is whether external Moreover, firms having reactive innovative strategies, that is, those
and internal innovation resources should be considered comple- that do not perceive the existence of a market for new products or
ments or substitutes (i.e. whether green innovators relying on processes (REACTIVE), are less likely to get involved in the reduc-
external resources are not investing in internal ones or vice versa). tion of their environmental footprint via the introduction of
In order to investigate this issue, a probit model was performed, innovations.
including in the analysis the interaction effect between variables As mentioned in the introduction, environmental innovation is a
capturing internal and external resources (Table 5). COOPxR&D is a broad category, including changes in product portfolios or pro-
dummy variable, which has a value of 1 if firms cooperate with duction processes and having different technological characteris-
external agents (COOPERATION) and perform R&D activities in- tics, which may require different resources to be employed. In order
house (R&D). While the selection equation highlights the pres- to account for such heterogeneity and check for the robustness of
ence of a substitution effect between the two innovation resources, the results discussed so far, Table 7 replicates the same base anal-
this effect is not confirmed when comparing green innovators to ysis reported in Table 3 (first column), focusing first on the sub-
non-green innovators. Indeed, the coefficient of COOPxR&D is group of product innovators (second column) and then on pro-
negative but not significant, similar to the findings of Ghisetti et al. cess innovators (third column). In other words, it was verified
(2013) in 11 European countries. Further investigation is needed to whether the differences between green and non-green innovators
determine whether this evidence is driven by the nature of the EI in the way they recur to the internal, external and hybrid resources

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
8 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10

Table 7 skilled personnel among the R&D employees does not appear to be
Average marginal effects of probit model analysis, with sample selection, explaining differentially relevant. Training personnel to increase a firm's
the factors correlated to the propensity to develop EIs distinguishing between
product and process innovators.
innovation stock of knowledge also positively supports the intro-
duction of innovation with lower impacts on the environment:
All innovations Product innovations Process innovations training is not only important to transmit technological knowledge
dy/dx t-values dy/dx t-values dy/dx t-values but also to increase awareness about the importance of tackling
R&D 0.151*** 4.47 0.132*** 3.59 0.168*** 7.20 environmental issues and motivating personnel at all levels to
TRAINING 0.112*** 4.84 0.095*** 3.70 0.142*** 5.40 address environmental challenges.
SKILL 0.0002 0.69 0.0004 1.26 0.0002 0.61 The econometric analyses performed suggest that external re-
INTRO 0.019 0.72 0.006 0.22 0.011 0.38
sources are more relevant for EIs than for the development of other
COOPERATION 0.054*** 3.14 0.060*** 3.00 0.040** 2.06
EXT-R&D 0.043** 2.52 0.024 1.21 0.039* 1.91 innovations: green innovators fuel their innovation efforts through
EXT-KN 0.021 0.52 0.031 0.70 0.018 0.42 inter-organizational relationships more intensively than other
EQUIPMENT 0.062*** 3.22 0.067*** 3.01 0.027 1.40 innovators, considering both the co-development of innovation
SIZE 0.055*** 9.61 0.057*** 6.52 0.062*** 9.39 through networking (external resources) and the acquisition of
REACTIVE 0.009* 1.81 0.009 1.59 0.003 0.62
externally developed resources (hybrid resources). The higher use
N. Obs. 4829 2905 2783
Censored Obs. 849 149 143 of relations in the case of EIs, which may involve both supply chain
Uncensored Obs. 3980 2756 2640 partners or knowledge providers like universities and KIBS, sug-
***Significant at 1%; ** significant at 5%; * significant at 10%; standard errors are gests the presence of technological interdependences on knowl-
robust. edge, skills and resources which arise in the development of EI, as
well as the need to complement the internal knowledge base with
competences coming from different domains. Such results support
that emerged in the previous analyses are valid when separately preliminary evidence on different empirical contexts, suggesting
considering product and process innovators. Interesting differences that networking may be more important than firms' structural
emerged, especially as far as hybrid resources are considered. In characteristics (see Mazzanti and Zoboli, 2005).
particular, the variable EXT-R&D is significantly more important in The acquisition of machineries or R&D services is significantly
explaining the process innovation of eco-innovators, whereas the correlated with the probability to develop new products or
contrary is true for EQUIPMENT, which is significant only for processes that reduce environmental impacts. Such hybrid innova-
explaining product eco-innovations vs. non-green innovations. tion resources are even more important than for other innovations.
Another interesting difference is the fact that COOPERATION is In fact, the development of EIs often requires the modification or
more significant and have a stronger impact when it comes to substitution of existing equipment to produce new products e
product rather than process innovations. However, the results are entailing, for example, the use of new materials or recyclability e
consistent overall across the three models, both in terms of the sign and to enable the reduction of environmental impacts in the pro-
and significance of the main variables, supporting the validity of the duction process (i.e. to improve eco-efficiency or reduce air or water
emerging evidence. emissions). The other hybrid resources measured (acquisition of
patents or non-patented inventions, know-how or other types of
5. Conclusions and future research knowledge developed by external partners), instead, are not sig-
nificant. Such mixed evidence supports the importance of intro-
The more that sustainable production and consumption systems ducing this third category of resources in the analysis and to study it
are advocated as a possible solution for escaping from the current disentangling its different components.
economic crisis while improving living conditions, the more it is The analysis reported entails an important policy implication e
important to understand under what circumstances firms adopt namely, the need to complement the analysis of the (external)
EIs. Extant contributions have focused mainly on the role of factors triggering environmental innovations with those of the
external and internal drivers in spurring the development of new resources the firm has access to in order to fully understand the
products or processes that reduce environmental impacts. This effectiveness of EI development. In this sense, the emerging evi-
paper attempted to enter the ‘black box’ of the firm and understand dence support that firms that are part of the same industry, and
how EIs are developed, irrespective of the drivers that motivated which are therefore subjected to the same policy and stakeholder
their adoption. In particular, the main contribution of the present pressures, have different attitudes toward the development of eco-
analysis lies in verifying the peculiar role of resources for the innovations. In this sense, at least when studying manufacturing
successful development of environmentally sound products or firms in the context of moderately innovative countries, consid-
processes, using a theory-driven framework that considers internal, ering the innovation resources the firms have access to is strategic
external and hybrid innovation resources. Further, while most for understanding whether they will succeed in becoming green.
existing contributions have focused solely on green innovators and Regardless of the policy instrument used to spur innovation or the
on cross-sectional data, this paper e based on a panel dataset of pressure from consumers, NGOs and the like, if the firms have no
Spanish manufacturing firms e enables the understanding of what internal R&D base or do not work with a network of external
is peculiar to EIs compared to other innovations and the testing for partners (other firms or institutions), such pressure will not effec-
causality relationships among the variables considered. tively translate to a new product or process that reduces the im-
Overall, the results speak for a greater importance of all the re- pacts on the environment.
sources considered to introduce products or processes that reduce An important limitation of this analysis, determined by data
environmental impacts, which confirms theories affirming that they constraints, is that it does not account for different types of EIs.
represent a technological frontier e becoming more and more Previous studies suggested that a firm's innovation strategy and the
important because of increasing policy pressure e which firms can drivers pushing its sustainable initiative might vary across EI types.
only address by exploiting new and complementary resources. It would be important to understand if the impact of the internal,
In terms of internal innovation resources, the presence of a R&D external and hybrid resources considered in the present analysis
structure is even more important than for other innovations in vary across EI types as well e for example, if distinguishing
spurring the development of EIs, whereas the presence of highly between end-of-pipe and cleaner technologies e and which of

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10 9

them show a common pattern in terms of relevant innovation re- Danneels, E., Kleinschmidt, E.J., 2001. Product innovativeness from the firm's
perspective: its dimensions and their relation with project selection and per-
sources. Furthermore, future research should better test the
formance. J. Prod. Innov. Manag. 18 (6), 357e373.
external innovation resource framework considering the different Darnall, N., Henriques, I., Sadorsky, P., 2010. Adopting proactive environmental
positions in the value chain (e.g. distinguishing between B2B and strategy: the influence of stakeholders and firm size. J. Manag. Stud. 47 (6),
B2C cases, where the role of innovation for the different partners, 1072e1094.
Das, T.K., Teng, B., 2000. A resource-based theory of strategic alliances. J. Manag. 26
for example clients, may be very different) and to identify more (1), 31e61.
detailed characteristics about the network (including, for example,  Junquera, B., 2003. A review of the literature on environmental
Del Brío, J.A.,
considerations of how many partners are included in each category innovation management in SMEs: implications for public policies. Technovation
23 (12), 939e948.
and where they are located). Case study analyses suggest that the Del Río, P., 2009. The empirical analysis of the determinants for environmental
more relevant partner may change from case to case e depending technological change: a research agenda. Ecol. Econ. 68 (3), 861e878.
Del Río, P., Taranco  n Mor  Albin
an, M.A., ~ ana, F.C., 2011. Analysing the determinants
on the industry but also on the specific strategy pursued by the
of environmental technology investments: a panel-data study of Spanish in-
company e and so the role of specific partners, like suppliers (see dustrial sectors. J. Clean. Prod. 19 (11), 1170e1179.
e.g. De Marchi et al., 2013), which informed the approach used in Del Río, P., Pen ~ asco, C., Romero-Jordan, D., 2013. Distinctive features of environ-
this paper of considering all of them together rather than verifying mental innovators: an econometric analysis. Bus. Strategy Environ. http://
dx.doi.org/10.1002/bse.1822.
the role of a specific category of partners at a time, considering for De Marchi, V., 2012. Environmental innovation and R&D cooperation: empirical
the data available. In-depth analyses equipped with data on such evidence from Spanish manufacturing firms. Res. Policy 41 (3), 614e623.
characteristics could therefore provide a useful complement to the De Marchi, V., Di Maria, E., Ponte, S., 2013. The greening of global value chains:
insights from the furniture industry. Compet. Change 17 (4), 299e318.
emerging evidence suggesting contingencies supporting the
De Marchi, V., Grandinetti, R., 2013. Knowledge strategies for environmental in-
specific importance of each cooperation partner so as to better novations: the case of Italian manufacturing firms. J. Knowl. Manag. 17 (4),
investigate the role of each of the hybrid resources analysed. 569e582.
The comparison of the emerging results with the existing De Medeiros, J.F., Ribeiro, J.L.D., Cortimiglia, M.N., 2014. Success factors for envi-
ronmentally sustainable product innovation: a systematic literature review.
empirical literature suggests that similarities and differences may J. Clean. Prod. 65, 76e86.
be driven not only by the use of different empirical methods but Demirel, P., Kesidou, E., 2011. Stimulating different types of eco-innovation in the
also by the different empirical context (e.g. Central-European UK: government policies and firm motivations. Ecol. Econ. 70 (8), 1546e1557.
Ghisetti, C., Marzucchi, A., Montresor, S., 2013. Does External Knowledge Affect
countries such as Germany versus Mediterranean ones such as Environmental Innovations? an Empirical Investigation of Eleven European
Spain and Italy), highlighting the importance of supporting cross- Countries. Ingenio Working Paper n. 2013-01.
country analyses and understanding which aspects of a country Gupta, A.K., Singhal, A., 1993. Managing human resources for innovation and
creativity. Res. Technol. Manag. 36 (3), 41e48.
drive the different attitudes toward EI at firms. Horbach et al. (2013) Halila, F., Rundquist, J., 2011. The development and market success of eco-
and Triguero et al. (2013) provide an interesting initial research innovations: a comparative study of eco-innovations and “other” innovations
effort in this regard. in Sweden. Eur. J. Innov. Manag. 14 (3), 278e302.
Handfield, R.B., Walton, S.V., Seegers, L.K., Melnyk, S.A., 1997. “Green” value chain
practices in the furniture industry. J. Operat. Manag. 15 (4), 293e315.
Horbach, J., 2008. Determinants of environmental innovation: new evidence from
References German panel data sources. Res. Policy 37 (1), 163e173.
Horbach, J., Oltra, V., Belin, J., 2013. Determinants and specificities of eco-
Albino, V., Balice, A., Dangelico, R.M., 2009. Environmental strategies and green innovations compared to other innovations: an econometric analysis for the
product development: an overview on sustainability-driven companies. Bus. French and German industry based on the ommunity Innovation Survey. Ind.
Strategy Environ. 18 (2), 83e96. Innov. 20 (6), 523e543.
Ambec, S., Cohen, M.A., Elgie, S., Lanoie, P., 2013. The Porter hypothesis at 20: can Horbach, J., Rammer, C., Rennings, K., 2012. Determinants of eco-innovations by
environmental regulation enhance innovation and competitiveness? Rev. En- type of environmental impact: the role of regulatory push/pull, technology
viron. Econ. Policy 7 (1), 2e22. push and market pull. Ecol. Econ. 78, 112e122.
Andersen, M., 1999. Trajectory Change through Interorganisational Learning: on the Ireland, R.D., Hitt, M.A., Vaidyanath, D., 2002. Alliance management as a source of
Economic Organisation of the Greening of Industry. Ph.D. series. Copenhagen competitive advantage. J. Manag. 28 (3), 413e446.
Business School, Copenhagen. Jaffe, A., Palmer, K., 1997. Environmental regulation and innovation: a panel data
Barney, J., 1991. Firm resources and sustained competitive advantage. J. Manag. 17 study. Rev. Econ. Stat. 79 (4), 610e619.
(1), 99e120. Kafouros, M.I., Buckley, P.J., Sharp, J.A., Chengqi, W., 2008. The role of internation-
Becker, W., Dietz, J., 2004. R&D cooperation and innovation activities of firms: alization in explaining innovation performance. Technovation 28 (1e2), 63e74.
evidence for the German manufacturing industry. Res. Policy 33 (2), 209e223. Kammerer, D., 2009. The effects of customer benefit and regulation on environ-
Berrone, P., Fosfuri, A., Gelabert, L., Gomez-Mejia, L.R., 2013. Necessity as the mother mental product innovation: empirical evidence from appliance manufacturers
of “green” inventions: institutional pressures and environmental innovations. in Germany. Ecol. Econ. 68 (8e9), 2285e2295.
Strat. Manag. J. 34 (8), 891e909. Kemp, R., Pearson, P., 2008. MEI Project about Measuring Eco-innovation: Final
Borghesi, S., Cainelli, G., Mazzanti, M., 2015. Linking emission trading to environ- Report. UNU-MERIT, Maastricht.
mental innovation: evidence from the Italian manufacturing industry. Res. Kerr, S., Newell, R., 2001. Policy-induced technology adoption: evidence from the US
Policy 44 (3), 669e683. lead phasedown. J. Ind. Econ. 51 (3), 317e343.
Brammer, S., Hoejmose, S., Marchant, K., 2012. Environmental management in SMEs Klewitz, J., Hansen, E.G., 2014. Sustainability-oriented innovation in SMEs: a sys-
in the UK: practices, pressures and perceived benefits. Business. Strategy En- tematic review. J. Clean. Prod. 65, 57e75.
viron. 21 (7), 423e434. Klewitz, J., Zeyen, A., Hansen, E.G., 2012. Intermediaries driving eco-innovation in
Brammer, S., Millington, A., 2006. Firm size, organizational visibility and corporate SMEs: a qualitative investigation. Eur. J. Innov. Manag. 15 (4), 442e467.
philanthropy: an empirical analysis. Bus. Ethics 15 (1), 6e18. Laursen, K., Foss, N.J., 2003. New human resource management practices, com-
Brunnermeier, S., Cohen, M., 2003. Determinants of environmental innovation in US plementarities and the impact on innovation performance. Camb. J. Econ. 27
manufacturing industries. J. Environ. Econ. Manag. 45 (2), 278e293. (2), 243e263.
Cainelli, G., Mazzanti, M., Montresor, S., 2012. Environmental innovations, local Laursen, K., Salter, A., 2006. Open for innovation: the role of openness in explaining
networks and internationalization. Ind. Innov. 9 (8), 697e734. innovation performance among UK manufacturing firms. Strat. Manag. J. 27 (2),
Carrillo-Hermosilla, J., del Río, P., Ko € nno
€ l€
a, T., 2010. Diversity of eco-innovations: 131e150.
reflections from selected case studies. J. Clean. Prod. 18 (10e11), 1073e1081. Lavie, D., 2006. The competitive advantage of interconnected firms: an extension of
Chiarvesio, M., De Marchi, V., Di Maria, E., 2014. Environmental innovations and the resource-based view. Acad. Manag. Rev. 31 (3), 638e658.
internationalization: theory and practice. Bus. Strategy & Environ. http:// Markusson, N., 2010. Unpacking the black box of cleaner technology. J. Clean. Prod.
dx.doi.org/10.1002/bse.1846. 19 (4), 294e302.
Cohen, W.M., Levinthal, D.A., 1990. Absorptive capacity: a new perspective on Mazzanti, M., Zoboli, R., 2005. The drivers of environmental innovation in local
learning and innovation. Adm. Sci. Q. 35 (1), 128e152. manufacturing systems. Econ. Polit. 22 (3), 399e438.
Cuerva, M.C., Triguero-Cano, A., Co rcoles, D., 2014. Drivers of green and non-green Mazzanti, M., Zoboli, R., 2009. Embedding environmental innovation in local pro-
innovation: empirical evidence in low-tech SMEs. J. Clean. Prod. 68 (1), duction systems: SME strategies, networking and industrial relations. Evidence
104e113. on innovation drivers in industrial districts. Int. Rev. Appl. Econ. 23 (2),
Daft, R.L., Lengel, R.H., 1984. Information richness: a new approach to managerial 169e195.
behavior and organization design. In: Staw, B.M., Cummings, L.L. (Eds.), Miotti, L., Sachwald, F., 2003. Co-operative R&D: why and with whom? an inte-
Research in Organizational Behavior. JAI Press, Greenwich, pp. 191e233. grated framework of analysis. Res. Policy 32 (8), 1481e1499.

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008
10 G. Cainelli et al. / Journal of Cleaner Production xxx (2015) 1e10

Nieto, M.J., Santamaria, L., 2007. The importance of diverse collaborative networks Sierzchula, W., Bakker, S., Maat, K., van Wee, B., 2012. Technological diversity of
for the novelty of product innovation. Technovation 27 (6e7), 366e377. emerging eco-innovations: a case study of the automobile industry. J. Clean.
Penrose, E.T., 1959. The Theory of the Growth of the Firm, first ed. Basil Blackwell, Prod. 37, 211e220.
Oxford. Tatikonda, M.V., Rosenthal, S.R., 2000. Technology novelty, project complexity, and
Pittaway, L., Robertson, M., Munir, K., Denyer, D., Neely, A., 2004. Networking and product development project execution success: a deeper look at task uncer-
innovation: a systematic review of the evidence. Int. J. Manag. Rev. 5e6 (3e4), tainty in product innovation. IEEE Trans. Eng. Manag. 47 (1), 74e87.
137e168. Teece, D.J., 1986. Profiting from technological innovation: implications for integra-
Porter, M.E., van der Linde, C., 1995a. Green and competitive: ending the stalemate. tion, collaboration, licensing, and public policy. Res. Policy 15 (6), 285e305.
Harv. Bus. Rev. 73 (5), 120e134. Thun, J.-H., Müller, A., 2010. An empirical analysis of green supply chain management
Porter, M.E., van der Linde, C., 1995b. Toward a new conception of the environment- in the German automotive industry. Bus. Strategy Environ. 19 (2), 119e132.
competitiveness relationship. J. Econ. Perspect. 9 (4), 97e118. Triguero, A., Moreno-Monde jar, L., Davia, M.A., 2013. Drivers of different types of
Rennings, K., 2000. Redefining innovation: eco-innovation research and the eco-innovation in European SMEs. Ecol. Econ. 92, 25e33.
contribution from ecological economics. Ecol. Econ. 32 (2), 319e332. Van Dijken, K., Prince, Y., Wolters, T., Frey, M., Mussati, G., Kalff, P., Hansen, O.,
Rennings, K., Ziegler, A., Ankele, K., Hoffmann, E., 2006. The influence of different Kerndrup, S., Søndergård, B., Rodrigues, E.L., Meredith, S., 1999. Adoption of
characteristics of the EU environmental management and auditing scheme on Environmental Innovations: the Dynamics of Innovation as Interplay between
technical environmental innovations and economic performance. Ecol. Econ. 57 Business Competence, Environmental Orientation and Network Involvement.
(1), 45e59. Kluwer, Dordrecht.
Sarkis, J., Gonzalez-Torre, P., Adenso-Diaz, B., 2010. Stakeholder pressure and the Vega-Jurado, J., Gutierrez-Gracia, A., Fernandez-de-Lucio, I., 2009. Does external
adoption of environmental practices: the mediating effect of training. knowledge sourcing matter for innovation? Evidence from the Spanish
J. Operations Manag. 28 (2), 163e176. manufacturing industry. Ind. Corp. Change 18 (4), 637e670.
Seuring, S., Müller, M., 2008. From a literature review to a conceptual framework for Wagner, M., 2007. On the relationship between environmental management,
sustainable supply chain management. J. Clean. Prod. 16 (15), 1699e1710. environmental innovation and patenting: evidence from German
Srivastava, M.K., Gnyawali, D.R., 2011. When do relational resources matter? manufacturing firms. Res. Policy 36 (10), 1587e1602.
Leveraging portfolio technological resources for breakthrough innovation. Acad. Wernerfelt, B., 1984. A resource-based view of the firm. Strat. Manag. J. 5 (2),
Manag. J. 54 (4), 797e810. 171e180.

Please cite this article in press as: Cainelli, G., et al., Does the development of environmental innovation require different resources? Evidence
from Spanish manufacturing firms, Journal of Cleaner Production (2015), http://dx.doi.org/10.1016/j.jclepro.2015.02.008