Journal of Cleaner Production 229 (2019) 974e988

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Does firm size matter? Evidence on the impact of the green innovation
strategy on corporate financial performance in the automotive sector
Woon-Leong Lin a, Jun-Hwa Cheah a, *, Mohamed Azali a, Jo Ann Ho a, Nick Yip b
a
Faculty of Economics and Management, Universiti Putra Malaysia, Serdang, Selangor, Malaysia
b
Norwich Business School, University of East Anglia, Norwich, UK

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

Article history: In the past few years, there has been increasing awareness regarding the significance of the Green
Received 27 January 2019 Innovation Strategy (GIS) in the academic and practical fields. Hence, it becomes important to determine
Received in revised form the correlation between the GIS and the Corporate Financial Performance (CFP). This study attempted to
2 April 2019
determine the dynamic correlation between the GIS and the CFP, with regards to the firm size. For this
Accepted 18 April 2019
Available online 7 May 2019
purpose, this study has collected data for 163 international automotive firms, from the CSRHub database,
for the period ranging between 2011 and 2017. Furthermore, we also used the dynamic panel data
system, i.e., the Generalised Method of Moment (GMM) method, for estimating this relationship. The
Keywords:
Green innovation strategy
empirical results indicated that the GIS positively affected the CFP. Interestingly, we also uncovered that
Corporate financial performance the firm size moderated the negative correlation between the GIS and the CFP. The small-sized firms
Automotive industry showed higher green innovation investments return than the larger-sized firms, which indicated that
GMM these smaller firms were more prone to seek variation and visibility, for accessing better resources.
Firm size Furthermore, due to the extensive scrutiny of the stakeholders, these small firms could generate higher
profits. The implications for managers and the theories in this regard are then discussed.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction quality, reducing emissions, and fuel consumption as follows:

Currently, the automotive industry faces many obstacles. This By 2021 the cars, which emit >95 g of CO2/km, would be dis-
sector generally relies on the technological paradigm of volume allowed in the market.
production, which has gradually become more unprofitable due to By 2020, the greenhouse gas intensity of all automobile fuels
the increase in the segmented niche markets. Furthermore, this must be reduced by 10%, for improving the fuel quality.
sector has to undergo many social and regulatory pressures, which By 2021, the automobile manufacturers should produce light-
can improve the sustainability of all its products and production duty vehicles that consume <3.6 l/100 km of diesel or <4.1l/
methods. Research conducted in this field showed that the auto- 100 km of petrol.
motive sector is facing these challenges and have to establish
techniques for developing a profitable and sustainable sector for In the past, the transportation sector was seen to be responsible
future generations (Smith and Crotty, 2008) After the publication of for 27% of the total global energy consumption and 33.7% of all
KPMG's report (2012) on environmental regulations in the auto- greenhouse gas emissions (Tie and Tan, 2013). These trends would
motive sector, various governments started imposing strict envi- change in the future, due to the scarcity of fossil fuels and
ronmental regulations on the OEMs (Original Equipment increasing environmental pressure (Nilsson et al., 2012). Because of
Manufacturers) for controlling the CO2 emissions. For example, the the increasing concern with regards to the environmental issues, by
European Commission implemented legislation for testing fuel the public, consumers, suppliers and the administration, a majority
of the firms have begun the development of environmentally-
friendly green products (Green et al., 2012; Tseng et al., 2013).
* Corresponding author. Hence, the techniques which save energy, or reduce CO2 emissions
E-mail addresses: linwoonleong@gmail.com (W.-L. Lin), jackycheahjh@gmail. and air pollution, in the automotive sector, are important chal-
com (J.-H. Cheah), azali@upm.edu.my (M. Azali), ann_hj@upm.edu.my (J.A. Ho), n. lenges and issues affecting the governments (Hui, 2010). In their
yip@uea.ac.uk (N. Yip).

https://doi.org/10.1016/j.jclepro.2019.04.214
0959-6526/© 2019 Elsevier Ltd. All rights reserved.
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 975

study, Shrivastava (1995) stated that firms must differentiate their that the agency costs (like the ineffective use of the corporate
products, lower production costs, improve product quality and funds) encountered while implementing the GIS in the larger firms
develop more innovation processes. Therefore, continuous inno- were the dominant factor that affected the strategic benefits which
vation was seen to be an important strategy which could help in these firms could derive after the green innovation investments.
overcoming the pressures implemented by the competitors, cus- Implementation of GIS could significantly upsurge the profitability
tomers and the regulators (Porter and Van der Linde, 1995a,b). of the smaller firms. Hence, the results indicated that the total
Due to the strict international rules, increasing consumer envi- benefits derived by the implementation of the GIS were not a one-
ronmentalism, and the conventions regarding green innovation, size-fits-all and were dependent on the firm characteristics.
the competition and the business-related tactics have undergone a Based on the study results, we have put efforts to make three
significant alternation in all global industries. These factors have key contributions to the literature with regards to this topic. First, as
also affected the business in the automotive industries. Hence, the per our knowledge, we are the first to examine the unique setting of
Green Innovation Strategies (GIS) have played a vital role (Russo GIS's role pertaining to firm size interactions. Also, prior studies
and Fouts, 1997). GIS is described as the development of green have not considered longitudinal aspect pertaining to GIS (e.g.,
process and green products-related innovation strategies and de- Aguilera-Caracuel and Ortiz-de-Mandojana, 2013; Gluch et al.,
cisions, that associated with the application of green activities and 2009; Chen et al., 2006; Ge et al., 2018). This research offers
environmental management systems (Eiadat et al., 2008; Tomomi, insight regarding the growing need to understand the impact of
2010; Dong et al., 2014). On the other hand, very few researchers firm size in justifying that the enhancement in green innovation
have investigated the effect of the GIS on the Corporate Financial improves firm-level financial performance.
Performance (CFP). CFP is defined as measuring the results per- Second, we put forward a significantly positive relationship that
taining to a firm's operations and policies in monetary aspects. The exists between CFP and GIS, which was crucial as it could help to
company's return on investment, value added and return on assets solve the existing perspective defining the relationship. Our finding
as depicted with these results. In this research study, CFP has been offers a holistic means to examine the firm's conditions that allow
employed as an instrument to measure economic performance as organisations to create green innovation initiatives as well as
well as integrating accounting-based measures, which includes simultaneously enhance their financial performance. Moreover, we
asset utilisation, firm's profitability, return on equity, the return of have stressed on the importance of the combination or configura-
investment and return on asset (Wu et al., 2006). An organisation's tion pertaining to the firm size, which could cast an impact on the
internal efficiency is represented by the accounting-based mea- automotive sector employing GIS. In this research study, we have
sures, which is impacted by the social performance of the organi- provided in-depth insights by considering all the factors that could
sation (Van Beurden and Go € ssling, 2008). So, it is important to have a role in simulating the GIS of an organisation. Also, we have
develop policies that can effectively implement GIS in the in- made an effort to aid governments and policymakers in designing
dustries (Petts, 1998). In this study, we have attempted to bridge impactful mechanisms and guidelines (instead of just creating
this gap and described many novel green innovation-related regulations), thereby allowing the development of environmentally
concepts. responsible attitudes.
In the past few years, GIS was seen to one of the major factors Finally, many of the corporate finance and empirical manage-
that affected environmental sustainability, financial growth and life ment researchers recognise at least two potential sources of
quality (Porter, 1981; Bansal and Gao, 2006; Dangelico and Pujari, endogeneity: simultaneity and unobservable heterogeneity
2010). Implementation of the GIS is a vital tool which increases (Wintoki et al., 2012). However, one source of endogeneity that has
the sustainable growth of the manufacturing industries due to an usually been ignored (explicitly or implicitly) comes from the
increasing environmental pressure, especially in the automotive possibility that the current values pertaining to firm performance
sector. Use of GIS embodies the idea of environmental protection variables are regarded as a function of previous performance of the
for designing and packaging products and improving the organisation. Overlooking this source of endogeneity could cast
differentiation-related advantage (Hart, 1995; Chen et al., 2006). serious impacts for inference. This study has applied a system GMM
The implementation of GIS could significantly increase the resource estimator (Arellano and Bover, 1995; Blundell and Bond, 1998) on
productivity of the companies (Porter and Van der Linde, 1995a,b). the dynamic panel data to resolve the issue pertaining to endoge-
Therefore, the development of GIS would help in resolving the neity between the CFP and GIS, which offers evidence confirming
conflicts between environmental protection and economic their relationship. Moreover, employing this technique has allowed
development. gaining an understanding of the unobservable heterogeneity as
In this study, we have also attempted to examine the connection well as a better depth pertaining to this study.
between the GIS and their effect on firm profitability in the auto- This study was structured as follows: Section 2 provided a brief
motive industry. Furthermore, this study also intended to deter- overview of the related literature with regards to the theoretical
mine the effect of firm size on the GIS and CFP in the automotive foundations. Section 3 describes the method, data sample collec-
industry. This study collected the data from the CSRHub for a period tion techniques and the measurements for all variables. Section 4
ranging between 2011 and 2017 and investigated whether the presents the descriptive statistics, all correlation coefficients be-
green innovation investment would increase the shareholders’ tween the variables, and also the outcomes of the interactive effect
wealth. We also studied whether the significance of these activities of the firm size. Section 5 discusses all results and offers directions
varied based on the firm characteristics (like firm size) employing for future studies. Finally, Section 6 presented the conclusions and
the dynamic panel data system GMM estimator. The empirical implications of this study.
outcomes highlighted the significantly positive relationship be-
tween GIS and CFP. This positive relationship was persistent when 2. Literature review and hypotheses development
this study attempted to control the endogeneity of GIS. We further
noted that the effect of the negative relationship between the GIS 2.1. Ecological modernisation theory (EMT)
and firm size on the CFP could be due to the fact that the GIS
spending by the company provides an overall limited tangible The ecological modernisation (EMT) theory deals with analysing
benefit, and helps the company obtain better profits. A small firm how contemporary industrialised societies handle environmental
size showed higher efficiency than the larger firms. This indicated crises (Mol and Sonnenfeld, 2000). The EMT theory that defines
976 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

environmental innovation is put forward as a possible solution to 2013). Hence, green innovative companies include those companies
resolve the conflict between environmental protection and indus- which are needed for implementing a process of improvement and
trial development (Murphy and Gouldson, 2000). As there is a need constant growth which can lead to better and concrete green
to achieve environmental performance and profitability simulta- strategies engagements (i.e., green technologies and products)
neously, as envisaged under EMT, green management has emerged (Marcus and Fremeth, 2009).
as a key management practice for organisations seeking to gain
competitiveness via environmental innovation (Hall, 2001). EMT 2.3. Green innovation strategy and corporate financial performance
postulates that continued industrial development, instead of
inevitably continuing to degrade the environment, provides the GIS is described as the development of green process and green
best choice to avoid the global ecological challenge (York and Rosa, products-related innovation strategies and decisions that associ-
2003). The perspective central to EMT can be attributed to the era of ated with the engagement of green activities and environmental
modernity that provides a promise that technology development, management systems (Eiadat et al., 2008; Tomomi, 2010; Dong
industrialisation, capitalism and economic growth are not just et al., 2014). The ecological modernisation theory has encouraged
potentially compatible with that of ecological sustainability but companies to implement novel technological and scientific pro-
also could act as the major drivers to bring environmental reforms cesses, which would help them strengthen the green processes and
(Mol, 1997). EMT also implies the chance of it being inherent to the green products (Mu et al., 2009; Zhu et al., 2012). Here, this study
process of late modernisation which could be self-referential has applied the definition presented by Huber (1985), who stated
mechanisms (e.g. the requirement to internalise environmental that ecological modernisation was a major economic theme
effects to guarantee future production inputs) that could possibly involved in the eco-social switchover, as it could lead to the
result in ecological sustainability (York and Rosa, 2003). modernisation of the production and consumption cycles using
EMT provides a theoretical lens to examine the relationship intelligent and novel technologies.
existing between innovation and environmental performance GIS helps in decreasing the negative effect on the environment
€nicke, 2008). It also motivates organisations to employ sustain-
(Ja and also enhance the competitive advantage of the various in-
able technology that allows decreasing the environmental impact dustries. The companies that advocate the implementation of
on their business. EMT focuses on the chance of achieving environmental innovation strategies would lead to the develop-
ecological-economic ‘win-win’ solutions, above all, by ensuring ment of a novel business model and alter the rules which help in
cost minimisation as well as competition for innovation. As per generating a better business opportunity (Chiou et al., 2011; Chen
EMT, the aim of the firm is to modify the direction of technological et al., 2012; Chang and Chen, 2013; Dong et al., 2014). Earlier
progress and to establish the compulsion pertaining to innovation studies showed that the implementation of GIS offered positive
benefitting the environment (Ja €nicke, 2008). Even though EMT of- firm benefits and economic developments. In one study, Huang and
fers a wide concept, in this study, we have emphasised on the Jim Wu (2010) observed that environmental innovation in high-
impact cast by environmental performance on financial perfor- tech firms could significantly improve the organisation's financial
mance. The theoretical insight pertaining to EMT states that tech- performance. Furthermore, Tomomi (2010) investigated many
nological innovation would aid firms in enhancing both their small or medium-sized Japanese companies and noted that the
economic and environmental performance. To this extent, EMT environmental strategies offer better opportunities to these com-
implies that firms can address environmental issues as barriers panies to improve their business activities and provide them with a
when technological change is complemented with organisational competitive advantage. Chiou et al. (2011) stated that if all suppliers
change (Park et al., 2010). We have contested that EMT needs to be implemented a green supply chain, they could easily fulfil the
regarded as a pertinent management theory that allows under- environmental design requirements and display a green innovative
standing and guiding management innovation and change that is performance. Fraj et al. (2013) mentioned that the use of the GIS
ecologically oriented, at the level of firm analysis. The core theo- positively affected the environmental and economic performance
retical underpinning that surrounds the EMT states that organisa- in a business-to-business context. Dong et al. (2014) noted that the
tional technological innovation, like GIS, will aid the firms in application of eco-innovative activities by the companies helps in
enhancing both economic and environmental dimensions. the determination of their competitiveness and environmental
performance. In their study, Yang et al. (2018) validated a new
2.2. Green innovation strategy model which indicated that GIS offers several firm benefits.
Many researchers argued that the implementation of the GIS
Driessen and Hillebrand (2002) defined the concept of ‘green can slightly increase the firm innovation portfolios (Hull and
innovation strategy’ and stated that this concept may not be Rothenberg, 2008). However, a lack of these GIS could be due to
developed with an aim to reduce the environmental burden. hindrances affecting this innovation, like the presence of a
However, it yields several important environmental benefits. In knowledge gap, inadequate governmental support and an aversion
their study, Chen et al. (2006) defined GIS as the software or to the risks in the capital markets (Runhaar et al., 2008). For
hardware innovative activities related to the green processes or instance, many green firms or ventures were seen to be vulnerable
products, like innovative technologies involved in preventing as they were developed based on the expectation that the constant
pollution, energy-saving, waste-recycling, designing green prod- governmental subsidies would soon diminish. These issues have
ucts, or even corporate environmental management. Organisation made the managers difficult to achieve competitive and environ-
for Economic Co-operation and Development (OECD, 2009) defined mental improvements in their firms (Hull and Rothenberg, 2008).
GIS as the implementation or creation of novel, better-quality ser- Though the firms can overcome all the barriers and develop GIS,
vices/goods, processes, marketing techniques or institutional ar- these innovations are unable to get translated to a higher financial
rangements, that intentionally or unintentionally, can offer better performance level (Link and Naveh, 2006). For instance, Ringer is a
environment compared to their other alternatives. This innovation manufacturer of nontoxic and natural pesticide, which decreases
includes many technological innovations required for preventing the ecological harm, however, it is more expensive and less effec-
pollution, energy-saving, waste-recycling, designing green prod- tive compared to the conventional pesticides. Hence, customers do
ucts, or even environmental management (Lai et al., 2003). It is not easily accept these novel products. Furthermore, the firms that
seen to extend beyond regulatory compliance (Arago  n-Correa et al., apply GIS can increase their product-quality, training and safety-
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 977

related costs (Gelb and Strawser, 2001). However, there will be display advanced management processes (Donaldson, 2001) and
more cost incurred during the preventive of risk and research and perceive or handle the exterior situation differently, based on their
development (Lo  pez-Nicola
s and Meron~ o-Cerdan, 2011). experience (Miles, 1986). Thus, the internal system necessary for
Conversely, many researchers believe that the GIS helps the handling the issues is more advanced, which shows better recep-
firms improve their overall life quality, are profitable and efficient tiveness to the environmental issues (Brammer and Millington,
(Hart, 1995; King and Lenox, 2002). They also increase the 2006).
requirement for the products amongst the environmentally- This is further summarised to indicate that the size of the firm
sensitive customers (Marcus and Fremeth, 2009). The imple- highlights the more complex phenomena which affect the green
mentation of the GIS helps in determining the performance of all innovation participation. Thus, broader conclusions can be derived
green processes and products so that they compete in the market, from the firm size, provided the interrelation between all related
which can be achieved by reducing the company's environmental attributes is also considered. This leads to the development of many
effects (Chen, 2008; Chiou et al., 2011; Yang et al., 2018). The GIS theories that are investigated using integrative contributions. In
helps in enhancing preventive pollution, which enables a company this study, we have made a small effort to scrutinize the influence of
to save the operational costs and enable material reuse by recycling the firm size on the correlation between the GIS and the CFP.
(Hart, 1995). Furthermore, a company that shows better environ- The different firm-level attributes can affect the engagement of
mental initiatives can gain better optimistic ecological image the GIS. Hence, it is significant to comprehend these effects, as
(Christmann, 2000), advantage from the premium pricing and these firms can develop strategic value from the GIS. Apart from
higher revenues, because of a higher societal endorsement (Bansal, these effects, the firm size was considered to be important and
2005). This societal endorsement helps the companies distinguish unexamined (Madden et al., 2006; Ho € risch et al., 2015). The firm
their services/products from their rivals (Rivera, 2002). Hence, the size affects the strategic motivation, which can positively affect the
ethical (environmental) and responsible initiatives were seen to be GIS (McElroy and Siegfried, 1985; Adams and Hardwick, 1998). The
a source of better and valuable opportunities (Porter, 2006; Porter larger firms showed a significant social effect, based on the scale of
et al., 2007). Also, GIS helps the firms to increase their efficient use all their activities (Cowen et al., 1987); hence, they are required to
of raw materials for decreasing the environmental costs and be more socially responsible than the small firms. On the other
increasing their waste recycling (Chiou et al., 2011; Zhu et al., 2012). hand, studies showed that small firms are involved in GIS activities,
Innovative activities cause the firms to develop new processes for especially by giving donations (Madden et al., 2006). Therefore, the
converting waste products into greener products which provide an query that rises is what inspires the small firms to apply GIS ac-
alternative income source. Along with the green products, the GIS tivities, and also if this was economically justified?
helps the firms to integrate the green concepts for reorganising and The firms which have to undergo financial or slack resource
improving their business tactics. constraint are likely to use the existing capitals for improving their
Additionally, the GIS is able to fundamentally alter the compe- competitive advantage using traditional ways of competition. The
tition in the industry. When the competition is fiercer, the firm is organisations with a higher cash flow show a better response to the
able to capitalise on the advanced technology for environmental stakeholder pressure, using discretionary activities like the GIS
innovation and address the environmental issues in the market. activities (McGuire et al., 1988), while the organisations with a low-
The GIS helps the firms to develop and reconfigure better and profit-margin cannot participate in this discretionary behaviour,
innovative processes for improving the competition and differen- based on the creditor and shareholder requirements (Brammer and
tiating them from their competitors (Eiadat et al., 2008; Tomomi, Millington, 2006). This inhibits the implementation of these com-
2010; Dong et al., 2014). Hence, the subsequent Hypothesis was panies in GIS actions. The resource-rich firms face a comparatively
developed: lesser constraint and are more likely to discharge their social
responsibilities.
Hypothesis 1. (H1): Green innovation strategy positively affects
The firm operations can affect their green innovation involve-
the corporate financial performance.
ment, at the functional and administrative level. The companies
with an established decision-making process or organisation
2.4. The moderating role of the firm size structure are more probable to participate in such activities since
they consist of developed systems that can handle the external is-
As shown in the above section, GIS positively affects the CFP. sues (Miles, 1986; Bhambri and Sonnenfeld, 1988; Donaldson,
However, this effect is an intricate and multi-faceted issue. Several 2001). Due to their organisational maturity, these firms display
factors can affect the firm performance such as the contextual and clear structures, especially related to their ability and expertise, and
environmental variables like firm size ( Yeung, 2008; Ramaswami were able to implement effective GIS activities. Furthermore, the
et al., 2009). The firms with varying sizes use different data man- firms make use of their firm's competencies for framing the GIS
agement strategies and can achieve a differing level of govern- activities (Hess et al., 2002). These companies are facing greater
mental benefit, thereby showing a different corporate performance. pressure to warrant that their environmental commitments do not
It is generally believed that the larger firms are more visible, and increase the organisational costs (Van de Ven and Jeurissen, 2005).
are more socially responsive. In contrast, the smaller firms have to Thus, they are seen to be specialists in implementing the GIS ac-
attain lesser pressures or acquire lesser environmental-related tivities, based on their firm's competencies. The firms with a higher
recognition, based on their lower visibility. It is also stated that operational scale are able to efficiently re-allocate and re-organise
the larger firms are less socially responsive and are more resistant their resources. These companies are very likely to initiate GIS ac-
to other effects (Meznar and Nigh, 1995), which is very contradic- tivities and show a distinguishable environmental transformation.
tory. The effect of the firm size on the GIS is based on the access to The scale-economies can increase the corporate environmental
all resources (Brammer and Millington, 2006). The large organisa- performance (Brammer and Millington, 2006), and the GIS activ-
tions are related to a superior financial or resources, and signifi- ities are more effective if they are implemented on a larger scale.
cantly influence their environmental initiative commitment This could deter the firms with a small-scale operation to imple-
(Johnson and Greening, 1999). The smaller companies have inade- ment such activities. Furthermore, these firms could be dissuaded
quate or constrained resources, which affects the GIS application. based on the probability that their involvement was not prominent
The final attribute is related to firm size. The larger organisations and would not generate benefits. The firms also tend to avoid any
978 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

based on their 2-digit SIC codes and the companies with less than 7
observations were eliminated. The final data sample consisted of
163 firms and annual 1194 observations between 2011 and 2017.

3.2. Definition of variables and measurement

3.2.1. Green innovation strategy

In this study, we defined the GIS performance evaluation” based
on the ISO 14031 standards, similar to that used in earlier studies by
Fig. 1. Research framework. Chen et al. (2006), Campos et al. (2015), and Nguyen and Hens
(2015). Thus, the performance of a GIS was defined as the perfor-
mance of the hardware and software involved in any innovative
activity that was implemented by the company with regards to the
participation in GIS activities, since ineffective or inadequate use of green processes or products. These also include the tech-
participation could negatively affect their reputation. Hence, the nologies required for preventing pollution, energy-saving, recycling
subsequent Hypothesis was proposed: of wastes, designing green products and corporate environmental
management. Hence, in this study, we measured the GIS using
Hypothesis (H2). Firm size moderates the relationship between
three main CSRHub databases as follows:
green innovation strategies and corporate financial performance,
especially when the firm size is big, the relationship between the
GIS and the CFP becomes stronger. 3.2.1.1. Energy and climate change subcategory scores. This param-
eter measures the company's effectiveness while addressing the
To test these hypotheses, the research framework (Fig. 1) dem- climatic changes using appropriate energy-efficient operations,
onstrates the relationships of GIS as part of the company vital strategies and policies, the development of better and renewable
strategy that effects on the CFP. This study also delves the moder- energy sources and alternative environmental technologies. This
ating role of firm size in order to assess their influence between the subcategory includes the energy usage, emission of CO2 and other
GIS to CFP path. greenhouse gases.

3.2.1.2. Environmental policies and reporting the subcategory scores.

3. Methodology
This subcategory includes the company's intentions and policies for
reducing its environmental effect and value streams to the levels
3.1. Data collection and samples
which are healthy for the environment, in present and in future.
This data comprises the firm's environmental reporting perfor-
In this study, we compiled all the data from two datasets, i.e., the
mance, its adherence to the environmental reporting standards like
CSRHub (https://www.csrhub.com/csrhub/), which consists of all
the Global Reporting Initiative, and its compliance with the in-
information regarding the measures of GIS. CSRHub is a leading
vestors', regulatory or stakeholders' request for transparency. This
research company that consists of the Environmental, Social and
compliance data comprises of a breach of the accidental releases
Governance (ESG) data. This approach is more advantageous as it
and regulatory limits.
addresses the limitations seen in other methods like the Viego and
KLD. The CSRHub1 database comprises data from more than 18,424
organisations from 132 countries in 10 regions. Thus, the CSRHub 3.2.1.3. Resource management subcategory scores. This category
provides data from nine sources from the premier Socially determines how effectively the company uses all resources for
Responsible Investment (SRI) firms, known as the ESG analysis manufacturing or delivering the products and services, like the
firms, like the EIRIS, Carbon Disclosure Project (CDP), ASSET4 company's suppliers. This also includes the firm's ability to
(Thomson Reuters), Governance Metrics International (merged decrease the usage of materials, water and energy and the deter-
with the Corporate Library), IW Financial, MSCI (ESG Intangible mination of effective solutions for improving the supply chain
Value Assessment and the ESG Impact Monitor), Trucost, RepRisk, management. Furthermore, this subcategory also contains the
and Vigeo. Thereafter, the data collected from 265 NonGovern- environmental performance with respect to its production size and
mental Organisations (NGOs) like the publications, associations, the manner in which it is monitored using the operation linked
foundations, activist groups, union groups, governmental data- Eco-Intensity Ratios (EIRs) for the energy and water resources,
bases, and research reports, was augmented using data from other well-defined as the resource consumption per released/formed
data sources. Hence, the CSRHub schema was seen to be associated unit. The resources contain the raw materials and the packaging
with the firm's achievement and was based on the 0e100 rating materials used for production and packaging of products and
scale. A higher score indicates a positive rating score (100 ¼ very similar other processes. The Resource Management data includes
positive rating). Though the CSRHub updates all the values waste and recycling performance. Furthermore, the recycling data
monthly, the Datastream updates all the financial data quarterly or was associated with the ratio of the amount of waste that is recy-
annually. Thus, it could be seen that if the changes in the GIS cled to the total amount of waste. The data also includes the
significantly affected the firm performance, the Datastream data manner in which the company manages all the operations for
undergoes an annual change. Here, this study estimated the annual benefiting the local watershed and air shed and the manner in
changes occurring in the GIS by taking an average of all the GIS which the company affects the land usage and local ecological
scores for the consecutive 12 months, and thereafter combined the stability. All water resource-related data also comprises the con-
value with the Datastream data. All the industries were classified sumption of drinking water, industrial waters and steam. For
deriving the GIS data, this study calculated the mean scores for the
three subcategories as follow:
1
See detail of the CSRHub Ratings Methodology: https://esg.csrhub.com/csrhub-
ratings-methodology.
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 979

Energy and Climate Change þ Environmental Policy and Reporting þ Resource Management
GIS ¼ ½ 
3

3.2.2. Corporate financial performance (CFP) company could lead to erroneous results. This could be because of
The CFP was measured using the accounting measures the following reasons: First, the process of the product differenti-
described below: ation includes the investments in all those R&D projects that add to
Return on Assets (ROA): It is defined as the percentage for the social or environmental attributes of a product, which can be
determining the profitable nature of the company, related to the easily acknowledged by the customers. Second, the advertising
total assets. For calculating the ROA, we collected all data from the helps in increasing consumer awareness regarding the
DataStream regarding the total earning of the company before the environmentally-friendly products and the manner in which they
interest taxes, repayments and also the total assets of every com- differ from the other products. Thus, advertising was seen to be an
pany during the study period. Thereafter, we computed the Return indicator of the environmental responsiveness of the company to
on Assets as follows: the market. Here, we have computed the R&D factor using the ratio

EBITA ðEarning Before Tax; Interest and AmortisationÞ

Return on Assets ¼
Total Assets

of the R&D expenditures to the total sales, whereas advertising

refers to the ratio of the advertising expenses to the total sales.
Another measure of profitability includes the Return on Equity Several studies attempted to control the firm risk. This study
(ROE). This parameter is expressed as a percentage and defined as investigated many reports (Waddock and Graves, 1997;
the net income which is returned as the percentage of the share- McWilliams and Siegel, 2001), before measuring the risk, which
holder's equity. This factor was manually calculated by collecting all was calculated as the ratio of the total debts to the total assets. This
the data from the DataStream based on the earnings before the study also included the slack resource, which was calculated as the
interest taxes, amortisations and shareholder equity, as follows: ratio between the free cash flow and the company's total assets.

EBITA ðEarning Before Tax; Interest and AmortisationÞ

Return on Equity ¼
Total Equity

3.3. Empirical model

The last accounting measure includes the Return on Sales (ROS), 3.3.1. System Generalisation Method of moment (GMM)
which refers to the ratio which is used for measuring the opera- According to this study, two major issues have to be resolved in
tional efficiency. This factor was also expressed as a percent value this study. First, we exploited the dynamic data structure and
and was manually computed by collecting all data from the Data- studied the past CFP for determining the current CFP (Surroca et al.,
Stream for the total revenue and net income as follows: 2010). Secondly, while investigating the relationship between the
GIS and the CFP, the existing CFP could be correlated with the un-
Net Income observable and the observable factors (like the unobservable and
Return on Sales ¼
Total Revenue observable heterogeneity), which helps in determining the GIS-
related decisions. Specifically, the firms which relied on the high-
quality products or processes showed a higher GIS commitment.
However, the contribution of the GIS to the CFP would be over-
3.2.3. Control variables
stated if the endogeneity issues were not properly calculated.
Here, this study included a set of variables for controlling the
This study used the system Generalised Method of Moments
potential effects on the relationship between the GIS and the CPF.
(GMM) estimator, proposed earlier by Arellano and Bover (1995)
The various control variables described in earlier studies included
and Blundell and Bond (1998). This estimator is particularly
the firm size, firm risk, research and development intensity,
formulated for circumstances with 1) “small T, large N00 panels,
advertising intensity, and slack resources. Firm size was seen to be a
meaning fewer time periods and various individuals; 2) a linear
significant control variable and used the total assets of the company
functional relationship; 3) a single left-hand-side variable that is
as the indicator variable regarding its size. In their study,
dynamic, depending on its own past realizations; 4) independent
McWilliams and Siegel (2001) stated that an omission of the
variables that are not strictly exogenous, meaning correlated with
advertising and R&D factors from the model which studies the
past and possibly current realizations of the error; 5) fixed
relationship between the social and financial performances of a
980 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

Table 1 individual effects; and 6) heteroskedasticity and autocorrelation

Descriptive statistic. within individuals, but not across them. This estimator helped in
Variable Obs Mean Std. Dev. Min Max overcoming issues like the dynamic panel bias and the potential
ROA 1129 0.1253 0.1152 2.2000 0.5774
endogeneity of the regressors. Hence, this estimator was used
ROE 1129 0.4847 1.3588 4.5000 29.0400 rather than the traditional panel OLS or Within Group estimations
ROS 1119 0.7183 2.2555 7.3333 14.0000 approach (Arellano and Bover, 1995; Blundell and Bond, 1998,
GIS 1129 55.1327 9.1333 28.1767 79.0690 2000; Blundell et al., 2001; Bond et al., 2001; Hoeffler, 2002).
R&D Intensity 1129 0.0191 0.1035 0.0000 1.5131
Furthermore, the OLS levels and Within Groups estimations were
Advertisement Intensity 1129 0.1087 0.0962 0.0000 0.5536
ln total assets 1129 8.4112 1.8941 2.7801 13.0843 inconsistent and biased, since (i) OLS levels often neglect the un-
Leverage 1129 1.2382 4.6004 0.6341 90.4000 observed time-invariant firm effects; and (ii) The Within Groups
Free Cash Flow 1129 0.0387 0.0800 0.4533 0.5177 approach considers the unobserved country-specific effects within
ln revenue 1119 8.3706 1.8334 0.1310 12.4798
a specific time period using the dynamic panel data model (Nickell,

Table 2
Bivariate correlation matrix.

1 2 3 4 5 6 7 8 9 10

ROA 1
ROE 0.1804 1
ROS 0.0189 0.0115 1
GIS 0.0383 0.0267 0.0205 1
R&D Intensity 0.0399 0.0318 0.0072 -.01202 1
Advertisement Intensity 0.0303 0.0236 0.0431 0.0510 0.0550 1
ln total assets 0.0318 0.0235 0.0404 0.0789 0.0429 0.4651 1
Leverage 0.0037 0.9206 0.0102 0.0137 0.0232 0.0118 0.0060 1
Free Cash Flow 0.2004 0.0879 0.0200 0.0651 0.0361 0.1857 0.2306 0.0278 1
ln revenue 0.0582 0.0112 0.1533 0.0014 0.0014 0.3295 0.6942 0.0158 0.01320 1

Table 3
The effect of GIS on CFP.

Variables Dynamic system GMM

Panel 1 Panel 2 Panel 3

ROA ROS ROE

ROAt-1 0.627***
(0.0926)
ROSt-1 0.791***
(0.000835)
ROEt-1 0.00246***
(0.000951)
GIS 0.000611** 0.00325*** 0.00158**
(0.000242) (0.000600) (0.000712)
R&D Intensity 0.0285 0.0440 11.24***
(0.191) (1.475) (1.617)
Advertisement Intensity 0.153 0.223 0.251
(0.100) (0.535) (0.245)
ln total assets 0.0251*** 0.0942* 0.0690**
(0.00575) (0.0660) (0.0276)
Leverage 7.61e-05 0.000124 0.289***
(0.000239) (0.00188) (0.00313)
Free cash 0.300*** 0.477** 0.745***
(0.0672) (0.823) (0.188)
Dummy R&D Intensity 0.0489 1.908*** 1.956***
(0.0716) (0.639) (0.398)
Dummy Advertisement Intensity 0.00424 0.0410 0.375**
(0.0334) (0.207) (0.177)
Constant 0.216*** 0.423 0.432
(0.0731) (0.695) (0.320)
Year Dummy Yes Yes Yes
Observations 966 955 966
Number of firms 163 162 163
No of Instruments 26 26 26
AR1 2.87 (0.004) 0.30 (0.062) 2.07 (0.039)
AR2 1.124 (0.263) 0.26 (0.797) 0.02 (0.984)
Hansen Test 16.45 (0.422) 13.76 (0.744) 30.49 (0.116)
Different in Hansen Test 4.052 (0.853) 9.02 (0.341) 22.26 (0.104)

Notes: All models are estimated by using the Blundell and Bond (1998) dynamic panel data system GMM estimations and Roodman (2009) - Stata xtabond2
command. The standard errors are reported in parentheses, except for Hansen test, AR (1), AR (2) and Difference-in-Hansen which are p-values. ***, ** and * indicate
significance at 1%, 5% and 10% levels, respectively. Time dummies are included in the model specification, but the results are not reported to save space. The in-
struments employed in the first-differenced equation are two or more lags of the levels of the endogenous variables, while one lag of the first-difference of the
endogenous variables is used as instrument in the difference equation.
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 981

Table 4
The contingency effect of firm size on GIS e CFP link.

Variables Dynamic System GMM

Model 1 Model 2 Model 3

ROA ROS ROE

ROAt-1 0.540***
(0.0853)
ROSt-1 0.735***
(0.00983)
ROEt-1 0.00404*
(0.00215)
GIS 0.00571*** 0.300*** 0.517***
(0.00140) (0.0725) (0.0799)
Size 0.0730*** 1.656*** 3.329***
(0.0125) (0.634) (0.569)
Size*GIS 0.000598*** 0.0350*** 0.0612***
(0.000153) (0.00905) (0.0101)
R&D Intensity 0.00259 0.550 0.308
(0.179) (0.496) (0.344)
Advertisement Intensity 0.127 0.650 1.278
(0.0799) (4.083) (1.892)
ln total assets 0.0111 0.0164 0.538***
(0.0145) (0.314) (0.201)
Leverage 0.000230 0.00101 0.286***
(0.000249) (0.00127) (0.00776)
Free cash 0.280*** 0.756 1.240***
(0.102) (0.829) (0.429)
Dummy R&D 0.0131 0.197 0.335
(0.0668) (2.256) (0.340)
Dummy Advertisement Intensity 0.0233 0.457 0.244
(0.0322) (2.403) (1.619)
Constant 0.530*** 14.33*** 22.91***
(0.105) (4.034) (4.401)
Year Dummy Yes Yes Yes
Observations 957 955 957
Number of Firms 162 162 162
No of Instruments 29 29 29
AR1 2.73 (0.006) 2.51 (0.012) 2.90 (0.004)
AR2 0.73 (0.468) 0.37 (0.712) 1.68 (0.093)
Hansen Test 22.82 (0.198) 17.44 (0.425) 13.37 (0.717)
Different in Hansen Test 8.74 (0.462) 8.75 (0.364) 5.82 (0.667)

Notes: All models are estimated by using the Blundell and Bond (1998) dynamic panel data system GMM estimations and Roodman (2009) - Stata xtabond2
command. The standard errors are reported in parentheses, except for Hansen test, AR (1), AR (2) and Difference-in-Hansen which are p-values. ***, ** and * indicate
significance at 1%, 5% and 10% levels, respectively. Time dummies are included in the model specification, but the results are not reported to save space. The in-
struments employed in the first-differenced equation are two or more lags of the levels of the endogenous variables, while one lag of the first-difference of the
endogenous variables is used as instrument in the difference equation.

1981; Hsiao, 2014). Furthermore, the coefficient estimates of the were weaker (Blundell and Bond, 2000; Hoeffler, 2002).
lagged dependent variables derived from the OLS levels and Within Tables 3 and 4 present the System GMM regression results for
Groups estimators were considered to be the approximate upper the automotive companies, derived using Eqs. (2) and (5) in the 7-
and lower limits, respectively (Bond et al., 2001; Hoeffler, 2002). year period, between 2011 and 2017. It was believed that the two-
This system GMM is seen to yield efficient and consistent esti- step system GMM estimator yielded efficient estimates compared
mates in the regression model, wherein the independent variables to a 1-step system GMM approach. It was noted that the efficiency
were not strictly exogenous, i.e., these estimates were correlated gain was very small and the asymptotic standard errors related to
with the past and the existing realizations of error, if the autocor- the two-step GMM estimators were seriously biased downwards in
relation and heteroscedasticity in the estimates are existent the finite samples (Blundell and Bond, 1998; Hoeffler, 2002). Since
(Roodman, 2009a). Furthermore, this estimator controls the there are extra groups in this study, we used a two-step system
endogeneity issues by instrumenting all lagged dependent and GMM estimation method. In the case of all the estimates, the lagged
other endogenous variables with the variables that are believed to dependent variables were presumed to be predetermined, while
be unrelated to the fixed effects (Nickell, 1981; Roodman, 2009a). the control variables were considered to be endogenous.
Compared to the difference GMM estimator, proposed earlier by The stability of the System GMM estimators was dependent on
Arellano and Bond (1991), the System GMM was more efficient as it the assumptions that the error terms do not show serious corre-
assumed that the initial differences between the instruments were lation issues, the validity of all instruments and additional moment
uncorrelated with all fixed effects, which, included additional in- restrictions. For verifying the validity of all assumptions, we further
struments (Roodman, 2009a). Furthermore, the System GMM applied the Arellano-Bond test for determining a no serial corre-
yielded effective estimates in the cases where the series were lation between the error terms, while we applied the Hansen test
similar to random walks, while the Difference GMM estimator was for all instruments, and a Difference-in-Hansen test for the addi-
subjected to large sample bias, in such scenarios (Blundell and tional moment restrictions. The table also reports the specification
Bond, 1998). The Difference GMM estimator was more biased test results for all System GMM estimations. Based on all the tests,
downwards than the Within Groups estimator if all instruments the System GMM equations were appropriately specified.
982 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

Furthermore, the Arellano-Bond test results need a lack of AR (2) 4. Results and discussion
serial correlation between all error terms. The Hansen test also
assesses if the instruments were uncorrelated with an error term; In Step 1 of the empirical study, we aimed to offer direct
while the Difference-in-Hansen test determines the validity of the empirical evidence for describing the dynamic correlation between
additional moment restriction in Eqs. (2) and (5). GIS and CFP. We used the ROA as a CFP measure (Table 3) and
thereafter, replicated these estimates using the ROE and ROS
measures (Table 3) for assessing if the results were sensitive to the
3.3.2. Model particular CFP indicators. Lastly, we tested the interactive effect of
The empirical model which was used in this report was an the GIS and firm size on the CFP (Table 4).
extension of the model described earlier by many researchers ( Li
et al., 2017; Wong et al., 2018). Thereafter, it is a common activity 4.1. Descriptive statistics and correlation results
to examine all empirical relationships between the GIS and the CFP,
with the help of the following linear growth equation. Based on Tables 1 and 2 present the descriptive and correlation results
various models, the CFP relationship for the firm, i, in time, t, was a noted in the study. Table 1 describes the mean and median values of
function of GIS and control variables as follows: the main and control variables. The average GIS scores for the
companies investigated in the study during the time period be-
ð tween 2011 and 2017, was 55.12, which indicated that the GIS
Corporate Financial Performance ¼ Green Innovation Strategy performance was even and all GIS initiated by the automotive
sector was optimistic. This result was consistent with that observed
(1)
by Vaz et al. (2017). With regards to the financial variables, the
We estimated the relationship between CFP (ROA, ROE or ROS), maximal and the minimal ROA values were 0.58 and 2.20,
based on its lagged value, CPFit-1, and the GIP variables (rating or respectively; while those for ROE were 29.04 and 4.5, respectively
scores defined earlier), GIP and a set of firm-level control variables and ROS were 14.00 and 7.33, respectively. Furthermore, the
(i.e., in total assets, leverage, R&D, advertisement costs, free cash respective average values were 0.13, 0.48, and 0.72. In comparison,
flow and annual dummies), labelled CONTROLit, using the following the effect of the mean firm size on the total assets and revenue was
regression equation: seen to be 8.41 and 8.37, which indicated that there was no sig-
nificant difference between the 2 proxies. For determining the
X
n
likelihood of the presence of multicollinearity between the vari-
þa
CFPit ¼ a ^CFPit1 þ a
~GISit þ dJ CONTROLit þ ıj þ 
ait (2)
ables, we investigated the degree to which every variable was
j¼5
explained using other model variables, using the Variance Inflation
^j<1. The disturbances, mit and åi, were not cross-correlated Factor (VIF) (Hair et al., 1998). The results showed that the VIF
where ja
values were below the maximal acceptable value of 10, with the
and showed the properties:
values ranging between 1.03 and 7.74, and the tolerance ranged
between 0.96 and 0.12. Also, the mean VIF value of 3.43 suggested
E (åi) ¼ 0; E(ìit) ¼ 0; E(åiìit) ¼ 0 (3)
that the data points showed no multicollinearity-related issue in
All time-varying errors were presumed to be uncorrelated: the study.

E(mitìis) ¼ 0 with t s s (4) 4.2. GIS-CFP relationship

i ¼ 1, …, 163; t ¼ 2011 …, 2017. Table 3 presents the system GMM estimates for Model 1. CFP
Based on the study, no additional conditions were imposed on was measured using ROA. Table 3 also presents the CFP results that
the ìit variance, since the moment conditions needed for model were determined using the ROE and ROS measures, for Models 2
estimation, requires no homoscedasticity. CFP represents the and 3. Using the system GMM estimator, we validated the standard
existing firm performance, GIS refers to the total GIS scores for firm, tests for misspecification, i.e., a 2nd-order serial correlation test
i, in the period, t; CFPt-1 denotes the firm performance with 1 period (i.e., AR (2) test); Hansen test for other-identifying restrictions and
lag; CONTROL refers to the control variables (ln total assets refer to a Difference-in-Hansen test that determines the validity of addi-
the log of total assets; leverage, free cash flow and time dummies); tional moment restrictions. We also controlled the no. of in-
mi refers to unobserved firm-specific fixed effects; while, åit.was an struments against the group. The positive coefficient of the lagged
error term. A robustness test was carried out using other dependent dependent variables showed that the CFP was persistent, i.e., CFP
variables like ROE and ROS. was dependent on its earlier realisation. Results indicated that
For confirming the moderating role of the firm size in the irrespective of the estimation techniques, the control variables
automotive sector, we established some models and also studied showed no difference. We noted that the total assets (e.g., firm size)
the relationship between the GIS and CFP. The model which studied and slack resources positively affected CFP. Factors like RnD,
the effect of the interactions between the GIS and firm size on CFP leverage or advertising ratio did not affect the ROA.
was: This study compared the ROA, ROE and ROS values, and noted
that these values were similar, except the ROE and the ROS nega-
X
n tively affected the R&D intensity. This could be due to the fact that a
CFPit ¼  ^CFPit1 þ a
aþa ~GISit þ g1 GIPit *SIZEit þ dJ CONTROLit higher R&D expense negatively affected the CFP, as it also increased
j¼5 the finances required for implementing the new strategies (Hall
þ ıi þ 
ait and Weiss, 1967). However, only the ROE showed a positive cor-
relation with the leverage, which indicated that the debt played a
(5)
positive role in decreasing the agency issues as it discouraged the
The above-mentioned variables accounted for all probable in- free cash flow over-investment by the self-serving managers
teractions between GIS and firm size, while the affiliation of the (Jensen, 1986; Stulz, 1990; Harvey et al., 2004).
product of variables with GIS was included as the regressor. For testing Hypothesis 1, we applied the regression Model 1 in
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 983

Table 3, which showed a positive correlation between GIS and ROA

~ ¼ 0.000666, p < 0.05). This confirmed Hypothesis 1 that GIS
(a
positively affected CFP. Table 3 also indicated that the GIS coeffi-
~ ¼ 0.0033,
cient positively affected the ROS and the ROE estimates (a
p < 0.001; a~ ¼ 0.0016, p < 0.05). These findings were not based on
the reverse causality and were consistent with Hypothesis 1. Thus,
all results supported the earlier evidence regarding the synergistic
correlation between GIS and CFP (Hart, 1995; Porter and Van der
Linde, 1995a,b; King and Lenox, 2002). Implementation of GIS ac-
tivities increases the CFP, which helps the companies display a
better corporate reputation, thereby highlight their social re-
sponsibility (Cordano and Frieze, 2000; Cronin et al., 2011; Sheng
et al., 2011). Also, the proactive green innovators attract several
clients, which further increase the market share of all companies.
Use of GIS activities helps the firms increase productivity, avoid
environmental protests/penalties, enhance corporate reputation,
develop new markets, foster a green awareness-related image, and
achieve a 1st-mover based competitive advantage (Chen et al., Fig. 2. Effects of GIS on ROA: Contingent on firm size.
2006; Mu et al., 2009; Lau et al., 2010). Zhu et al. (2012) stated
that these eco-innovative activities help the firms decrease their
waste and increase brand promotion, which stimulates their mar-
ket shares and generate new business opportunities. This was
supported by the Toyota Prius Hybrid case, which was a status
symbol vehicle and used green-labelling product strategies (Bonini
and Oppenheim, 2008). Based on a resource-based view, the
corporate reputation was an intangible asset and source of
n-Correa
competition, as it was rare, inimitable and valuable (Arago
and Sharma, 2003). Based on a financial view, the market investors
offered a higher premium to the firms with a good image (Konar
and Cohen, 2001).

4.3. Moderation effect of the firm size on the correlation between

the GIS and CFP

Table 4 presents the model, which describes the interaction

between the GIS and firm size (SIZE*GIS). Brambor et al. (2006)
stated that the variables need not be individually interpreted, as
they were not important. Table 4 showed that the coefficient values Fig. 3. Effects of GIS on ROS: Contingent on firm size.
related to GIS and firm size were negative, while the interactive
terms were seen to be statistically significant determinants of the
CFP for all 3 models. For example, in the case of Model 1, the benefit after implementing their GIS which Fig. 3, and Fig. 4 also
interaction coefficient value between the firm size and GIS showed displayed the same results. Figs. 2e4 also indicated that the smaller
a significantly negative moderating effect on ROA (a ^ ¼ 0.000598, firms, which implemented GIS, lowered the damage inflicted onto
p < 0.01). Such empirical results showed that the firm size played a the organisation, and also helped in increasing the support from
vital role in moderating the effect of GIS on the CFP. However, if the the high-identification stakeholders. Furthermore, the small firms
interactive term showed a negative sign, and was significant, the
moderating effect of the GIS on the CFP weakened with increasing
firm size. Thus, the GIS showed a higher detrimental effect on the
CFP. Hence, this result supported the view that smaller firms could
easily recognise better opportunities. They were seen to be more
flexible while adjusting their research plans or during the imple-
mentation of their GIS activities. Furthermore, the smaller firms
were better able to adjust the employee incentives for providing
optimal innovative efforts, and they also allowed a lesser rigid
management structure which helped the important employees
devote more time for innovative activities, and not for
management-related activities (Rogers, 2004). This disproved Hy-
pothesis 2.
For interpreting and understanding the nature of this interac-
tion, this study presented the moderating correlation graphically
(Fig. 2) (Aiken et al., 1991). As shown in the figure, the slope of GIS
on the ROA for the larger firms was negative and significant
(0.0036, p < 1), however, it was still positive and slightly signifi-
cant for the smaller firms (0.0164, p < 1). Hence, compared to the
larger firms, the smaller firms could derive a higher financial Fig. 4. Effects of GIS on ROE: Contingent on firm size.
984 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

were probably close-structured, which indicated a differing appropriately estimate the innovative input of the small firms.
governance structure compared to the larger firms (Demsetz, 1983). Secondly, Acs et al. (1994) showed that small firms could effectively
Though a majority of the smaller firms face a performance-based take better advantage of the knowledge spill-over from the
issue, they offer better benefits than the larger firms. Their size corporate or university R&D departments. Thirdly, the economic
also makes them more flexible and independent from the institu- value of all the innovative activities differed between the smaller
tional bureaucracy (Leal-Rodríguez et al., 2015). These results were and larger firms, as noted by Cohen and Klepper (1992), who noted
in direct contrast to those presented by the Schumpeterian Hy- that under specific stochastic conditions, the large firms produce
pothesis (Scherer, 1970), which stated that the larger firms showed lesser innovations for every dollar spent on the R&D activities;
higher technological progress than the smaller firms, and therefore, however their innovations were of a better average quality.
displayed a better firm performance. Furthermore, due to the fact that the small firms generated
Here, this study noted that the firm size significantly affected more innovations compared to their input, Zenger (1994) stated
the performance of the automotive companies. Thus, the firm size that the apparent organisational diseconomies of scale far out-
was an external environment indicator and affected the business weighed the technological economy of scale in the R&D. Based on
performance. Even though GIS can drive firm's sustainability and the above-mentioned explanations regarding the organisational
growth, despite the fact GIS is implemented, an incur of higher cost characteristics depending on the firm size, it was concluded that
is unavoidable and an alteration in organisational normal practises the small or large firms were not better innovators. Rather, the
will transform an organisation becoming more fragile, and even small or the larger firms were better at differentiating the various
tarnish the sustainable competitive advantages. The traditional innovations, or their role varied in the industry cycle, in a “dynamic
economic trade-off debate suggests that companies impose higher complementary” manner (Nooteboom, 1994). The larger firms were
costs to enhance better environmental performance and that these better at innovating as they made good use of economy of scope or
costs surpass the monetary benefits gained from them (e.g., scale, or consisted of a large team of experts and specialists, for
Friedman, 1970; Greer and Bruno, 1998). Moreover, by improving conducting basic, science-based innovations or large-scale appli-
environmental performance a firm is simply transferring societal cations, which were of a higher economic value (Cohen and
costs to the firm (e.g., Bragdon and Marlin, 1972). Consequently, this Klepper, 1992). The smaller firms were better innovators since
approach proposes that engaging environmental initiatives might their effect of scale were not (yet) vital and they could make good
be both lossmaking money and unsuitable for firms. use of their proximity and flexibility to the market demands, like
Many researchers stated that the larger firms were effective developing new products or new market combinations, modifying
innovators. Any firm, which already possesses monopoly power, the existing products for the niche markets, or developing small-
was less motivated towards innovation, as it felt threatened by its scale applications. Furthermore, their efficiency in generating
competitors, or due to the fact that the sale of new products could these innovative products was also improved by their capability to
affect the sale of the existing products. Some studies (Mansfield, take advantage of the knowledge spill-over from the larger com-
1968; Mansfield et al., 1971) indicated that the larger firms, pany's R&D departments (Acs et al., 1994).
which included many people in the decision-making process and
consisted of a long chain of command, showed lower flexibility and 5. Conclusions
inefficient managerial coordination. It was stated that as the firm
size increased, the firms became very bureaucratic. Furthermore, In this study, this study has developed a novel theoretical model
this study would also be less motivated to investigate the larger which examined the relationship between the GIS, firm size and
firms, since their efforts would not yield a higher personal benefit the CFP. The results obtained from this study could make important
as the smaller firms. Also, the unexpected results would be lost in contributions to the existing literature regarding the sustainable
the shuffle, in the larger firms than the smaller firms. Thus, the development into the innovation and the strategy management
relative strength of the smaller firms was based on their behav- (Zhu et al., 2012; Dong et al., 2014; Fargnoli et al., 2014; Pekovic
ioural characteristics. For example, higher the motivation displayed et al., 2016). In the past few years, many companies have begun
by the management and workers, the better the variation and developing and implementing ecological modernisation tech-
improvisation in all tasks performed by the workers, tacit knowl- niques which helped them conduct their operations in an
edge resulting in specialised skills, and higher the flexibility and environmentally-innovative manner. This study noted that the
communication (Nooteboom, 1994; Rothwell and Dodgson, 1994). application of the GIS activities positively affected the CFP. The
As the advantages displayed by the larger firms were the limita- results could also contribute to the existing green management
tions of the smaller firms and vice versa, they could be summarised literature and offer more empirical support to the ecological
as the advantages offered by the smaller and larger firms. modernisation theory, which stated that the companies must
A majority of the empirical findings showed that the small and recognise the issues which hinder the environmental adaptation of
the medium-sized firms conducted more efficient R&D than the the industrial development and the economic growth. The
larger firms. The small firms and the independent investors were ecological modernisation theory was seen to be an important
disproportionately responsible for the major innovations (Acs and theory for the environmental innovation as it could offer solutions
Audretsch, 1990), which was similar to the observations made by for resolving the conflicts between the industrial development and
Vossen (1998), who stated that the smaller firms were more cost/ the environmental protection (Zhu et al., 2012). The ecological
profit efficient. The smaller firms showed a higher innovative modernisation theory stated that the green enterprises consider
output compared to their innovative inputs due to many reasons. the implementation of the environmental innovative activities as
Firstly, the R&D activities of small firms are usually underestimated an effective opportunity for asserting their social role and re-
in several standard surveys, since the formal R&D carried out in sponsibilities (Dong et al., 2014; Pekovic et al., 2016). Furthermore,
different R&D departments is generally measured (Kleinknecht and an environmental commitment was seen to stimulate the green
Reijnen, 1991). Many researchers investigated the different com- activities and the environmental innovation strategies. These re-
ponents of the innovation costs and noted that the large firms sults were in line with other studies which observed that the
showed a higher R&D investment compared to the small firms environmental GIS activities help the firms derive many benefits,
(Archibugi et al., 1995; Felder et al., 1996). If this was different from like economic performance, corporate reputation, and novel
the manner in which it was measured, the R&D activities cannot product-related success (Chiou et al., 2011; Fraj et al., 2013; Dong
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 985

et al., 2014; Yang et al., 2018). The firms can incorporate green organisational practices and constantly depending on new infor-
concepts into their processes and products for improving resource mation to diversify practices further, thus laying down the foun-
efficiency, reducing waste and increasing resource recovery for dation for enterprises that allows them to implement green
improving performance and sustainability. activities efficiently.
Furthermore, this study could make significant contributions to Finally, this study considers firm size to be a situational variable
the literature since it applied the concept of firm size to the GIS and examines the action path as well as the impact of a GIS on an
activities. The results offer empirical evidence which highlights the organisation's sustainable performance that were of various sizes,
correlation between the firm size and CFP. This study emphasised and a comparison was made for the different results for these
the moderating role of the firm size. The larger firms showed a different levels. At the micro-level, this study considers firm size
higher negative effect of the GIS and the CFP. Hence, the companies since it can influence GIS and moderate the association among GIS
must always consider their firm size before transitioning from the and CFP. For example, small firms and large firms might vary in GIS
GIS to the firm performance. perceptibility and decision-making preferences. GIS need consid-
erable investment and might be a high cost (Hull and Rothenberg,
5.1. The implications of the study 2008). Larger firms are usually the main target of environmental
complaints from government, societies, social media and cus-
This new trend of the environmental legislation for novel tomers, and they regularly become distinctive cases and references
product development has been increasing. The automobile manu- in dealing with environmental matters (Welch et al., 2000;
facturers are under high pressure for developing appropriate stra- Nishitani, 2009). Therefore, better understanding of whether and
tegies for meeting the challenges occurring due to an uncertain under what circumstances GIS improves CFP is monetarily mean-
business environment (Huang and Jim Wu, 2010; Cheng et al., ingful to managers who have been engaging or are being advised to
2014). Many environmental regulations like WEEE, RoHS and adopt GIS. This study claims that firms practices GIS activities to
REACH, have triggered the firms to increase the environmental convey a good message and send green signals to external stake-
sustainability of their processes and products (Fargnoli et al., 2014). holders, resolve the information asymmetry, and obtain positive
The car manufacturers need to constantly comply with environ- feedback from various stakeholders. For example, public listed
mental standards (Cheng et al., 2014). There is a higher demand for firms (larger organisations) are required to declare their CSR re-
developing novel car parts based on the green services and prod- ports annually.
ucts; hence the car manufacturers must implement GIS activities This study can be of significant practical value by allowing en-
for complying with the regulations and legislation for environ- terprises to understand how and when to implement a GIS. The
mental protection (Chiou et al., 2011; Dong et al., 2014; Fargnoli conclusion can aid enterprise managers in understanding the
et al., 2014). This study showed that the GIS played a vital role in meaningfulness of context-based green innovation. This means
the automotive industry. The results also indicated that this study that enterprise managers should not engage in green innovation
fulfilled the objection and showed that the implementation of GIS blindly, and GIS implementation needs to be done appropriately for
activities could help the companies show a superior performance each of the specific situations.
by managing the environmental risks and developing better ca-
pabilities for a constant improvement of the green processes and 5.3. Implications for the policy developers
products.
The implementation of GIS activities can prove to be advanta-
5.2. The implication for the managerial staff geous to the firms and even the society at large. These activities
must be encouraged by all policy-makers and governmental bodies.
This study offered several managerial implications. The imple- Though the GIS at every level could positively affect the CFP, this
mentation of GIS activities affected the competitiveness and firm was not universally true for the large firms. In this study, we stated
profits. The managers can resolve many environmental manage- that interactive term (GIS*Size) showed a negative sign, and was
ment problems during the strategic planning stage, i.e., managers significant, the moderating effect of the GIS on the CFP weakened
can develop better environmental GIS for integrating the ecological with increasing firm size. Thus, the GIS showed a higher detri-
activities in their business operations (Eiadat et al., 2008; Zhu et al., mental effect on the CFP. Hence, the policymakers should take care
2012; Fraj et al., 2013; Cheng et al., 2014). The managers have to both side of the firms either large or small, the governmental pol-
identify the ecological issues and implement environmental inno- icies must encourage green innovations strategies in the firms that
vative activities for addressing these issues. The managers need to due to the development of progressive measures like rebates,
understand the manner in which the environmental incentive grants or other punitive measures like quotas or tariffs. These ac-
programs can be executed, which would help them promote the tivities increase the significance of green innovation strategies
sustainable development of the green processes and products. The amongst the managers, who can help in resolving environmental
environmental innovation can help the firms achieve a waste management-related issues. For instance, The Paris Climate
reduction or elimination, recovery of resources and demateriali- Agreement was signed by many countries who pledged to decrease
sation and reuse of resources. These factors can positively affect the the emissions and environmental pollution. Development of GIS
GIS. was seen to be an important step in reducing emissions. More
This study also highlights that the firm employing a GIS enjoys effective tools need to be established by the government that go
performance as long term and perpetual and not short lived. It is beyond green subsidies or grants to support and encourage green
crucial for organisations to acknowledge that there is a rise in cost strategies. They also need to assess if these green subsidies were
or the short-term loss pertaining to advantages led by GIS imple- not as effective as anticipated e for example, access to cheaper
mentation in the early stage that result in long-term advantages. capital to conduct green projects (notably via direct participation
Also, this paper shows that the implementation path pertaining to a by the government and subsidised loans), subsidies to green R&D,
GIS would aid enterprise managers to gain a better understanding consumer mandates, feed-in tariff policies pertaining to renewable
regarding the change brought by GIS in the original organisational energy and green public procurement rules. Firms that have
practices and structure. Thus, enterprise managers need to focus on received grant or green subsidies need to be put under scrutiny to
organisational practices, which include integrating flexible improve the probability of subsidies being utilised effectively (Lin
986 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

et al., 2015). Next, as an alternative to framing policies derived from evidence and an application to employment equations. Rev. Econ. Stud. 58 (2),
277e297.
a particular theory, government authorities need to focus on
Arellano, M., Bover, O., 1995. Another look at the instrumental variable estimation
listening to practitioners to gain a better understanding regarding of error-components models. J. Econom. 68 (1), 29e51.
what issues firms are facing when employing green strategies, Bansal, P., 2005. Evolving sustainably: a longitudinal study of corporate sustainable
particularly those large firms that can perform well even when GIS development. Strat. Manag. J. 26 (3), 197e218.
Bansal, P., Gao, J., 2006. Building the future by looking to the past: examining
is in place. research published on organizations and environment. Organ. Environ. 19 (4),
A consequential practical implication is that if smaller com- 458e478.
panies decide to gain greater CFP through GIS, formerly they will Blundell, R., Bond, S., 1998. Initial conditions and moment restrictions in dynamic
panel data models. J Econom 87 (1), 115e143.
need also to pay attention to priority in resource distribution to the Bhambri, A., Sonnenfeld, J., 1988. Organization structure and corporate social per-
GIS engagement. This study results also propose that such strategic formance: a field study in two contrasting industries. Acad. Manag. J. 31 (3),
resource allocation is not incompatible with the pursuit of CFP 642e662.
Blundell, R., Bond, S., 2000. GMM estimation with persistent panel data: an
objectives that many smaller companies face the resource and application to production functions. Econom. Rev. 19 (3), 321e340.
knowledge constraints and understate the problems small firms Blundell, R., Bond, S., Windmeijer, F., 2001. Estimation in dynamic panel data
can meet in managing the inconsistent pressures involved in the models: improving on the performance of the standard GMM estimator. In:
Nonstationary Panels, Panel Cointegration, and Dynamic Panels. Emerald Group
concurrent pursuit financial and environmental objectives. The Publishing Limited, pp. 53e91.
environmental policy-makers necessity identifies the limitations Bond, S., Hoeffler, A., Temple, J., 2001. GMM Estimation of Empirical Growth Models.
which resource and knowledge constraints enact on the attainment Bonini, S.M., Oppenheim, J.M., 2008. Helping ‘green’ products grow. McKinsey Q. 3
(2), 1e8.
of environmental performance goals controlled for smaller com-
Bragdon, J.H., Marlin, J., 1972. Is pollution profitable. Risk Manag. 19 (4), 9e18.
panies and offer suitable funding and training support programs Brambor, T., Clark, W.R., Golder, M., 2006. Understanding interaction models:
that help in the development of the capability for smaller com- improving empirical analyses. Polit. Anal. 14 (1), 63e82.
panies to innovate and enhance their environmental and CFP. Brammer, S., Millington, A., 2006. Firm size, organizational visibility and corporate
philanthropy: an empirical analysis. Bus. Ethics Eur. Rev. 15 (1), 6e18.
Campos, L.M., de Melo Heizen, D.A., Verdinelli, M.A., Miguel, P.A.C., 2015. Envi-
5.4. Limitations and future suggestions ronmental performance indicators: a study on ISO 14001 certified companies.
J. Clean. Prod. 99, 286e296.
Chang, C.H., Chen, Y.S., 2013. Green organizational identity and green innovation.
Some limitations were noted during the interpretation of all the Manag. Decis. 51 (5), 1056e1070.
results of this study. Firstly, the self-reporting data could lead to a Chen, Y.S., 2008. The driver of green innovation and green imageegreen core
common method variance. This study recommends that future competence. J. Bus. Ethics 81 (3), 531e543.
Chen, Y.S., Lai, S.B., Wen, C.T., 2006. The influence of green innovation performance
studies must adopt objective data for evaluating green innovative on corporate advantage in Taiwan. J. Bus. Ethics 67 (4), 331e339.
performance. Secondly, the cross-sectional data could cause the Chen, Y.S., Chang, C.H., Wu, F.S., 2012. Origins of green innovations: the differences
occurrence of the firm-specific effects (Fraj et al., 2013; Pekovic between proactive and reactive green innovations. Manag. Decis. 50 (3),
368e398.
et al., 2016). In future, the researchers must use a longitudinal Cheng, C.C., Yang, C.L., Sheu, C., 2014. The link between eco-innovation and business
research design for validating the causal inferences. Furthermore, performance: a Taiwanese industry context. J. Clean. Prod. 64, 81e90.
this study only focused on the automotive sector in the world, Chiou, T.Y., Chan, H.K., Lettice, F., Chung, S.H., 2011. The influence of greening the
suppliers and green innovation on environmental performance and competitive
which could have many limitations. However, focusing only on the advantage in Taiwan. Transport. Res. E Logist. Transport. Rev. 47 (6), 822e836.
automotive sector could ensure positive innovation-related results. Christmann, P., 2000. Effects of “best practices” of environmental management on
On the other hand, in future, the researchers must focus on cost advantage: the role of complementary assets. Acad. Manag. J. 43 (4),
663e680.
different industries in other countries, for deriving additional in- Cohen, W.M., Klepper, S., 1992. The tradeoff between firm size and diversity in the
sights and comparing all the results. pursuit of technological progress. Small Bus. Econ. 4 (1), 1e14.
Cordano, M., Frieze, I.H., 2000. Pollution reduction preferences of US environmental
managers: applying Ajzen's theory of planned behavior. Acad. Manag. J. 43 (4),
Acknowledgments 627e641.
Cowen, S.S., Ferreri, L.B., Parker, L.D., 1987. The impact of corporate characteristics
This research was supported by Grant Universiti Putra Malaysia on social responsibility disclosure: a typology and frequency-based analysis.
Account. Org. Soc. 12 (2), 111e122.
(GUPM) (GP-IPS9536600).
Cronin, J.J., Smith, J.S., Gleim, M.R., Ramirez, E., Martinez, J.D., 2011. Green marketing
strategies: an examination of stakeholders and the opportunities they present.
Appendix A. Supplementary data J. Acad. Mark. Sci. 39 (1), 158e174.
Dangelico, R.M., Pujari, D., 2010. Mainstreaming green product innovation: why and
how companies integrate environmental sustainability. J. Bus. Ethics 95 (3),
Supplementary data to this article can be found online at 471e486.
https://doi.org/10.1016/j.jclepro.2019.04.214. Demsetz, H., 1983. The structure of ownership and the theory of the firm. J. Law
Econ. 26 (2), 375e390.
Donaldson, L., 2001. The Contingency Theory of Organizations. Sage.
References Dong, Y., Wang, X., Jin, J., Qiao, Y., Shi, L., 2014. Effects of eco-innovation typology on
its performance: empirical evidence from Chinese enterprises. J. Eng. Technol.
Acs, Z.J., Audretsch, D.B., 1990. Innovation and Small Firms. Mit Press. Manag. 34, 78e98.
Acs, Z.J., Audretsch, D.B., Feldman, M.P., 1994. R & D spillovers and recipient firm Driessen, P., Hillebrand, B., 2002. Adoption and Diffusion of Green Innovations.
size. Rev. Econ. Stat. 336e340. Eiadat, Y., Kelly, A., Roche, F., Eyadat, H., 2008. Green and competitive? An empirical
Adams, M., Hardwick, P., 1998. An analysis of corporate donations: United Kingdom test of the mediating role of environmental innovation strategy. J. World Bus. 43
evidence. J. Manag. Stud. 35 (5), 641e654. (2), 131e145.
Aguilera-Caracuel, J., Ortiz-de-Mandojana, N., 2013. The Financial Performance of Fargnoli, M., De Minicis, M., Tronci, M., 2014. Design Management for Sustainability:
Green Prospector Firms: a Contingent Approach (No. 13.02). an integrated approach for the development of sustainable products. J. Eng.
Aiken, L.S., West, S.G., Reno, R.R., 1991. Multiple Regression: Testing and Interpreting Technol. Manag. 34, 29e45.
Interactions. Sage. Felder, J., Licht, G., Nerlinger, E., Stahl, H., 1996. Factors determining R&D and
Arago n-Correa, J.A., Martín-Tapia, I., Hurtado-Torres, N.E., 2013. Proactive environ- innovation expenditure in German manufacturing industries. In: Determinants
mental strategies and employee inclusion: the positive effects of information of Innovation. Palgrave Macmillan, London, pp. 125e154.
sharing and promoting collaboration and the influence of uncertainty. Organ Fraj, E., Martínez, E., Matute, J., 2013. Green marketing in B2B organisations: an
Environ 26 (2), 139e161. empirical analysis from the natural-resource-based view of the firm. J. Bus. Ind.
Arago n-Correa, J.A., Sharma, S., 2003. A contingent resource-based view of proac- Mark. 28 (5), 396e410.
tive corporate environmental strategy. Acad. Manag. Rev. 28 (1), 71e88. Friedman, M., 1970. A theoretical framework for monetary analysis. J. Political Econ.
Archibugi, D., Evangelista, R., Simonetti, R., 1995. Concentration, firm size and 78 (2), 193e238.
innovation: evidence from innovation costs. Technovation 15 (3), 153e163. Ge, B., Yang, Y., Jiang, D., Gao, Y., Du, X., Zhou, T., 2018. An empirical study on green
Arellano, M., Bond, S., 1991. Some tests of specification for panel data: Monte Carlo innovation strategy and sustainable competitive advantages: path and
 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988 987

boundary. Sustainability 10 (10), 3631. and discovery in the ethical pharmaceutical industry. In: Research and Inno-
Gelb, D.S., Strawser, J.A., 2001. Corporate social responsibility and financial disclo- vation in the Modern Corporation. Palgrave Macmillan, London, pp. 157e185.
sures: an alternative explanation for increased disclosure. J. Bus. Ethics 33 (1), McElroy, K.M., Siegfried, J.J., 1985. The effect of firm size on corporate philanthropy.
1e13. Q. Rev. Econ. Bus. 25 (2), 18e26.
Gluch, P., Gustafsson, M., Thuvander, L., 2009. An absorptive capacity model for McGuire, J.B., Sundgren, A., Schneeweis, T., 1988. Corporate social responsibility and
green innovation and performance in the construction industry. Constr. Manag. firm financial performance. Acad. Manag. J. 31 (4), 854e872.
Econ. 27 (5), 451e464. McWilliams, A., Siegel, D., 2001. Corporate social responsibility: a theory of the firm
Green Jr., K.W., Zelbst, P.J., Meacham, J., Bhadauria, V.S., 2012. Green supply chain perspective. Acad. Manag. Rev. 26 (1), 117e127.
management practices: impact on performance. Supply Chain Manag.: Int. J. 17 Meznar, M.B., Nigh, D., 1995. Buffer or bridge? Environmental and organizational
(3), 290e305. determinants of public affairs activities in American firms. Acad. Manag. J. 38
Greer, J., Bruno, K., 1998. Greenwash: the Reality behind Corporate Environmen- (4), 975e996.
talism. IBON Foundation Incorporated. Third World Network. Miles, R.H., 1986. Managing the Corporate Social Environment: A Grounded Theory.
Hair Jr., J.F., Anderson, R.E., Tatham, R.L., Black, W.C., 1998. Multivariate Data Anal- Prentice Hall.
ysis. Prentice-Hall, New Jersey. Mol, A.P., 1997. Ecological modernization: industrial transformations and environ-
Hall, B.K., 2001. Organic selection: proximate environmental effects on the evolu- mental reform. Int. Handbook Environ. Sociol. 138e149.
tion of morphology and behaviour. Biol. Philos. 16 (2), 215e237. Mol, A., Sonnenfeld, D., 2000. Ecological modernisation around the world: an
Hall, M., Weiss, L., 1967. Firm size and profitability. Rev. Econ. Stat. 319e331. introduction. Environ. Pol. 9, 3e14.
Hart, S.L., 1995. A natural-resource-based view of the firm. Acad. Manag. Rev. 20 (4), Mu, J., Peng, G., MacLachlan, D.L., 2009. Effect of risk management strategy on NPD
986e1014. performance. Technovation 29 (3), 170e180.
Harvey, C.R., Lins, K.V., Roper, A.H., 2004. The effect of capital structure when ex- Murphy, J., Gouldson, A., 2000. Environmental policy and industrial innovation:
pected agency costs are extreme. J. Financ. Econ. 74 (1), 3e30. integrating environment and economy through ecological modernisation.
Hess, D., Rogovsky, N., Dunfee, T.W., 2002. The next wave of corporate community Geoforum 31 (1), 33e44.
involvement: corporate social initiatives. Calif. Manag. Rev. 44 (2), 110e125. Nguyen, Q.A., Hens, L., 2015. Environmental performance of the cement industry in
Hoeffler, A.E., 2002. The augmented Solow model and the African growth debate. Vietnam: the influence of ISO 14001 certification. J. Clean. Prod. 96, 362e378.
Oxf. Bull. Econ. Stat. 64 (2), 135e158. Nickell, S., 1981. Biases in dynamic models with fixed effects. Econometrica: J. Econ.
Ho€risch, J., Johnson, M.P., Schaltegger, S., 2015. Implementation of sustainability Soc. 1417e1426.
management and company size: a knowledge-based view. Bus. Strateg. Environ. Nilsson, M., Hillman, K., Magnusson, T., 2012. How do we govern sustainable in-
24 (8), 765e779. novations? Mapping patterns of governance for biofuels and hybrid-electric
Hsiao, C., 2014. Analysis Of Panel Data (No. 54). Cambridge university press. vehicle technologies. Environ. Innov. Soc. Trans. 3, 50e66.
Huang, Y.C., Jim Wu, Y.C., 2010. The effects of organizational factors on green new Nishitani, K., 2009. An empirical study of the initial adoption of ISO 14001 in Jap-
product success: evidence from high-tech industries in Taiwan. Manag. Decis. anese manufacturing firms. Ecol. Econ. 68 (3), 669e679.
48 (10), 1539e1567. Nooteboom, B., 1994. Innovation and diffusion in small firms: theory and evidence.
Huber, J., 1985. Die Regenbogengesellschaft: okologie und Sozialpolitik (The Small Bus. Econ. 6 (5), 327e347.
Rainbow Society: Ecology and Social Politics). Fisher Verlag, Frankfurt am Main. Organisationl for Economic Co-operation and Development (OECD), 2009. Eco-
Hui, S., 2010. Multi-objective optimization for hydraulic hybrid vehicle based on innovation in industry e enabling green growth. http://www.oecd.org/
adaptive simulated annealing genetic algorithm. Eng. Appl. Artif. Intell. 23 (1), document/62/0,3746,de_34968570_34968855_44452670_1_1_1_1.00.html.
27e33. Available at:
Hull, C.E., Rothenberg, S., 2008. Firm performance: the interactions of corporate Park, J., Sarkis, J., Wu, Z., 2010. Creating integrated business and environmental
social performance with innovation and industry differentiation. Strat. Manag. J. value within the context of China's circular economy and ecological moderni-
29 (7), 781e789. zation. J. Clean. Prod. 18 (15), 1494e1501.
€nicke, M., 2008. Ecological modernisation: new perspectives. J. Clean. Prod. 16 (5),
Ja Pekovic, S., Rolland, S., Gatignon, H., 2016. Customer orientation and organizational
557e565. innovation: the case of environmental management practices. J. Bus. Ind. Mark.
Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeovers. 31 (7), 835e848.
Am. Econ. Rev. 76 (2), 323e329. Petts, J., 1998. Environmental responsiveness, individuals and organizational
Johnson, R.A., Greening, D.W., 1999. The effects of corporate governance and learning: SME experience. J. Environ. Plan. Manag. 41 (6), 711e730.
institutional ownership types on corporate social performance. Acad. Manag. J. Porter, M.E., 1981. The contributions of industrial organization to strategic man-
42 (5), 564e576. agement. Acad. Manag. Rev. 6 (4), 609e620.
King, A., Lenox, M., 2002. Exploring the locus of profitable pollution reduction. Porter, M.E., Reinhardt, F.L., Schwartz, P., Esty, D.C., Hoffman, A.J., Schendler, A., et al.,
Manag. Sci. 48 (2), 289e299. 2007. Climate Business. Business Climate. Harv. Bus. Rev. 1.
Kleinknecht, A., Reijnen, J.O., 1991. More evidence on the undercounting of small Porter, T.B., 2006. Coevolution as a research framework for organizations and the
firm R&D. Res. Pol. 20 (6), 579e587. natural environment. Organ. Environ. 19 (4), 479e504.
Konar, S., Cohen, M.A., 2001. Does the market value environmental performance? Porter, russo M.E., Van der Linde, C., 1995a. Toward a new conception of the
Rev. Econ. Stat. 83 (2), 281e289. environment-competitiveness relationship. J. Econ. Perspect. 9 (4), 97e118.
KPMG, 2012. Expect the Unexpected: Building Business Value in a Changing World. Porter, M., Van der Linde, C., 1995b. Green and Competitive: Ending the Stalemate.
Executive Summary (accessed 24.11.18). https://www.kpmg.com/ The Dynamics of the Eco-Efficient Economy: Environmental Regulation and
dutchcaribbean/en/Documents/KPMG%20Expect_the_Unexpected_Exctve Competitive Advantage, vol. 33.
Smmry_FINAL_WebAccessible.pdf. Ramaswami, S.N., Srivastava, R.K., Bhargava, M., 2009. Market-based capabilities
Lai, S.B., Wen, C.T., Chen, Y.S., 2003. The exploration of the relationship between the and financial performance of firms: insights into marketing's contribution to
environmental pressure and the corporate competitive advantage. December. firm value. J. Acad. Mark. Sci. 37 (2), 97.
In: Paper Presented at the 2003 CSMOT Academic Conference, Hsinchu, Taiwan. Rivera, J., 2002. Assessing a voluntary environmental initiative in the developing
Lau, A.K., Tang, E., Yam, R.C., 2010. Effects of supplier and customer integration on world: the Costa Rican Certification for Sustainable Tourism. Pol. Sci. 35 (4),
product innovation and performance: empirical evidence in Hong Kong man- 333e360.
ufacturers. J. Prod. Innov. Manag. 27 (5), 761e777. Rogers, M., 2004. Networks, firm size and innovation. Small Bus. Econ. 22 (2),
Leal-Rodríguez, A.L., Eldridge, S., Rold an, J.L., Leal-Milla n, A.G., Ortega-Gutie
rrez, J., 141e153.
2015. Organizational unlearning, innovation outcomes, and performance: the Roodman, D., 2009. How to do xtabond2: an introduction to difference and system
moderating effect of firm size. J. Bus. Res. 68 (4), 803e809. GMM in Stata. STATA J. 9 (1), 86e136.
Li, D., Cao, C., Zhang, L., Chen, X., Ren, S., Zhao, Y., 2017. Effects of corporate envi- Rothwell, R., Dodgson, M., 1994. Innovation and size of firm. Handbook Indus. Innov.
ronmental responsibility on financial performance: the moderating role of 324.
government regulation and organizational slack. J. Clean. Prod. 166, 1323e1334. Runhaar, H., Tigchelaar, C., Vermeulen, W.J., 2008. Environmental leaders: making a
Lin, H., Zeng, S.X., Ma, H.Y., Chen, H.Q., 2015. How political connections affect difference. A typology of environmental leaders and recommendations for a
corporate environmental performance: the mediating role of green subsidies. differentiated policy approach. Bus. Strateg. Environ. 17 (3), 160e178.
Hum. Ecol. Risk Assess. Int. J. 21 (8), 2192e2212. Russo, M.V., Fouts, P.A., 1997. A resource-based perspective on corporate environ-
Link, S., Naveh, E., 2006. Standardization and discretion: does the environmental mental performance and profitability. Acad. Manag. J. 40 (3), 534e559.
standard ISO 14001 lead to performance benefits? IEEE Trans. Eng. Manag. 53 Scherer, F.M., 1970. Industrial Market Structure and Economic Performance, first ed.
(4), 508e519. Rand McNally College Publishing Company, Chicago.
pez-Nicola
Lo s, C., Meron
~ o-Cerd 
an, A.L., 2011. Strategic knowledge management, Sheng, S., Zhou, K., Li, J.J., 2011. The effects of business and political ties on firm
innovation and performance. Int. J. Inf. Manag. 31 (6), 502e509. performance: evidence from China. J. Mark. 75 (1), 1e15. https://doi.org/
Madden, K., Scaife, W., Crissman, K., 2006. How and why small to medium size 10.1509/jmkg.75.1.1.
enterprises (SMEs) engage with their communities: an Australian study. Int. J. Shrivastava, P., 1995. The role of corporations in achieving ecological sustainability.
Nonprofit Voluntary Sect. Mark. 11 (1), 49e60. Acad. Manag. Rev. 20 (4), 936e960.
Marcus, A.A., Fremeth, A.R., 2009. Green management matters regardless. Academy Smith, M., Crotty, J., 2008. Environmental regulation and innovation driving
of Management Perspectives 23 (3), 17e26. ecological design in the UK automotive industry. Bus. Strateg. Environ. 17 (6),
Mansfield, E., 1968. Industrial Research and Technological Innovation; an Econo- 341e349.
metric Analysis. Stulz, R., 1990. Managerial discretion and optimal financing policies. J. Financ. Econ.
Mansfield, E., Rapoport, J., Schnee, J., Wagner, S., Hamburger, M., 1971. Innovation 26 (1), 3e27.
988 W.-L. Lin et al. / Journal of Cleaner Production 229 (2019) 974e988

Surroca, J., Tribo, J.A., Waddock, S., 2010. Corporate responsibility and financial to the reduction of carbon dioxide. J. Policy Anal. Manag. 19 (3), 407e425.
performance: the role of intangible resources. Strat. Manag. J. 31 (5), 463e490. Wintoki, M.B., Linck, J.S., Netter, J.M., 2012. Endogeneity and the dynamics of in-
Tie, S.F., Tan, C.W., 2013. A review of energy sources and energy management ternal corporate governance. J. Financ. Econ. 3 (105), 581e606.
system in electric vehicles. Renew. Sustain. Energy Rev. 20, 82e102. Wong, C.W., Miao, X., Cui, S., Tang, Y., 2018. Impact of corporate environmental
Tomomi, T., 2010. Environmental management strategy for small and medium- responsibility on operating income: moderating role of regional disparities in
sized enterprises: why do SMBs practice environmental management? Asian China. J. Bus. Ethics 149 (2), 363e382.
Bus. Manag. 9 (2), 265e280. Wu, F., Yeniyurt, S., Kim, D., Cavusgil, S.T., 2006. The impact of information tech-
Tseng, M.L., Wang, R., Chiu, A.S., Geng, Y., Lin, Y.H., 2013. Improving performance of nology on supply chain capabilities and firm performance: a resource-based
green innovation practices under uncertainty. J. Clean. Prod. 40, 71e82. view. Ind. Mark. Manag. 35 (4), 493e504.
Van Beurden, P., Go €ssling, T., 2008. The worth of valuesea literature review on the Yang, D., Wang, A.X., Zhou, K.Z., Jiang, W., 2018. Environmental strategy, institu-
relation between corporate social and financial performance. J. Bus. Ethics 82 tional force, and innovation capability: a managerial cognition perspective.
(2), 407. J. Bus. Ethics 1e15.
Van de Ven, B., Jeurissen, R., 2005. Competing responsibly. Bus. Ethics Q. 15 (2), Yeung, A.C., 2008. Strategic supply management, quality initiatives, and organiza-
299e317. tional performance. J. Oper. Manag. 26 (4), 490e502.
 2017. Sustainability and innovation in the automotive
Vaz, C., Rauen, T., Lezana, A., York, R., Rosa, E.A., 2003. Key challenges to ecological modernization theory:
sector: a structured content analysis. Sustainability 9 (6), 880. institutional efficacy, case study evidence, units of analysis, and the pace of eco-
Vossen, R.W., 1998. Relative strengths and weaknesses of small firms in innovation. efficiency. Organ. Environ. 16 (3), 273e288.
Int. Small Bus. J. 16 (3), 88e94. Zenger, T.R., 1994. Explaining organizational diseconomies of scale in R&D: agency
Waddock, S.A., Graves, S.B., 1997. The corporate social performanceefinancial per- problems and the allocation of engineering talent, ideas, and effort by firm size.
formance link. Strat. Manag. J. 18 (4), 303e319. Manag. Sci. 40 (6), 708e729.
Welch, E.W., Mazur, A., Bretschneider, S., 2000. Voluntary behavior by electric Zhu, P., Chen, Y., Wang, L.Y., Zhou, M., 2012. Treatment of waste printed circuit board
utilities: levels of adoption and contribution of the climate challenge program by green solvent using ionic liquid. Waste Manag. 32 (10), 1914e1918.