Dynamic Capabilities, Environmental Management Capabilities, Stakeholder Pressure and Eco-Innovation of Chinese Manufacturing Firms: A Moderated Mediation Model

: Growing social and academic concerns toward environmental sustainability are constantly driving attention toward eco-innovation as an effective solution to environmental problems. Extant studies on drivers of eco-innovation have not fully investigated the interaction mechanisms between different factors. Drawing upon the resource-based view, stakeholder theory, and environmental literature, this paper aims to explore the mechanisms of how ﬁrm capabilities and perceived stake-holder pressure interact to facilitate eco-innovation. Based on survey data collected from 169 Chinese manufacturing ﬁrms, the results of regression analysis based on bootstrap resampling method show that environmental management system (EMS) completely mediates the relationship between dynamic capabilities (including absorptive capability and reconﬁguring capability) and eco-innovation (including eco-process and eco-product innovation). Furthermore, the relationship between absorp-tive/reconﬁguring capabilities and EMS, as well as the direct and indirect relationship between absorptive capability and eco-process innovation, is contingent upon stakeholder pressure. The direct and indirect effects become stronger when managers perceive higher stakeholder pressure. This paper contributes to answering why ﬁrms undertake environmental activities beyond compliance through specifying the interaction between ﬁrm capabilities and stakeholder pressure.


Introduction
Social concerns and academic interests toward sustainable development keep growing as environmental problems, such as climate change and environmental deterioration, draw ever-more attention.Eco-innovation provides a promising solution for improving environmental sustainability without impeding economic activities [1,2].However, firms lack motivation for eco-innovation due to the "double externalities" feature; that is to say, eco-innovation generates positive social and technology spillovers, while the costs are solely borne by firms [3].
Existing research mainly focuses on the effects of regulatory push/pull, technology push and market pull [4,5], while firm-specific factors such as capabilities receive less attention [6].The institutional theory emphasizes how firms become homogenous in pursuit of gaining and maintaining legitimacy [7,8], failing to explain why firms act differently when confronted with similar environmental contexts [5].On the contrary, the resource-based view (RBV) focuses on heterogeneity of firms and posits that valuable, rare, inimitable, and non-substitutable resources (VRIN resources) are sources of competitive advantages [9].The RBV explains how heterogenous firm resources lead to different strategic decisions, even under similar circumstances [10].Furthermore, in response to regulatory pressures, firms prefer to undertake reactive actions and minimum investments to comply with regulation and cope with external scrutiny, while firm resources facilitate proactive activities and increased investment in eco-innovation [11,12].
With regard to the influence of resources on eco-innovation, evidence is not only limited but also mostly in a static view [13], and mainly focuses on operational capabilities, such as technological capabilities, marketing capabilities, and environmental management capabilities [6,14].In fact, as the environmental condition changes frequently, firms have to constantly renew resources and capabilities in order to sense and seize new opportunities, thus achieving sustainable competitive advantages [15].Such capabilities are referred to as dynamic capabilities [16].Dynamic capabilities are essential for developing and reconfiguring necessary capabilities for eco-innovation, among which environmental management capability is considered as a particularly relevant factor due to its sustainability-oriented characteristics [6].
Furthermore, prior studies on drivers of eco-innovation are mostly scattered, inconclusive [5], and mainly focus on one single driver or simply compare the effects of multiple drivers [4,17].Empirical studies that combine different perspectives and investigate the mechanisms between different factors and eco-innovation are still insufficient.Besides RBV, the perspective of stakeholder also provides potentially important insights.The stakeholder theory indicates that stakeholders are able to impose important influences on firm activities, and stakeholder pressure on environmental issues facilitates proactive actions by firms [18,19].
Therefore, the reason that firms undertake environmental activities and pursue ecoinnovation beyond compliance may well lie in internal capabilities and stakeholder pressure.On the basis of the RBV, stakeholder theory, and environmental literature, this paper attempts to fill the above-mentioned gaps through exploring the influence mechanisms between capabilities, stakeholder pressure, and eco-innovation.Specifically, this paper aims to answer how firm capabilities, including dynamic capabilities and environmental management capabilities, facilitate eco-innovation, and how stakeholder pressure influences the relationship between them.Based on survey data collected from 169 manufacturing firms in eastern China, the results show that dynamic capabilities facilitate eco-innovation indirectly, while the environmental management system (EMS), as a reflection of environmental management capabilities, fully mediates the relationship between them.Furthermore, stakeholder pressure positively moderates the relationship between dynamic capabilities and EMS, as well as the direct and indirect effects between absorptive capability and eco-process innovation.
This paper contributes to the literature via combing capabilities and stakeholder perspectives to explain drivers of eco-innovation.To our best knowledge, this is the first empirical study that explores how dynamic capabilities and perceived stakeholder pressure interact to stimulate eco-innovation.
The remainder of this paper is structured as follows.Section 2 discusses the theoretical background, and Section 3 develops research hypotheses.Section 4 presents the data collection process and measurement of variables.Section 5 presents empirical results.Section 6 discusses the findings, theoretical contributions, and managerial implications of this paper, as well as limitations and directions for future research.

Theoretical Background 2.1. Dynamic Capabilities
The RBV emphasizes the heterogeneity of firms and considers VRIN resources as sources of competitive advantage [9].Firm resources contain assets, capabilities, organizational processes, information, and knowledge.Among these various resources, dynamic capabilities are argued to be particularly important in rapidly changing markets for firms to evolve promptly in order to gain and maintain competitiveness [20].Others prove that dynamic capabilities are also effective in less dynamic environments [21].In their seminal work, Teece, Pisano, and Shuen [16] (p.516) define dynamic capabilities as "the firm's ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments"; this definition has been broadly accepted by researchers [22][23][24][25].
Although researchers define dynamic capabilities from different perspectives as capabilities [16], processes [26], routines [27], or behavioral orientation [28], consensus has been reached that dynamic capabilities are higher-order capabilities which acquire, integrate, and reconfigure lower-order capabilities and assets, such as operational (ordinary) capabilities [29].Ordinary capabilities are the abilities required to solve a problem, while dynamic capabilities change the way the firm solves its problems [20].Therefore, dynamic capabilities can be seen as higher-order capabilities that absorb and reconfigure internal and external capacities.
Dynamic capability is regarded as a multidimensional construct in the majority of current studies.For instance, Eisenhardt and Martin [26] elaborate three important processes of dynamic capabilities that integrate, reconfigure, gain, and release resources.Wang and Ahmed [28] consider adaptive, absorptive, and innovative capability as important components of dynamic capabilities.Teece [15] delineates the nature and microfoundations of the capacity to sense and seize opportunities, and the ability to maintain competitiveness through reconfiguring.
Pandza and Holt [30] emphasize on the importance of absorptive and transformative capabilities to innovation systems.Absorptive capability is defined as "the ability of a firm to recognize the value of new, external information, assimilate it, and apply it to commercial ends" [31] (p.128), and transformative/reconfiguring capability refers to "the capability to constantly redefine a portfolio of product or service opportunities based on knowledge endogenous to the firm" [30] (p.350).These two dimensions are conceptually distinct but mutually reinforcing components of dynamic capabilities [32].Absorptive capability is outward-looking, and the inward-looking transformative process is dependent on the former capability, which concerns assimilation of external novel knowledge.The ability to reconfigure and update the resource base also feeds back into the development of absorptive capability.

Eco-Innovation
Consensus has not been reached considering the terminology and definition of ecoinnovation.Terms including "green", "eco/ecological", and "environmental" innovation are used interchangeably and largely synonymously in the literature [33].Scholars define eco-innovation either broadly or specifically.Rennings [3] (p. 322) adopts a broad definition of eco-innovation as "all measures of relevant actors (firms, politicians, unions, associations, churches, private households) which develop new ideas, behavior, products and processes, apply or introduce them and which contribute to a reduction of environmental burdens or to ecologically specified sustainability targets", while Cooke [34] specifically emphasizes the "recombinative" and "systemic" characteristics of innovation.Kemp and Pearson [35] note that the novelty of eco-innovation is not restricted to something that is new to the world, but also refers to something new to the firm or user.The distinct feature of ecoinnovation is that it brings environmental benefits, which can either be the goal of it, or incidental result of an innovation not deliberately aimed at reducing environmental harms.
Eco-innovation not only generates environmental improvement, which reduces external environmental costs in the diffusion phase, but also brings positive technology spillovers in the innovation phase.These peculiarities of eco-innovation relative to other types of innovation are referred to as "double externalities" [3].Due to the double-externality problem, firms lack motivation for eco-innovation since the costs are borne by the innovator alone, even though the whole society benefits from it [1].Consequently, from the perspective of neoclassical environmental economics, researchers argue that environmental regulation is negatively related to profits since the benefits are insufficient to offset the costs [36].However, Porter and van der Linde [37] criticize the neoclassical model as being static and ignoring innovation, and put forward a dynamic perspective based on innovation, often referred to as the Porter Hypothesis, which postulates that eco-innovation may lead to a "win-win" situation.On one hand, eco-innovation often contributes to reduction in resources and energy use, thus helping to save costs and increase profits.On the other hand, eco-innovation is also critical for firms to establish early-mover advantages in expanding markets for green products.
As for the typology of eco-innovation, drawing from the innovation literature, researchers distinguish between environmental-oriented technical, organizational, and social innovation.Focusing on technical innovation, Shen et al. [38] provide a measurement of green product and process innovation; the latter concept can be further divided into clean production technology and end-of-pipe treatment.The OECD's Guidelines for Collecting and Interpreting Technological Innovation Data document differentiates between technical and organizational innovation [39].Rennings [3] argues that the nature of ecoinnovations can be technological, organizational, social, or institutional.Hellström [40] puts forward a typology in terms of Schumpeterian innovation type and innovation mode, which identifies incremental-component, incremental-architectural, radical-component and radical-architectural modes of innovation.
This paper focuses on environmental technology innovation, and categorizes it into eco-process and eco-product innovation.Eco-process innovation involves the introduction of a more environmentally benign component to firms' production processes, which enables the reduction in input for a given amount of output.Eco-product innovation encompasses environmental improvements to products or introduction of a new green product to the market [39,41].

EMS
EMS, as the manifestation of environmental management capabilities, refers to "voluntary organizational frameworks that detail the procedures used to manage the impacts of the organization on the natural environment" [12] (p.864).It is also considered as voluntary environmental programs [41], environmental organizational innovations [17], or soft environmental policy instruments [42] by other scholars.Although researchers have different opinions regarding EMS, they commonly reveal the distinct feature of it; that is to say, EMS includes voluntary internal activities and processes which tackle with environmental problems.
Several management standards were introduced to assist organizations in developing EMS.ISO14001 is the most prevailing program, comprising the standards and requirements of EMS, and helping organizations to systematically manage their environmental responsibilities and enhance their environmental performance.Issued in 1996, nowadays there are over 420,000 organizations in more than 170 countries and regions adopting the ISO14001 standards, among which over 220,000 are certificated in China, according to the ISO survey [43].EMAS (the EU Eco-Management and Audit Scheme) is another instrument that helps organizations to evaluate, report, and improve their environmental performance.

Dynamic Capabilities and Eco-Innovation
The RBV asserts that VRIN resources are fundamental to competitive advantages [9], and also crucial for innovation [41].On the basis of RBV, Hart [44] develops the naturalresource-based view, which considers the challenge posed by the natural environment.Environmental practices and eco-innovation are influenced through routines, capabilities, and knowledge associated with environmental issues [41,45].Firms need to accumulate and recombine resources to boost eco-innovation and generate sources of sustained competitive advantage under environmental constraints [6].
There have been several past empirical studies that explore the relationship between capabilities and eco-innovation.Cainelli, De Marchi, and Grandinetti [46] prove that internal resources have a greater impact on eco-innovation than external and hybrid resources.Cai and Li [14] suggest that eco-innovation is driven via technological capabilities and environmental organizational capabilities.A meta-analysis conducted by Liao and Liu [11] reveals that technological capabilities and knowledge sources are positively related to eco-innovation.
Existing studies mainly adopt a static view of capabilities and focus on lower-order operational capabilities [13], despite the fact that higher-order dynamic capabilities, which build and reconfigure operational capabilities, play an important role in innovation.On one hand, compared to general innovations, for eco-innovation, firms have to consider not only economic interests, but also environmental and social benefits [11].Resources needed for eco-innovation are often quite different from those that the firm currently possesses [47].Dynamic capabilities enable firms to absorb and assimilate new external knowledge, and reconfigure internal resources so as to acquire and develop necessary resources needed for eco-innovation.On the other hand, eco-innovation deviating from current technology of the firm is associated with greater risks and uncertainty, and has a longer term of return on investment (ROI) [38].Dynamic capabilities provide sufficient resources that provide firms with a buffer against uncertainty and encourage riskier actions, and, consequently, facilitate investment in new environmentally friendly projects and products [48].Therefore: Hypothesis 1 (H1).Dynamic capabilities (a.absorptive capability, b. reconfiguring capability) are positively related to eco-process innovation.

EMS and Eco-Innovation
Despite the wide implementation and research interest toward EMS, prior studies on the relationship between EMS and eco-innovation remain inconclusive [49,50].There are some evidences supporting the impact of EMS on eco-innovation.For instance, Li, Tang, and Jiang [47] find that ISO14001 certified firms adopt more eco-innovation than noncertified firms.Horbach [51] demonstrates that introduction of environmental management tools triggers eco-product innovation based on panel databases.Kesidou and Demirel [12] highlight that EMS influences not only decisions to invest in eco-innovation, but also the level of investment.Another study on EMAS certified firms proves that maturity of EMS and strong participation of specific departments have positive impacts on ecoprocess innovation, while learning process via EMS is positively related to eco-product innovation [52].
However, EMS, as a voluntary environmental program, has been criticized by other scholars for lacking obligations or enforcements [41], and for being symbolic, legitimacyseeking, and based on "greenwashing" [47].The mere existence of EMS or ISO14001 certification does not necessarily lead to improved environmental performance [49].In some cases, EMS may only be a superficial tool to build a green image and manage stakeholder pressure without substantial efforts to mitigate environmental influence [42].For instance, King and Lenox [53] reveal that voluntary environmental programs do not guarantee environmental performance.Wagner [54] finds no evidence that supports the influence of EMS on eco-product innovation.Frondel, Horbach, and Rennings [55] find no association between EMS adoption and abatement activities, including end-of-pipe technologies and cleaner production innovations.
These mixed results can be attributed to inconsistency in measurement.Since the level of EMS implementation varies significantly across firms, it is important to capture the commitment and comprehensiveness of EMS.Dichotomous measurement, which simply classifies firms into EMS adopters and non-adopters, fails to distinguish between superficial efforts and authentic commitment.Firms may implement a minimum level of practices only to improve their green image and avoid scrutiny from stakeholders [56].Voinea et al. [42] note that the effect of EMS on environmental performance becomes consistently significant when EMS comprehensiveness is taken into account.Phan and Baird [56] prove that EMS comprehensiveness is associated with environmental performance.Demirel and Kesidou [6] argue that the quality of EMS implementation is the most important factor.Deep involvement from executive levels and years of experience in adopting EMS are essential to boost eco-innovation.
EMS facilitates eco-innovation through helping firms to overcome information asymmetry, as well as coordinating firm activities and accumulating necessary resources.On one hand, the Porter Hypothesis assumes that firms lack experience in detecting potential benefits of eco-innovation [17,37].EMS enables firms to detect information gaps and identify opportunities related to eco-innovation, which reduces cost or satisfies customer needs for green products [11,17].On the other hand, EMS contributes to coordination of firm activities and generation of necessary resources.Deep implementation of EMS triggers learning processes to accumulate necessary knowledge, and skills to initiate eco-innovation that continuously seeks new ways of environmental improvement [12,47].EMS also acts as a coordination mechanism that helps to set plans, coordinate inner activities, and guide environmental actions [50,54].Furthermore, EMS strengthens the environmental orientation of the firm and integrates environmental considerations into business strategy, creating the necessary conditions for environmental capabilities to be more efficient [49,57], thus making it possible to proactively improve production processes to raise efficiency, and redesign products to reduce environmental influence [49].Therefore: Hypothesis 3 (H3).EMS is positively related to (a) eco-process innovation and (b) eco-product innovation.

Mediation Role of EMS between Dynamic Capabilities and Eco-Innovation
It is argued that capabilities with specific environmental initiatives are more directly related to eco-innovation compared to general capabilities [6].Demirel and Kesidou [6] highlight that eco-innovation is driven via sustainability-oriented capabilities, such as EMS, that allow firms to tackle external pressures.Capabilities specific to the environmental domain may be more important to successful eco-innovation than general capabilities [6,58].Liao and Liu [11] reveal in their meta-analysis that the motivational effect of EMS on eco-innovation is stronger than technological capabilities and knowledge sources.Thus, the relationship between dynamic capabilities and eco-innovation can be indirect, mediated through environmental capabilities.EMS is the reflection of environmental capabilities [12,49], namely operational capabilities that allow firms to solve environmental problems [20].Dynamic capabilities as higher-order capabilities influence such lower-order operational capabilities via modifying the way of solving problems in response to the changing environment [20,24,29].As Collis [59] points out, dynamic capabilities govern the rate of change in operational capabilities [28].Several empirical studies tried to investigate the relationship between dynamic capabilities and environmental management activities.Hofmann, Theyel, and Wood [60] demonstrate that three types of dynamic capabilities-capabilities for adopting advanced technology, inter-firm collaboration, and product innovation-facilitate environmental management practices.Lee and Klassen [61] reveal that learning capabilities positively affect carbon management practices.Dangelico, Pujari, and Pontrandolfo [62] suggest that sustainability-oriented dynamic capabilities facilitate environmental capabilities.
Implementation of EMS is a complex and time-consuming task which requires various expertise and resources [42,63].Indeed, deficiency of competences is one of the major obstacles of EMS [64].Dynamic capabilities generate necessary resources via absorbing external knowledge and reconfiguring internal assets.Kabongo and Boiral [13] state that in order to promote efficient environmental activities, firms have to make major changes in different areas, such as employee training and supply chain management, that require cross-functional capabilities.They reveal that environmental activities are dependent upon several functional dynamic capabilities.
Moreover, the decision to proceed environmental activities is affected by the resources available.When faced with the same level of uncertainty and climate change, firms act differently since heterogeneous internal capabilities influence their perception of external environment, which, in turn, affects the optimal strategy for managing environmental problems.Firms unable to develop necessary resources prefer to "wait-and-see", avoiding taking risks and preferring a reactive response.On the contrary, firms with sufficient resources are more confident with the consequences of precautionary actions, thus becoming more willing to take risks and proactively implement environmental management activities to achieve differentiation [61].Dynamic capabilities are critical for firms to perceive potential competitive benefits of proactive environmental actions, including lower costs, improved reputation, and better relationships with stakeholders [65].Empirical results support the notion that firms adopting EMS and related practices are those who have more resources to devote to environmental improvement [66].Hence, firms develop sufficient environmental capabilities through absorbing external novel knowledge as well as reconfiguring internal resources, and abundant resources enable them to proactively identify and respond to new environmental challenges.Therefore: Hypothesis 4 (H4).The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and eco-process innovation is mediated by EMS.

Hypothesis 5 (H5).
The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and eco-product innovation is mediated by EMS.

Moderation Role of Stakeholder Pressure
Besides internal capabilities, stakeholder pressure also has an important influence on environmental activities.Stakeholder refers to "any group or individual who can affect or is affected by the achievement of the organization's objectives" [18] (p.46).The stakeholder theory states that firms should expand their objectives from solely realizing shareholder interests to satisfying expectation of various stakeholders [67].It is argued that stakeholder pressure facilitates implementation of proactive environmental activities [19,68].Furthermore, the institutional theory also supports the idea that firms have to meet stakeholders' expectation of corporate environmental legitimacy in order to improve their access to resources and insulate themselves from scrutiny [69].Corporate environmental legitimacy can be improved through demonstrating strong commitment to sustainable development.
Stakeholders can be classified as primary and secondary stakeholders based on their relationship with firms [67].Primary stakeholders, such as customers and employees, are crucial for firms to survive and can influence firm activities directly; firms have formal relationship with them, which often involves special duties.Secondary stakeholders, such as non-governmental organizations (NGOs) and media, have no formal direct relationship with firms; however, firms may have regular moral duties to them [70,71].Buysse and Verbeke [67] further classify stakeholders into internal stakeholders, external primary stakeholders, external secondary stakeholders, and regulatory stakeholders.Stakeholder pressure influences the goals and strategies that firms adopt [72].As environmental problems are drawing increasing attention of various stakeholders, firms paying more attention to stakeholder pressure are more likely to adopt proactive environmental strategies to identify environmental risks and opportunities, and react in a timely manner [73].Perceived pressure from various stakeholders influences environmental activities in several ways.
Firstly, with regard to external primary stakeholders, customers are becoming increasingly better informed about environmental information, and more concerned with health, clean production, and green products [74].They can force firms to take more environmental responsibility through "buy-cotting" green products and boycotting products of high-polluting firms [75].Secondly, with regard to external secondary stakeholders, NGOs and the media are capable of facilitating information availability and leading public opinion toward environmental issues.They influence firms' environmental activities indirectly via improving public awareness, as well as via protesting and industry calls [68].Thirdly, with regard to internal stakeholders, employees that support the sustainable development goals of firms are more devoted to work and pivotal for the development of environmental capabilities.On the contrary, firms lacking effective environmental management will find it difficult to attract or retain employees who are concerned with environmental problems, and even risk being publicly exposed for their negligent behaviors [67].Shareholders are also concerned with potential liabilities and risks related to ineffective environmental management [73].Fourthly, with regard to regulatory stakeholders, the government establishes environmental standards and monitors firm behaviors.Firms violating environmental regulations will be punished via penalties, taxes, or even having their licenses revoked, while firms promoting environmental sustainability will be rewarded with tax reductions or stimulated demand for green products [75].The increasing complexity and stringency of environmental regulation raised the risks associated with non-compliance [67].
Therefore, stakeholder pressure poses challenges as well as creating potential opportunities.It is beneficial for firms to respond to stakeholder expectations via strengthening environmental orientation and promoting environmental activities.For one thing, proactive environmental management is an important approach to gain and maintain competitive advantages derived from differentiation.When the majority of competitors are paying more attention to environmental management, firms that act poorly will lose their advantages and face boycotts [67].Moreover, efforts to meet stakeholder expectations also improve firms' reputations, while environmental incidents tarnish their reputations and harm economic benefits [76].
In short, firms implementing higher levels of environmental management activities in order to address stakeholder pressure are able to benefit from improved competitive advantages and reputations, while firms failing to meet expectation of stakeholders on environmental issues suffer from higher risks.Thus, the higher the stakeholder pressure that a firm perceives, the more likely it will utilize dynamic capabilities to improve environmental management capabilities, in order to benefit from environmental activities and avoid risks.On the contrary, the lower the perception of stakeholder pressure, the lower the motivation of firms to improve environmental management capabilities, which requires absorption and reconfiguring of resources.Therefore: Hypothesis 6 (H6).The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and EMS is moderated through stakeholder pressure.The greater the stakeholder pressure, the stronger the relationship.
Similarly, when firms perceive higher stakeholder pressure, they are more likely to absorb new information and reconfigure existing resources to develop eco-innovation, which satisfies stakeholder expectation as well as bringing competitive advantages, especially in cases where substantial first-mover advantages from environmental technologies are present [67], or when customers are environmentally conscious and willing to pay higher prices for green products.Therefore: Hypothesis 7 (H7).The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and eco-process innovation is moderated through stakeholder pressure.The greater the stakeholder pressure, the stronger the relationship.

Hypothesis 8 (H8).
The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and eco-product innovation is moderated through stakeholder pressure.The greater the stakeholder pressure, the stronger the relationship.
In line with the previous hypotheses, it is also hypothesized that the indirect effect of dynamic capabilities on eco-innovation through EMS is contingent upon perception of stakeholder pressure.Therefore: The research framework is presented in Figure 1.
Hypothesis 7 (H7).The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and eco-process innovation is moderated through stakeholder pressure.The greater the stakeholder pressure, the stronger the relationship.
Hypothesis 8 (H8).The relationship between dynamic capabilities (a.absorptive capability, b. reconfiguring capability) and eco-product innovation is moderated through stakeholder pressure.The greater the stakeholder pressure, the stronger the relationship.
In line with the previous hypotheses, it is also hypothesized that the indirect effect of dynamic capabilities on eco-innovation through EMS is contingent upon perception of stakeholder pressure.Therefore: The research framework is presented in Figure 1.

Data Collection
We conducted a questionnaire survey in order to better reflect the detailed information regarding firms dynamic capabilities, intensity of eco-process and eco-product innovation, comprehensiveness of EMS, and perceived stakeholder pressure.Survey data was collected from manufacturing firms operating in Zhejiang, China, including chemicals, construction material, electronic components, energy, food, furniture, machinery, metal, paper, textiles and clothing, transportation, and other industries.Questionnaire was translated into Chinese following Brislin s [77] procedure.Prior to the survey, short interviews with 10 managers were conducted to identify ambiguities and make sure that the questionnaire was well understood, thus assuring content validity.
We contacted 350 firms via telephone to inform all respondents of the aim and confidentiality of the survey, as well as seek the consent of senior and middle-level managers familiar with each firm s capabilities and environmental issues to participate.After gaining permission, questionnaires were sent to managers, and 181 completed responses were obtained with a response rate of 51.7%.After screening out 12 invalid questionnaires, 169 effective responses remained, representing a 48.2% effective response rate.
Results of t-test comparison between early and late respondents indicated no significant differences for variables, suggesting that non-response bias did not appear to present a problem [78].

Methodology 4.1. Data Collection
We conducted a questionnaire survey in order to better reflect the detailed information regarding firms' dynamic capabilities, intensity of eco-process and eco-product innovation, comprehensiveness of EMS, and perceived stakeholder pressure.Survey data was collected from manufacturing firms operating in Zhejiang, China, including chemicals, construction material, electronic components, energy, food, furniture, machinery, metal, paper, textiles and clothing, transportation, and other industries.Questionnaire was translated into Chinese following Brislin's [77] procedure.Prior to the survey, short interviews with 10 managers were conducted to identify ambiguities and make sure that the questionnaire was well understood, thus assuring content validity.
We contacted 350 firms via telephone to inform all respondents of the aim and confidentiality of the survey, as well as seek the consent of senior and middle-level managers familiar with each firm's capabilities and environmental issues to participate.After gaining permission, questionnaires were sent to managers, and 181 completed responses were obtained with a response rate of 51.7%.After screening out 12 invalid questionnaires, 169 effective responses remained, representing a 48.2% effective response rate.
Results of t-test comparison between early and late respondents indicated no significant differences for variables, suggesting that non-response bias did not appear to present a problem [78].

Measurement 4.2.1. Dynamic Capabilities
Following Pandza and Holt [30], this paper conceptualized dynamic capabilities as consisting of two dimensions, namely absorptive and reconfiguring capability.As mentioned above, they were conceptually distinct but mutually promoting components of dynamic capabilities that formed integral parts of the construct [32].Absorptive capability was measured using three items, and reconfiguring capability was measured using six items adapted from Wang, Senaratne, and Rafiq [32].Respondents were asked to answer on a five-point Likert scale, ranging from "1 = strongly disagree" to "5 = strongly agree".

Eco-Innovation
Some empirical studies operationalized eco-innovation as a binary variable, considering whether a firm implemented cleaner technology or provides environmentally friendly products [4,51,54].This approach reflected the choice of environmental investments, but was unable to capture detailed information, such as intensity of eco-innovation.The majority of current research used environmental R&D expenditures [12] or patent data [79] as proxies of eco-innovation [38].While they indicated the level of environmental investment and performance, it was difficult to differentiate between various types of eco-innovation.Moreover, the measure of patent number attributed the same importance to all patents, ignoring the fact that value of different patents may vary [80].
Consequently, researchers have developed survey questions in order to better reflect the intensity of different aspects of eco-innovation.For instance, Horbach, Rammer and Rennings [17] assess eco-innovation with 12 items measuring different areas of environmental impacts, and Long and Liao [2] use 4 items to measure eco-product innovation.In this paper, eco-process innovation was measured using two items adapted from Frondel, Horbach, and Rennings [39], and eco-product innovation was measured using three items adapted from Amores-Salvadö, Martin-de Castro, and Navas-López [49] on a five-point Likert scale, ranging from "1 = strongly disagree" to "5 = strongly agree".

EMS
Dichotomous measurement of EMS is criticized since EMS might be motivated by "greenwashing" rather than intention to mitigate environmental effects [49].Others take a more comprehensive approach and scrutinize different aspects of EMS.For example, Anton, Deltas, and Khanna [81] looked into the adoption of thirteen environmental management practices.Phan and Baird [56] considered not only the number of environmental management practices, but also the level of implementation of nine practices.Similarly, Amores-Salvadö, Martin-de Castro, and Navas-López [49] took into account the degree of development of seven distinctive elements that constitute EMS.
As a result, in order to more accurately evaluate EMS, it was necessary to look into the degree of implementation of various environmental management practices which represents the comprehensiveness of EMS.Following Phan and Baird [56], this paper measured the extent to which firms implemented each of nine environmental practices, using a five-point Likert scale ranging from "1 = not at all" to "5 = to a great extent".

Stakeholder Pressure
In line with prior empirical studies [19,82], when measuring stakeholder pressure in the context of environmental issues, what really mattered was managers' perception of environmental stakeholder pressure, since implementation of environmental practices and environmental strategies are highly dependent on managers [76].
Murillo-Luna, Garcés-Ayerbe, and Rivera-Torres [19] demonstrated that managers regard stakeholder pressure as a whole, with stakeholders' expectation for environmental issues perceived by managers as one dimension, rather than different demands originating from various groups of stakeholders.Therefore, this paper measured stakeholder pressure as one dimension, reflecting the overall collective perception of managers.Respondents were asked to indicate the level of four types of stakeholder pressure [67] they perceived to reduce environmental influences on a five-point Likert scale, ranging from "1 = very low pressure" to "5 = very high pressure".

Control Variables
Firm size, age and, industry were controlled in the following analyses.Firm size was measured as the natural logarithm of the total number of employees in the firm.Firm age was assessed as the natural logarithm of the firm's established years.With regard to industry, firms were classified into high pollution industries and low pollution industries on the basis of the Guidelines for Environmental Information Disclosure of Listed Companies, issued by Ministry of Ecology and Environment of the People's Republic of China [83].High pollution industries included chemical, construction, energy, metal, paper, and textiles.Low pollution industries included machinery, transportation, electronic components, food, furniture, and others.
spondents were asked to indicate the level of four types of stakeholder pressure [67] they perceived to reduce environmental influences on a five-point Likert scale, ranging from "1 = very low pressure" to "5 = very high pressure".

Control Variables
Firm size, age and, industry were controlled in the following analyses.Firm size was measured as the natural logarithm of the total number of employees in the firm.Firm age was assessed as the natural logarithm of the firm s established years.With regard to industry, firms were classified into high pollution industries and low pollution industries on the basis of the Guidelines for Environmental Information Disclosure of Listed Companies, issued by Ministry of Ecology and Environment of the People s Republic of China [83].High pollution industries included chemical, construction, energy, metal, paper, and textiles.Low pollution industries included machinery, transportation, electronic components, food, furniture, and others.

Validity and Reliability
Confirmatory factor analysis (CFA) was employed to test construct validity.Anderson and Gerbing [84] suggested that a minimum sample size of 150 was needed for structural equation modeling.Moreover, based on the analysis suggested by MacCallum,

Validity and Reliability
Confirmatory factor analysis (CFA) was employed to test construct validity.Anderson and Gerbing [84] suggested that a minimum sample size of 150 was needed for structural equation modeling.Moreover, based on the analysis suggested by MacCallum, Browne, and Sugawara [85], the minimum sample size for tests of fit on the basis of the root mean square error of approximation (RMSEA) index was calculated.The minimum sample size required to achieve the desired power of 0.8 was 89 (degree of freedom (df) = 309, significance level = 0.05; H0: RMSEA = 0, H1: RMSEA = 0.05).The CFA fit indices were: chi-square/df ratio (χ 2 /df) = 1.786 < 2, comparative fit index (CFI) = 0.927 > 0.9, and goodness-of-fit index (GFI) = 0.810 > 0.8, RMSEA = 0.068 < 0.08, showing that the measurement model was acceptable [85][86][87].Standardized loadings were statistically significant, ranging from 0.639 to 0.937 (see Appendix A), indicating convergent validity.The average variance extracted (AVE) for the individual variables exceeded the recommended threshold of 0.5, and the Pearson correlation coefficients between each variable were lower than the corresponding square roots of AVE, demonstrating discriminant validity (see Tables 1 and 2) [88].Common method variance (CMV) was tested via Harman's single-factor test [89].The results of unrotated principal component factor analysis revealed that no single factor emerged, and none of the factors accounted for the majority of variance.The most influential factor accounted for 31.88% of the total variance.CMV was also tested via CFA analysis.A single-factor model was generated in which all 27 items joined into a single latent construct.The model fit of the six-factor measurement model was significantly better than the single-factor model (∆χ 2 = 1772.919,p < 0.001).Consequently, the presence of common method bias was not supported [90].
Reliability was assessed via both composite reliability (CR) and Cronbach's α coefficients.The results showed that CR and Cronbach's α coefficients of all variables were larger than the recommended level of 0.7 [91].

Hypotheses Testing
Hypotheses were tested via multiple linear regression analysis.Variables and interaction items were mean centered to avoid potential multicollinearity.The values of variance inflation factor (VIF) were lower than 1.416 (see Table 3), much lower than the threshold of 10 suggested by Cohen et al. [92], indicating that multicollinearity was not a problem in regression models.As Model 1 in Table 3 shows, both absorptive and reconfiguring capability are significantly positively related to EMS.Model 2 shows that absorptive capability and reconfiguring capability are positively and statistically significantly related to eco-process innovation, supporting H1a and H1b.The F value of Model 5 is not statistically significant, showing that H2a and H2b are not supported.Model 3 and Model 6 show that EMS has statistically significant relationship with eco-process innovation and eco-product innovation; thus, H3a and H3b are supported.Model 4 shows that when dynamic capabilities and EMS enter simultaneously, the relationship between dynamic capabilities and eco-process innovation becomes insignificant, while the impact of EMS on eco-process innovation remains statistically significant.Model 7 shows that reconfiguring capability and EMS have a statistically significant relationship with eco-product innovation, while absorptive capability does not have a significant impact.Thus, EMS fully mediates the relationship between dynamic capabilities and eco-innovation [93].
We implemented the bootstrap resampling method recommended by Preacher and Hayes [94] using the PROCESS macro (model 4) to further test indirect effects.The indirect effects are statistically significant when the 95% bootstrap confidence intervals (CI) do not include zero.As Table 4 shows, the mediation role of EMS between the two types of dynamic capabilities and eco-process/eco-product innovation is supported.Moderation effects were also tested utilizing the bootstrap resampling method and using the PROCESS macro (model 8).As Model 8 of Table 5 shows, the interaction between absorptive capability and stakeholder pressure is positively and significantly related to EMS, as is the interaction of reconfiguring capability and stakeholder pressure.Furthermore, results of conditional effects in Table 6 show that the direct effects of absorptive/reconfiguring capability on EMS at a low level of stakeholder pressure are insignificant, while the effects are significant at medium and high levels of stakeholder pressure, supporting H6a and H6b.Model 9 shows that the interaction between absorptive capability and stakeholder pressure is positively and significantly related to eco-process innovation.Moreover, the direct effect of absorptive capability on eco-process innovation at a low level of stakeholder pressure is insignificant, while the effects are significant at medium and high levels of stakeholder pressure.Therefore, H7a is supported.However, the interaction of reconfiguring capability and stakeholder pressure is not significantly related to eco-process innovation; thus, H7b is not supported.
Model 10 shows that both of the interaction terms are statistically insignificant; therefore, H8a and H8b are not supported.
The results of conditional indirect effects show that the indirect effect of absorptive capability on eco-process innovation via EMS is insignificant at a low level of stakeholder pressure, while the indirect effects are significant at medium and high levels of stakeholder pressure.Furthermore, the index of moderated mediation is significantly different from zero [95]; therefore, H9a is supported.H9b, H10a, and H10b are not supported.
As Figure 4 shows, stakeholder pressure positively moderates the relationship between absorptive/reconfiguring capability and EMS.The effect of absorptive/reconfiguring capabilities on EMS is stronger when firms perceive greater stakeholder pressure.Figure 5 shows that stakeholder pressure positively moderates the relationship between absorptive capability and eco-process innovation.The effect of absorptive capability on eco-process innovation is stronger when firms perceive greater stakeholder pressure.As Figure 4 shows, stakeholder pressure positively moderates the relationship between absorptive/reconfiguring capability and EMS.The effect of absorptive/reconfiguring capabilities on EMS is stronger when firms perceive greater stakeholder pressure.Figure 5 shows that stakeholder pressure positively moderates the relationship between absorptive capability and eco-process innovation.The effect of absorptive capability on ecoprocess innovation is stronger when firms perceive greater stakeholder pressure.

Discussion and Contributions
This paper aims to explore the mechanisms controlling how firm capabilities and perceived stakeholder pressure interact to facilitate eco-innovation.Based on questionnaire survey of 169 Chinese manufacturing companies, we explore the relationship between dynamic capabilities and eco-innovation, the mediation role of EMS, and the moderation role of stakeholder pressure.
Firstly, the impact of dynamic capabilities on eco-innovation is confirmed, which is in accordance with Ellonen, Wikström and Jantunen [96], who discuss the relationship between dynamic capabilities and innovation, and Verona and Ravasi [22], who indicate that knowledge absorption and reconfiguring are essential for continuous innovation.Dynamic capabilities facilitate eco-innovation through acquiring and developing necessary resources as well as allowing riskier investments.Firms with stronger dynamic capabilities will be more flexible, and more capable of addressing opportunities and challenges in the ever-changing environment.
Secondly, higher level of EMS implementation facilitates eco-innovation, which is in line with Rennings et al. [52], who reveal that maturity of EMS is positively related to ecoprocess innovation.EMS as a manifestation of environmental management capabilities facilitates information flows and reduces information asymmetry [17], and acts as a coor-

Discussion and Contributions
This paper aims to explore the mechanisms controlling how firm capabilities and perceived stakeholder pressure interact to facilitate eco-innovation.Based on questionnaire survey of 169 Chinese manufacturing companies, we explore the relationship between dynamic capabilities and eco-innovation, the mediation role of EMS, and the moderation role of stakeholder pressure.
Firstly, the impact of dynamic capabilities on eco-innovation is confirmed, which is in accordance with Ellonen, Wikström and Jantunen [96], who discuss the relationship between dynamic capabilities and innovation, and Verona and Ravasi [22], who indicate that knowledge absorption and reconfiguring are essential for continuous innovation.Dynamic capabilities facilitate eco-innovation through acquiring and developing necessary resources as well as allowing riskier investments.Firms with stronger dynamic capabilities will be more flexible, and more capable of addressing opportunities and challenges in the ever-changing environment.
Secondly, higher level of EMS implementation facilitates eco-innovation, which is in line with Rennings et al. [52], who reveal that maturity of EMS is positively related to eco-process innovation.EMS as a manifestation of environmental management capabilities facilitates information flows and reduces information asymmetry [17], and acts as a coordination mechanism for internal activities [54].
Thirdly, EMS fully mediates the relationship between dynamic capabilities (including absorptive and reconfiguring capabilities) and eco-innovation (including eco-process and eco-product innovation).Thus, dynamic capabilities are indirectly related to ecoinnovation via EMS.These results are in line with environmental scholars, such as Demirel and Kesidou [6], who argue that capabilities specific to the environmental domain, instead of capabilities with general purposes, are more likely to facilitate eco-innovation.The findings also correspond with the dynamic capabilities perspective, which contends that "the higher the dynamic capabilities a firm demonstrates, the more likely it is to build particular capabilities over time" [28] (p.41).Thus, dynamic capabilities as higher-order capabilities are essential for firms to develop and reconfigure lower-order environmental capabilities, which, in turn, facilitate eco-innovation.Firms demonstrating stronger absorptive capability are more able to identify, assimilate, and exploit new external information [31], and are more adept in learning from various partners, sharing information, exhibiting long-term resources commitment, and developing abundant knowledge and skills for new technology [28].Reconfiguring capability involves sensing the need to modify and recombine capabilities in response to environmental changes [97], thus reducing organizational inertia [98] and establishing new capability configurations [99] that facilitate eco-innovation.
Fourthly, the relationship between absorptive/reconfiguring capability and EMS, as well as the direct and indirect effects between absorptive capability and eco-process innovation, are contingent upon perceived stakeholder pressure.The effects become stronger when managers perceive higher stakeholder pressure.Thus, firms that perceive stronger stakeholder pressure will be more concerned with environmental issues and undertake more environmental activities.They will endeavor to absorb and reconfigure resources toward the goal of improving environmental management capabilities and developing eco-innovation, meaning that they can fulfill stakeholder expectations on sustainable development.By doing so, firms are able to benefit from achieving differentiation, gaining competitive advantages, improving reputation of being environmentally responsible, and reducing risks.Therefore, efforts should be made not only on capability building, but also on understanding stakeholder expectations and facilitating stakeholder engagement.
However, the moderation effect of stakeholder pressure between reconfiguring capability and eco-process innovation is not confirmed.This result may be due to the different characteristics of the two types of dynamic capabilities.Absorptive capability is outwardlooking, constantly searching for new opportunities and reliant on interaction between external and internal actors.In contrast, reconfiguring is inward-looking [32] and, therefore, could be less sensitive to stakeholder pressure.Moreover, the direct and indirect effects between dynamic capabilities and eco-product innovation are not moderated through stakeholder pressure.The reason for this finding may be that eco-product innovation is generally riskier and requires larger investments than eco-process innovation.Green products are often more expensive than less environmentally friendly competitors.Although environmental concern among stakeholders keeps growing, it is probably still not strong enough to promote development and purchasing of green products in the Chinese context.This paper makes three major contributions to the current literature.Firstly, this paper provides empirical evidence that contributes to the ongoing debate around whether implementation of EMS facilitates eco-innovation.It is confirmed that substantive implementation of EMS drives eco-process and eco-product innovation.Although superficial implementation of EMS is criticized as being symbolic, legitimacy-seeking, and "greenwashing" [47], a higher level of EMS implementation is beneficial for firms to overcome information asymmetry and identify potential benefits of eco-innovation [11,17], as well as coordinate environmental activities to accomplish eco-innovation [50,54].
Secondly, this paper contributes to the resource-based view and environmental literature through specifying how firms deal with environmental problems from the perspective of internal capabilities, exploring the mechanism of how different capabilities influence eco-innovation.Prior studies about the effects of internal capabilities on eco-innovation are not only limited, but also mostly have a static view [13].From a dynamic perspective, this paper demonstrates the influence of dynamic capabilities on eco-innovation through EMS.
Thirdly, to our best knowledge, this is the first empirical study that explores how dynamic capabilities and perceived stakeholder pressure interact to stimulate eco-innovation.Extant research on drivers of eco-innovation is short of studies that combine different perspectives and elucidate the relationship between different factors.The RBV and dynamic capabilities perspective emphasizes the importance of gaining and maintaining important resources and capabilities [9,16], while the stakeholder theory considers the influence of various stakeholders on firm activities and strategy [68].Combining these perspectives, this study sheds some light on how dynamic capabilities influence environmental capabilities and eco-innovation through demonstrating that stakeholder pressure plays a moderation role between them.

Practical Implications
This paper has several implications for firm managers and public policy makers.For firm managers, as environmental problems are drawing ever-more attention from stakeholders, eco-innovation ahead of competitors can bring not only improved efficiency and reduced cost, but also competitive advantages and a better reputation.The findings suggest that in order to sense and seize environmental opportunities, it is essential to nurture dynamic capabilities as well as environmental management capabilities through increasing investment, promoting employee training, facilitating cross-functional collaboration, and cooperating with various organizations, including firms, universities and research institutions.It is beneficial for firms to proactively identify environmental opportunities and implement environmental activities to profit from eco-innovation, including economic profits and social benefits gained through reducing energy and material consumption, cutting emissions, lowering costs, recycling, seizing market opportunities for green products, and achieving differentiation.Furthermore, since stakeholder pressure incurs both challenges and opportunities, managers should enhance communication with stakeholders to better understand their expectations and deliver information about firms' commitment to sustainable development.
For policy makers, this paper suggests that besides environmental regulations that control firm behaviors, it is also important to support voluntary environmental activities and encourage firms to comprehensively implement EMS.Public policy could promote firms' resource building and reconfiguring, which improves environmental capabilities through providing subsidies, building innovative platforms, and facilitating coordination between organizations.

Limitations and Future Research
Several limitations of this paper should be acknowledged.Firstly, this paper adopts cross-sectional survey data, which is a widely accepted approach in business studies.However, since firms are constantly developing and transforming capabilities in response to environmental changes, future research could adopt a longitudinal dataset to provide further insights into evolution of capabilities and their impact on eco-innovation.
Secondly, the sample is limited to manufacturing firms in eastern China.Future research could analyze how firms' capabilities affect eco-innovation in different regional and national contexts.

Hypothesis 9 (
H9). Stakeholder pressure moderates the indirect effect of dynamic capabilities (a.absorptive capability, b. reconfiguring capability) on eco-process innovation via EMS.The greater the stakeholder pressure, the stronger the indirect effect.Hypothesis 10 (H10).Stakeholder pressure moderates the indirect effect of dynamic capabilities (a.absorptive capability, b. reconfiguring capability) on eco-product innovation via EMS.The greater the stakeholder pressure, the stronger the indirect effect.

Hypothesis 9 (
H9). Stakeholder pressure moderates the indirect effect of dynamic capabilities (a.absorptive capability, b. reconfiguring capability) on eco-process innovation via EMS.The greater the stakeholder pressure, the stronger the indirect effect.Hypothesis 10 (H10).Stakeholder pressure moderates the indirect effect of dynamic capabilities (a.absorptive capability, b. reconfiguring capability) on eco-product innovation via EMS.The greater the stakeholder pressure, the stronger the indirect effect.

Figure 2 .
Figure 2. Size and age of sample firms.(a) Number of employees (b) Firm age (years).Figure 2. Size and age of sample firms.(a) Number of employees (b) Firm age (years).

Figure 2 . 23 Figure 3 .
Figure 2. Size and age of sample firms.(a) Number of employees (b) Firm age (years).Figure 2. Size and age of sample firms.(a) Number of employees (b) Firm age (years).Sustainability 2023, 15, x FOR PEER REVIEW 12 of 23

Figure 4 .
Figure 4. Stakeholder pressure as a moderator of effect of dynamic capabilities on EMS.(a) Absorptive capability on EMS; (b) Reconfiguring capability on EMS.Figure 4. Stakeholder pressure as a moderator of effect of dynamic capabilities on EMS.(a) Absorptive capability on EMS; (b) Reconfiguring capability on EMS.

Figure 4 . 23 Figure 5 .
Figure 4. Stakeholder pressure as a moderator of effect of dynamic capabilities on EMS.(a) Absorptive capability on EMS; (b) Reconfiguring capability on EMS.Figure 4. Stakeholder pressure as a moderator of effect of dynamic capabilities on EMS.(a) Absorptive capability on EMS; (b) Reconfiguring capability on EMS.Sustainability 2023, 15, x FOR PEER REVIEW 16 of 23

Figure 5 .
Figure 5. Stakeholder pressure as a moderator of effect of absorptive capability on eco-process innovation.

Table 1 .
Means, standard deviations (SD), square roots of AVE, and Pearson correlation coefficients.

Table 2 .
Results of Cronbach's α, CR, and AVE.(Item details in Appendix A).

Table 3 .
Results of regression analysis.

Table 4 .
Results of indirect effect.

Table 5 .
Results of moderation effects.

Table 6 .
Conditional effects and index of moderated mediation.