Driving sectoral sustainability via the diffusion of organizational eco-innovations

Using insights from institutional literature, the resource-based theory of the firm, and internationalization, we explain variations in the diffusion of organizational eco-innovations. Studies have previously reported that the drivers of eco-innovation are regulatory pressures, technology push, market pull, and firm factors. But relatively little attention has been paid to nontechnological forms of eco-innovation, such as environmental management systems (EMS). Consequently, how exactly to encourage EMS adoption across sectors is still unclear. We attempt to address this question by combining sectoral panel data (2009–2014) from a number of sources in Spain. The econometric analysis reveals that environmental policy is driving the adoption of ISO 14001 largely due to differences across sectors in energy and pollution intensity. In addition, the adoption of ISO 9001 increases the use of ISO 14001 in industry because of complementarities between the two systems. Third, in highly internationalized sectors, firms adopt a greater amount of ISO14001.


INTRODUCTION
How organizations manage their environmental performance has become a strategic issue for many companies, and this reflects the extent to which the environment is now viewed as a valuable asset. Managers today are expected not only to reduce lead times, improve quality, reduce costs, and enhance flexibility, but also to be more environmenta lly responsible (Montabon, Melnyk, Sroufe, & Calantone, 2000). During the past ten years an increase in research on eco-innovation has shed light on ways in which companies could progress along the path towards environmental sustainability (e.g. Horbach, Rammer, & Rennings, 2012;Kesidou & Demirel, 2012;Marzucchi & Montresor, 2017; Zubeltzu-Jaka, Erauskin-Tolosa, & Heras-Saizarbitoria, 2018). Yet less attention has been given to different types of eco-innovation (Triguero, Moreno-Mondéjar, & Davia, 2013). For instance, eco-innovations might take the form of eco-products or ecoprocesses (Triguero et al., 2013). There could be technological eco-innovations or nontechnological types of eco-innovation (Demirel & Kesidou, 2019). This paper focuses on nontechnological eco-innovations. Organizational eco-innovatio ns are vehicles of corporate environmental self-regulation that facilitate the introduction of significantly different organizational structures, corporate environmental strategies, and new management methods (OECD, 2009;Ziegler & Nogareda, 2009). Well-known examples of organizational eco-innovations include environmental management systems (EMSs), which, since the 1990s, have become standard environmental protection policy tools due to their greater flexibility and costs that are lower than those of traditio na l regulatory tools (Demirel, Iatridis, & Kesidou, 2018;Frondel, Horbach, & Rennings, 2008).
Previously EMSs were usually seen as drivers of environmental investment and effort.
Over the past twenty years, environmental management has moved from being simply a matter of "command and control" to increased awareness of the need for self-regula tio n and an acceptance of public accountability and consultation . Some companies are actually now using environmental issues to strengthen their market position and access new markets. In recent decades, voluntary proactive approaches to environmental protection have been accepted as proper additions to traditional mandatory command-and-control regulation and economic incentives (Khanna & Damon, 1999).
However, some criticism has been directed at these practices on the grounds that proactive actions of this kind merely represent symbolic decisions (Aravind & Christmann, 2011;Gavronski, Ferrer, & Paiva, 2008).
The purpose of this research is to analyse the reasons underlying differences in sustainability among sectors by taking note of their respective levels of certified forms of organizational eco-innovation (i.e. ISO 14001). Enhancing sustainability may require a systemic approach at the sectoral level. For instance, corporate social responsibility (CSR) tools, which are adopted by many companies, have been criticized as not comprehensive enough to bring about transformative change. Instead, sectoral approaches, such as the Sustainable Apparel Coalition or the Sustainability Accounting Standards Board, are increasingly seen as the key to driving sustainability (Hohnen, 2013), which they do by applying organizational eco-innovations across a whole sector and along the supply chain. Using insights from institutional theory, resource-based theory, and the internationalization literature, we attempt to explain variations in adopting organizational eco-innovations. First, we contend that coercive institutional pressures, such as government mandates (DiMaggio & Powell, 1983;Oliver, 1997), are driving the adoption of EMSs. Considering that sectors differ in the intensity of their energy use and levels of pollution, the pressures to comply with regulations across sectors also differ (Agnolucci & Arvanitopoulos, 2019;Cole, Elliott, & Shimamoto, 2005). Second, we argue that EMSs are more readily adopted in sectors that have already impleme nted quality management systems such as ISO 9001. This can be explained in terms of the benefits that arise from resource complementarities (Teece, 1986), so that the adoption of ISO 9001, for example, allows organizations to develop intangible processes and routines that facilitate the adoption of eco-innovations such as EMS (Darnall & Edwards, 2006). Third, this paper postulates that organizational eco-innovation is driven by differe nces across sectors concerning their need to signal to international markets that they do actually abide by the required safety and environmental standards (Christmann & Taylor, 2006;Yeung & Mok, 2005).
We gather together data from the following Spanish sources for the period 2009-2014: the International Organization for Standardization (ISO) survey, the Statistics on R&D activities, the Technological Innovation Panel, the Industrial Companies Survey, the Environmental Protection Survey, and the Environmental Tax Account. By using data from these multiple sources we can include information on a broad range of factors that may be driving the adoption of EMS across all industrial sectors. We also gain the advantages of using panel data for our econometric estimations.
The paper is organized as follows. First, we review the literature on the drivers of organizational eco-innovation across sectors and develop the conceptual framework of this research. Next, we present the empirical framework and the panel dataset. Finally, we discuss the results of the analysis and underline the main contributions made by our study.

Organizational eco-innovations: Certified environmental management systems
An EMS is a set of processes and practices that make it possible for an organization to reduce its environmental impact and increase the efficiency of its operations. It is a form of organizational eco-innovation as it allows an organization to achieve its environme nta l goals through the consistent monitoring, evaluation, and improvement of its environmental footprint. An EMS itself does not prescribe a level of environme nta l performance that has to be achieved; rather, each organization's EMS is tailored to its own individual objectives and targets. Because of this, an EMS helps an organiza tio n meet regulatory demands in a systematic and effective way. This kind of proactive approach can be useful in reducing the risk of noncompliance and improve health and safety practices for employees and external stakeholders alike.
Many firms have now chosen to introduce some form of organizational eco-innovation to address their environmental concerns, and some firms also choose to be certified according to recognized international standards. This paper focuses on ISO 14001 as a certified form of organization eco-innovation. 3 As Testa, Rizzi, Daddi, Gusmerotti, Frey, and Iraldo (2014)  In general, firms are under no legal obligation to obtain ISO 14001or any type of environmental certification, for that matterfrom either the ISO or another EMScertifying organization. It may be for this reason that many studies to date have examined the effectiveness of adopting an EMS for a firm's environmental performance (e.g., Montabon et al., 2000;Montobbio & Solito, 2017;Testa et al., 2014), and an increasing number of studies ask why firms participate in voluntary environmental programmes (Arora & Cason, 1995;Blackman, 2007;Quazi, Yee-Koon, Chin-Meng, & Poh-Seng, 2001), and why firms opt to certify their EMSs under various voluntary certifica tio n schemes (Klassen & McLaughlin, 1996;Peiró-Signes, Segarra-Oña, Verma, Mondéjar-Jiménez, & Vargas-Vargas, 2014).
In sum, this paper contends that sectors can become more sustainable via the diffusion of certified forms of organizational eco-innovation. By contrast, prior literature on ecoinnovation has paid more attention to technological forms of eco-innovation (Horbach, Rammer, Rennings, 2012, Kesidou & Demirel, 2012: Marzucchi & Montresor, 2017. However, the drivers as well as the impact of technological versus organizational ecoinnovations may be different (Triguero et al., 2013).

Institutional perspectives of eco-certification: Environmental policy
Institutional theory holds that organizations often behave in a similar way (i.e., are isomorphic) when functioning within similar social structures (Scott, 2001).
Organizations abide by prevailing institutionalized norms, values, and assumptio ns (Meyer & Rowan, 1977) because of the risk that if they do not, they will lose legitimac y (DiMaggio & Powell, 1983). This theory claims that institutional compliance may be as important as, or even more important than, motives of profit maximization (Scott, 2001), since a lack of legitimacy may threaten an organization's survival (Suchman, 1995).
Research on institutional theory contends that institutional compliance may not only reduce potential threats to the organization (e.g., sanctions, fines, and boycotts) but may even improve organizational performance (Deegan & Rankin, 1996;Oliver, 1997).
It is necessary for organizations to respond to a range of institutional pressures, be they coercive, mimetic or normative (DiMaggio & Powell, 1983;Oliver, 1997). Coercive isomorphism includes government mandates or contractual obligations with which an organization is forced to comply; mimetic isomorphism occurs when organizatio ns imitate the structures or practices of successful organizations in their sector; and normative isomorphism is driven by the prevalence of professional standards within a sector.
Previous studies of EMSs, and particularly eco-certification, use institutional theory to explain how EMSs and certified forms of EMS are used by firms to fulfil their responsibilities to their stakeholders, thus sustaining and enhancing their legitimac y (Delmas, 2002;Demirel et al., 2018;González-Benito & González-Benito, 2005;Jiang & Bansal, 2003;Khanna & Anton, 2002). However, the fact that coercive, mimetic, and normative pressures might affect firms in different sectors in different ways has received less attention.
The focus of this paper is on coercive pressures. Coercive isomorphism might be apparent at the sectoral level, as sectors differ with respect to their energy and pollution intens it y, and consequently pollution regulations and taxes could be more or less stringent in specific sectors (Agnolucci & Arvanitopoulos, 2019;Cole et al., 2005). This would, in turn, affect compliance with, and adoption of, eco-certification. This allows us to formulate our first hypothesis:

H1: Firms operating in sectors with stricter regulatory pressures exhibit a greater
likelihood of eco-certification.

Resource complementarity perspectives on eco-certification
Previous research on variations across countries in the adoption of eco-certification shows that the implementation of ISO 9001 facilitates the diffusion of ISO 14001 (Corbett & Kirsch, 2001;Vastag, 2004). Research at the firm level also indicates that firms with quality-based management systems are more likely to adopt EMSs because the costs are lower (Darnall & Edwards, 2006).
The joint adoption of ISO 9001 and ISO 14001 can be explained theoretically by drawing on insights from the resource-based theory of the firm (Barney, 1991) and the literature on asset complementarity (Ozusaglam, Kesidou, & Wong, 2018;Teece, 1986). When firms implement quality management systems (e.g. ISO 9001), they develop intangib le and knowledge-based processes that can make the adoption of eco-certification easier (Demirel & Kesidou, 2019;Darnall & Edwards, 2006). This is so because firms that adopt ISO 9001 have already built tacit capabilities and routines across their organizations and have embraced a systematic process of organizational change based on monitor ing, assessment, and action that is very similar to that needed for ISO 14001. For instance, Zhu, Cordeiro, and Sarkis (2013) emphasize the path-dependent character of organizational learning and show that Chinese firms that have prior experience with ISO 90001 are more prone to also adopting ISO 14001.
To sum up, based on the above insights, it would be expected that sectors characterized by high levels of adoption of ISO 9001a highly institutionalized quality manage me nt systemalso present high rates of adoption of ISO 14001.

H2: Sectors with high rates of ISO 9001 adoption exhibit a greater likelihood of eco-
certification.

Eco-certification and internationalization
Eco-certification may also be motivated by a need to signal to international capital and export markets that a firm does in fact abide by required safety and environme nta l standards (Christmann & Taylor, 2006;Yeung & Mok, 2005). Pressures from powerful global suppliers or multinational corporations may push firms to adopt eco-certificatio ns, especially if they want to be integrated into global supply chain networks (King, Lenox, & Terlaak, 2005;Withers and Ebrahimpour, 2000). Both these factors imply that sectors that are export intensive may be more likely to meet to eco-certification requirements so as to access foreign markets (Bodas Freitas & Iizuka, 2012).
Internationalization might affect eco-innovations via different mechanisms. For instance, a recent study by Chiarvesio, De Marchi, and Di Maria (2015) explores, in the context of Italy, three channels via which internationalization could drive eco-innovation, namely, outsourcing, exporting, and being part of a multinational corporation. The role of multinational corporations in reinforcing eco-innovation is also stressed in Cainelli, Mazzanti, and Montresor's (2012) study. In this paper we focus on one channel of internationalization, i.e. exporting.

H3: Highly internationalized sectors exhibit higher rates of eco-certification.
[Insert Figure 1 around here]

EMPIRICAL FRAMEWORK
Previous empirical studies describing the determinants of the adoption of environme nta l certification often mention the characteristics of environmental practices or the impacts of these practices on a firm's strategy and performance. Quazi et al. (2001), for example, propose a model, tested on a small sample of firms in a number of specific industries, that can predict the intentions or motives of a company in seeking ISO 14001 certificatio n.
Recent empirical research examining the drivers of environmental certification has increasingly attempted to include the factors that, according to different theoretical approachesmost notably institutional theory and the resource-based view of the firmexplain the motivations and drivers that encourage firms to adopt these environme nta l practices (Cole, Elliot, & Shimamoto, 2006;Neugebauer, 2012;Nishitani, 2009;Singh et al., 2015). However, the empirical approach is usually based on cross-sectional survey data for certain industries (González-Benito and González-Benito, 2005). This study differs in that we use panel data for 16 manufacturing sectors for the period 2009-2014.
Sectors differ in their energy use and pollution intensity (Agnolucci & Arvanitopoulos, 2019;Cole et al., 2005), and therefore, we argue in this paper, their approaches to sustainability may be different.

Data and variables
Data availability is a common constraint of empirical analyses in environme nta l economics. For this reason, information must be collected from various databases in order to examine the drivers of environmental certification. In this paper we depart from the database constructed by Costa-Campi, García-Quevedo, and Martínez-Ros (2017) Table 1.
[Insert Table 1 around here]

Dependent variable
The main agency for environmental certification is the ISO. To measure environme nta l certification, we use ownership of approved ISO 14001 certification, which, as Kesidou and Demirel (2012) and Testa et al. (2014) confirm, is one of the most usual forms of EMS. Information about ISO 14001 accreditation in the Spanish manufacturing sector is provided directly by the ISO, but has only been available since 2009. This information is now published every year on the ISO website. All firms that wish to set up an EMS can use the ISO 14001 to certify their processes. The ISO website also contains informa tio n about other certificates, including ISO 9001 (quality management) and ISO 5001 (energy management systems). In addition, and as a robustness test, we also use information from EMAS. The European Commission maintains the original source of this data.

Independent variables
The Spanish Institute of Statistics (INE) carries out various surveys that provide the information used for the independent variables. They include the following: the Industria l Companies Survey, an annual report covering the main features of both firms and sectors (e.g., the number of firms in each sector, the number of employees, and sales and export figures); the Survey on Industry Expenditure on Environmental Protection, which reports on spending by firms on environmental protection in a given industry, distinguis hing between current expenditure and investment (divided between "end-of-pipe" solutio ns and integrated equipment); the Environmental Tax Account, from which information is taken about the pollution taxes paid by each industrial sector; and the Statistics on R&D Activities, which provide information about environmental research and development.
All these surveys are anonymous and mandatory for the firms involved. They all use the same classification of economic activities. INE publishes the information at industr ylevel and guarantees its representativeness. In addition to these surveys, we have used the Technological Innovation Panel, also compiled by the INE, which provides informa tio n of sales at firm level, to calculate the Herfindahl-Hirschman index (HHI) of concentratio n for the industries of our analysis. In Table 2, we present the descriptive statistics. 4 [Insert Table 2 around here] Table 3, which shows the percentage of firms with an EMS, underlines the fact that the number of firms with an ISO 14001 is much higher than those with EMAS registratio n.

An initial appraisal of Spain's EMSs by sector
Second, and more importantly with regard to the objectives of this analysis, there are considerable differences between different industries in their use of EMS, and especially in the case of ISO 14001.
[Insert Table 3] Coke and refined petroleum products, chemicals, pharmaceutical products, and motor vehicles are the four sectors in which the implementation of ISO 14001 is most common.
ISO 14001 has been adopted by more than 15% of firms in these four sectors, the highest proportion being in the coke and refined petroleum industry. This sector contains very few firms, all of which have adopted ISO 14001.
The main reasons for the greater adoption of this certificate in these sectors are related to the characteristics of each sector. First, these industries display a high level of environmental commitment. Many firms participate in multilateral agreements, particularly in the chemical industry. Second, the environmental impacts of all four sectors are governed by common and extremely stringent regulations.
Third, these industries are highly internationalized, with regard not only to sales to foreign markets, but also to the share of foreign capital in their ownership. Quality manage me nt systems are also prevalent in these industries: that is, in 100.0% of firms in the coke and petroleum sector and 60.7% of firms in the pharmaceutical sector. At the same time 38.0% and 36.8% of firms respectively in the chemicals and motor vehicles sectors have adopted ISO 9001. These rates are well above the average reported for manufacturing industr ies.
Already having quality management systems in place makes the adoption of an EMS easier, primarily because of the knowledge acquired, but also because these manage me nt systems require a higher degree of integration.
Close behind these four sectors, the computers, electronics, optical products, and electrical equipment sector (14.5%) and the machinery and equipment sector (11.7%) are two more sectors in which there is a considerable degree of adoption of ISO 14001. These two industries share many of the characteristics of the other four industries, especially as regards their degree of internationalization and the earlier adoption of quality management systems.
Finally, there are sectors with a very low degree of adoption of ISO 14001. These are textiles and wearing apparel; leather products; furniture and other manufactur ing activities; and wood, and wood and cork products. In these sectors, fewer firms have adopted an EMS. These industries are also the four sectors that present the lowest degree of adoption of ISO 9001. They are characterized by a particularly high number of small low-technology-content firms.

The empirical specification
We test the three hypotheses empirically via the following model specification: ECit = β0 + β1Institutionalit + β2Complementaryit + β3Internationalizationit + β4 Zit + µi + eit has also shown that government policy is an important factor. Therefore we include taxes with environmental objectives, pollution, and resources as a potential factor that may explain environmental certification.
Complementary refers to a sector's experience in obtaining other types of certificatio n.
This includes experience in obtaining ISO-approved certificates, which may be an important driver thanks to the learning process already undergone in applying for previous certification from this international organization or others. The total number of ISO 9001 certifications in each sector, certifying quality management systems, is specifically included.
Internationalization refers to a sector's capacity to sell on international markets. This driver is measured using a variable that captures the destination of exports: the European environmental commitment at the firm level and sectoral level (Bodas Freitas & Iizuka, 2012).
A set of control variables, Z, is also included in this specification as control. These are drivers identified in the literature as being determinants of environmental certification at the industry level. First, we use the volume of sales to control for demand. Second, we include firm characteristicsalbeit at the industry levelthat may be drivers of certification, such as the average firm size. We also consider the role of management as a driver of environmental certification, and therefore include three types of environme nta l investment and pollution prevention measures. These variables -R&D expenditure for environmental purposes, investment in end-of-pipe solutions, and investment in the production processcapture the environmental strategies firms develop in relation to environmental certification. End-of-pipe investment refers to technological solutions that firms incorporate into existing manufacturing processes but are not essential parts of them, while investments in production processes correspond to new, or substantia lly modified, production facilities that represent an integral part of the production process aimed at reducing pollution (Demirel & Kesidou, 2011).
Finally, we consider random time-invariant characteristics µi and time-effect dummies to control for macroeconomic effects (e.g. the business cycle) common to all industries.  (2013)  Our results regarding the control variables show that firm size is a major driver of ISO 14001 adoption. As the literature in this field emphasizes, a minimum size threshold has to be reached in order to be able to implement an EMS. In contrast, most of the other control variables were found not to be significant.

RESULTS AND DISCUSSION
[ Table 4 around here] To check the robustness of our results we have carried out some complementar y Finally, we have used random models to carry out our estimations. Some of the independent variables of our model show very little variation over time and, in this situation, fixed effects models do not perform well and can lead to imprecise estimates.
Nevertheless, there is still the problem with the random effects model that some of the regressors may be correlated with the error term. To deal with that we have included in the robustness section an estimation using the endogeneity-robust approach developed by Mundlak (1978). Mundlak (1978) suggested that one way to overcome endogeneity is by including the means of the repressors in the equation. Again, the results remain unchanged and confirm our hypotheses.
[Insert Table 5 around here]

CONCLUSIONS
The aim of this analysis has been to contribute to the literature on eco-innovation by explaining the drivers of organizational eco-innovations across sectors using ISO 14001 certification, one of the main reference standards for EMSs. Prior theoretical and empirical studies have shown the need to take a broad range of firm characteristics and motivations into accountinternal as well as externalwhen examining the factors that drive the diffusion of ISO 14001.
We have constructed a panel database with information from several industry sources concerning ISO certification and innovation and including details about prevailing economic and environmental characteristics in order to examine these determinants in Spain. Although the limitations of using industry-level data as opposed to firm-level data are well known, this has enabled us to include a broad range of variables that the literature identifies as potential drivers of environmental certification. We have been able to include all industrial sectors in the estimations with this information, whereas most empirica l studies limit themselves to either one or only a few sectors.

Complementarity benefits arising from the adoption of two practices
International markets may pressure firms to consider their enviromental impact Environmental policy may enforce stricter sectoral regulations for firms to abide by