p ー エッ ーイ ウ エィ ョョゥカ イウ ッョヲ イ ョ c ッー ョィ ァ ョ オョ e ャ ッーエ エ

Transcription

1 p ー エッ ーイ ウ ョョゥカ イウ ッョヲ イ ョ c ッー ョィ ァ ョ L@j e ャ ッーエ ウエ ョ イ ウ Z@r イッャ ィ イ エ イゥウエゥ ー ゥャゥエゥ ウ ウ イ コオウ ァャ ュ l ョゥカ オウゥョ ィッッャ c ー イ ィ ウ イ ッコオウ ァャ ュ ` ァュ ゥャ N ッュ s エ ーィ ッ ゥョ u ョゥカ ョエィ ッイ ッョョ s ッイ ョ ァ ュ ョエウエ ーィ ョ N イッ ゥョ ` オョゥカ M ー イゥウ QN ヲイ c ッョァ l ョゥカ オウゥョ ィッッャ c ー イ ィ ウ イ N ケ N キッョァ ` ャ ウ N a ウエイ エ e ョカゥイッョュ ョエ ョ ァ ュ エ ョ イ ゥューッイエ カッャオョエ ュ ョ ァ ュ エィ イ オ ョカゥイッョュ ョエ ゥュー エ N@r カ エィゥ イ クー ヲゥョ ョ ゥ ャ ョ ゥョ イ ーイ ウウオイ イ ァオャ オエィッイゥエゥ ウ N@e カ エィッオァィ L@ エィ ュッエゥカ イ ウゥァョゥヲゥ ュ ッイァ ョゥウ エゥッョウ L@ ャゥエ イ ウッュ キィ ヲ ューゥイゥ ャ カゥ クーャ ウー ッイァ ョゥウ エゥッョ ー ゥャゥエゥ ウエイオ エオイ ィ イ エ イゥウエゥ ウゥァョゥヲゥ イゥカ イッー エ ュ ッョ ウ ッヲ f イ ョ ュ ョオヲ ウ ヲゥイュウ L@ ゥョカ ウエゥァ キィ ヲゥイュ ウエイオ エオイ ィ イ エ イゥウエゥ ョ ッイァ ョゥウ エゥッョ ー ゥャゥエゥ イ ャ エゥカ ョ O ャ ー イエゥ オャ イ L@ エ エィ ヲヲゥ ゥ ョ ウ ウ ャヲ M イ ァオャ ゥョ ヲゥイュ ッュー エゥエゥカ ョ ャ ーイッ オ エゥカゥエケ N エィ エ ィョッャッァゥ ッューャ ャ エィ オイッー ュ イォ エィ ィ カ クー イゥ ッエィ アオ ュ ョ ァ ュ ウエ ョ イ オウエッュ ウ イカゥ ャゥォ ョ ems@ イ ャ エゥカ イャゥ イ N@t ィョッャッァゥ ャ ッューャ ャ エィ ィ クー イゥ ッエィ アオ ャゥエケュ ョ ァ ュ ウエ ョ イ イ L@ ィッキ カ ャゥォ ウエ ョ イ ゥョ ャ エ ウエ ァ エィ エ ャ ッーエ エ ィ ィゥァィ ャ ーイッ オ ッュー ョッョ M ッーエ イ L@ キィゥ ュ ケ

2 Powered by TCPDF ( ゥョ ゥ エィ ュ ーイッカゥ キゥョ M ッーーッイエオョゥエケ N j ャ ッ ウ ZoSRLoSQ

3 Early and late adopters of ISO14001-type standards: Revisiting the role of firm characteristics and capabilities Abstract Environmental Management Standards (EMS) are important voluntary management tools that aim to reduce environmental impact. Reasons to adopt an EMS can vary from truly ethical reasons to expected financial benefits and to increasingly high pressures from the regulatory authorities. Even though, these motivations are significant to many EMS adopting organisations, the literature somewhat fails to provide sound empirical evidence to explain how and why some firm specific organisational capabilities and structural characteristics could be significant drivers to adopt an EMS. By using the Propensity Score Matching method on a sample of French manufacturing and service firms, we investigated whether a firm s structural characteristics and organisational capabilities promote relatively early and/or late adoption of an EMS. In particular, we tested the efficiency of the EMS as a self-regulatory tool to increase a firm s competitiveness through labour productivity. We found that technologically more complex large firms that are active in the European markets and that have existing experience in other quality management standards and customer services are more likely to adopt an EMS relatively earlier. Technologically less complex large firms that have experience only in other quality management standards are, however more likely to adopt the standards in a later stage. We also found that both early and late adopters tend to have higher labour productivity compared to non-adopter, which may indicate that adopting an EMS may provide a win-win opportunity.

4 Introduction The adoption of environmental management standards (EMS) such as ISO has been explained from broadly three perspectives. In terms of motives, EMS adoption may be driven by firm strategies (Bansal and Hunter, 2003), external factors such as regulations and citizen pressures (e.g., Nakamura et al., 2001; Anton et al., 2004; Porter and van der Linde 1995; Lim and Prakash, 2014) and the usual economic factors such as cost, finance and market requirements (Ayres and Walter 1991; Kirkpatrick and Pouliot 1996, Khanna and Zilberman 1997, Albertini 2013). Firm characteristics such as size, ownership structure and employee provide the second explanation (Harter and Homison, 1999; Nakamura et al., 2001; King and Lenox, 2001; Welch et al., 2002). The third perspective refers to firm capabilities such as complementary assets, export experience, ISO 9001 or TQM experience, R&D expenditures and technologies (Melnyk et al., 1999; Christmann 2000; Nakamura et al., 2001; Bowen 2002; Lenox and King 2004; Grolleau and Thomas, 2007; Grolleau et al. 2007a, b). As pointed out by Grolleau and Thomas (2007), due to conflicting findings, the roles of firm characteristics and capabilities such as firm size, export experience, ownership structure and R&D expenditures are still unclear. Furthermore, few studies distinguish the roles of firm characteristics and capabilities in case of early, late or non-adoption of an EMS,despite the potential first-mover advantages in cost, learning curve and profitability (Christmann, 2000; Bansal and Hunter, 2003; Grolleau and Thomas, 2007). The Mori and Welch s (2008) study of ISO adoptions in Japan show that the timing of the adoption might be affected by an organisation s resources and attitude toward environmental protection, indicating firm size and capability as possible factors. Early adoption of ISO 9001 is more a reinforcement strategy for firms with existing environmental legitimacy (or capability) with a strong international presence (Bansal and Hunter, 2003). These studies show that more insights can be created by comprehensively revisiting the roles of firm characteristics and capabilities in distinguishing early from late adopters. This research contributes to the environmental management literature by investigating whether a firm s structural characteristics and organisational capabilities promote relatively early and/or late adoption of voluntary EMS. The research revisits some of the popular factors (e.g., firm size, ISO 9001 experience) but also investigates a firm characteristic not previously examined (i.e., being part of a group) and several capabilities (i.e., labelling service, delivery service, market exposure, and R&D intensity). It provides insights to policy makers by highlighting crucial firm characteristics that may speed up the diffusion of

5 voluntary proactive approaches for environmental protection. It identifies capabilities that accelerate EMS adoption. To ascertain the benefits of being first-mover, the research also examines if early adoption leads to better labour productivity. A brief description of ISO EMS provides firms with a management structure and framework to minimise their environmental impact, ensure compliance with environmental regulation, and reduce wasteful uses of natural resources. Implementing an EMS such as ISO generally involves institutional change in the management and an internally motivated environmental selfregulation (Anton et al., 2004). ISO is part of ISO s broader environmental management standards family (ISO 1400). The number of ISO certified firms exceed 300,000 in year 2013 (iso.org). ISO prescribes principles for firms to form their own internal environmental management. They are considered efficient tools to support the market-based instruments (i.e. tradable permits, emission taxes, subsidies etc.). They do not impose environmental performance requirements (Khanna and Damon, 1999; Alberini and Segerson, 2002). There is no need to prescribe how to improving adopters business performance. However, they may help firms to introduce product improvements, to reduce their carbon foot-print and input cost, and increase employee, consumer, investor, shareholder and insurer trust by reflecting environmentally friendly image (ISO). This greener image may also provide substantial marketing advantages and help to capture additional market share, provide access to green markets, reduce insurance charges and input costs, which in turn may have a significant impact on a firm s business performance. The adoption of an ISO requires several steps to be taken and successful implementation of the standard depends on how well the Plan, Do, Check, Act (PDCA) model is applied. This feature of the standards ensure that organisations are responsible for setting their own environmental targets and make the standards applicable to wide variety of organisations. The standards covers the following components: 1) environmental policy, 2) performance goals: objectives and targets, 3) environmental programs: action plan to meet objectives, 4) roles and responsibilities, 5) training and awareness, 6) communication internal and external, 7) documentation of the system, including procedures or operational controls, 8) monitoring, measurement and record keeping, 9) procedures for corrective and preventive action, 10) Environmental Management System audits and 11) management

6 reviews (adapted from ISO 1 ). These steps make sure that the organisations identify and minimise the negative effect of their operations, comply with existing laws and continually improve in the above. Certification of ISO also involves an extensive third-party auditing process. In the absence of experienced and knowledgeable managers, firms may also have to hire additional staff to plan and monitor their environmental management process. Therefore, the expected time and cost for the implementation can vary considerably between organisations. ISO is often adopted voluntarily to reduce their environmental impact, facilitate sustainable development and foster international trade (Bansal and Hunter 2003; Simpson et al. 2012). Thus, its adoption can be relevant to wide variety of organisations. Some firms however, tend to be more proactive and adopt the standards earlier than their competitors because their managers may think that improving product and service quality through the standard adoption can provide substantial competitive advantages over their competitors. A recent study by Russo (2009) shows that early adoption of ISO and experience with the standard is associated with lower emissions. Bansal and Hunter (2003: 290) show that firms reinforcing their current strategy are more likely to look for the competitive advantage associated with being a first mover with ISO Empirical evidence So far, firm characteristics such as firm size, ownership structure and employee have been examined by the ISO adoption literature. Firm size is found dominantly positively related to ISO adoption (Melnyk et al., 1999; Nakamura et al., 2001; King and Lenox, 2001; Welch et al., 2002) but no significant link is reported by Harter and Homison (2001). Bowen (2002) shows that the significant link between firm size and ISO adoption disappeared after adding visibility and organizational slack into the regressions. A revisit of the role of firm size is thus required. Studies of ownership influence is rare. In a study private and public ownerships were found not related to ISO adoption. Foreign ownership matters among manufacturers in the USA (Melnyk et al., 1999) but it is not relevant to those in Japan (Nakamura et al., 2001). To our knowledge, there is no evidence suggesting that firms size and ownership structure differentiate early from late adopters. 1 Environmental Management Systems, Requirements with Guidance for Use, ISO 14001

7 There is an argument that the pre-existence of some specific complementary capabilities that help reduce both time and cost of implementation. Such capabilities allow firms to innovate and implement modification in production processes (Christmann, 2000). The existing literature has explored some such capabilities. For example, prior experience in ISO 9001 and TQM have been found positively related to the adoption of ISO (Menlyk et al., 1999; Harter and Homison, 2001; Nakrmura et al., 2001; King and Lenox, 2001). Experience in export matters in Japan (Nakamura et al., 2001) but not in the USA (Melnyk et al., 1999). R&D expenditures are thought to be positively related to ISO but negative result has been reported (Nakamura et al. 2001). Bansal and Hunter (2003) argue that firms without existing environmental legitimacy and capability are better off becoming followers and adopt EMS such as ISO later. In light of these arguments, it is thus interesting to find out of such capabilities differentiate early and late adopters. Firm Characteristics This research argues that the decision to adopt ISO standards earlier or later depends on specific firm characteristics such as firm size and being part of a group. Firm size is one of the most common firm characteristics being examined. Large firms attract increasing attention from customers and regulatory authorities as they are more likely to operate in highly polluting industries and feel more pressure for performance improvement (Simpson et al. 2012; Bansal and Roth 2000; Darnall 2006; Delmas and Montes-Sancho 2010; Murillo- Luna et al. 2008). Due to higher public visibility larger firms also experience significant pressure from stakeholders to demonstrate environmental performance improvement (Henriques and Sadorsky 1999; Brammer and Millington 2008) and accordingly, they also expect higher benefits from enhanced company image (Arora and Cason, 1995; 1996). Most empirical studies found that the probability of implementing environmental standards increases with firm size and large firms could lower the cost of adoption owing to scale economies and through learning (Arrora and Cason 1995; King and Lenox 2001; Nakamura et al. 2001; Delmas and Montiel, 2009; Grolleau et al. 2007a, b). Montiel and Husted (2009) argue that many large firms tend to be a part of smaller industry associations and those firms have more economic, intellectual, and political resources to adopt a voluntary EMS. They also argue that these firms may act as institutional entrepreneurs and not only certify early but also play an active role in promoting voluntary EMS. By doing so, they could virtually set a new entry barrier for the small-size new entrants, which could help larger

8 firms to consolidate their strategic position within the industry. However, one should not overstate the impact of a firm s size on adoption decision because many large firms may have created their own environmental standards. In either case, we still expect that larger firms have more incentives and available financial resources for EMS adoption. Accordingly, we propose that: Hypothesis 1: Large firms are more likely to adopt an EMS earlier compared to smaller firms. Previous research asserts that ownership structure is another important factor that shapes organizational strategy and institutional entrepreneurship (Mascarenhas, 1989; Gedajlovic, 1993; Darnall, 2003; Darnall and Edwards 2006). The ownership structure, in its basic sense, could indicate a firm s proximity to financial resources and therefore it can be crucial for adoption decision. Firms that are part of a larger group tend to have relatively easier access to a greater amount of financial resources to develop their internal environmental expertise (Bowen, 2002; Russo and Fouts 1997; Pekovic 2010; Zyglidopoulos 2002). By sharing financial and technical resources with parent company and the other sibling companies, group firms could also take advantage of the economies of scale and reduce the cost of adoption. However, we can argue that having an easier access to financial resources is only one side of the coin. One should also consider the importance of environmental competencies and organisational challenges pertinent to the adoption of an EMS (Christmann, 2000; Henriques and Sadorsky, 1996; Hart, 1995; Russo and Fouts, 1997). Group companies traditionally share one source of control and a group may have several sibling companies that engage in relevant and/or irrelevant businesses. Sibling firms are heterogeneous in their development of complementary capabilities or resources (Darnall and Edwards, 2006). If the parent company s and its sibling companies align in favour of environmental protection strategies, then we might expect a group-wide adoption. However, such an alignment might take some time to realize. If there is a mismatch between parent and sibling firms environmental strategies then, being part of a larger group of firms may hinder group-wide adoption of the EMS. Instead, being not part of a group means swifter decisions can be made. Therefore, we propose that: Hypothesis 2: Group firms are more likely to adopt an EMS later than firms not being part of a group.

9 Firm Capabilities In the case of an EMS adoption, a firm s complementary capabilities may have an impact on the early adoption decision. Previous researches show that firms can obtain complementary capabilities by implementing other types of quality management standards. For example, ISO 9001 quality management standards are one of the most widely used quality management tools in the world today. The ISO 9001 standards provide guidance and tools for organizations to ensure that their products and services consistently meet customer s requirements and that quality is consistently improved. Darnall and Edwards (2006) show that experience with quality-based (i.e. ISO 9001) and inventory control management systems reduce the cost of EMS adoption because existing experience could help organisations to acquire crucial knowledge-based processes with fewer resources. Firms that possess such capabilities may also operationalize their EMS within a shorter time and with much less effort (King and Lenox 2001; Delmas 2003; Lopez-Gamero et al., 2009) and they can even reduce the cost of implementing the standards (Bansal and Hunter, 2003; Nakamura et al., 2001). Hormozi (1997) posits that the implementation costs of ISO 9001 are a good indicator of the costs that can be expected with ISO standards adoption. Accordingly, having ISO 9001 quality management standards provide a good basis to adopt an ISO type standards earlier as it can help the adopters to incorporate environmental management systems into practice and to extend the boundaries of TQM. Hence, we propose that: Hypothesis 3: Firms with previous experience in implementing management standards are more likely to adopt an EMS earlier than those without such experience. Similarly, customer-related services are the other efficient tools to obtain feedbacks that would allow firms to increase product and service quality, maintain customer satisfaction, attract new customers and increase efficiency by reducing time and cost. We argue that labelling goods and services as customer-related services can encourage early adoption decision. Such customer-related services are the main sources to obtain invaluable customer feedback about the products and services. Efficient use of these feedbacks may help firms to differentiate their products and services and alter customers perception. Recent investigations show that corporate reputation can be affected by information about an organisation s environmental performance (Arora and Gangopadhyay 1995; Konar and Cohen 1997; Marshall and Mayer 1992). Labelling would be an efficient way to convey this information. For example, labelling may assist firms to inform their customers about their greener and more sustainable products and services. Arora and Cason (1996) showed that

10 downstream firms that are close to final consumers are more likely to adopt an EMS. These firms could use the logo of these eco-labels on their products, website, brochures, business cards etc. and inform their customers about the content of the label and how it effects their products and services. This would not only help firms to inform current customers but also to attract potential customers attention. Hence, we expect that firms that already have labelling services are more likely to adopt an EMS earlier: Hypothesis 4a: Firms that label its products and services are more likely to adopt an EMS earlier than those do not use label. In the same vein, providing delivery services is another customer-related service that requires knowledge-based capabilities and close relationship with clients. Contractually undertaking to deliver or supply goods or services in a fixed deadline requires firms to take responsibility of delivering different components, products, materials to materials suppliers, distributors, retailers and customers. On the one hand, poor organisation of such systems and mechanisms can hinder organizational efficiency, endanger customer satisfaction, create waste (in the form of unused and/or accumulated inputs) and increase delivery cost and time. On the other hand, adequate organisation of these systems and mechanisms can increase customer satisfaction that leads to higher profitability and competitiveness (Lai and Cheng, 2009). Nishitani (2009) states that firms engaging close links with their end-customers may have strong incentives to demonstrate goodwill by adopting and implementing successful environmental management systems. Darnall and Edwards (2006) argue that skills required for adopting management systems (i.e. inventory control) are complementary to the basic capabilities required for the adoption of an EMS because these management systems aim at reducing resources use and the organisations environmental impacts. Firms that are looking to differentiate themselves from their competitors that have similar product and service quality could adopt an EMS to maintain their market position or even to improve it. Hence, firms that provide delivery services are more likely to adopt an EMS earlier as these tools can facilitate the adoption of an EMS. In view of that, we propose that; Hypothesis 4b: Firms that have own delivery service are more likely to adopt an EMS earlier than those do not have delivery service. Foreign customers or in other words a firm s wider international scope might be another crucial factor affecting firm s adoption decision (Corbett and Kirsch 1999; Bansal and Hunter 2003). External pressures can drive firms to adopt standards more than their technical

11 capacity to employ them (Simpson et al., 2012) and these pressures are all the more important if a firm intends to enter new markets such as European markets (Potoski and Prakash 2004b). Firms that do business in the EU are more likely to be early adopters as the EU markets generally demand more comprehensive and transparent information about firms operations and their impact on the environment. Increasing pressures to reduce carbon footprint and along with it, increasing consumer awareness within international markets has created perceivable pressures on firms to take preventive measures. Wojan and Bailey (2000) argue that when domestic environmental regulations are in place, customers might have greater confidence to their domestic suppliers while in the absence of legitimate environmental compliance proof foreign suppliers might struggle to obtain acceptance. When this is the case, certifying an internationally recognized EMS can help firms to overcome legitimacy problems, generate trust (Zucker, 1986) and ensure better relationships with customers, regulators and governments (Bansal and Hunter, 2003). Accordingly, we propose the following hypothesis: Hypothesis 5: Firms with wider international scope are more likely to adopt an EMS earlier than those with narrower scope. A firm s Research and Development (R&D) intensity ratio of R&D expenditures to total turnover- can be another factor that may have an impact on firm s EMS adoption decision. Firms invest in R&D activities to be able to produce significant knowledge and gain invaluable experience that would help the firm to maintain their presence in the market. However, the sustainability of a firm cannot be maintained by relying only on its R&D intensity. Many firms are able to maintain their market position without investing in R&D. Therefore, it is plausible to expect that R&D-intense firm would look for other means to develop social and technical capabilities that are not imitable by their competitor. We posit that especially R&D-intense firms would adopt EMS earlier as these standards may help to re-orient firms management practices to a more sustainable one and help them to reduce their carbon foot-print. By doing so, it could be possible to stay ahead of regulations and gain legitimacy in the eyes of external stakeholders (Meyer and Rowan, 1977; Di Maggio and Powell, 1983) and improve the identification of products with unique characteristics (Mueller et al. 2009; Jamali 2010). Therefore, we expect that R&D intense firms are more likely to adopt the standards earlier compared less R&D intense firms. Hence, we propose that:

12 Hypothesis 6: Firms with higher R&D intensity are more likely to adopt an EMS earlier than those with lower R&D intensity. Firm performance In a conventional sense, an EMS is not designed to directly improve a firm s business performance. However, the relationship between the EMS and firm performance is a significant research question as its potential advantages can benefit industries and society at large. This relationship may also be one of the most significant drivers for the adoption decision. Lim and Prakash (2014) assert that appropriately designed voluntary programs can support both the environmental and the economic goals suggesting that if successfully implemented, the standard may provide substantial financial return via different channels. Adoption of an EMS has been proven positively associated with financial performance (Melnyk et al., 1999; King and Lenox, 2002) but this finding is challenged by the work of Nakaruma et al. (2001). Frondel et al. (2007) and, Delmas and Pekovic (2013) show that by adopting an EMS firms may reduce input cost (raw materials, energy, water, labour), mitigate environmental liability and create a sustainable brand image that can attract new customers and/or reduce trade barriers to reach certain markets (i.e. European markets). Delmas and Pekovic (2013) also found that sustainable brand image has a positive impact on employees work attitudes. Ambec and Lanoie (2008) stated that employees that are proud of their company s good reputation are more likely to have higher job satisfaction and that they are also more likely to disseminate the good reputation during their daily encounters. Even though an EMS does not specifically try to achieve such an objective, resulting in increased labour productivity could be a crucial driver for its adoption and an important incentive for the early adopters in particular. However, the advantage of the adoption is countered by the cost of implementation (Harter and Homison, 1999). In addition, increasing employees work satisfaction cannot be achieved over night as it requires perceivable improvements in working conditions and on environment. Early adopters are therefore, in a unique position to achieve higher labour productivity compared to late and non-adopters. Because, at best they would have longer time to efficiently operationalise their EMS and hence, maintain employees satisfaction and loyalty. Therefore, it is essential to prove that ISO adoption indeed improve cost efficiency and helps to offset the implementation costs. Hypothesis 7: Firms that adopted EMS earlier are more productive in terms of labour than those with late adoption.

13 Empirical analysis The data This research is based on two data sources. The first one is the French Organisational Changes and, Information and Communication Technologies use 2006 (COI 2006) survey. The COI survey provides a matched employer / employee dataset on organisational change and ICT use and it compiled through the joint efforts of the INSEE (National Institute for Statistics and Economic Studies), DARES (Ministry of Labour) and CEE (Centre for Labour Studies). This comprehensive survey contains various questions related to changes and is addressed to companies and to their employees. The latest survey was conducted in 2006 on a representative sample (7700 firms) of the population of French firms from all industries except agriculture, forestry and fishing. Each of these firms filled in a questionnaire concerning the use of ICT and organisational practises in 2003, in 2006 and during the period in between. The COI 2006 survey is a retrospective survey in which firms answer in 2006 a questionnaire that includes questions pertaining to 2003 and to the period in between. Furthermore, the survey design allows for an easy matching with administrative databases. In order to obtain information about value added, number of employees and some other accounting variables, we merged the COI survey with another French administrative dataset; the Annual Enterprise Survey (EAE 2006). As a result of this merger, we worked with a panel of 11,168 observations (5584 firms observed in 2003 and in 2006) of firms with more than 11 employees. We tried to eliminate the outliers by excluding firms that employ less than 11 employees and have more than 100% annual growth rate (in terms of totals sales). Table 1 presents before merger and after merger means and the standard deviation of all the key variables used in our empirical analyses. The test for differences in means indicates that merging datasets did not lead to severe selection biases and there aren t any significant changes in mean values of the key variables before and after merger.

14 Table 1. Summary of key variables before and after merging the datasets COI COI and EAE P- value Variables: Mean Std. Dev. Mean Std. Dev. Firm Size 691, ,84 653,7 3735,85 0,29 Turnover , , ,21 Firm part of a Group in ,34 0,48 0,35 0,48 0,38 Firm part of a Group in ,36 0,48 0,36 0,48 0,37 Firm active in Local or Regional markets in ,13 0,34 0,13 0,34 0,48 Firm active in National market in ,3 0,46 0,3 0,46 0,24 Firm active in European market in ,54 0,5 0,53 0,5 0,25 Firm active in International market in ,64 0,48 0,64 0,48 0,41 Firm active in Local or Regional market in ,14 0,34 0,14 0,34 0,46 Firm active in National market in ,31 0,46 0,3 0,46 0,2 Firm active in European market in ,55 0,5 0,55 0,5 0,26 Firm active in International market in ,65 0,48 0,65 0,48 0,39 Firm has ISO 9001 certificate in ,49 0,5 0,47 0,5 0,1 Firm has ISO 9001 certificate in ,53 0,5 0,52 0,5 0,08* Firm has an EMS certificate in ,8 0,4 0,79 0,41 0,2 Firm has an EMS certificate in ,85 0,36 0,84 0,37 0,19 Firm belongs to High-tech. manufacturing 0,03 0,18 0,04 0,19 0,27 Firm belongs to High-medium tech. manufacturing 0,09 0,29 0,1 0,29 0,16 Firm belongs to Low-medium tech. manufacturing 0,1 0,3 0,1 0,31 0,14 Firm belongs to Low-tech manufacturing 0,14 0,35 0,15 0,36 0,08* Firm belongs to Knowledge intensive services 0,18 0,39 0,18 0,38 0,32 Firm belongs to Other services 0,46 0,5 0,44 0,5 0,02* Disaggregated industry affiliations: Textile, clothing, accessories 0,02 0,14 0,02 0,15 0,32 Wood 0,01 0,09 0,01 0,09 0,37 Paper and printing 0,03 0,17 0,03 0,17 0,26 Coal, petroleum and nuclear products 0 0,05 0 0,05 0,43 Chemical products 0,03 0,17 0,03 0,17 0,28 Rubber and plastic 0,02 0,15 0,02 0,15 0,31 Non-metallic mineral products 0,02 0,13 0,02 0,13 0,32 Metal and metal products 0,06 0,23 0,06 0,24 0,23 Machinery and equipment 0,03 0,18 0,03 0,18 0,3 Electrical and electronic equipment 0,04 0,19 0,04 0,2 0,28 Transportation equipment 0,02 0,14 0,02 0,15 0,3 Other manufactured products; recycling 0,02 0,13 0,02 0,13 0,34 Electricity, gas and water 0 0,07 0,01 0,07 0,4 Construction 0,07 0,26 0,07 0,26 0,21 Car dealing and repair 0,03 0,16 0,03 0,17 0,27 Wholesale trade 0,08 0,27 0,08 0,27 0,15 Retail trade 0,09 0,29 0,1 0,3 0,13 Hotels and restaurants (horeca) 0,03 0,18 0,04 0,18 0,29 Transportation & communication services 0,09 0,29 0,1 0,29 0,16 Housing and real estate 0,02 0,14 0,02 0,14 0,34 Rental 0,02 0,13 0,02 0,13 0,37 ICT services, R&D 0,04 0,21 0,03 0,18 0,* Services to firms 0,12 0,32 0,12 0,33 0,09* Culture, entertainment, sports 0,01 0,1 0,01 0,1 0,45 Financial and insurance services 0,04 0,2 0,04 0,2 0,2

15 Estimation method: Propensity Score Matching In this research, we employ Propensity Score Matching method to identify firm specific characteristics and capabilities that facilitate relatively early and late adoption of EMS (hypotheses 1-6) and to discover the effect of relatively early and late EMS adoption on adopters labour productivity (hypothesis 7). Even though, there are several performance measures such as material and energy productivity, specific issues and problems call for an appropriate measure of performance. In our approach, we use labour productivity (here, the ratio of total Value Added VA to firm size) as the performance indicator. We measure the VA as the efficiency in the use of resources approximated with total price of sales minus the total production cost including factor costs (i.e., material cost along with subventions, taxes and subsidies). In order to estimate the impact of adoption on relatively early and late adopters labour productivity, we need to compare the average productivity of these firms to the average productivity that these same firms would have achieved had they not adopted the standards. However, since a firm either adopts the standards or does not, the average productivity that firms would have achieved had they not adopted the standards remains an unobserved counterfactual since only one outcome is observed (i.e., we only observe Y 1 if the firm has been treated (adopted the standards, D = 1) or Y 0 if she has not (D = 0). More precisely, what would have resulted had the firm not been adopted the standards (treated) cannot be observed, which gives a rise to a case of what is known in the econometric literature as the Evaluation Problem. The evaluation problem consists in providing unbiased estimates of this average counterfactual through the use of appropriate methods and usually untestable assumptions (Goodman and Sianesi, 2005). In this research, relying on Ordinary Least Square method (OLS) may produce biased results when considering the fact that there would be some firms adopting the standards, which are not comparable to the firms in the non-adopting sample. In this sense, performing OLS might hide the fact that we are actually comparing incomparable firms by using the linear estimation (Goodman and Sianesi, 2005). Hence, we used the Propensity Score Matching (PSM) method proposed by Rosenbaum and Rubin (1983) and further developed by Heckman et al. (1997; 1998). The matching method is a non-parametric alternative to Instrumental Variable (IV) and Heckman-type models for estimating a causal effect net of endogeneity bias. The PSM method accounts for selectivity bias and the differences in subjects receiving treatment by estimating the probability to receive treatment given these background variables. This probability (Pr (Di =1 Xi) is called the propensity score. The

16 PSM offers a unique advantage by excluding the firms from the sample that are not comparable to any other firm in the non-adopting sample, leaving only firms from both group that have same features and hence, comparable. In our case, the ATT (average treatment effect on the treated) is defined as the expected difference between Y 1 and Y 0 conditional on D = 1. TT = E(Y 1 Y 0 X, D = 1) = E(Y 1 X, D = 1) E(Y 0 X, D = 1) Where Y 1 is the performance indicator measured by labour productivity in case of adoption of an EMS, Y 0 is the productivity of the same firm in case of non-adoption of an EMS and D={1,0} is an indicator of adoption of an EMS. However, we can only observe E(Y 1 X, D = 1) and E(Y 0 X, D = 0), whereas E(Y 1 X, D = 0) and E(Y 0 X, D = 1) are remain unobserved counterfactual to be estimated. We can estimate these counterfactuals by relying on the Conditional Independence Assumption (CIA), which implies that selection occurs only on observables and is eliminated when accounting for X. Under the CIA, E(Y 0 X, D = 1) = E(Y 0 X, D = 0) and thus we can use the latter to estimate the counterfactual. However, conditioning on all pre-treatment variables X is computationally burdensome and we can, therefore, introduce the propensity score P (X) = Pr(D = 1 X ) i.e. probability to participate in the treatment group. In order to match individuals on the basis of their propensity scores we use the kernel estimator, which attributes weights to control observations according to their relative proximity to the treated observation where exact matches get a large weight, and poor matches get a small weight (Vandenberghe and Robin, 2004). We have also performed an alternative approach to the PSM method, which is known as Fully Interacted Linear Regression Model (a.k.a. FILM). The FILM approach is used to complement the OLS and matching analyses. One of the unique advantages of this method over PSM method is that it allows the impact of adoption to vary for each observable factor. Moreover, the FILM allows us to test the presence heterogeneous effects (Goodman and Sianesi, 2005). If the treatment effect is driven only specific set of control variables, then the FILM could be able to identify that. Consequently, the FILM provides an in-between solution between OLS and matching this method and, it provides a base for comparison and check for the consistency of PSM method results.

17 Dependent variable and Treatment group In this research, we distinguish early and late adopters of voluntary EMS according to timing of their EMS adoption. Table 2 below represents frequency of the adopters and non-adopters with respect to observation periods. In our approach, firms that already adopted voluntary environmental standards and obtained the certificates in 2003 are considered early adopters and similarly, firms that did not have the certification in 2003 but obtained the certification between 2003 and 2006 are considered late adopters. Firms that did not have an EMS certificate either in 2003 or 2006 are considered non-adopter. In addition, firms that had an EMS certification in 2003 but not in 2006 are excluded from our sample. Creating these categories help us identifying a firm s structural characteristics and capabilities that facilitates the adoption of an EMS by contrasting i) non-adopters vs. early adopters and ii) non-adopters vs. late adopters. These comparisons should also point out whether early and late adopters differ in terms of structural characteristics and capabilities even if we do not directly compare them with each other. Table 2. Frequency of adopters and non-adopters according observation periods Yes No # of early adopter 873 (%15.6) 4711 (%84.4) # of late adopter 300 (%5.4) 5284 (%94.6) Total in both year 1173 (%21) 4411 (%79) We may assume that a certain amount of time is required before the standards can have significant effect on firms performance. Empirical evidence (Hart and Ahuja, 1996) asserts that there may be a positive relationship between emissions reductions and financial performance but one may have to wait as long as two-years to have any significant financial return. Ambec et al. (2013) claims that in many of the previous studies that investigate the determinants of productivity, researchers have often regressed productivity at time 0 on proxies of environmental regulation stringency at time 0 as well, which does not allow time for the innovation process to occur. Therefore, introducing a clear distinction between the early and late adopters of voluntary EMS also allows us to test this claim. Explanatory Variables and Descriptive Statistics We are first interested in observing the impact of a firm s size on voluntary EMS adoption decision. We measure the firm size with its actual number of employees. We also included Size variable, which is measured by taking the square of a firm s size. These two variables

18 together could point out the direction of the relationship between a firm s size and their propensity to adopt the standards. Being part of a Group is another control variable included in our analyses and it would help us to understand whether a firm`s ownership structure facilitates the EMS adoption. This variable is a binary indicator of a group ownership and it takes the value of 1 if a firm is part of a larger group of firms and 0 otherwise. We also would like to control for the impact of having other types of quality standards (ISO 9001) and customer-related services (Labelling and Delivery) on a firm s propensity to adopt voluntary EMS. ISO 9001, Labelling and Delivery are three distinct dummy variables and each take the value of 1 if a firm has these management standards and customer services, and they take the value of 0 otherwise. We have also included a categorical variable indicating whether a firm is active in National, European and International markets, where the local and local markets category is the base category for comparison. By including this variable we aim to observe the respective significance of different business markets on a firm s propensity to adopt voluntary EMS. Table 3. Summary statistics on the variables used in the econometric analysis Variables Mean Std. Dev. Min Max Firm size (in numbers) Firm size (in numbers) Firm part of a Group Having ISO9001certificate Having Delivery services Having Labelling services Firm active in Local market Firm active in National market Firm active in European market Firm active in International market Firm belongs to High-tech manufacturing Firm belongs to Medium-high tech. manufacturing Firm belongs to Medium-low tech. manufacturing Firm belongs to Low tech. manufacturing Firm belongs to Knowledge intensive services

19 The industry classification dummies are also included in our econometric model. We created these dummies following the OECD s technological level classification and reorder the NACE codes accordingly. The OECD identifies industries largely on the basis of their level of research and development intensity (ratio of research and development expenditure to value added). According to this approach firms are considered to be active in (1) Hightechnology industries if their R&D intensity is above 5%, (2) Medium-high technology industries if the R&D intensity is between 5% - 3%, (3) Medium-low technology industries if the R&D intensity is between 3% %, (4) Low technology industries if the R&D intensity is between 0.09% - 0% and, (5) Knowledge-intensive services. These dummy variables then, help us to investigate whether a firm s R&D intensity has a significant impact on the adoption decision. Table 3 shows the summary statistics on the variables used in the econometric analysis, while table 4 presents the classification of the technological levels by the industries and their respective Nace codes.

20 Table 4. Aggregated Nace codes according to OECD s technology classification Technology Nace codes classification High-technology 24.4 Manufacture of pharmaceuticals, medicinal chemicals and manufacturing botanical products industries 30 Manufacture of office machinery and computers 32 Manufacture of radio, television and, communication equipment and apparatus 33 Manufacture of medical, precision and optical instruments, watches and clocks 35.3 Manufacture of aircraft and spacecraft Medium-high technology manufacturing industries Medium-low technology manufacturing industries Low technology manufacturing industries Knowledgeintensive services (KIS) 24 Manufacture of chemicals and chemical product, excluding 24.4 Manufacture of pharmaceuticals, medicinal chemicals and botanical products 29 Manufacture of machinery and equipment n.e.c 31 Manufacture of electrical machinery and apparatus n.e.c 34 Manufacture of motor vehicles, trailers and semi-trailer 35 Manufacture of other transport equipment, excluding 35.1 Building and repairing of ships and boats and excluding 35.3 Manufacture of aircraft and spacecraft. 23 Manufacture of coke, refined petroleum products and nuclear fuel 25 to 28 Manufacture of rubber and plastic products; basic metals and fabricated metal products; other non-metallic mineral products 35.1 Building and repairing of ships and boats. 15 to 22 Manufacture of food products, beverages and tobacco; textiles and textile products; leather and leather products; wood and wood products; pulp, paper and paper products, publishing and printing 36 to 37 Manufacturing n.e.c. 61 Water transport 62 Air transport 64 Post and telecommunications 65 to 67 Financial intermediation 70 to 74 Real estate, renting and business activities 80 Education 85 Health and social work 92 Recreational, cultural and sporting activities

21 Results and discussions Table 5 displays the estimates of the Probit model used to compute the Propensity Score (Columns 1 and 2) as well as the marginal effects of the Probit model (Columns 3 and 4). The estimation results imply that a firm s size is one of the most significant factors that facilitate the adoption of an EMS (be it early or late). The negative sign of size-square (size) variable however, assert that this probability tends to decrease as the firm size reaches to a threshold. Therefore, the relationship between adoption of an EMS and a firm s size can be represented by an inverted U-shape. Even though our results do not support the Hypothesis 1 and state that both the early and the late adopters are more likely to be large size firms, the results also indicate that not every large firm may adopt the standards. Recent researches (Simpson et al. 2007; Mueller et al. 2009) showed that some large firms, such as Toyota, have developed their own EMS or used independently designed EMS because ISO type standards were not adequate for their specific requirements. This may explain why the adoption is not relevant for the major large size firms. The other reason might be related to implementation problems that large firms have to deal with. The adoption of such standards generally involves extensive employee participation, training and extensive documentation of the organisation s processes, which can cause increased implementation costs (Kollman and Prakash, 2001). Accordingly, some large organisations might seek other alternatives to reduce their environmental foot-print where more generic standards are insufficient and/or involve high implementation costs. Our results also reveal an important insight about the relationship between voluntary EMS adoption and the ownership structure. The coefficient of the group variable is negative and significant both for early and late adopters which indicate that individual firms rather are more likely to adopt ISO14001-type standards than subsidiaries of a larger group. These results provide supports for Hypothesis 2. This relationship might point out that subsidiaries shy away from the adoption of the EMS not only because of the financial burden but also the organisational burden associated with the adoption. The results may also suggest that subsidiaries do not undertake strategic facility-level decisions independently from the parent company (Oliver, 1991). Accordingly, the parent company s organisational and technical capabilities and general business and environmental strategy can significantly affect the group-wide adoption decision. If subsidiaries have enough managerial space to plan and execute strategic actions, they may choose to adopt the standards. But if they have little or no space for facility-level decision making and will have to bound the strategic decisions

22 (adopting or not adopting the standards) made by the parent company, they would not adopt the standards. Our estimation results indicate that subsidiaries may adopt voluntary environmental standards only if the parent companies promote a corporate culture that encourages facility-level adoption (Darnall, 2006). The results also show that existing experience in implementing ISO 9001 quality management standards has a significant and positive impact on early and late adoption decision. Although this result does not support Hypothesis 3, it implies that having ISO 9001 standards is an important complementary asset that can assist firms in operationalising other type of management standards, i.e. ISO type standards. Moreover, we can also observe that firms having labelling and delivery services are more likely to adopt an EMS. These two factors are significant only for the early adopters, which support our Hypotheses 4a and 4b. These results may indicate the significance of early adopters survival concerns. More precisely, when the adoption of standards is widespread within an industry some firms would adopt the standards to survive in challenging market conditions by approximating the characteristics of the industry leaders and gain legitimacy to continue doing business. These results all together suggest that quality management standards (Bansal and Roth, 2000; King & Lenox, 2001; Kitazawa & Sarkis, 2000) and customer services are not only assisting organisations to improve the quality of their product and provide customer satisfaction, but also improve a firm s organisational capability to accommodate new sets of routines and standards. Furthermore, we observe that firms active in National markets are less likely to adopt an EMS while firms that are active in European markets are more prone to adopt them earlier. Interestingly, the marginal impact of being active in the international markets is insignificant but has a positive sign. These results may indicate that the adoption of the standards is becoming increasingly important to conduct business in European markets and that other international markets still accommodates firms without a certified EMS. Recently, Nishitani (2010) showed that countries exporting more to environmentally conscious European countries such as Finland, Germany, Sweden, Denmark and the UK are more likely to have a higher number of ISO adoptions because the environmental preferences and pressures of customers in these countries are stronger. Accordingly, this author argues that customers within these countries are more likely to require both domestic and foreign suppliers to adopt ISO Given that our sample represents French manufacturing and services firms, the results make sense in this light. Existing environmental regulations and standards in France aim at maintaining certain product, process and service quality and therefore, many