Untangling Drivers for Supplier Environmental and Social Responsibility: An Investigation in Philips Lighting’s Chinese Supply Chain

This study unpacks the environmental and social dimensions of supplier responsibility and links each dimension to distinct drivers. Using stakeholder theory and the relational view, we distinguish between two main drivers: stakeholder pressures (i.e., from regulatory agencies, customers and NGOs) and relational mechanisms offered by multinational companies (MNCs) (i.e., lean trainings and relational capital). We used a multi-method research design to study how these drivers uniquely influence supplier responsibility in an emerging-country context. An in-depth case study with Philips Lighting and ten of its Chinese suppliers reveals causal inferences that link stakeholder and relational drivers with each responsibility dimension (environmental versus social). Audit and survey data from Philips Lighting’s 134 Chinese suppliers, complemented with four archival databases, bolster these inferences. Overall the results show that supplier environmental responsibility can be fostered through both stakeholder pressures and relational drivers; supplier social responsibility is much harder to address. The integrated methods offer a fuller, more comprehensive understanding of the specifics of supplier responsibility in China and also provide recommendations for MNCs that seek to improve it.


Introduction
Increasing pressures on global brands to promote responsible practices in their supply chains have driven some multinational companies (MNCs) to adopt supplier development programs that seek to minimize environmental harm and improve working conditions at supplier factories. However, suppliers do not always comply with these environmental and social requirements to the same degree. According to recent reports by the Electronics Industry Citizenship Coalition (EICC) 1 , audited suppliers have steadily improved their environmental efforts but not their labor practices (e.g., overtime, fair wages). For example, the detrimental labor practices by Foxconn's subsidiary Taiyuan have been blamed for multiple employee suicides, even as it gained ISO 14001 certification which reflected some commitment to environmental protection (Foxconn Social and Environmental Reponsibility Report, 2011). Most studies have focused exclusively on drivers of either environmental (e.g., King & Lenox, 2001;Klassen & Vachon, 2003) or social responsibility of suppliers (e.g., Locke, Qin, & Brause, 2007;Short, Toffel, & Hugill, 2016). This makes systematic comparison between the two dimensions difficult. Other studies have lumped together social and environmental requirements (e.g., Ioannou & Serafeim, 2015;Jiang, 2009;Porteous, Rammohan, & Lee, 2015;Tang, Hull, & Rothenberg, 2012). Because suppliers should comply with both social and environmental requirements, we need enhanced insights into how MNCs can stimulate progress in each dimension.
In this paper, we argue that suppliers' environmental and social responsibilities are distinctly different in nature and, thus, can be linked to distinct drivers. Using stakeholder theory (Agle, Mitchell, & Sonnenfeld, 1999;Mitchell, Agle, & Wood, 1997) and the relational view (Dyer & Singh, 1998), we argue that suppliers respond to pressures exerted by stakeholders (e.g., regulatory agencies, customers, non-governmental organizations [NGOs]) and to relational drivers offered by MNCs that purchase from them (e.g., lean training and relational capital). However, their responses may well differ across social and environmental responsibilities. Based on past studies, we define supplier responsibility in terms of a supplier's compliance with its global customers' environmental and labor standards (Chen & Lee ,2004;Guo et al., 2016;Plambek & Taylor, 2016).
Such standards are often considered as forms of private regulation (Vogel, 2010) because they are formulated and promoted by MNCs in light of the limited ability (and perhaps unwillingness) of Electronic copy available at: https://ssrn.com/abstract=3681079 identify critical social and environmental issues in Chinese factories; we also gain in-depth understanding of the idiosyncratic nature of each dimension of supplier responsibility. To interpret the findings of our in-depth case study, we ground our analysis in stakeholder theory and the relational view within the domain of supplier responsibility and in an emerging-country context, such that we derive causal inferences between stakeholder/relational drivers and supplier responsibility. Then to bolster these causal inferences, we perform quantitative testing, using Through these multi-method efforts, we establish several notable contributions. First, using stakeholder theory and the relational view, this study proposes a comprehensive framework of drivers of supplier responsibility. In addition to stakeholder pressures that are typically studied for MNCs (e.g., Delmas & Toffel, 2008;Surroca, Tribó, & Zahra, 2013), we include two relevant relational drivers that MNCs use to support their emerging-country suppliers' capability-building for more responsible practices (Distelhorst, Hainmueller, & Locke, 2016;Locke et al., 2007;Locke, Rissing, & Pal, 2013). That is, stakeholder theory and the relational view together help explain suppliers' motivation and ability to respond to stakeholder pressures for responsible practices (Durand et al., 2019). Second, we uncover the idiosyncratic nature of suppliers' social and environmental responsibilities in an emerging-country context and reveal their unique links to distinct drivers. In particular, we show that supplier environmental responsibility can be fostered by stakeholder pressures (especially regulatory ones) and relational drivers, whereas the same drivers are not effective for supplier social responsibility. To address labor issues in China (e.g., excessive overtime, no social insurance), requires that relevant stakeholders (e.g., regulators, buyers) must present those concerns as legitimate and urgent, then exercise their power to induce supplier compliance. Third, from a methodological perspective, our research design demonstrates an effective approach to combining qualitative and quantitative methods to provide a more integrated, complete view of suppliers' environmental and social responsibility in China. By leveraging the strengths of each method, we mitigate concerns associated with singular methods and establish strong validation for our research insights.

Theoretical Background
Although often treated as one construct, initial evidence suggests that corporate responsibility's environmental and social dimensions are only weakly correlated (Chatterji, Levine, & Toffel, 2009;Wilhelm et al., 2016). According to , corporate social commitment and corporate environmental commitment differ, such that the latter is technical in nature, as it requires technical skills and innovations to implement, and features internal control, so firms can make internal changes (e.g., product design) to achieve it. In contrast, social commitment depends more on external stakeholders' interests and actions, which are harder for any individual firm to control.
We leverage these insights, while noting a key difference in our approach: Bansal et al.'s (2014) corporate commitment construct refers to do-good practices pursued by publicly listed, U.S. firms, whereas we focus on do-no-harm practices by emerging-country suppliers. These suppliers (cf. MNCs) often lack the expertise or resources needed to address environmental and labor concerns, so they likely focus on do-no-harm practices to deal with negative externalities, before they can even consider starting to engage in do-good practices to further their positive externalities (Crilly, Ni, & Jiang, 2016).
Regulatory bodies are obvious external stakeholders for supplier firms (Zhu & Sarkis, 2007); suppliers that fail to comply with governmental regulations face legal actions, penalties, or fines.
As Wu et al. (2014) show, state-owned (cf. non-state-owned) suppliers in the Chinese electronics industry respond more swiftly to regulatory pressures to implement energy-efficiency programs. However, as our example of local pollution versus climate change demonstrates, the legitimacy of regulatory stakeholder claims also could vary with the focal issue. In China for example, environmental and labor issues are regulated by different authorities, and their enforcement varies, because regulatory authorities have unique resources and law enforcement capabilities (Van Rooij & Lo, 2010).
MNCs also influence supplier firms (Christmann & Taylor, 2006;Huq et al., 2016;Klassen & Vereecke, 2012). They might demand that suppliers comply with the environmental and social requirements that are standard in their home countries, as well as monitor their compliance through factory audits. However, audits cannot be effective if emerging-country suppliers often lack the basic capacities to implement their buyers' environmental and labor requirements (Huq et al., 2016;Jiang, 2009;Locke et al., 2007). In other cases, suppliers actively attempt to trick or cheat auditors (Clifford & Greenhouse, 2013;Plambeck & Taylor, 2016), such as by creating false inspection or maintenance records, related to both safety standards and pollution prevention (Pruett, 2005;Walsh & Greenhouse, 2012). Despite these caveats, audits remain the main control mechanism that buyers use to monitor supplier compliance.
Emerging-country suppliers might be relatively less responsive to pressures from NGOs, because they have less motivation to protect their reputation (particularly if they are private firms) Electronic copy available at: https://ssrn.com/abstract=3681079 (Lee et al., 2012) and often suffer no immediate consequences for not addressing environmental and labor issues (Villena & Gioia, 2018). Consider IPE, an NGO in China that has created a comprehensive database of Chinese firms' environmental violations. It uses pressure tactics, primarily on MNCs, in the hope that they will shift those pressures onto suppliers. Although firms such as Nike, Walmart, and Esquel use IPE's database to screen suppliers for compliance with environmental standards, IPE's achievements have been limited (Lee et al., 2012). Similarly, several electronics suppliers in Mexico have been cited by CEREAL, an NGO that fights for factory workers' rights, for sexual and job harassment. If it does not receive responses from the suppliers, CEREAL makes the allegations public and informs major electronics brands about its concerns (Villena & Gioia, 2018). Yet in contrast with evidence showing NGOs can influence MNCs (e.g., Briscoe & Gupta, 2016;Delmas & Toffel, 2008;Doh & Guay, 2006;Eesley & Lenox 2006), they might be less able to exert strong pressures on emerging economy suppliers. To date, we know of no systematic exploration of pressures from these different stakeholders on suppliers in emerging countries or their distinct effects on their environmental and social responsibility.

Relational drivers of supplier responsibility
In addition to the pressures exerted by external stakeholders, suppliers might respond favorably to efforts by customers to enter into strategic partnerships (Christmann & Taylor, 2006). According to the relational view, firms can achieve superior performance with their strategic partners through the combination, exchange, or investment in idiosyncratic assets, knowledge and resources/capabilities (Dyer & Singh, 1998). The creation of relational rents is based on four sources, namely, relation-specific investments; knowledge-sharing routines; complementary resources and capabilities; and effective governance. Few operations and supply chain Electronic copy available at: https://ssrn.com/abstract=3681079 management studies adopt a holistic relational view though (cf. Chen et al., 2013;Nair et al., 2011;Walker et al., 2013). Rather, they tend to integrate specific elements of this theory to explore the relational benefits that result when a buyer engages in a strategic partnership with its suppliers.
First, relationship-specific investments might involve physical or human assets (Dyer, 1996). Chen et al. (2013) investigate idiosyncratic investments by a hospital and its key suppliers in IT integration, revealing that this integration investment enhanced their supply chain performance.
The level of assistance that a buyer provides its suppliers also may be an intangible investment that can lead to lower costs, higher product quality, and more flexibility (e.g. Dyer & Hatch, 2006;Inemek & Matthyssens, 2013). Second, knowledge-sharing routines increase economic performance (e.g., Cheng, 2011;Dyer & Hatch, 2006), because they facilitate the transfer, recombination, or creation of specialized knowledge (Dyer & Singh, 1998). For instance, Dyer and Nobeoka (2000) describe how Toyota's sophisticated knowledge-sharing routines with suppliers encouraged their holistic implementation of the Toyota Production System, which produced lasting productivity benefits. Third, complementarity of resources and capabilities might provide benefits, yet it is not studied often, seemingly because in a buyer-supplier relationship, the parties' contributions to their vertical partnerships are, by definition, complementary. Fourth, trust, reciprocity, and respect, as the norms that constitute relational capital-serve as a form of self-enforcing governance, which effectively lowers operational risks by controlling opportunism and promoting cooperation (e.g., Cousins et al., 2006;Kale et al., 2000;Lawson et al., 2008). Selfenforcing governance is generally preferred over third-party enforcement (e.g., legal contracts), because it safeguards and stimulates relationship-specific investments without invoking heavy transaction and monitoring costs (Dyer & Singh, 1998). Self-enforcing governance rests on the "shadow of the future" (Poppo et al., 2008) and foregone future benefits if the partnership were to Electronic copy available at: https://ssrn.com/abstract=3681079 be terminated.
Furthermore, the relational view mainly appears in studies that seek to explain superior economic gains, not environmental or social benefits in buyer-supplier relationships (Touboulic & Walker, 2015). Still, without explicitly referring to the relational view, several studies show that knowledge-sharing practices (e.g., on-site consultation, training programs, temporary personnel transfers) can improve supplier responsibility (Gimenez & Tachizawa, 2012;Porteous et al., 2015).
According to Klassen and Vachon (2003), buyer-initiated knowledge-sharing programs reduce uncertainty for suppliers and thus alleviate their resistance to investing in environmental technologies. Similarly, Distelhorst et al. (2016) find that Nike's lean production training helped build the capabilities of its global suppliers by improving employee-management relationships.
Because the lean production efforts focused on decentralizing decision-making and upskilling workers, training along these lines effectively motivated suppliers to improve the employment terms (e.g., higher wages and less overtime) to retain skilled workers.
Similarly, the effectiveness of governance for encouraging supplier responsibility rarely has been studied explicitly, perhaps because trust asymmetry is common between MNCs and suppliers in relation to environmental and labor issues (Villena & Gioia, 2018). Yet studies of related concepts, such as collaborative governance (Jiang, 2009) and procedural fairness (Soundararajan & Brammer, 2018), indicate that suppliers reciprocate MNCs' efforts to establish self-enforcing governance mechanisms by addressing environmental and labor issues, even in weak institutional settings.
Finally, we found only few studies that investigate how complementary resources or relationship-specific investments influence environmental and social outcomes in buyer-supplier relationships (cf. Lee & Klassen, 2008;Simpson & Power, 2005). Considering this foundation in prior literature, we attempt to apply the relational view to study non-economic outcomes, such that we explore which sources of relational rents might be the most effective drivers of suppliers' environmental and social responsibility.

Research Design
To investigate the drivers of suppliers' environmental and social responsibility, we combine two research methods to support a single, unified causal inference (Seawright, 2016). First, we conducted an in-depth case study with Philips and 10 of its Chinese suppliers in 2014. Through this analysis, we identified severe environmental and social issues for Chinese suppliers and gained in-depth insights into the idiosyncratic nature of each responsibility dimension. Guided by our theoretical framework, we also identified relevant drivers of suppliers' environmental and social responsibilities. Second, we collected the Chinese suppliers' audit data for 2015-2017, provided by an independent auditing firm, and complemented these data with a survey and four archival datasets. Thus, we combined two methods in an integrative manner, such that we used the case study method to draw causal inferences about the links of stakeholder pressures and relational drivers with suppliers' social and environmental responsibilities, then adopted quantitative methods to test and bolster the causal inferences (Seawright, 2016). Our goal was to develop a thorough, complete assessment of the effects of stakeholder and relational drivers on supplier responsibility.
We selected the consumer electronics industry in China as our empirical setting for two reasons.
First, China is the world's largest supply base of electronic products (Distelhorst et al., 2016).
Rapid industry growth has contributed to its economic development but also created serious social and environmental problems. High-profile incidents, such as the Foxconn suicides, have exposed Electronic copy available at: https://ssrn.com/abstract=3681079 the unacceptable suffering of workers (Chan & Pun, 2010). Increased fine particulate matter concentrations, at least four to five times the levels found in developed countries, also are regularly reported in major Chinese cities (Wang, 2013). As a result, Chinese suppliers come under growing pressures from other stakeholders, such as regulatory bodies, to address their environmental and labor issues (Li et al., 2019). Second, because most Chinese electronics manufacturers have limited resources and lack sustainability expertise (Xiao et al., 2019), they rely on MNCs' support. A few global electronics brands provide this support, but these investments' payoffs are still unknown.

Stage One: Building Causal Inferences Through an In-Depth Case Study
Philips ranks as one of the leading companies in the Dow Jones Sustainability Index, partially due to its efforts to improve its suppliers' responsibility. With its assistance, we selected 10 Chinese suppliers that participated in Philips' Supplier Sustainability Development Program. A dedicated corporate office is responsible for this program, which is designed to enable the company to work closely with suppliers to improve their environmental and labor practices. Philips selects suppliers for this program using two main criteria: (1) the supplier is located in a high risk country, such as China, India, Brazil, or Indonesia; and (2) the annual purchasing from the supplier is at least 1 million Euro. For our selection process, we sought suppliers that operate in different Chinese provinces, have maintained a business relationship with Philips for at least three years, have varying relevance for Philips (e.g., preferred vs. commercial), and differ in size. All participating firms and interviewees were guaranteed anonymity. Table 1 provides an overview of the 10 selected suppliers and interviewees.
---Insert Table 1 about here---We collected data in two phases. In the first phase (January 2014), we visited Philips' headquarters in China and interviewed nine supplier account managers (SAMs) and ten supplier sustainability experts (SSEs), who oversaw the 10 selected suppliers. The interviews focused on Philips' supplier environmental and labor policies, supplier training, and supplier audit process. In the second phase (April-May 2014), we visited the selected suppliers and interviewed each firm's key informants. These semi-structured interviews, conducted in Mandarin, focused on the major environmental and labor issues with which the suppliers had been struggling and lasted about 60 minutes each on average, during site visits that lasted two to three days per supplier. All but 2 of the 57 interviews were recorded and transcribed verbatim; we provide the interview protocol in Appendix A. We also obtained 22 audit reports for these 10 suppliers from Philips. Finally, we gathered archival documents from Philips (e.g., audit preparation training material), suppliers (e.g., internal audit reports), and NGOs (e.g., IPE's environmental violations) for triangulation purposes.
Three co-authors conducted the data analysis; one of them is fluent in both Mandarin and English and was responsible for identifying relevant interview excerpts, in accordance with our theoretical framework. Relevant quotes from these excerpts were translated into English, and the accuracy of the translation was checked by another bilingual speaker. Two authors then coded the excerpts; in any case of disagreement between these authors, we initiated an extensive discussion that included the third author to refine the analysis. Playing the role of devil's advocate (Gioia, Corley, & Hamilton, 2013), the third author also challenged some interpretations of the other two authors, and the coding was repeatedly discussed in bi-weekly team meetings until any disagreements were resolved.
From the first coding round, we identified the most common forms of non-compliance exhibited by the 10 Chinese suppliers, in terms of frequency and severity, using their audit reports.
These critical, recurring forms of non-compliance arguably offer the best potential to represent each dimension's specific nature. We identified environmental permits, management of hazardous substances, and disposal of wastewater and solid waste as the most severe environmental issues; excessive overtime and lack of social insurance were the most critical labor issues. Using these thorny issues as a starting point, we inductively coded the interview data to identify the main reasons for such non-compliance. The results are presented in Table 2.
---Insert Table 2 about here---Next, we iterated between the main constructs in the theoretical framework and the qualitative data to develop a coding scheme; it reflects both stakeholder theory, and its related salience considerations, and the relational view. With this detailed evidence, we can connect both theories to idiosyncrasies of supplier responsibility in China (see Appendix B). In particular, with regard to stakeholder theory-based drivers, suppliers ascribed high legitimacy to environmental claims, because they perceived a clear cause-and-effect relationship between their manufacturing activities and environmental consequences. No such legitimacy emerged for the social issues. In terms of the relational drivers, knowledge-sharing routines appeared relevant, in the form of regular training offered by Philips to improve environmental or social responsibility. These training methods pertained to lean production and IDH (which reflects a Dutch-language abbreviation of Sustainable Trading Initiative). Moreover, the results highlighted effective governance efforts, based in trust, respect, and reciprocity. The two other sources of relational rents-complementary assets and relation-specific investments-were engrained in our research context.
In the final coding round, we examined the relationships of the specified stakeholder groups and relational drivers with the outcome of interest (i.e., supplier responsibility) to describe their causal relationships. From this close examination, we distilled our findings into five broad areas or observations, which provide foundations for our subsequent empirical testing.

Stakeholder pressures for supplier responsibility
As we established previously, stakeholder saliency is jointly determined by three attributes: the legitimacy and urgency of the claim and the power of the stakeholder making the claim (Eesley & Lenox, 2006;Mitchell et al., 1997). According to stakeholder theory, legitimacy derived from societal norms and values (Suchman, 1995) can be captured best by perceptions among the general public, whereas urgency and power tend to be specific to a stakeholder group. Notably, in China, stakeholder claims related to environmental responsibility generally evoke stronger legitimacy than those related to social responsibility. The managers we interviewed often cited the popular opinion that China is paying the price for 40 years of rapid economic development, which was partly fueled by weak environmental regulations. Their perceptions also reflected the clear causeand-effect relationship between their manufacturing activities and environmental problems (e.g., pollution). Instead, they regarded labor issues generally as less legitimate; the supplier firms accepted little accountability for outcomes such as excessive overtime or low wages and even mentioned workers' own requests for overtime shifts in their effort to maximize their take-home income. The managers we interviewed asserted that strict compliance with EICC labor practices would be detrimental to workers' well-being, because at the low minimum wages, without overtime, most workers would not be able to make ends meet. Other interviewees indicated that the labor law was unreasonable, noting that China's Labor Contract Law (passed in 2008) mainly sought to show Western countries that China had been catching up in terms of labor standards, even though it remained a relatively poor country, with a gross domestic product (GDP) per capita of less than US$10,000 in 2018 (World Development Report, 2019).
Despite the consistently greater perceived legitimacy of environmental (cf. labor) issues, the urgency of the claims, as defined by specific stakeholder groups, and the power that those groups Electronic copy available at: https://ssrn.com/abstract=3681079 have to enforce their requests still vary. We consider pressures from three relevant stakeholder groups: regulatory agencies, buying firms, and NGOs.
Regulatory agencies. The urgency of environmental versus social claims by regulatory agencies in China differ notably. Rampant environmental pollution has given rise to a general perception that "we need to take immediate actions to tackle environmental issues, before it is too late" (General Manager, Supplier Alpha). This urgency also is evident in China's national fiveyear environmental plan, which sets ambitious targets for environmental protection, renewed periodically (e.g., the 13 th five-year plan goes from 2016 to 2020). Therefore, top managers of the suppliers were very aware of the need for immediate actions, to avoid hefty fines imposed by local environmental protection bureaus, as one noted: "Our prime minister has made it very clear that environmental protection is the top priority of this administration. Environmental violations will be levied with hefty fines and punishments. As Prime Minister Li put it, 'there is no upper limit'." (Director of EH&S, Supplier Gamma) The central government also has formulated strict targets for reductions in emissions of sulphur dioxide, chemical oxygen, and heavy chemicals (Van Rooij et al., 2017), and to achieve them, it has granted environmental protection bureaus at the national, provincial, and local levels more In contrast, non-compliance with labor standards seemingly stems from the perceived aspirational character of the Labor Contract Law, which includes many clauses that do not reflect China's current socio-economic environment. This divergence also might explain why the Electronic copy available at: https://ssrn.com/abstract=3681079 enforcement of these laws tends to be lax. In turn, managers perceive little urgency to address existing labor issues: "It is not a problem to schedule overtime work, as long as workers are not forced to work overtime and legal wages are paid out in full. Local bureau of labor protection would not bother us for issues of overtime works." (General Manager Assistant, Supplier Eta) The local labor protection bureaus, compared with their environmental counterparts, also have limited sanctioning power. Confronted with violations of labor laws, such as non-payment of wages or excessive working hours, labor inspectors can only impose less serious penalties, such as warnings, correction notices, fines, or compensation orders. They cannot order a factory to close, confiscate earnings, or detain the proprietor (Cooney, 2007). The interviewees thus admitted that they did not expect serious punishments if they failed to comply with the legal overtime limit (36 hours per month), as long as the overtime work was voluntary and workers received adequate monetary compensation. This expectation is reasonable; factory owners also rarely are punished for inadequate social insurance for workers (Chung, 2015). Thus, Observation 1: Differing levels of perceived claim urgency and sanctioning power held by local environmental and labor protection bureaus, in combination with the greater legitimacy ascribed to environmental issues, mean that pressures by environmental (compared to labor) regulatory agencies are perceived as more salient by Chinese suppliers.
Buying firms. Many electronics brands, such as HP, Intel, Apple, and Microsoft, require suppliers to comply with the EICC code of conduct, which establish standards for safe working conditions, treating workers with respect and dignity, and environmentally responsible business operations. These brands use audits to monitor supplier compliance. The suppliers we interviewed reported that they are subject to five audits per year on average, signaling how often buyers use these tools to exert compliance pressure on suppliers.
A common method involves a three-color scheme to indicate the urgency of issues identified by the audit, from red to orange to yellow, in decreasing order. Most environmental issues, such as toxic waste and environmental permits, earn red labels, so they must be resolved within three months. Severe labor rights abuses, such as physical punishments, forced labor, or inadequate remuneration, also might receive a red label. However, recurring labor issues, such as overtime and social insurance gaps, are often labeled orange or even yellow, so non-compliant suppliers have more time to resolve them. In turn, the Chinese managers we interviewed regarded addressing labor issues as a long-term process, especially because local labor protection bureaus do not strictly enforce labor laws. Accordingly, they prefer gradual improvements, as one manager noted: "Compared to Shanghai, Kunshan is less strict in enforcing the law on social insurance, as the local labor protection bureau knows that strict enforcement of this law will drive many companies into bankruptcy…. We are required to come up with a five-year plan to gradually involve all workers in the social insurance system." (HR Manager, Supplier Epsilon) Moreover, buyers might use their power to ensure suppliers' compliance with their social and environmental standards. If suppliers display unsatisfactory levels of compliance after an audit, they could face consequences, such as reduced orders or termination of a business relationship (Lund-Thomsen & Lindgreen, 2014). All 10 participating suppliers sell products to multiple buyers, and it is noteworthy that all the interviewees also identified Philips as their strictest buyer when it comes to enforcing responsibility standards through the exercise of its buying power: "Auditors conducted a full-scope audit in 2012, which identified several serious non-compliances. By the time I went there to conduct the first resolution audit, I was surprised to find that those identified issues were not closed yet. [The supplier] did not resolve those issues in the second resolution audit either. I had no choice but to involve our purchasing manager who told them that Philips would postpone payments until all pending issues were completely resolved." (Supplier Sustainability Expert for Epsilon, Philips) Suppliers' environmental (cf. labor) non-compliances also pose greater risks to buyers because environmental violations are more likely to trigger responses from regulatory agencies in the form of fines, suspended production, and even permanent closure, which would significantly disrupt its buying firms' supply chains. For instance, Philips is willing to exercise power to force suppliers to rectify serious environmental violations, before they come to the attention of local environmental protection bureaus, as our quote above illustrated. In another example, mentioned by Supplier Beta, Disneyland imposed considerable pressure on it to improve its air purification system in an injection workshop, after auditors found that emission levels exceeded the legal standards in Guangdong Province. But buyers appear to exert their power less in relation to recurring labor issues. For instance, several suppliers reported that electronics MNCs (some of which are affiliated with the EICC) would continue doing business with them, even if they failed to meet the 60-hour workweek limit established by the EICC. The buyers may be concerned that strict labor enforcement would drive most suppliers out of business, as one Philips manager noted: "It is not really realistic to require all suppliers to comply with the law on social insurance. The Chinese manufacturing industry is not ready for such a high standard. Take Shanghai for example; strict compliance with this law means at least 500 YUAN of additional labor cost per month for employers. That will drive many small-and medium-sized suppliers to bankruptcy." (Supplier Sustainability Expert for Delta, Philips) Observation 2: Due to the higher levels of urgency and power exercised by buying firms in relation to environmental issues, combined with the greater legitimacy ascribed to environmental issues, environmental (compared to labor) pressures by buying firms are perceived as more salient by Chinese suppliers.

NGOs.
To leverage their power, NGOs usually mobilize public pressure and engage other relevant stakeholders (Hall, Schmitz, & Dedmon, 2019). Although prior research suggests NGOs can effectively "name and shame" global brands for failing to take responsibility for supply chain operations (e.g., Doh & Guay, 2006;Vasi & King, 2012), such pressures might be less effective in relation to emerging-country suppliers. These suppliers seem less concerned about reputational damage due to local NGOs' allegations, because they do not deal directly with end-customers (Villena & Gioia, 2018). In line with some recent findings (Li, Xia, Chen, & Sun, 2018;Lo et al., 2018), we observed the growing influence of IPE, a Chinese environmental NGO, over the course of our research. IPE gathers information about environmental incidents from various government sources and news agencies, then compiles a public database. When we visited Philips Supplier Sustainability Office in 2014, its director had just initiated a collaboration with IPE, recognizing its increasing influence, such that "IPE is becoming an influential NGO in China. It has been developing a database which collects and lists environmental violations of manufacturers in China. If IPE does not get a positive response from the listed manufacturers, it will reach out to their main customers." Since its inception in 2006, IPE also has developed several new strategies to contribute to environmental protection in China, such as encouraging investors to penalize large public Chinese firms that it has blacklisted (Lo et al., 2018). However, persistent evidence indicates that firms have not fully addressed environmental issues. For example, Youngor -a Chinese public textile company-received a complaint, but did not respond immediately to the IPE's demands to clean up its polluting manufacturing processes in 2011. Only after the third consecutive annual complaint did it make some corrections (Davis & Moosmayer, 2015). Private Chinese firms may be even less responsive to NGOs' concerns about pollution, unless buyers (i.e., MNCs) also intervene. Although IPE has encouraged many brands, including Apple, Panasonic, Microsoft, and Dell, to use its data to inform their procurement decisions (IPE Annual Report, 2015, 2019), the ways each brand uses IPE data in its dealings with suppliers vary widely. Still, by collecting, analyzing, and disseminating pollution information, IPE has enhanced public awareness and facilitated public participation in environmental governance in China (Li et al., 2018).
In contrast, labor NGOs in China face more limitations on their daily operations (Chan, 2013), including state repression, lack proper governance, and difficulty in gaining trust from (particularly migrant) workers (Franceschini, 2014;He & Huang, 2015). The NGOs that have gained some legitimacy, such as China Labor Watch, tend to focus on a select group of large Chinese manufacturers, such as Foxconn. As a result, few of the managers we interviewed regarded labor NGOs in China as relevant stakeholders. However, Philips' SSEs noted that one of its suppliers, VTech had been criticized by the Institute for Global Labor and Human Rights (IGLHR) for its labor practices in 2012. In response, VTech denied the accusations vehemently and threatened to sue IGLHR for defamation; it did not take substantial actions to address the labor issues until Philips and other buyers intervened. Thus, environmental (cf. labor) NGOs have gained some influence in China, but they continue to build advocacy power over MNCs and Chinese suppliers.
Observation 3: Compared with labor NGOs, environmental NGOs (and IPE in particular) are developing growing, albeit limited, influence in China, and pressure from them is perceived as more salient by Chinese suppliers.

Relational drivers for supplier responsibility
In addition to stakeholder pressures, Chinese suppliers might respond to drivers that are specific to their relationships with Philips. Although the relational view distinguishes between four sources of relational rents, we only observed two in our data. The other two are more inherent. That is, complementary assets reflect the division of labor between Philips and Chinese suppliers, such that the latter are in charge of manufacturing products, while the former is responsible for design and engineering. Relationship-specific investments are reflected in Philips' Supplier Sustainability Development Program, in which it maintains 19 full-time personnel and an organizational structure that allows for monitoring and offers training to the 220 Chinese suppliers enrolled in the program. Therefore, our discussion below focuses on the other two sources of relational rents: knowledgesharing routines and effective governance.
Knowledge-sharing routines. Philips' knowledge-sharing efforts include a lean training program. Classroom trainings focus on the exchange of best practices for waste reduction, waste handling, and automation. Shortly after the classroom training session, a team of Philips' product and quality engineers visits the supplier's site for about a week and works closely with the supplier's engineers to identify areas for improvement. As a result of such a collaboration, one assembling supplier redesigned its facility layout, which shortened transportation time and reduced the amount of work-in-process; it also reduced fuel consumption for transporters, which diminished CO2 emissions. As a side effect, worker productivity improved, which helped free up financial resources that the supplier invested in other energy-reduction technologies. A Philips SSE reported that another supplier cut its operating costs by approximately US$1 million by applying lessons learned from Philips' lean training program, then invested about 30% of the savings to improve workers' salaries, benefits, and welfare. This reinvestment decision was at the discretion of the supplier firm's top managers; Philips had not formulated any rules regarding how to use these savings. However, Philips' emphasis on efficiency improvements encourages the closer consideration of how human resources enable lean practices, as one manager noted: "We experienced many challenges in promoting lean production. As one way of reducing waste, our workers were required to stand rather than sit. We received a large amount of complaints from workers about this change. They were exhausted from standing long hours. We should not simply aim for productivity maximization, but also consider ergonomics." (HR Manager, Supplier Gamma) Other types of training focus on improving employee-management relations or labor issues.
For example, the IDH program sought to facilitate communications across hierarchical levels within supplier firms, so that management can better understand employees' interests and preferences, and simultaneously, workers can make more substantial contributions to productivity and quality improvement efforts. The IDH program consisted of monthly training sessions, conducted in suppliers' facilities; managers and workers receive coaching by third-party trainers regarding how to engage in productive dialogue about themes, such as food quality, accommodation, remuneration, and production scheduling. However, Philips struggled to convince suppliers to participate in this program. Because the IDH program did not produce evident, significant improvements for the few participating suppliers, it was prematurely terminated in 2016. Reflecting on the program's limitations, one Philips manager explained, "To be honest, I think it is going to be rather difficult to make this program a success. You cannot create a new organizational culture in which managers and workers will feel free to communicate with each other. There must be mutual trust and respect in the first place. Otherwise, it will be oneway instruction rather than mutual dialogue." (Supplier Account Manager for Delta, Philips) Observation 4: Knowledge-sharing routines between Philips and its suppliers disseminate technical knowledge and thus can help suppliers address environmental issues. Knowledgesharing routines are less meaningful for the dissemination of organizational and interpersonal knowledge (e.g., employee-management communications) that could evoke deeper changes throughout the supplier's organization and thereby address labor issues.
Effective governance. Suppliers with closer relations with Philips were more likely to commit to meet Philips' environmental and labor requirements, because they sensed more trust and respect in the relationship. These suppliers expressed their belief that Philips' request to participate in its Supplier Sustainability Development Program was a signal that this customer was committed to helping them overcome their environmental and labor issues. For example, one manager noted: "They involved our company in the Supplier Sustainability Development Program in 2007 because they treat us as a strategic supplier, and they want to prepare us for the increasingly strict enforcement of laws related to labor protection, health & safety, and environmental protection." (General Manager, Supplier Alpha) Because suppliers can count on a long-term partnership with Philips, they tend to reciprocate by increasing their commitment to comply with Philips' environmental and labor requirements. Supplier Zeta, for example, institutionalized internal audits after receiving on-site training from Philips. The relational capital accumulated induced the supplier to address various environmental and labor issues, even with the knowledge that the investments it made to do so might not be equally valued by other customers. Suppliers also benefited from trust, reciprocity, and respect with Philips because these relational norms safeguard their investments. As one manager indicated, "We are quite responsive toward Philips' requests and suggestions. Of course, we would like to maintain our long-term trustworthy collaborations with Philips. Based on such a consideration, we have decided to participate in its supplier development program." (HR Manager, Supplier Alpha) This enabling effect of relational capital on inducing the supplier's investments in its own responsibility efforts might differ across social and environmental dimensions though, because improving social responsibility generally calls for more uncertain investments, whereas investments to enhance environmental responsibility are less uncertain and generally involve material replacements or technological changes. For example, Supplier Gamma's environmental director proudly noted the installation of a water purification system in 2013, which provided immediate results: "Last year our company made serious investments in cleaning wastewater, and the equipment cost us more than 1 million US dollars…. Now we are in full compliance with the water emission standards in Guangdong Province." In collaboration with Philips, Supplier Gamma also initiated a product redesign project in 2012, for which Philips assigned product engineers and quality managers from supply chain partners across three tiers to assist. Within a couple of months, the collaborative project proved very successful. The redesigned product met strict environmental standards, its production process involved less pollution, and unit manufacturing costs decreased by more than 40%. Several other suppliers (e.g., Alpha, Delta, Epsilon) also noted their relatively certain environmentally oriented efforts, which they made while retaining full control over the amount invested, and the immediate outcomes, including enhanced product quality and cost savings.
But to deal with thorny labor issues, the investments may risk moving outside the supplier firm's control, and their results are far more uncertain. In particular, addressing labor concerns usually demands organization-wide changes, and even if the investments pay off, that outcome likely will be far in the future. When Philips introduced the IDH program in 2012, even suppliers with good relationships with Philips were reluctant to join, even after Philips and IDH offered to cover 75% of the program participation fee. Two suppliers that eventually participated, Alpha and Gamma, complained that they had yet to observe any concrete outcomes: "It is really difficult to see any concrete outcomes, at least so far. We have started to doubt the potential values of this program, but the trainers are asking us to be more patient. They are trying to assure us that our investments [in the IDH program] will pay off." (HR manager, Supplier Alpha) Such uncertainty associated with investments in social responsibility seemingly cannot be overcome with the relational capital embedded in close buyer-supplier relationships.
Observation 5: Effective governance between Philips and suppliers helps safeguard the suppliers' investments in environmental responsibility. However, because investments in social responsibility have more uncertain outcomes, they cannot be sufficiently safeguarded by relational capital. Figure 1 provides a visual summary of the main findings of our qualitative study. Three attributes of stakeholder salience-legitimacy, urgency, and power-explain why suppliers are more responsive to pressures from three groups of stakeholders to address environmental (cf. social) issues. The relational drivers that Philips adopts explicitly seek to enhance the sustainability capabilities of suppliers, but they turn out to be more impactful with regard to environmental issues, whereas social responsibility requires deeper, organization-wide changes and even then offers only uncertain outcomes. Overall though, understanding the drivers for supplier environmental and social responsibility requires consideration of both stakeholders and relational drivers. In Table 3, we transform these findings into testable hypotheses and depict the proposed causal chain by which stakeholder and relational drivers lead to supplier environmental and social responsibility.

Sample and data collection
To test the hypotheses, we gathered primary information on Philips' suppliers, using audit, archival, and survey data. Every three years, a third-party auditing firm assesses suppliers' compliance with Philips' labor and environmental standards. We collected these independent audit data for 2015- Statistical Yearbook, and IPE. In addition, we developed a survey, based on our literature review and insights from the case study. The questionnaire was first developed in English, then translated into Chinese. Next, the Chinese version was reviewed by two native Chinese academics, who were independent of the author team. We pilot tested the survey with three academics and four Chinese practitioners, who offered several suggestions for improving the survey's wording, design, and administration.
We targeted all the Chinese suppliers (220) that had participated in Philips' Supplier Sustainability Development Program, which collectively account for 98% Philips' spending in China. With the support of the Philips' Supplier Sustainability Office, we distributed the survey in January 2018. To boost the response rate, we promised a benchmarking report to participants and asked Philips' supplier sustainability manager to invite suppliers' participation directly. With three follow-up contacts, we gathered 134 completed responses, for a response rate of 60.9% (134/220).
For missing values in two cases, pertaining to the number of audits, we substituted an average value. The participating suppliers represent firms of different sizes, industries, ownership types (foreign vs. locally owned), and locations across China (see Appendix C).
We took several steps to ensure respondents were knowledgeable about the key topics. First, Philips' Supplier Sustainability Office knows the best contact person, because its SSEs regularly conduct training at supplier facilities and work closely with suppliers to address environmental and labor non-compliance concerns. We leveraged this knowledge by asking the office to distribute the survey to appropriate respondents representing all 220 suppliers. Second, the cover letter provided specific instructions to choose the best informant to answer the questionnaire (e.g., EH&S manager). Third, we included two questions about the respondent's work experience, with the supplier firm and in the electronics industry. The respondents had been working for their firm and in the electronics industry for averages of 8 and 12 years, respectively.

Triangulation
Philips' Supplier Sustainability Office provided demographic information about the suppliers, such as their industry, size, and management systems. We triangulated this data with those provided by the suppliers on the questionnaire. The results revealed high consistency, in terms of assembly operations (r = 0.78, p < 0.001), chemical operations (r = 0.93, p < 0.001), mechanical operations (r = 0.74, p < 0.001), number of employees (r = 0.75, p < 0.01), ISO 9001 (r = 0.79, p < 0.001), ISO 14001 (r = 0.79, p < 0.001), and OSHAS 18001 (r = 0.86, p < 0.001). These comparative results establish some validity for the information collected through the survey. Table 4 provides complete details related to the variables, data sources, and timing.

Measurements
---Insert Table 4  In terms of NGO pressures, we turned to IPE databases. If IPE had blacklisted a supplier, it would likely be exposed to stakeholder demands to resolve the reported violation (Plambeck, Lee, & Yatsko, 2012). When reviewing the 2014 IPE database, we found a range of 0-4 reported violations for the suppliers, so we created a dummy variable, such that 1 indicates a supplier firm has had at least one IPE violation, and 0 otherwise. We conducted an exhaustive search to identify similar information from a labor-oriented NGO (e.g., Chinese Labor Watch), but none of them have compiled systematic information about labor violations, possibly due to the weak governance structures surrounding labor NGOs in China (Franceschini, 2014). Thus, we cannot include social NGO pressure in our analysis.
For the buyer pressure measure, in the survey, we asked suppliers for the average number of environmental and labor audits they were subjected to each year. We also asked about the number

Control variables.
We included three sets of control variables. First, we controlled for firm features, including its size, gender ratio, integrated management systems, ownership type, and relationship duration. Larger firms might enjoy economies of scale and have more resources to invest in environmental and labor improvements (Darnall, Henriques, & Sadorsky, 2010;Oka, 2010). For the firm size measure, we consider the number of employees. Female employees tend to resist exploitative workplace conditions and actively exercise their rights more, so suppliers with a higher percentage of female workers might feature better labor conditions (Bird, Short, & Toffel, 2019), which is why we include % of female worker, measured by dividing number of female employees by total employees. The preceding variables all came from the survey data. We also accounted for management systems on quality (ISO 9001), environmental (ISO 14001), and health and safety (OSHAS 18001). Based on supplier certification data provided by Tronics in 2017, we created integrated management system, ranging from 0 (no management system) to 3 (ISO 9001, ISO 14001 and OSHAS 18001). A supplier with multiple management systems is better equipped to respond to its buyers' sustainability requirements. In addition, we determined whether suppliers were owned by a foreign company (1) or not (0) to craft the dummy variable foreign-owned firms. International firms tend to offer better pay (Oka, 2010) and can exchange best environmental and labor practices with their parent firms. We gathered this information from National Enterprise Credit Information Publicity System of China. 3 Furthermore, suppliers with longer-term relationships with Philips might be more responsive to its requests, including those for environmental and labor compliance. Relationship duration reflects the number of years the supplier had maintained a business relationship with Philips, as reported in the survey. Finally, we controlled for supplier relevance because suppliers capturing a high share of Tronics' purchasing expenditure could be more interested in meeting its environmental and labor requirements. Based on spending categories that Tronics provided, we created supplier relevance, indicating that Tronics had placed orders with that supplier for more than 5 million euros.
Second, different manufacturing processes create specific challenges. Even though the participating suppliers all serve an electronics manufacturer, each has its own manufacturing processes. Therefore, we asked each supplier to report its industry, using categories defined by Philips (e.g., assembling, mechanical and chemical processing). Assemblers perform laborintensive processes and often demand excessive overtime (Yu, 2008); suppliers that rely on mechanical processes often use heavy machinery, which can pose safety risks for their workers.
We thus created two dummy variables, assembling and mechanical, which represented 34% and 23% of our sample, respectively. Some Chinese suppliers also are required to comply with the EICC code of conduct, according to their procurement agreements, which should leave the more willing to address labor and environmental issues. We included an item that measured the degree (5-point Likert scale) to which a supplier perceived it would risk its reputation if it failed to comply with EICC rules, labeled EICC compliance.
Third, we controlled for each province's average economic development in the year of the audit. The province's annual per capita GDP in 2017 dollars is provided by the International Monetary Fund in its World Economic Outlook reports. We used the log value of GDP to reduce skewness and labeled this control variable province GDP. A more developed province may have more resources to enforce environmental and labor regulations .
We also applied several remedies for potential common method bias (Podsakoff et al., 2003).
Namely, we guaranteed respondent anonymity in the survey, conducted a pilot test to minimize item ambiguity, and included several objective survey questions (e.g., number of buyer audits, number of lean trainings, number of years the supplier had worked with Philips). We also relied on different information sources for the independent variables (i.e., survey and archival public data sets) and two dependent variables (i.e., audit reports by an independent auditing firm). Thus by design, our study is unlikely to be subject to common method bias. Table 5 contains the descriptive statistics and correlations. We examined the data for violations of assumptions of normality and multicollinearity. The variables exhibited approximately normal distributions with the exceptions of firm size, province GDP, and relationship length, which were log transformed.

Results
---Insert Table 5 & Table 6 about here ---Because our dependent variables consist of count data, we used Poisson regression with robust standard errors to estimate the statistical models. We subsequently estimate environmental and labor violations simultaneously through seemingly unrelated regression (SUR) analysis. The main reason for performing a SUR analysis is that it allows us to formally compare the regression coefficients of independent variables in the environmental violations with their counterparts in the labor violations. We standardized all independent and continuous variables before they were entered in our SUR regression analysis. The results are reported in Table 6. A significant, negative coefficient of environmental regulatory pressure on environmental violations (β = -0.17, p = 0.04; IRR = 0.84) suggests that regulatory pressure can prompt suppliers to reduce the number of environmental violations by up to 16%. The coefficient of labor regulatory pressure is not significant though (β = 0.03, p = 0.66; IRR = 1.02). The Wald test between equations indicated that the effects of environmental regulatory pressure on environmental violations is significantly more negative than the effects of labor regulatory pressure on labor violations (p < 0.05), offering further support for hypothesis 1. The buyer audits coefficient is not significant for environmental or social violations, indicating that hypothesis 2 is not supported. With regard to environmental NGO pressure, the coefficient is significant in the opposite direction (β = 0.12, p < 0.01; IRR = 1.13). This suggests that suppliers blacklisted by the IPE database do not appear to address their environmental issues. Due to the lack of data regarding labor NGOs in China, we could not fully test hypothesis 3.
Among the relational drivers, the significant, negative coefficient of lean training on environmental violations (β = -0.22, p < 0.05, IRR = 0.80) indicates that it is effective in reducing up to 20% of environmental violations. Although this coefficient is also negative for social violations, it is not significant (β = -0.01, p = 0.49; IRR = 0.99). Also, the negative coefficient of relational capital on environmental violations (β = -0.18, p = 0.01, IRR = 0.83) suggests that suppliers that share norms of trust, respect, and reciprocity with Philips can reduce such violations by up to 17%, whereas the coefficient is not significant in relation to social violations (β = 0.03, p = 0.68; IRR = 1.03). The Wald test between equations indicated that the effects of both lean training and relational capital on environmental violations are significantly more negative than the same effects on labor violations (p < 0.05). Thus, hypotheses 4 and 5 receive support.
Regarding the control variables, we find a positive, significant coefficient of firm size and province GDP on social violations. Larger suppliers are more likely to address their labor violations perhaps due to their larger resource base. Suppliers located in richer provinces have more labor violations. These provinces might rely on more migrant workers, who tend to be more exposed to labor breaches (e.g., lack of social insurance). Conversely, the coefficients for ISO certifications, EICC compliance and supplier relevance are negative and significant. Earlier studies similarly show that ISO-certified plants achieve greater employee productivity and morale and decrease absenteeism (Elmuti & Kathawala, 1997). Also, suppliers that concentrate Philips' expenditures are more likely to address their labor violations. Finally, suppliers who believe that complying with the EICC code of conduct reduces their risk of reputational damage are more likely to address labor violations. The EICC code has been adopted by various electronics brands, which collectively might pressure suppliers to show more social responsibility (Villena & Gioia, 2020).

Alternative measures
We consider alternative measures for both environmental and labor regulatory pressures, as detailed in Appendix E. First, we replaced the number of environmental fines with the value of those fines (in $), as reported in the China Environment Statistical Yearbook (2013; see also Van Rooij et al., 2017). When we reran the analysis, the results remain the same. Second, we replaced the number of labor lawsuits with the number of labor protection bureau offices, an alternative measure that captures the monitoring capacity of a province for assessing compliance with labor standards . It also came from the China Labour Statistical Yearbook (2013).
The results match the main results, except that labor regulatory pressure becomes significant. A province with more allocated resources might prompt supplier firms to address their labor issues.

Robustness checks
We conducted several robustness checks. First, suppliers listed on the Shanghai or Shenzhen Stock Exchange could experience contrasting pressure from investors to deliver short-term results (and ignore long-term responsibility goals) while also being exposed to more public scrutiny (implying more pressure to address environmental and labor issues) (Bansal & DesJardine, 2014;Lo et al., 2018;Wang & Choi, 2013). Nine percent of participating suppliers were listed in these two stock markets. We reran the models, controlling for listing on either exchange (1 = yes, 0 = no), and the results remain the same. Second, suppliers that have a bigger stake in their relationship with Philips (i.e., Philips represents a greater share of their revenues) could be more motivated to address environmental and labor issues, but the results do not change when we control for suppliers' percentage of production output devoted to Philips. Third, suppliers competing mainly on cost might find it especially challenging to deal with environmental and labor issues. In the survey, we asked suppliers to designate their competitive capabilities in terms of cost, flexibility, and innovation (Paulraj, Jayaraman, & Blome, 2014). If suppliers prioritized cost, by assigning it more than 50 points (out of 100), we designated them as exhibiting high cost competitiveness. But when we included the related dummy variable (1 = cost competitive, 0 = otherwise), the results do not change.
Finally, we note that most of the independent variables (e.g., regulatory pressure, NGO pressure, lean training) were measured at least two years before the dependent variables, which reduces reverse causality concerns. Pressures from regulators, NGOs, and buyers also are exogenous (i.e., Philips cannot control them). The relational variables-lean training and relational capital-could be of more concern, if such managerial decisions reflect responses to the audit results. For example, Philips might require a supplier to participate in lean training after it reaches some threshold level of non-conformance. However, this argument does not seem to apply to our study setting. Philips provides lean training to all its strategic suppliers, regardless of audit results.
Another concern might arise because we did not capture Philips' trust in suppliers, and Philips might calibrate its level of relational capital with a supplier according to the supplier's audit results.
Instead, we measure the supplier's perceptions of its relational capital with Philips and reduce the potential endogeneity problem that way. Moreover, the interviews indicate that the suppliers' relational norms with Philips have remained stable over time.

Discussion
Our goal with this study was to provide more in-depth understanding of the nature of suppliers' environmental and social responsibilities in China and their distinct drivers. Building on both stakeholder theory and the relational view, we have identified two set of drivers, which we test using multiple methods. The results show that Chinese suppliers respond to both stakeholder pressures (and regulatory pressures in particular) and relational drivers by increasing their environmental responsibility efforts. However, the same drivers are not effective for encouraging supplier social responsibility.
The results regarding regulatory pressures as a key driver for environmental responsibility among Chinese supplier firms are in line with recent studies that note the government's strong emphasis on the need to deal with environmental pollution (Li et al., 2019;Wang et al., 2018).
Strong environmental protection laws and enforcement efforts have led to increasing numbers of environmental inspections and stricter sanctions, in response to China's severe air, water, and soil pollution (Van Rooij et al., 2017;Wang et al., 2018). However, despite being the world's largest country by population and the biggest manufacturer, China has not addressed its prevalent labor rights issues (e.g., excessive overtime, lack of social insurance). Government agencies in charge of enforcing labor-related laws should be equipped with more sanctioning power and resources comparable to those granted to the government agencies in charge of enforcing environmental laws if the goal is to get supplier firms to address environmental and labor issues to the same degree.
Notably, in Chinese provinces with more labor bureau offices, supplier firms appear to address labor issues to a greater degree, suggesting that some regions may feature enhanced monitoring capacity, even if not necessarily stronger enforcement.
Buyer pressures induce mixed outcomes. Some suppliers started to consider environmental issues after being audited by strict buyers, such as Philips. However, the quantitative analysis shows that environmental and labor violations do not decrease with growing numbers of audits.
Furthermore, prior studies indicate that buyers rarely punish suppliers, even after they fail an audit (Lund-Thomsen & Lindgreen, 2014), and instead continue their business relationships with suppliers despite persistent evidence of their non-compliance (Barrientos & Smith, 2007;Distelhorst et al., 2016). China Labor Watch (2009) also alleges that some Chinese suppliers use bribery to pass audits or else have learned how to trick auditors. Yet audits remain the most common tool buyers adopt to monitor suppliers' environmental and social responsibility (Porteous et al., 2015;Villena, 2019). In our study, the suppliers were subject to an average of five audits per year while a few were subject to up to eleven audits per year. Thus, our results imply that buyers may need to reconsider their monitoring approaches and enforce more penalties on non-compliant suppliers. For instance, buyers might consider conducting joint audits, which could reduce suppliers' burden of being subjected to multiple audits (Caro et al., 2018;Chen, Qi, & Dawande, 2020). Alternatively, buyers could take a more hands-on approach, such that they start by investing in suppliers' sustainability capabilities, then monitor their subsequent progress via audits. Solely conducting more independent audits puts MNCs at risk of significant economic waste.
Pressure by NGOs in China-in contrast to those in Western countries (Doh & Guay, 2006)is growing but still appears ineffectual. Even IPE, the main environmental NGO in China (Lee et al., 2012), with some impact on public firms (Lo et al., 2018), cannot induce the supplier firms in our study (most of which are not public) to address their water and air pollution violations by blacklisting them. The second study's results show that IPE-blacklisted suppliers display a higher number of environmental violations. Thus, our participating Chinese suppliers still do not seem to recognize IPE as a source of pressure, likely because few electronics buyers incorporate IPE data into their supplier selection and monitoring processes. Although the 2019 IPE Annual Report cites an increasing number of brands that use its data, only a small percentage of them score high enough on a measure designed to indicate if the brand engages in high-functioning oversight to develop a portfolio of suppliers that take their environmental responsibilities seriously (IPE Annual report, 2019). A limitation of our study-we could not capture the impact of Chinese labor NGOs on supplier sustainability, because they do not systematically record labor rights violations-offers further evidence of the lack of prominence of NGO pressure. Our in-depth interviews in the first study suggest that the pressure of labor NGOs in China is too weak to induce them to address labor issues.
Turning to relational drivers, such as lean training and relational capital, we find pivotal influences with regard to overcoming a lack of technical knowledge and organizational skills among Chinese suppliers. However, they are effective only in relation to environmental efforts.
The suppliers we interviewed identified major changes to their manufacturing processes (e.g., plant layout) and progress in tackling severe environmental issues (e.g., waste disposal), based on the lean training they received. This outcome was confirmed in our quantitative study. However, with its strong emphasis on productivity and the dissemination of technical knowledge, the lean training offered by Philips does not mitigate suppliers ' labor violations. Similarly, Distelhorst et al. (2016) find that Nike's Chinese suppliers did not improve their social performance after participating in Nike's lean program. Western buyers arguably could help fill this gap, but thus far, their supplier development programs have not focused on improving worker rights and benefits (e.g., collective bargaining) and instead have centered on compliance with minimal labor standards (e.g., minimum wage), as a signal of corporate legitimacy (Anner, 2012).
Suppliers that have greater relational capital with Philips also are willing to invest in addressing environmental issues, but because they regard investments in social efforts as uncertain, they do not undertake such initiatives. Philips struggles to motivate suppliers to make structural and organization-wide investments that might help reduce excessive overtime or increase social insurance, a challenge that our quantitative analysis confirms, in showing that shared trust, respect, and reciprocity leads suppliers to address environmental issues but not labor ones. Thus, relational capital can safeguard suppliers' investments in responsibility practices, but only to a certain extent.

Theoretical contributions
This study integrates insights from both stakeholder theory and the relational view. Whereas most responsibility research focuses on stakeholder pressures as drivers of firm responses (e.g., Crilly Delmas & Toffel, 2008), evidence also indicates that supplier firms must be both willing and able to respond to those pressures (Durand et al., 2019). That is, they need support from buyers to respond to stakeholders' demands. Thus, in addition to stakeholder pressures, we consider relational drivers with likely enabling roles for supplier responsibility (Gimenez & Tachizawa, 2012;Huq et al., 2016;Lee & Klassen, 2008). By integrating these two literature streams, and conducting multi-method research, we derive a more comprehensive answer to a critical question that confronts academics and practitioners alike: How do stakeholder pressures and relational drivers differ in their impact on emerging-country supplier environmental and social responsibility?
Our results show that the impact of these drivers differs for environmental versus social responsibility domains. In this sense, our study offers an initial response to calls to "unpack the dimensions of CSR" to reflect the knowledge that "each [responsibility] dimension has its unique attributes" (Wang et al., 2016: p. 357). Our case study shows why Chinese supplier firms might be more likely to respond to regulatory pressures and relational drivers in environmental domains, whereas their social responsibility is harder to influence. Our quantitative analysis, involving 134 Chinese firms, further shows that environmental regulatory bureaus' sanctioning power, lean training, and relational capital all drive suppliers to address environmental (but not labor) violations. The combination of two methods means our research provides a robust, theory-based portrait of the drivers of emerging country suppliers' environmental versus social responsibility and offers a more integrated, complete view of such challenges that still remain unaddressed This study also advances applications of stakeholder theory to supplier responsibility topics. This theory has been used previously to investigate how MNCs respond to stakeholder pressures, but suppliers in emerging economies likely respond differently, due to their distinct environment, including a lack of public visibility (Bowen, 2002;Villena & Gioia, 2018). With our fine-grained, comparative analysis of different stakeholder groups and their distinct influences on Chinese suppliers' responsibility, we show that Chinese regulatory agencies pressure supplier firms to address their environmental deficiencies, but Chinese NGOs still lack enough legitimacy to exert influences. This insight in turn implies that MNCs should make more active use of IPE's data to select and monitor their suppliers' performance. They should not rely solely on audits, considering our finding that this type of buyer pressure does not ensure supplier responsibility.
Finally, this study features a novel consideration of the relational view too, moving beyond applications to explain economic benefits in interfirm relationships (e.g., Cousins et al., 2006;Kale et al., 2006;Lawson et al., 2008). We show that it can be extended to explain environmental and social benefits. In particular, the development of knowledge-sharing routines can help build and strengthen suppliers' environmental capabilities; relational capital can safeguard suppliers' investments in efforts to improve their environmental responsibility that they otherwise would perceive as too risky. However, relational drivers are not effective for safeguarding investments to address labor issues, which entail higher levels of uncertainty.

Limitations and research directions
Along with these contributions, this study has some limitations. The theoretical framework differentiates three groups of external stakeholders, which represents an advancement beyond studies that treat stakeholder pressures as uniform (e.g., Crilly et al., 2012;Surroca et al., 2013).
However, pressures also come from internal stakeholders and could exert complementary encouragements to improve supplier responsibility. For example, factory workers in China are increasingly vocal with regard to labor rights (Zhang & Luo, 2013); younger workers in particular appear more willing to act to protect their rights. We also do not consider the supplier firms' organizational capabilities and their potential effects. That is, we controlled for the firms' size, manufacturing processes, and management systems, but internal factors such as top managers' commitment to responsibility (Lee & Klassen, 2008) or amount of organizational slack (Bansal, 2005) could increase supplier firms' ability and motivation to address environmental and labor concerns. Further research thus should explore a wider range of influential factors. Furthermore, we investigate pressures exerted by NGOs in China, not the potential influence of international NGOs. Yet these stakeholders might encourage responsibility at the supply chain level, through the development of cross-sector partnerships (Johnson et al., 2018). The different roles that NGOs adopt (e.g., advocacy versus developmental) also might determine their effectiveness for improving supplier environmental and social responsibility, which offers another interesting avenue for research.
Continued research could consider other responsibility dimensions too. In particular, we did not include occupational health and safety (OHS) issues, largely because their ambiguous nature would complicate the comparative analysis. In particular, some studies include OHS issues within conceptualizations of environmental health and safety (Corbett & Kleindorfer, 2001), but the majority of prior literature treats OHS as part of the social dimension (e.g., Distelhorst et al., 2016;Locke et al., 2007Locke et al., , 2009).
Our integrated, mixed-method study overcomes many of the limitations associated with exclusively qualitative or quantitative methods. However, the sample consists solely of Philips' Chinese suppliers, and our factory visits confirm that most of them work with other global electronics brands. In this sense, our findings should be interpreted with caution in other countries where labor and environmental issues might differ. For example, child and forced labor problems are less common in China's manufacturing sector than in some other emerging economies, such as Venezuela or Cambodia. 4 Because the nature of the environmental and labor challenges thus varies from country to country (and even from industry to industry), (Wilhelm et al., 2016), we call for continued research in various supply chain settings, to improve understanding of diverse responsibility considerations that arise in global supply chains. Notes: SAM = supplier account manager, SSE = supplier sustainability expert at Philips. GM = general manager, GMA = general manager assistant, HRD = HR director, HRA = HR assistant, ED = engineering director, FW = factory workers, FD = finance director, SCM = supply chain manager, EHS = environmental health and safety specialist, MSD = management system director, SD = safety director, QM = quality manager, PM = production manager at suppliers. "It is very common in our country, especially in the Pearl Delta Area and the Yangtze Delta Area, that factories lease workshops from local enterprises, which constructed the workshops before they had obtained all the necessary environment and safety permits from the local government." (Philips' Supplier Account Manager) Costly and lengthy application procedures Working hours "Average hours worked in a workweek does not exceed 60 hours or the legal limit (whichever is stricter)." "When the workers noticed that their monthly incomes went down, because of overtime control, they started to complain. As you can understand, the workers have to support their families. When they cannot earn enough, they leave us to work for another company." (General manager assistant, Eta) Diverging stakeholder pressures "Almost all manufacturers in our country have constructed certain facilities illegally, in the sense that not all the necessary permits are readily available... Applying for all the necessary environment and safety permits is a tedious process in which dozens of governmental agencies are involved." (Philips Supplier Sustainability Expert) Corruption and bureaucracy "There is a problem of labor shortage in Guangzhou. It is becoming more and more difficult to recruit shop-floor workers. Meanwhile, we also have to cope with expedited orders with very short lead times… How can we deliver on time without scheduling overtime work?" (HR manager, Beta) Maintain operational flexibility while controlling staff size "There are not many qualified service providers for disposing toxic wastes in Guangzhou. Service fees are extremely high. During the internal audit, we thought finding one qualified service provider would be enough, and we did not differentiate between different types of toxic wastes. It turned out that SW12 is quite different from SW19, and we needed to find two service providers for these two types of toxic wastes." (HR manager, Beta) Lack of qualified service providers that can handle different types of toxic waste Social insurance "Social insurance schemes (e.g., pension, unemployment and medical) and other benefits as required by local law is provided to all workers." "China started to develop its social security system in 1994, but the system is still characterized by a lack of integration across its 31 provinces. Because it is still very difficult to transfer social securities across provinces, workers generally were not willing to contribute to the social security system in their host provinces. Recurring scandals such as corruptions and embezzlements in the past two decades undermined worker's trust in the social security system." (Supplier Sustainability Expert, Philips) Regional fragmentation and malfunctioning of national social insurance system Wastewater and solid waste "Effluent discharges (industrial/process wastewater, sewage and storm water) meet the discharge limits for regulated constituents." "The SGS auditors reported that the level of copper particles was not clearly monitored in our emitted water. That was indeed a very serious issue because we did not have the necessary equipment and technology to accurately monitor that." (Management Representative, Theta) Lack of equipment "We contacted the Bureau for social insurance. We told them that our customer was asking for official proof that [Epsilon] was complying with the local regulation on social insurance. However, the government official told us it would be very difficult to provide such an official proof. The local government in Suzhou is not strictly enforcing this law." (HR manager, Epsilon) Low enforcement of social security laws of local labor agencies "During the first audit, they told us that our emitted water contained hexavalent chromium, we did not know how to deal with that. We sent the sample to the research institute in Hangzhou, and the analysis showed that the level of hexavalent chromium was 3 times higher than the legal level… The researchers suggested us to purchase a specialized equipment to deal with this problem, but the equipment can cost as much as 800,000 Yuan." (General manager, Kappa) Lack of knowledge of permitted limits and insufficient resources for technological investments   Electronic copy available at: https://ssrn.com/abstract=3681079 We report unstandardized coefficients (B), robust standard error (r.s.e.), and incidence rate ratio (IRR). Sample size is 134 Chinese supplier firms.
Electronic copy available at: https://ssrn.com/abstract=3681079 "A couple of weeks ago, officials from the local environmental protection bureau dug four wells around four corners of our plant, and they took sample water from these wells to test whether the underground water has been contaminated or not. They further installed electronic monitoring devices next to our water emission tunnels to have real-time access to our emission status…" (Director of Environmental Safety, Supplier Gamma) "Nowadays, factories have to meet strict standards on pollution emission before they are allowed to locate in this industry zone, for example, waste water has to be processed before being discharged in the sewage system. It was not like this by the time when our factory moved to this industry zone [in 2007]…… It is clear that the whole society has started to accept and embrace environmental protection." (General Manager, Supplier Epsilon) "The central government has started to pay attention to labor protection in recent years. In the past, they did not pay that much attention… Labor Contract Law was enacted [in 2008]. However, the enactment of this law has caused many issues. The new labor law has many clauses and requirements that are not really reasonable, at least in the current context of China." (General Manager, Supplier Epsilon) "During the training session, experts highlighted that employers have a responsibility of protecting the physical, mental, and financial wellbeing of employees. To be honest, I cannot see how complying with the code on overtime control will benefit our workers. The minimum wage is still very low. Without having adequate overtime works, most workers will not even be able to meet their needs." (General Manager, Supplier Kappa).

Urgency of claims by regulatory agencies
Operationalization Degree to which request from regulatory agencies calls for immediate attention (time  criticality) "This is a sustainable industry zone, probably the best in Guangzhou and even in China. It has the highest environmental standards, occupational health and safety, and also labor standards. It is a well-managed industry, as the local environmental protection bureau is quite active…… Companies that cannot meet these standards are not allow to operate in this industry zone." (HR Manager, Supplier Beta) "We need to take immediate actions to tackle environmental issues, before it is too late" (General Manager, Supplier Alpha) "Actually, local bureaus of labor protection are generally sympathetic with our suppliers' struggles in complying with the law on social insurance. Rather than requiring companies to immediately comply with this law, these bureaus would allow companies to develop a longterm plan to gradually involve all workers in the social insurance system, step by step." (Supplier Sustainability Expert for Eta, Philips) "I had lengthy discussions with them [managers of Supplier Beta] about the issue of dispatched workers. It is still a grey area, as the local government of Guangdong has yet to publish its policies on dispatched workers, even though the labor law has a clause which specifies that dispatched workers can only take temporary, supportive, and substitutable positions. I asked Supplier Beta to at least identify such positions as the first step." (Supplier Sustainability Expert for Beta, Philips)

Power of regulatory agencies
Operationalization Degree to which regulatory agencies exert punitive power in case of non-compliance "The local bureau of environmental protection is strictly monitoring all companies located in this industry zone. It is very proactive in conducting on-site inspections.
If we cannot meet all the environmental laws and regulations, we may be forced to move out of this industry zone." (HR Manager, Supplier Beta)" "They [officials from the local environmental protection bureau] have installed monitoring devices at our water emission points, and those devices are running on 24 hours per day… Violating the laws and regulations on environmental protection will have serious consequences." (Director of Environmental Safety, Supplier Gamma) On-site observation (April 28 th , 2014): The quality manager of Supplier Epsilon reported the following incident from another factory in the region. Shortly before International Labor Day, many workers complained about overtime works scheduled on that weekend. The general manager threatened these workers that they would face termination of their labor contracts if they were to defy the production schedule. These workers filed an official complaint to the local labor protection bureau, which levied a fine of 4,000 YUAN (approximately 500 EURO) to this factory. Workers were unhappy as they perceived the penalty as trivial.
"Officials from the bureau of labor protection rarely inspect our company. When a worker has filed a complaint at the bureau, I will need to report to the bureau and provide explanations." (HR manager, Supplier Theta) Urgency of claims by buying firms Operationalization Degree to which request from buying firms calls for immediate attention (time criticality) On-site observation (May 22 nd , 2014): SGS auditors found that one supplier of metal components located in Dongguan did not have environmental permits for one of its workshops, and contacted Philips's SSE in charge by phone. According to this expert, although most local governments were not strictly enforcing the laws and regulations on environmental permits at that time, Philips foresaw stricter enforcement of these laws and regulations in China. Philips was thus exerting pressures on suppliers to remedy related issues proactively.
On-site observation (Jan. 10 th , 2014): During lunch time, a group of supplier sustainability experts and supplier account managers in Shanghai shared several stories related to water purification systems. In 2012, Philips started to press its "It is not really realistic to require all of our suppliers to comply with the law on social insurance. The Chinese manufacturing industry is not ready for this high standard. Take Shanghai for example, strict compliance with this law means at least 500 YUAN of additional labor cost per month for employers. That will drive many SMEs to bankruptcy." (Supplier Sustainability Expert for Delta, Philips) "Once the SGS audit is concluded, customers pay less attention to overtime control because on-time delivery has the highest priority." (HR Manager, Supplier Alpha) suppliers to upgrade their water purification systems to meet higher national standards, although at that time meeting provincial standards was the standard practice. The main motivation was to prepare Philips' suppliers, particularly those that have high environmental impact, for stricter enforcement of laws and regulations on water purification. This proactive strategy turned out to be rather timely, as the Chinese Central Government started to promote higher standards for water emission as of January of 2014. Power of buying firms Operationalization Degree to which buying firms monitor non-compliances and exert punitive power "SGS auditors conducted the full-scope audit in 2012, which identified several serious non-compliances [related to the environmental permit of one cleaning workshop]. By the time I went there to conduct the first resolution audit, I was surprised to find that those issues were not closed yet. (…) they did not resolve those issues in the second resolution audit. I had no choice but to involve our purchasing manager who told the general manager of Epsilon that Philips would postpone paying for the products, until all pending issues were completely resolved." (Supplier Sustainability Expert for Epsilon, Philips) "That caused huge pressures for all of us because the postponement in cash injection from Philips caused serious problems for our company. We have learned a lesson from this incident that we have to comply with Philips' responsibility standards. Otherwise, the survival of our company is at risk." (HR Manager, Supplier Epsilon) "Our practices [faked attendance records]? I think it is okay, and I do not think it is going to cause any serious problems between our company and customers." (HR Manager, Supplier Alpha) -The interviewee believed that buying firms would not punish Supplier Alpha, even if the faked attendance records were exposed.
"If you want my honest opinion, most customers will prioritize on-time delivery over overtime restrictions. Standards like this are floating bottom lines." (General Manager, Supplier Alpha)

Knowledge-sharing routines
Operationalization Degree to which Philips offers regular trainings (e.g. lean, IDH) to its suppliers that can improve supplier capabilities with regards to their environmental/social responsibility "We have helped with automation. Automation is a key program that helps these suppliers to reduce labour costs, which are rising in Asia, specifically in China. This month, we are going to integrate lean principles in the supplier's operations for semi-automatic production. We have identified some gaps there." (Supplier Account Manager for Alpha, Philips) "They would provide us with many suggestions on reducing waste and improving productivity. To be honest, those suggestions are pretty theoretical, and we need to translate them into our daily practices. Of course, I appreciate their suggestions and contributions, especially the ones that are really valuable. I hope that they will be more engaged with us in really identifying room for improvement, because I trust they have adequate knowledge and expertise." (General Manager, Supplier Delta) "The IDH program is designed to improve management-worker relations. So far, it is really hard to assess whether these trainings are effective or not. In my opinion, the content of these training sessions is mostly theoretical, but has rather limited practical values." (HR Manager, Supplier Alpha) "Through participating in the IDH program, we want to better understand our workers' requests. We could see the value of this program, and we need to make long-term investments into this program. We need to be responsive toward workers' requests, otherwise, this program will only generate complaints rather than improvements." (HR Manager, Supplier Gamma)

Effective governance
Operationalization Degree to which suppliers perceive trust, respect, and reciprocity in their relationship with Philips that can safeguard their investments in environmental/social responsibility "A couple of years ago, we started to promote higher environmental standards in our supply base in China. It was two or three years before the Chinese government started to enforce the laws and regulations on environmental protection… Supplier [Gamma] was quite responsive to our suggestions, because they trust us. They invested more than one million US dollars in upgrading their water processing system… In doing so, they have been complying with the environmental laws and regulations ahead of their main competitors…" (Supplier Account Manager for Gamma, Philips) "[Supplier Gamma] knows that Philips is very much valuing supplier responsibility. They are one of the first suppliers that pass supplier responsibility audit almost six years ago. They have applied our responsibility requirements on their environmental management… And they are adhering to that. They are very much dedicated to that, and they are following these requirements in their businesses." (Supplier Account Manager for Alpha, Philips) "In general, we are quite responsive toward Philips' requests and suggestions. Of course, we would like to maintain our long-term collaboration with Philips. Based on these considerations, we have decided to participate in the IDH program, although it is not really clear what we can gain from this program." (HR Manager, Supplier Alpha) "Philips tried to persuade us to make use of this opportunity [to participate in the IDH program]. Of course, we always respect Philips and would heed to its requests, even though we know that participation in this program will cause us serious disruptions. We were already aware of the potential issues. Still, in the end we decided to participate in this program. We will try our best to make this program a success, although we can see a lot of challenges ahead." (HR Manager, Supplier Alpha) "From our point of view, the invitation to participate in the IDH program shows that Philips is satisfied with us, in terms of performance, evaluation, and cooperation. Of course, Philips has also invited other suppliers, but not that many. We feel quite honored to be invited to participate." (HR Manager, Supplier Gamma)  (16)